National Academies Press: OpenBook

Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214 (1987)

Chapter: 1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation

Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 1
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 2
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 3
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 4
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 5
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 6
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 7
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 8
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 9
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 10
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 11
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 12
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 13
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 14
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 15
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 16
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 17
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 18
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 19
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 20
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 21
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 22
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 23
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 24
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 25
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 26
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 27
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 28
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 29
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 30
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 31
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 32
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 33
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 34
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 35
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 36
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 37
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 38
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 39
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 40
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 41
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 42
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 43
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 44
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 45
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 46
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 47
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 48
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 49
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 50
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 51
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 52
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 53
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 54
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 55
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 56
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 57
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 58
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 59
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 60
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 61
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 62
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 63
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 64
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 65
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 66
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 67
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 68
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 69
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 70
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 71
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 72
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 73
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 74
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 75
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 76
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 77
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 78
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 79
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 80
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 81
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 82
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 83
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 84
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 85
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 86
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 87
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 88
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 89
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 90
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 91
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 92
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 93
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 94
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 95
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 96
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 97
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 98
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 99
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 100
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 101
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 102
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 103
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 104
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 105
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 106
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 107
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 108
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 109
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 110
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 111
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 112
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 113
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 114
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 115
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 116
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 117
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 118
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 119
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 120
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 121
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 122
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 123
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 124
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 125
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 126
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 127
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 128
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 129
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 130
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 131
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 132
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 133
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 134
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 135
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 136
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 137
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 138
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 139
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 140
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 141
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 142
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 143
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 144
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 145
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 146
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 147
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 148
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 149
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 150
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 151
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 152
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 153
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 154
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 155
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 156
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 157
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 158
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 159
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 160
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 161
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 162
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 163
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 164
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 165
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 166
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 167
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 168
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 169
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 170
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 171
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 172
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 173
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 174
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 175
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 176
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 177
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 178
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 179
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 180
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 181
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 182
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 183
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 184
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 185
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 186
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 187
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 188
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 189
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 190
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 191
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 192
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 193
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 194
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 195
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 196
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 197
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 198
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 199
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 200
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 201
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 202
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 203
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 204
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 205
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 206
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 207
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 208
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 209
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 210
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 211
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 212
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 213
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 214
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 215
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 216
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 217
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 218
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 219
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 220
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 221
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 222
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 223
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 224
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 225
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 226
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 227
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 228
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 229
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 230
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 231
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 232
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 233
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 234
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 235
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 236
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 237
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 238
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 239
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 240
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 241
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 242
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 243
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 244
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 245
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 246
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 247
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 248
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 249
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 250
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 251
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 252
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 253
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 254
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 255
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 256
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 257
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 258
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 259
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 260
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 261
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 262
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 263
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 264
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 265
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 266
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 267
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 268
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 269
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 270
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 271
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 272
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 273
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 274
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 275
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 276
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 277
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 278
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 279
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 280
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 281
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 282
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 283
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 284
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 285
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 286
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 287
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 288
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 289
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 290
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 291
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 292
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 293
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 294
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 295
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 296
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 297
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 298
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 299
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 300
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 301
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 302
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 303
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 304
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 305
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 306
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 307
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 308
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 309
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 310
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 311
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 312
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 313
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 314
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 315
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 316
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 317
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 318
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 319
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 320
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 321
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 322
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 323
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 324
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 325
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 326
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 327
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 328
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 329
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 330
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 331
Suggested Citation:"1. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation." National Academies of Sciences, Engineering, and Medicine. 1987. Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214. Washington, DC: The National Academies Press. doi: 10.17226/11357.
×
Page 332

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

SPECIAL REPORT 214 DESIGNING SAT'ER ROADS Practices for Resurfacing, Restoration, and Rehabilitation I I Tlansportation Research Board National Research Council Washington, D.C. 1987

Tiansportation Research Board Special Report 2f4 mode t highwaytrânsportation subject areas 2l facilities design 5l transportation safety 52 human facton Transporratiør Research Board publications are available by ordering directly from TRB. They may also be obtained on a regular basis through organizational or individual affiliation with TRB; affüiates or library subscribeis are eligible for subsøntial discounts. For further information, write to the Transportation Research Board, National Research Council, 2101 Constitution Avenue, N.W., Washington, D.C. 20418. Printed i¡ the United States of America NOTTCE The projecr that is the subject of this publication was approved by the Goveming R_oard of the Narional Academy of Sciences, the National Academy of Engineering, and the Institute of Medicine. The members of the committee responsible for the publication were chosen for their special competence and with regard for appropriate balance. - This report has been reviewed by a group other than the authors according to procedures approved by a Report Review Committee consisting of members of the National Academy of Siiences, the National Academy of Engineering, and the Institute of Medicine. This srudy was sponsored by the Federal Highway Administration of the U.S. Department of Transporlation. Library of Congress Cataloging-in.Publication Data National Research Council (U.S.). Transportation Research Board. Designing safer roads : practices for resurfacing, restoration, and rehabiliration. p. cnr, - (Special repon / TranspoÍ¿tion Reseafch Board, National Research council ; 214) Includes bibliographies. ISBN 0-309-04453-7 1. Roads-Maintenance and repair. 2. Roads-Design. I. Title. II. Series: Special report [Narional Research Council (U.S). Transportation Research Boardl ; 214. T8220.N37 1987 625;7'6-4c19 87-t5331 CIP

COMMITTEE FOR THE STTJDY OF GEOMETRIC DESIGN STANDARDS FOR HIGHWAY IMPROVEMENTS Peran G. Kor-rNow, American Trucking Associations, Alexandria, Virginia Co-Chnirnnn Hnnsrnr H. Rrc¡¡anosoN, The Texas A&M University System, College Station, Co-Chairman Rov }V. ANorRsoN, TranSafety, Inc., Springfleld, Virginia LroNann EvaNs, General Motors Research Laboratories, Warren, Michigan Jo¡rN C. GleuNoN, John C. Glennon Chartered, Prairie Village, Kansas Ezna Hauan, University of Toronto, Onta¡io W. Ro¡t¡lo HuosoN, University of Texas, Austin Jacr T. KasseL, Sacramento, Califomia Javss L. MeRrn¡, Fresno, California Bnoors O. ñcnors, Arkansas State Highway and Transportation Department, Little Rock Bn¡aN O'NnLL, The Insurance Institute for Highway Safety, Washington, D,C. Rosenr H. R¿yvo¡¡o, Jn., Pennsylvania Department of Transportation, Hanisburg Jou¡'¡ H. SHaneR, New York State Department of Transportation, Albany Rrc¡ano R. Srnnuen, Jn., Mansñeld Asphalt Paving Company, Mansfreld, Ohio Ja¡¡es I. Tnyr-oR, University of Notre Dame, Incliana E. DEaN Trsonr-e, Idaho Transportåtion Department, Boise Liaison Representatives D,qvn J. Herusnic, American Association of State Highway and Transportation Officials, Washington, D.C. Mrcs¡le A. McMunrnv, National Transportation Safety Board JaaN Scnnac-L¡uven, Senate Environment and Public Works Commifiee Ssppo I. SrLLaN, Federal Highway Administration, U.S. Depârtmènt of Transportation Rrcn¿.no V. Tsnzu-e, House Committee on Public Works and Transportation JcNr.rtFER WIsHanr, Congressional Budget Offìce D¡vro K. Vy'mrrrono, Transportation Research Board Crvna E. Woonle, Jn., House Committee on Public Works and Transportation i i i I t I

Tiansportation Research Board Staff Rossnr E. Sxn¡¡æn, Director for Special Projects H¿nnv S. ConEi.r, Senior Program Officer Josrpn R. Monnrs, Senior Program Officer Jom¡ A. DaacoN, Consulønt Rrc¡¡¿no MaRcrorrn, Research Associate Mar-cor-¡"r Quwr, Research Associate Envme Tn¡yron Chuvp, Senior Editor

Preface In response to a provision in the Surface Transportation Assistance Act of 1982, the Secretary of Transportation, acting through the Federal Highway Administration, requested the National Academy of Sciences to study the safety cost-effectiveness of highway geometric design stândards and recom- mend minimum st¿ndards for resurfacing, restoration, and rehabiliøtion ßRR) projects on existing federal-aid highways, except freeways. Specifi- cally, the act called for the Secreøry of Transportation to enter into Íurange- ments with the National Academy of Sciences to conduct a study of the safety cost-effectiveness of geometric design criteria of standards currently in effect for construction and reconstruction of highways, other than highways access to which is fully controlled, to determine the most appropriate minimum standards to apply to resurfacing, restoration, and rehabilitation projects on such highways . . . and to propose standards to pre- serve and extend the service life of such highways and enhance highway safety. To carry out the study, the National Research Council, the principal operat- ing agency of the National Academy of Sciences and the National Academy of Engineering, assembled a committee of 16 experts in the various disciplines needed to develop and apply geomefric design standards and assess their impact on safety, highway serviceability, cost, environment, and system administration. Committee members included individuals with experience in highway design, traffic engineering, highway safety, accident analysis, high- way construction, statistics, economics, highway administration, and law. The committee began its work with a review of RRR practices in state and local highway agencies. Committee staff visited the state highway agency and

vl PREFACE the Federal Highway Adminisnation offices in each of the 15 states selected for case studies and conducted telephone interviews with local highway officials representing 16 counties, 20 cities, and 3 metropoliøn planning organizations. Federal, state, and local offlcials provided valuable information on the types of projects funded with federal aid, procedures used to select RRR projects, current design standa¡ds and their use, and the ways in which safety needs a¡e taken into account. The study committee sponsored critical reviews of prior research on the safety effects of key highway features and special research projects on - t-- t---- ^-à -^^Å-:Å^ ^^f^a-. TL^ ^-:]:^^l - .,:^,.,- â-,¡ €ñJ:ñ^ôpavelllc'llt ËuBç ulups at¡u tu¿lusruç ¡ralçLy. r¡¡9 vlrllv4l lvvrvwù 4rru u¡rsr¡¡öù from the special resea¡ch projects were used to make judgments about rela- tionships between safety and key highway features. For several design fea- tures, the committee found suftcient evidence to support quantitative relation- ships between safety and design improvements. However, these relationships must be viewed as approximate in nature. Although the relationships are based on the best available data, they could be subsøntially changed by the results of future research. In addition, the study committee developed relationships between cost and key highway features. These relationships are based on an examination of published cost data, cost records, and cost-estimating procedures for a sample of highway agencies throughout the country. The cost relationships provide estimates of typical costs for making geometric design improvements on RRR projects. However, the cost for a given improvement can vary considerably from site to site because of variations in site conditions, labor and material costs, design practices, and project scale. Thus, actual costs could be much g¡eater or less than estimates developed using the cost relationships. The safety and cost relationships were used to assess the safefy cost- effectiveness of geometric design stândards. The added cost per accident eliminated that can be expecæd for improvements to highway geometry was estimaæd for illusnative projects. When system data were available for existing highway conditions, the study committee examined the effects of alternative RRR sønda¡ds on systemwide safety and the total expenditure needed to meet the standard on a nationwide basis or for selected states. Drawing primarily on case studies of current RRR practices and analyses of safety cost-effectiveness, the committee has recommended a variety of prac- tices that encompass the entire RRR process but with special focus on design. In selected instances, federal, state, and local highway agencies can use the recommendations, along with published manuals, design aids, and local expe- rience to develop or modify minimum design st¿ndards for RRR projects. For federal-aid RRR work, the Secretary of Transportation is required by statute to ensure thatprojects are designed and constructed in accordance with standards that extend the service life of highways and enhance highway safety. To

PREFACE vll accomplish this, the Secretary, acting through the Federal Highway Admin- isEation, must either set nationwide RRR standards or approve standards adopted by individual states. In either case, ttre committee's recommendations provide guidance. In addition, the committee has recommended various research and raining activities that federal and staæ highway agencies can use to improve their ability to enhance safety through RRR projects. The study was performed under the overall supervision of Dr. Damian J. Kulash and Robert E. Skinner, Jr., the former and current Directors for Special Projects. Robert E. Skinner, Jr., directed the project staff. Dr. Harry Cohen, Joseph R. Morris, D¡. John A. Deacon, Richard Margiofta, and Malcolm euint made signifrcant confibutions. Special appreciation is expressed to Nancy A. Ackerman, TRB publications Manager, and Edythe T. Crump, Senior Editor, for editing the final report and to Marguerite E. Schneider, Frances E. Holland, and Margaret M. Sheriff for typing the many drafts and the ûnal manuscript.

Contents EXECUTIVE SUMMARY 1 Gao*rr:rtrc DFsrcN Sraxn¡.nls ron Rrsrrn¡'¿,crNc, RnsroRATIoNt a¡¡o Rn¡ranrr.rranroN Pno.rÈcrs: BacxcnornD añD IssuEs. 14 Introduction, 14 Evolution of Federal Highway Policy, 16 Federal-Aid Highway Program, 18 Geometric Design Standards and Federal Rulemaking, 24 Key Issues, 30 References, 32 2 Srarp aNo Locar. Pnocnuunes ron Sn,r,rcuoN, DrstcNr.nNo CoNsrnuctloN oF Hrcnwav Irrpnovenmxt PnoJrcrs .. . . 35 Review of RRR Practices: Information Sources, 37 Ståte RRR Programs, 38 Local RRR Programs, 66 Summary of Findings, 72 References, 75 3 Rrlar¡oNsnrps BerwpBN Slnli-rv lNn GBovrnrnrc DrsrcN 76 Application of Safety Relationships to Design Søndards, 76 Relationships Between Safety and Key Road Features, 78 Low-Cost Safety Measures, 100 Effect of Changing Vehicle Fleet, 102 Roadway Consistency, 104 Summary, 105 References, 106

4 Rrr.arroNsgrps BmwBBx Hrcnwav Cosrs eNn Gnorr-rnrc Drs¡cx ...... : Cost Relationships-Problems and Limiøtions, 110 Typical RRR Project Costs, 113 Added Project Costs for Geometric Improvements, 116 Right-of-Way Requirements, 125 Maintenance Cost Implications, 126 Summary, 129 References, 129 5 Srnrrv Cosr-EFFncrrvENEss op Gpop¡vrnrc DssrcN SraNn¿,nos .,. Ea¡lier Studies of Safety Cost-Effectiveness in Highway Design, 132 Scope and Framework of Cos¡Effectiveness Analyses, 133 Safety-Cost Trade-Offs, 136 Safety-Preservation Trade-Offs, 166 Summary of Findings, 170 References, 172 6 Tonr Lresrr.rry ¿.Nu Gn'ovrr,"rnrc DBsrcN Background on Tort Liability, 175 Implications for RRR Design St¿ndards and hactices, 178 Summary, 183 References, 184 7 FrNorNcs ¿.Np Rncom,rENDED Drsrcx Pnacrrcrs non RBsunrncrNc, RnsroutloN, AND ReiHnrlrranroN PnoJrcrs ., . . . . Findings, 186 Safety-Conscious Design Process, 190 Design hactices for Key Highway Features, 193 Other Design Procedures and Assumptions, 2M Planning and Programming RRR Projects, 207 Safety Research and Training, 208 References, 212 ApraNorx A Summary Comparison of Nonfreeway Geometric Design St¿nda¡ds and Guidelines. AppnNorx B Case Study State andLocalRRRhograms ........ : Appnxnx C Summary of Detailed Safety Relationships. ........, AprrNorx D Relationship Between Accidents and HorizontalCurvature ........: Appprunrx E Relationship Between Accidents and Sight Distance at Crest Vertical Curves

Arrr,Nnrx F Relationship Between Accidents and Speciflc Roadside Featuresr.... r ,. ... 270 AppnNmx G Physical and Operational Features AffectingSafetyatlntersections .......286 Arpe¡mrx H Highway Accidents on the Federal-Aid System. .... 292 AppaNnrx I Initial Cost to Flatten Highway Curves ... ... 296 Aprei.¡n¡x J Relationship Between Cost per Accident Eliminated andBeneût-CostRatioApproaches... ....301 Apprxnx K Effects of Lane and Shoulder Widths on Thavel Time. . . . . . 303 Appnnnrx L Alternative Lane and Shoulder Width StandardsUsedinSystem-LevelAnalyses. .....308 Srrmv Counrrtra Brocnapslcar, fxnon*ranroN . . .... .. 372

Executive Summary In response to a provision of the Surface Transportation Assistance Act of 1982, the Secretary of Transportation, acting through the Federal Highway Adminisnation, requested the National Academy of Sciences to study the safety cost-effectiveness of geometric design standards and recommend mini- mum standards for resurfacing, restoration, and rehabilit¿tion (RRR) projects on existing federal-aid highways, except freeways. RRR projects may include resurfacing and other pavement repairs, minor widening of lanes and shoul- ders, minor alterations to vertical and horizontal alignment, bridge improve- ments, and removal of roadside hazards. Until 1976, federal highway funds could be used only for the construction of new highways or complete reconstruction of existing highways. This policy was changed by the Federal-Aid Highway Act of 1976, which authorized state and local highway agencie" to use federal aid for RRR projects on existing federal-aid highways. RRR projects can extend the service life of existing highways through pavement and other repairs and at the same time improve highway safety by making selective improvements to highway geometry and other roadside features. Striking a balance between preservation and safety improvements on RRR projects has proved controversial, however. The conûoversy has centered on which minimum geometric design stan- dards should be applied to RRR projects to qualify for federal aid. Some highway organizations have contended that pavement repairs alone enhance safety and that additional safety improvements would greatly increase project costs and delay improvements to many miles of deteriorating highways.

2 DESIGNIT.{G SAFER ROADS Safety organizations, on the other hand, have viewed the federal RRR pro- grarn as an opportunity to make long-needed safety improvements to older highways at the same time as pavement repairs a¡e made. These organizations have viewed ttre flexible RRR standards proposed by some highway agencies as too lenient and have favored a more rigorous, safety-oriented design process. A study committee of 16 individuals with expertise in highway safety, design, and administration conducted case studies of curent RRR design practices, reviewed cl¡rrent knowledge about relationships between geometric design anci safery, an<i anaiyzed rhe cosi and saieiy irade-oüs of geomeirjrc improvements to existing highways. These activities led Ûo a number of flndings concerning the effects of RRR projects on highway safety' On the basis of these findings, the study committee has recommended a variety of practices that will increase the safety cost-effectiveness of RRR projects. FINDINGS Resurfacing, restoration, and rehabilitation projects enable highway agencies to improve highway safety by selectively upgrading existing highway and roadside features without the cost of full reconstruction. For example, widen- ing lanes and shoulders on two-lane rural highways on the federal-aid systems alone could save approximately 1,000 lives and prevent nearly 30,000 injuries each year. Federal-aid RRR projects usually enhance safety. Moreover, since 1982 when Congress declared that RRR project objecúves include both the exten- sion of highway sewice life and the enhancement of safety, highway agencies generally have paid increasing attention to safety. Nevertheless, many oppor- tunities for low-cost safety improvements are neglecæd, and RRR funds currently spent for safety improvements could be redirected for greater sys- temwide safety gains. A number of factors are responsible for this situation: r RRR design practices vary widely ftom agency to agency- Some highway agencies follow exemplary practices to address safety needs; others do not place enough emphasis on safety. o RRR projects are initiated prímarily to adiress pavement repair and rehabilitation needs. Safety needs a¡e often not addressed until little time remains úo accommodate geometric improvements that require additional time for design or right-of-way acquisition. c Fed¿ral-aid RRR projects frequently widen lanes and shoulders but seldom reconstruct sharp curves or replace bridges with narrow decks. Because there is a higher concentration of accidents at curves and bridges,

WECWNE SUMMARY improvemenfs at these locations can sometimes be more safety cost-effective than routine cross-section improvements. . Not enough is known abow the s$ery gains that will occur after the geometry of existing highways is improved or other s$ety-oriented improve- ments are nnde. Avulable information is not always in the hands of designers, or in a form that can be applied without ambiguity. Also, past studies of the safety effects of geometric design improvements frequently lacked rigorous statistical controls, a shortcoming that severely limits the accuracy of study results. . Engineers who ødminister state traffic and safety programs seldam participate in the design of RRR projects. They are usually the agency søff members most knowledgeable about accident data and special safety mea- sures. Design standards alone cannot address these factors that collectively limit the safety gains of federally funded RRR projects. Within the overall process of planning, selecting, and designing RRR projects, the influence of safety standards is small. RRR sønda¡ds, which can affect only a few key design features, cannot be tailored to ût all possible, or even most, circumstances encountered in a given state or at a specifrc site. Consequently, a variety of practices are recommended that encompass the RRR process but with special focus on design. In selected instances, federal, ståte, and local highway agencies can use the recommendations, along with published manuals, design aids, and local experience to develop or modify minimum design standards for RRR projects. For federal-aid RRR work, the Secretary of Transportation is required by statuæ to ensure that projects are designed and constructed in accordance with standards that extend the service life of highways and enhance highway safety. To accomplish this, the Secre- tary, acting through the Federal Highway Administration, must either set nationwide RRR standards or approve standards adopted by individual states. In either case, the committee's recommendations provide guidance. The recommended practices also provide guidance on the planning and programming of RRR projects, existing conditions that warrant special design analyses, safety improvements that should routinely be considered on RRR projects, and training and research. These practices are intended !o develop more safefy-conscious design. This will enhance highway safety nationwide by taking advantage of low-cost opportunities to improve safety and selecting the most safety cost-effective improvements. If these recommendations are followed for federal-aid projects on nonfree- way highways, project spending for lane and shoulder widening will generally decline and spending for alignment, bridge, roadside, and intersection improvements, as well as project design, should increase. In some states these

4 orsrcumc sAFER RoADS shifts may decrease RRR project costs; in others they will increase costs. Nationwide, the typical project cost will probably increase slightly but not enough to measurably affect RRR pavement repair and preservation activities. RECOMMENDATIONS Study recommendations are organized into five cafegories (Table ES-l): 1. Safety+onscious design Drocess, 2. Design pracfices for key highway features, 3. Other design procedures and assumptions, 4. Planning and programming RRR projects, and 5. Safety research and raining. TABLE ES-l Organization of Study Recommendations Safety Consc ious D esign P roce s s L Assessment olSite Conditions Affecting Safety 2. Determination olProject Scope 3. Documentation of the Design Process 4. Review by Traffic and Salety Engineers Design Practicesfor Kèy Highway Features 5. Minimum Lane and Shoulder Widths 6, 7. Horizontal Curvature and Superelevation 8. Vertical Curvature and Stopping Sight Distance 9. Bridge Width 10. Sideslopes and Clear Zones I I . Pavement Edge Drop and Shoulder Type 12. Intersections 13. Normal Pavement Crown Other Design Procedures and Assumpüons t4. Tiaffic Volume Estimates for Evaluating Geometric Improvements I 5. Speed Estimates for Evaluating Geometric Improvements 16. Design Values lor Geometric Improvements 17. Design Exceptions Planning and Programming RRR Projects 18. Screening of Highways Programmed for RRR Projects 19. Assessment of the Systemwide Potential lor Improving Safety Safety Research and TÌaining 20. Special Täsk Force to Assess Highway Safety Needs and Priorities 2 I . Compendium of Inlormation on Safety Effects of Design Improvements 22. Increased Research on the Relationships Between Safety and Design 23. Salety Training Activities for Design Engineers These recommendations apply to nonfreeway RRR projects whether or not they are funded with federal aid. For federal-aid RRR projects in particular, the Secretary of Transportåtion, through the Federal Highway Adminishation,

ÐGCWNE SUMMARY 5 should øke the necessary steps to implement the recommendations in the ûrst three categories-safety-conscious design process, design practices for key highway features, and other design procedures and assumptions. Taken together, these recommendations comprise a practical national policy on RRR project design that will be more safety-cost effective and comprehensive than an extensive set of rigid minimum standards. The fourth category, planning and programming RRR projects, is direcæd to state and local highway agencies that have ttre authority to perform ttrese functions for federal-aid projects without federal oversight. The flnal category safety research and training, is direcæd to the larger highway community with specific recommendations intended for the Congress, the Federal Highway Administation (FHWA), the American Association of StatÊ Highway and Transportation Offlcials (AASHTO), and staæ and local highway agencies. Safety-Conscious Design Process Significant improvements in safety are not automatic by-products of RRR projectsi safety must be systematically designed into each project. Highway designers must deliberaæly seek opportunities specific to each project and apply sound safety and trafûc engineering principles. Designers of RRR projects work with existing highways whose design and operational charac- teristics can be observed and measured; yet not all highway agencies take advantage of these favorable circumstances. Greater attention to safet¡ along with greater document¿tion of the design process improves design decisions-. Highway agencies should review and revise their design practices to incorpo- rate the following steps. o Ass¿ss current conditions: At the beginning of RRR project design, highway designers should assess existing physical and operational conditions affecting safety by using accident datå, site inspections, and measurement of existing design and raffic characteristics. o Determine project scope:In addition to pavement repairs and geometric improvements, designers of RRR projects should consider and, where appro- priate, incoqporate other intersection, roadside, and traffic control improve- ments that may enhance safety. o Document the design process: Before developing construcúon plans and specifications, designers should prepare a safety and design report that covers existing design and operational characæristics, accident history, applicable design standards, identified safety problems and related design options, rationale for any proposed design exceptions, and the recommended design. o Review the design: Traffic and safety engineers should review safety and design reports, as well as proposed RRR designs, before final approval.

6 ossrcNl.Ic sAFER RoADs Although many state highway agencies already incorporate one or more of these steps in their design process, most will have to modify their process to include ttrem all. Design Practices for Key Highway Features Designers use minimum RRR geometric design standards to determine whether a particular geometric feature must be upgraded âs paft of a RRR ==--!- -¡ \t==:^,:-^r -:-:-,-- DDD ^;ññ¡^r,t. a¡a "'o*o¡to¡7 fia¡ natinnrrri¡lp ¡tqcproJtçt. I\ulrrglIL;an ltltllltlll¡ut r\r\l\ ù6rruquù av w@¡orùw ¡v¡ ¡¡ssvrilr ¡sv srv when the following conditions are met: o Trade-offs between safety and performance against cost can be evaluated quantitåtively, and conclusions can be drawn about the safety cost-effective- ness of different sfandards generally applicable regardless of the project. . Standa¡ds would help refocus RRR expenditures on more safety cost- effective geometric improvements. . Standards would simplify parts of the design process and FI{WA approval procedures, freeing design resources for the analysis of site improve- ments not covered by numerical standards. Lane and shoulder widths on fwo-lane rural highways meet these condi- tions, and minimum values are recommended. Two-lane rural highways account for about three-fourths of all nonfreeway, federal-aid highway mileage, and lane and shoulder widths are particularly imporønt because ttrey can affect highway safety and cost over the length of the highway. When these condiúons are not met, for other key features or categories of highways, other design practices are recommended that will help achieve the same safety objectives as minimum standards. These recomtnended pracúces specify threshold conditions that warrant detailed evaluation of particular improvemenß, improvements that should routinely be made or evaluated on RRR projects, or design policies that should be developed on a state-by-state basis. Minimum Lane and Shoulder Widths Minimum lane and shoulder width values ¿ìre recommended that FFIWA and state highway agencies can use to set minimum RRR design søndards (lable ES-2). These recommended values are simila¡ to the minimum lane and shoulder width values proposed by ttre FHWA in 1978 but include several modifications to improve safety cost-effectiveness. Most important, the aver-

øßCWNE SUMMARY TABLE ES-2 Recommended Minimum Lane and Shoulder Width values for Two-Lane Rural Highways l0 Percent or More Tiuckså Less Than l0 Percent Tiucks Design Year Running Volume Speed,(ADT) (mph) Combined Lane and Shoulder Lane widrh. widrh Combined Lane and Shoulde¡ Widrh. Lane widrh 1 -750 75 1 -2,000 Over 2,000 Under 50 50 and over Under 50 50 and over All l0 l0il T2 T2 t2 t2 15 r8 9 r0 t0 ll ll ll t2 T2 t4 t7 oHighway segments shouìd be classifled as "under 50" only ilmost vehicles have an average speed ofless than 50 mph over the length of the segment. 'For this comparison, trucks are defrned as heavy vehicles with six or more tires. 'One foot less for highways on mountainous terrain. age daily Eaffic (ADT) ranges arc adjusted so that a larger number of roads with high ADT and fewer roads wirh low ADT would be improved. Lane and shoulder width improvements are more cost-effective on high-volume roads than on low-volume roads. In terms of cost per accident eliminated, the recommended values are more cost-effective than other standards proposed for nationwide use. For all federal-aid, two-lane rural highways combined, the recommended minimum values imply approximately the same overall investment as the FFIWA stan- dards proposed in 1978-a total of roughly $13 billion if all of rhe lane and shoulder improvements were made a[ current cost levels. Application of these values, however, would eliminate about 10,000 (40 percent) additional acci- dents annually. l¿ss is known about the safety cost-effectiveness of widening urban and multilane rural highways, and minimum values that highway agencies can adopt as standards have not been proposed. Horizontal C urvature and Superelevation current RRR súandards and practices generally emphasize lane and shoulder width improvements and do not pay enough attention to alignment improve- ments. At traffrc volumes geater than 750 vehicles per day, reconstruction of horizonøl curves can be more safety cost-effective than lane and shoulder widening and can reduce vehicle operating costs and travel time. Because ofthe variability in project costs (and safery cost-effectiveness) for reconsrucdng similar curves, minimum geometric standards are inappropri-

8 ossrCNS{GSAFERROADS aúe. Nevertheless, highway agencies should evaluate the safety benefits and added costs of curve reconstruction when there is a reasonable possibility that reconstruction will be safety cost-effective. The study recommends that highway agencies ¡ Evaluate the reconstruction of horizontal curves when the design speed of the existing curve is more than 15 mph below the running speeds of approaching vehicles and the average daily faffic volume is gteater than 750 vehicles per day. . i4crease ihe supereievaiion Oí horizorriai ci¡rves whenever tlie design speed of an existing curve is below the running speeds of approaching vehicles and the exisúng superelevation is below the allowable maximum specified by AASHTO new construction policies. In many cases, safety can be improved at horizontal curves without costly reconstruction. Where reconstruCtion is unwa¡ranted, highway agenCies should evaluate less costly safety measures such as widening lanes, widening or paving shoulders, flattening steep sideslopes, removing or relocating road- side obsøcles, and installing traffic control devices. Vertical Curvature and Stopping Sight Distarnce Reconstruction of vertical curves at hill crests to increase stopping sight distance may be safety cost+ffective at average daily Eafflc volumes grcater than 1,500 vehicles per day depending on site conditions' Generally, to be safety cost-effective, vertical curve reconstruction must correct a substantial sight distance restriction that affects drivers' ability to anticipate a hazardous situation-{uming vehicles, sharp curves, or other conditions that demand speciñc driver responses. Highway agencies should evaluate the reconstruction of hill crests when (a) the hill crest hides from view major hazards such as intersections, sharp horizontal curves, or narrow bridges; (b) the average daily traffic is greaær than 1,500 vehicles per day; and (c) the design speed of the hill crest (based on the minimum stopping sight distance provided) is more than 20 mph below the running speeds of vehicles on the crest. Whether or not an evaluation of hill crest reconstruction is required, designers should examine the natûe of potential hazards hidden by a hill crest and consider other options such as removing the hazards or providing warning signs.

UECWNE SUMMARY Bridge Wídth The safety cost-effectiveness of bridge widttr improvements depends on the usable width of the bridge, the width of approach lanes, fraffic volumes, and the length of the bridge. Highway agencies should evaluafe bridge replace- ment or widening in siruations in which bridge width improv"."nL might bejustified on the basis of safety cost-effectiveness--åridges less than 100 ft long with usable widths less than the values given in Table ES-3. At low úaffrc volumes the recommended values are similar to those proposed by the 1978 FHWA standards, and at high Fafûc vorumes they are simitar to those specified by the AASHTO policy for bridges to remain in place on arterial highways. TABLE ES-3 usable Bridge widths Below which Bridge Replacement orWidening Should Be Evaluared (If Bridge is Less Than fõO niò"e) ---- -^ Design Year Volume (ADT) Usable Bridge Widrh (fr), 0-7s0 751-2,000 2,00 I -4,000 Over 4,000 Width of approach lanes Width of approach lanes plus 2 ft Width olapproach lanes plus 4 ft Width of approach lanes plus 6 ft alflane wideningis planned as part ofth,e RRR project, the usable bridge width should be compared withthe planned width ofthe approaches after thev are wídened. \vhether or not evaluation of bridge widening is warranted, designers should consider installing transition guardrails at bridge approaches, rehabili- tâted or new bridge rails, and warning signs. Sideslopes and Clear hnes Roadside characteristics are important in determining the overall level of safety provided by a highway. Accident rates are lowerand accidents a¡e less severe on highways wittr gentle sideslopes and few obstacles near the road- way' Despite these findings, the study revealed no basis for nationwide standards addressing either sideslopes or clear zone width. The safety cost- effectiveness of particular roadside improvements appeårs highly dependent on site-specific conditions and interactions between different roadside fea_ tures. The study recommends that highway agencies develop and apply their own procedures for identifying and selecúng sideslope and clear ione width

10 DESIGNINGSAFERROADS improvements on RRR projects. These procedures should encourage the following: o Flatten sidesþes of 3:1 or steeper at locations where run-off-road accidents are likely t0 occur (e.g., on the outside of sharp horizontal curves); o Retain cu.enl slope widths (without steepening sideslopes) when widen- ing lanes and shouldeis unless warranted by special circumstances; and 1 Remoue, relocate, or shield isolaæd roadside obstacles' Pavement Edge Drop and Shoulder Type pavement edge drops (i.e., vertical drops or ruts) often develop between the pavement s*iu.. and adjacent unpaved shoulders or roadsides' These drops ã* pr"n"nt drivers whosã vehicles cross over the lane edge from successfully returning to their original lane without encroaching on an opposing lane or losing control. Reiearch sponsored as pafi of ttris study indicated that pavement edge drop hazards r" gf"ut"t than previously believed. However, no basis exists for estimatng hõw often pavement edge drops contribute to accidents or the cost and safety trade-offs involved in preventing or correcting them' Oepenáing on the type of shoulder construction, resurfacing can increase the likelihooã that edge drops will develop later and require repeated mainte- nance to coÍect. To reduce pavement edge drop hazards on highways with narrow unpaved shoulders, highway agencies should either r Selectively pave shoulders at points where out-of-lane vehicle excursions and pavement edge drop problems are likely to develop (e.g., at horizontai curves); or . C|nsfuct a beveled or tâpered pavement edge shape at these points. Intersections Reliable information about the cost and safery nade-offs of individual inter- section improvements is generally unavailable because of the large number of physical and operational feâtufes affecting intersection safety and because intersection projec6 typically address multiple intersection safety problems simuløneouJly. Nevertheless, many intersection improvements can be made at relatively low cost and are safety cost-effective, particularly at higher rafflc volumes. Nationwide numerical standards are inappropriate for RRR projects. There- fore, designers musi tailor inærsection improvements to site-specific condi-

Ð{ECWNE SUMMARY 11 tions and rely heavily on professional judgment and experience. To faciliøte this, st¿te highway agencies should deveþ criæria for identifying int€rsec- tions that warant careful evaluation and checkliss of improvements to be considered. Other Design Procedures and Assumptions Different highway agencies may design RRR projects differently even when their minimum RRR standards a¡e the same and project conditions are prac- tically identical. Four procedures are recommended that will encourage a more uniform application of RRR standards and a more consistent approach to safety. c Design trafic volume: The design naffrc volume for a given highway feature should match the average naffic anticipated over the expected perfor- mance period of ttrat feature. Most state highway agencies use current-year traffic even though the expected performance period of the pavement rehabiliøtion work is 5 to 15 years and the performance period for geometric improvements may exceed 25 years. . Speed: When determining whether geomeFic improvements are war- ranted for featu¡es for which vehicle speed is a factor, highway agencies should estimate actual running speeds in a manner appropriate for the feature under consideration. For example, for horizontal curves, designers should use the 85th percentile speed of vehicles approaching the curve, estimated at a point where drivers have not yet reduced speed. o Design values: When selecting design values for geometric improve- ments, highway agencies should estimate the incremental safety cost-effec- tiveness of improvements that exceed the minimum søndard and should consider overall highway geometry, design of adjacent sections, and expected trends in trafflc growth and truck use. Improvements beyond the RRR mini- mum standards may not be cost-effective and may create inconsistencies between the level of safety provided by the features improved on the RRR project and the features unimproved. c Design exceptions:Highway agency requests for an exception to a design stândard should explicitly address the expected safety consequences, along with cost and other impacts. The ciæd justifications for exceptions are often imprecise and vary from state to state. Planning and Programming RRR Projects Highway agencies select RRR projects primarily on the basis of pavement repair needs and seldom consider safety needs until preliminary design begins.

12 DESIGNING SAFER ROADS As a result, most highway agencies have not deærmined where geometric improvements to existing highways would have ttre greatest safety payoffs and be the most safety cost-effective. Moreover, when additional right-of-way is needed for a RRR geometric improvement, highway agencies must often delay the project or neglect the improvement. Given current budget levels and existing highway conditions, pavement repair needs will continue to be the dominant factor in the selection and scheduling of RRR projects. Nevertheless, highway agencies should begin to take safety into account earlier in the overall RRR process. . Systemwide safety planning: Highway agencies should periodically assess the systemwide potential for improving safety through upgraded design to help guide project programming and design practices. o Expedite right-of-way acquisitions: Highway agencies should review highways programmed for RRR projects to identify locations where design and right-of-way acquisition should be expedited' Safety Research and Ttaining Despite more than one-half centuy of modern road building, knowledge of the safeti consequences of highway design decisions is limited. Except for a modest FF{-A research program and occasional resea¡ch studies sponsored by the National Cooperative Highway Research Program, few opportunities exist for coordinated, purposeful safety research aimed at improving this knowledge. To a large extent, the highway community has relied on uncoordi- nated research without rigorous stâtistical controls to expand knowledge about the safety effecs of highway design. Conflicting research findings are often left unresolved. Equally serious, current knowledge about the safety effects of highway design is inconsistently applied. Although designers can rely on stândards and design aids in many instances, some decisions must be based on site-specific ci¡cumstances and judgment. Designers often lack the capability or time to apply the existing safety knowledge. Additional resources must. be devoted to safety research, training, and design to expand knowledge about the relaúonships between safety and highway design and take steps to assure that this knowledge is properly applied. The study committee believes that the payoff in long-term highway safety gains will be worth the added cost. The recommendations that follow offer the flrst. step toward meeting long- term ¡esearch and training needs and also suggest steps that can be taken now to improve research and raining.

Ð(rcWME SAMMARY 13 o special s$ety taskforce.' congress should di¡ect the secretary of Trans- porøtion to establish a special task force to assess highway safety engineering needs and to esablish research, education, and funding priorities to meet these needs. o Safety compendium; The Federal Highway Administration should develop, distribute, and periodically update a compendium that reports the most probable safety effects of improvements to key highway design featu¡es and identiûes the principal gaps in curent knowledge. o Increased reseørch: The Federal Highway Administration and the National cooperative Highway Research program should increase research on the relationships between safety and highway design. c Training: The Federal Highway Administration, tle American Associa- tion of state Highway and rransporøtion officials, state and local highway agencies, and other organizations of public works professionals should suppoft continuing training activities to keep design engineers abreast of safety-conscious design.

Geometric Design Standards for Resurfacing, Restoration, and Rehabilitation Projects : Background and Issues INTRODUCTION Since 1976 when the U.S. Congress first authorized federal aid for resurfacing, restoration, and rehabilitation (RRR) work, questions about geomeEic design standards have persisted: r What improvements !o existing highways yield the greatest safety gains in relation to cost?¡ How can the overall process of selecting, designing, and constructing RRR projects take advantage of these opportunities for safety improvements? o How much federal aid should be used for resurfacing and other pavement repairs that preserve and extend the service life of existing highways? Many state highway organizations have viewed the federal RRR program primarily as a means of addressing critical pavement repair needs. During the 1970s these needs mounted as construction costs escalated and state highway revenues leveled off or declined. As a result, flexible geometric design standards were preferred for RRR projects because stringent standards such as those required for new construction or full reconsfuction would, if followed rigorously, dramatically increase project costs. Stâte highway organizations believed that such increases would inevitably lead to the concentration of t4

BACKGROAND AND TSSUES 15 available funds on a small number of projects, leaving unattended many federal-aid highways in need of pavement repair and meeting neither preser- vation nor safety objectives. Safety organizations, on the other hand, have viewed ttre federal RRR program as an opportunity to make long-needed safety improvements to older highways at the same time as pavement repairs are made. These organizations have viewed the flexible RRR søndards proposed by some highway agencies as too lenient and have favored a more rigorous, safety-orienæd design process. RRR projects may include resurfacing, pavement structural and joint repairs, widening of lanes and shoulders, selected alterations to vertical and horizontal alignment, bridge repairs, and removal of roadside hazards. The federal government considers more extensive improvements reconsEuction; lesser repairs are viewed as maintenance and therefore ineligible for federal aid. Nearly $5 billion, about 10 percent of highway expenditures by all levels of government, are spent annually on non-Interstate RRR repairs, a share that will likely increase as highway programs continue to shift emphasis from construction to preservation. Federal-aid highways account for approximately one-half of the $5 billion spent. From a safety standpoint, about 30,000 persons are killed each year on non-Interstate federal-aid highways, which amounts to two-thirds of all u.s. traffrc fatalities, ând almost 2 million persons are injured each year. until 1976, state and local govemments were responsible for undertaking and financing RRR work, and they made their own judgments about the priorities of pavement repair and geometric improvements. Reacting to wide- spread concem over the deteriorating condition of the nation's highway system in1976, congress authorized the use of federal-aid construction funds for RRR projects. Because no other standards existed, the Federal Highway Adminisration (F[IWA), which oversees the federal highway program, ini- tially applied new construction standards, expecting to later adopt special RRR standards for older nonfreeway highways. Design exceptions were permitæd on a case-by-case basis to accommodate difficult situations such as widening roads in urban areas or straightening roads in mountainous regions. Because many of the roads involved were built many years ago to different standards, geometric features such as shoulder widths or curve radii a¡e not uniform. Imposing new construction standards on a nationwide basis resulted in a large number of design exceptions in cases in which upgrading to these ståndards would have been exftaordinarily expensive. Because of the divergent views on standards and ttre conEoversy that arose, FHWA never adopted a set of special RRR standards for nationwide use. Insæad it adopted a flexible approach that permits stâtes to develop and apply

16 DESIGNING SAFER ROADS their own RRR standatds subject to federal approval, or to continue to use new construction standards. This action failed to silence the debate over RRR standards as evidenced by the congrcssional mandate for this study, as well as the restated program objectives cont¿ined in the same legislation ". . . to preserve and extend the service life of highways and enhance highway safety." The background and origins of this conEoversy are discussed in gÌeater deøil in the remainder of this chapter. Fi¡st, the evolution of federal involve- ment in funding highway construction is described, and the resulting federal- aid highway program is discusseci. Discusse<i next is the division oi respon- sibility between federal and state govenìments for setting highway design ståndards and the specific federal rulemaking efforts directed toward mini- mum geomefric standards for RRR projects. Finally, a number of key issues related to the development and application of nationwide standards for RRR projects are summarized. In Chapær 2 the following procedures are discussed: use of federal aid for RRR work by state and local highway agencies, project selection, type of projects undertaken, design standa¡ds and practices used, and the overall role of safety. Identified in Chapters 3 and 4 are the relationships between key geometric features and safety and those between key geometric features and cost. These relationships are identiûed as a preliminary step to ttre evaluation of the safety cost-effectiveness of geometric design standards. Wherever reliable quantitative relationships are identified, the safety and cost trade-offs of alternative stândârds are evaluated in Chapter 5, with an examination of project-level effects, as well as the implicadons for systemwide effects on safety, highway condition, and funding. The tort liability implications of geometric design standards and other RRR design practices are examined in Chapter 6. Finally, the study committee's key flndings and recommendations for improved highway design practices are given in Chapær 7. EVOLUTION OF FEDERAL HIGFTWAY POLICY Most highways in the United States a¡e constructed, administered, and main- tained by state and local governments. These tasks were accomplished with little federal assistance or involvement until the passage of the Federâl-Aid Road Act of 1916. This act provided substantial financial assistance for highways, and in doing so, eståblished the following key principles of federal highway policy: o State highway agencies: specifled that state highway departments would be the usual coordinator and contåct point for all federal assistance;

BACKGROUND AND ISS¿/AS 17 o Fed¿ral-state relationship; afflrmed the responsibility of state and local governments to construct, own, and maintain highways while committing the federal govemment to share in the financing of highway construction; o Fed¿ral-øid apportionmenr: prescribed distribuûon of federal highway funds through a formula, which initially considered area, population, and rural postal route mileage; o Project cost sharing: on federal-aid projecc, required that federal funds be matched with state (or local) funds initially, with a maximum federal share of 50 percent; . Federal oversight: specified that the federal government, originally through the Secreøry of Agriculture, approve the plans, specifications, and estimates used for federal-aid highway projects l/-3). In succeeding legislation over the next 40 years, Congress continued to shape the federal-aid highway program, adding other major principles of federal policy. o Designatedfederal-aid highways: In 192I Congress directed that federal aid be limited to a designated system of interconnected highways. The system designated at the time was the forerunner of today's federal-aid primary system. Late¡, other legislation established additional designated federal-aid systems-the secondary system and urban primary extensions in 1934, the Interstate system tn 1947 , and the urban system in 1970. . User fee rtnancing: Inl932 the fi¡st federal excise tax on motor fuels was enacted. Although [ax revenues were directed into ttre general fund, Congress in its deliberations began to link ttre tax rate to the level of federal highway expenditures. . Special categorical programs: Through various highway and economic recovery legislation in the 1930s, Congress began lo supplement federal aid for highway construction (regular federal aid) wirh categorical funds ea¡- marked for speciûc pu{poses. The initial progÌams focused on eliminating the problems of railroad at-grade crossings, a federally funded activity that con- tinues today. By 1956 the basic principles of the federal-aid highway progmm were in place, but widespread concern existed over adequacy of U.S. highways, particularly in view of d¡amatic postwar increases in automobile and truck traffic. Through landmark legislaúon (the Federal-Aid and Highway Revenue Acts of 1956), the federal government responded to this concern, providing new funding to accelerate construction of the Interst¿te system and increased funding for otler federal-aid systems. To finance this greatly expanded federal-aid program, the federal government increased existing highway-

18 DESIGNING SAFER ROADS related excise taxes and established additional taxes whose revenues were totally or partly funneled into the Highway Trust Fund. (The rust fund was established as a holding mechanism for tax revenues earmarked for highway purposes.) The legislation required that the trust fund be used on a pay-as- you-go basis---cffectively prohibiting federal deficit spending for highway construction. From the outset financial assistance for the federal highway progËm (except for special categorical programs) was conûned to relatively large, nonroutine projecs-projects refened to in the Federal-Aid Road Act of 1916 as "substantial in character." In practice, this meant that federal aid was available only for constructing new roads or fully reconst¡ucting existing roads to higher design standards. l¡sser improvements to existing roads, including costly resurfacing or minor widening of roads constructed earlier with federal aid, were considered maintenance, and therefore were ineligible for federal aid until 1976. Congress changed the distinction between construction and maintenance when it enacted the Federal-Aid Highway Act of 1976.ln response to mount- ing concern over fhe condition of the nation's highway system and the need to shift emphasis from consructing new facilities to preserving existing ones, this act amended the U.S. Code to include resurfacing, restorâtion, and rehabilitation within the deûnition of "construction" as the term was used in the fèderal-aid highway progam. Tt.e 1976 act authorized, for the first time, the use of federal funds for major repair work on the federal-aid highway system. The act required that at least 20 percent of the regular federal aid for the primary and secondary systems be spent on RRR work. I¿ter in the Surface Transporøtion Assistance Act of 1982, Congress modified this provision so that at least 40 percent of the primary, secondary, and urban system funds could be used on the combination of RRR work and reconstruction. FEDERAL.AID HIGHWAY PROGRAM The federal-aid highway progarn focuses on a designated system of highways and provides financial aid earmarked for particular components of this sysæm or specific types of highway improvements. Congress authorized federal aid for RRR work within the context of the existing system of federal-aid highways and established funding categories. Federal-Aid System The designated federal-aid highway system accounts fot 22 prcent of the nation's total highway mileage but carries 81 percent of all traffic and accounts

BACKGROUNDIND/SSUES 19 for 77 percent of all traffic fat¿lities (Table 1-1). It contains high proportions of the more heavily traveled expressways and arterial highways and few local roads and streets. For example, about 95 percenf of all rural arterial and major collector highways are included in the federal-aid system. TABLE l-l U.S. Highway Fatalities, Road Mileage, and Tiavel, 1985 1985 Fatalities Mileage(thousands) (thousands) Vehicle-Miles Tiaveled (billions) Federal-aid systems Interstate Primary Secondary Urban Subtotal Off-federal systemsa Total Number Percent Miles Percent Miles Percent 44 257 398 144 843 3,0 l9 3,862 Souncr: FHWA,, Highway Statistics 1985 (4) and Highway Safety Perþrmance, 1985 (5). alncludes oñsystem local travel. The federal-aid highway system comprises fow components: . Interstate system: Consists of about 44,000 mi of multilane expressways wittr fully contolled access traffic. The Interstate system is part of the federal- aid primary system but receives separate funding. c Primary system: Excluding Interstates, consists of 257,000 mi of high- ways, of which 198,000 (77 percent) are two-lane ru¡al roads. c Secondary system: Initially designated in1944, consists of 398,000 mi of highways classified as major collector roads in rural areas, nearly all of which (99 percen$ are two-lane roads. . Urban system: Includes 144,000 mi of urban arterial and collector roads, about 75 percent of which are two-lane roads. This snrdy is concerned with improvements to all systems except the Interstate, which in total account for about 95 percent of all federal-aid mileage, 60 percent of the nation's highway traffic, and two-thirds of its trafûc fatalities. Two-lane rural highways alone account for about 75 percent of all nonfreeway federal-aid mileage and serve approximately one-fourth of vehi- cle miles taveled in the United Søtes l4). The Federal Highway Adminisradon maintains the Highway performance Monitoring system (IIPMS) that keeps track of the design characteristics and 4.2 14.2 6.5 8.9 33.8 10.0 43.8 370 519 156 383 t,428 347 t,77 5 2t 29 9 22 8l 19 100 1 7 t0 4 22 78 r00 l0 3L t5 20 77 23 100

20 DESIGNING SAFER ROADS conditions of the U.S. highway system, including the federal-aid systems, based on a nationwide sample of highway segments. using this system, FTIWA fó) reported to Congess that approximately one-half of the federal-aid primary and secondary mileage has some type of significant geometric design defrciency relaæd to either cross section (e.g., lane and shor¡lder widths) or alignment (e.g., curves, grades, and sight disønce) (Table 1-2). About 16 porcent of the mileage on these systems contains pavement deficiencies. In estimating these deficiencies, FIIWA assumed "minimum Ûolerable condi- tions" that varied by functional classification, terrain, and traffic level. TABLE l-2 Estimated Defrciencies in Federal-Aid Systems lól Federal-Aid System (Percentage of Paved Miles) lnterstate Primary Secondary Urban Pavemento Horizontal and vertical alignmentb Cross section' Operational'/ Access control" Miles with single defrciency Miles with more than one deficiency Total miles with defìciency t0.4 t.2 4.6 I r.0 0.1 17.8 16.1 t4.4 45.9 19.2 0.1 41.5 23.9 6 5.5 16.2 39.6 38.5 3.1 40. I 26.5 66.6 r 5.0 iq.z t4.3 0 40.'1 4.5 22.3 tt.2 51.9 NorE: Estimates based on HPMS data for 1981. dpavement condition is based on the Present Serviceability Rating (PSR). A pavement with a PSR of2.0 or lower (2.5 lor lnterstates) is considered deficient. ,Highway segments with curves that require reduced speed, or any grades with insufficient sight distance thairequire irucks to slow on arterials and major collectors, or more than an occasional such curve or grade on minor collectors. Zlane widths less than I I ft on principal arterials, l0 lt on minor arterials and major collectors, or 8 to 9 lt on minor collectors; or shoulder widths on arterials and major collectors ofless than 4 to 8 ft, depending on terrain and traffic. lOperational defìciencies in rural areas occur when operating speed drops below a threshold that is a function ofclassifrcation, average daily traffic (ADT), and terrain. In urban areas the defrnition is based on the peak period volume-to-capacity ratio. ,'A defiiiency results when a segment olhighway that should be access controlled is notl this applies to all Interslates and certain primary and urban routes. Comparatively, the Interst¿te system contains fewer defrciencies owing to its more recent design and construction, aS well as the higher maintenance priority it generally receives. The FFIWA estimated that less than 10 percent of Interståte mileage contains geometric defrciencies and approximately 10 per- cent of its pavement is deûcient. The federal-aid systems include approximately 267,000 bridges, which account for about one-half of all U.S. bridges. The FTIWA (7) estimaæs that

BACKGROANDANDTSSUES 2T 23 percent of these (more than 60,000 bridges) are eligible for special federal bridge replacement and rehabilitation funds because of existing deflciencies in design and condition. About 80 percent of these bridges are on either the primary or secondary federal-aid system. Funding Programs The overall federal-aid highway programt consists of more than 40 separate funding assistance programs (8). However, t¡e construction programs for the designated federal-aid systems (regular federal aid plus minimum state alloca- tions) account for about 80 percent of all federal assistance (Table 1-3). In general, these funds are apportioned to the søtes by formulas that vary depending on the system. Although most highway programs have some direct or indirect effect on safet¡ over the years Congress has established tttree programs that fund safety TABLE 1-3 Expenditure of Federal-Aid Highway Funds Administered by the Federal Highway Administration During 1985 Expenditures paid from the Highway Tiust Fund($millions) (percent) Primary Secondary Urban Interstate Interstate resurfacing 85 percent minimum allocation funds Planning and research Highway salety Bridge replacement Other fbtal ) )7q 17.4590 4.6833 6.53,923 30.4 2,322 316 r 8.0 2.5202 1.6423 3.3t,5u It.j 4.1i2,884 100.0 Souncr: H i g hw ay St at i s t ic s 19 I 5 ( 4 ), -fab\e FA-3 lïhis section contains a description of federal funding programs as authorized through fiscal year 1986. Shortly beforc this writing (in April 1987), the Surface Transportation and Uniform Relocation Assistance Act of 1987 became law. This act authorizes expenditures qr federal-aid highway projects and safety programs for fiscal yean 1987 through 1991. Although the act provides some changes in funding priorities, it do€s not sub'stantially change those elements of the federal-aid highway program that arc of interest in this study. 525

22 DESIGNING SAFERROADS improvements at specific locations, often involving construction or rehabilita- tion: o Bridge replacement and rehabilitation: Since 1979 this program has funded bridge replacement and rehabiliøtion projects on and off the federal- aid system. Before the 1978 Highway Act, bridges on fhe federal-aid system could be partially replaced (e.g., new deck) under the special bridge replace- ment pro$am. t Rail-highway grade crossings; Established by the Highway Safety Act of r^ãô .t,t_ ---_, tl__ î-__l_ 1^- -^f-L_ :--,-_ --l -cclylJ, trus prograln Pruvrucs lul¡us rur sargty u¡lpruvçtrrçltLs aL ull- al¡u ulr- system rail crossings, including grade separation, relocations, automatic gates, and warning devices. . Høzard elimination: Since 1974 this program and its predecessors have funded on- and off-system spot safety improvements, frequently at locations with histories of high accident rates. Through flscal year 1982, program funds were directed to improvements such as Eaffic signals and intersection chan- nelization (34 percent), guardrail installation (15 percent), signs and break- away supports (2 percent), and pavement skid treatments (4 percent) (9). Just as it has done with regular federal-aid funds, Congress has authorized some transferability benveen the bridge replacement and rehabiliøtion, rail- highway grade crossings, and hazañ elimination caægorical progÍrms. In recent years federal aid from all programs has accounted for approx- imately 25 to 30 percent of ¡otal highway disbursements ($54 billion in 1985) at all levels of government. During the years when major Interstate con- struction was underwa¡ capital outlays, which include most federal aid, represented more than one-half of total disbursements; but in recent yeârs, capital outlays have shrunk to less than one-half of the total while mainte- nance and operational disbursements have increased as a percentåge. More- over, the buying power of capital funds decreased dramatically during periods of rapid inflation in the 1970s so that, despite recent improvements in con- strucfion buying power, the F[{WA 15) reported t}rat 1982 capiøl outlays were expected to be less in constant dollars than any year since the early 1950s. This trend was reversed, at least temporarily, by the Surface Transportation Assis- t¿nce Act of 1982, which increased 1983 federal-aid highway system authoriz- ation levels about 40 percent over the 1982 level. Further, it authorized annual increases in the Interstate resurfacing, restoration, rehabilitation, and recon- struction (4R) and primary system categories each year between 1983 and 1986. Fiscal year 1986 Interstate 4R authorizations were nearly four times the 1982 level, and primary authorizations were nearly fwice the 1982 amount.

BACKGROUND.A.ND/SSUES 23 Resurfacing, Restoration, and Rehabilitation Under the classiflcation of resurfacing, restoration, and rehabilitation, the federal-aid highway program funds the following types of improvements to existing federal-aid highways: resurfacing, pavement structural and joint repair, minor lane and shoulder widening, minor alterations to vertical gades and horizontal curves, bridge repair, and removal or protection of roadside obstacles. h making RRR work eligible for regular federal aid, Congress made federal funds available for the heavier, more costly t¡Des of mainte- nance and at fhe same time provided highway agencies the opportunity to use federal funds for incremental geometric and safety improvements short of full reconstruction. RRR is not a separate federal-aid pro$am but rather a collec- tion of improvement types that became eligible for regular federal aid in 1976. Cunently, in distinguishing between RRR work and ineligible routine mainte- nance, FIIWA classifies as maintenance any overlays less than 3/+ in., pave- ment repairs on short segments, and patching and repair of minor pavement failures (10). Between 1977 and 1982, when federal law required at least 20 percent of regular primary and secondary funds be used for RRR, FHWA also had to distinguish between RRR work and reconstruction so that compliance could be verified. In general, FHWA views the complete removal and replacement of pavement structure or the addition of new continuous traffic lanes as reconstruction, rather than RRR (11). After 1983, when the law changed to require at leasf 40 percent of non-Interstate regular federal aid be used for the combination of RRR and reconstruction, the distinction befween the two became less important because the treatment was identical in the revised program. Not all federal-aid RRR work is funded with regular federal-aid con- struction funds. Some RRR-type projects may be funded with either regular federal aid or separate categorical aid. For example, state or local highway agencies might seek federal funds for a bridge rehabiliøtion project through either regular federal aid (as an eligible RRR project) or through the bridge replacement and rehabiliøtion program. Similarly, roadside hazard removal and guardrail installation might, in some circumstances, be funded with either regular federal aid or hazard elimination program funds. State highway agencies spent $2.9 billion on RRR projects in 1985, includ- ing $1.7 billion federal-aid and state matching funds and $1.2 billion state funds (4). Counties and municipalities spent approximaæly $2.1 billion on

24 DESIGNINGSAFERROADS RRR projects.z Total spending on RRR projects in 1985 was $5.0 billion, about 10 percent of expenditures for highways by all levels of government combined. Mâny state highway agencies use RRR federal aid as a means of addressing critical pavement preservation and repair needs while making selective improvements to road geometry and roadside features. For example, a state mignt use federal aid for a resurfacing project that would also widen the shoulder and replace obsolete guardrail. The extent of such geometric and roadside improvements varies among the states and is at the center of the controversy that ius existed reganiing geomeiric design Starrdãids for federal- aid RRR projects. GEOMETRIC DESIGN STANDARDS AND F'EDERAL RULEMAKING Responsibility for Standards-Federal and State Roles Historicall¡ the primary responsibility for developing and adopting design standards has rested with st¿te legislatures, søte highway agencies, and local governments. TÏese standards generally set minimum values for key geo- metric features and sometimes call for specific design procedures or practices as well. over the years highway design ståndafds have been developed principally for the construction of new highways or the complete reconstruc- tion of existing highways. v/ith the passage of the Federal-Aid Road Act of 1916, the federal govern- ment became involved with highway design standards for the first time. Initially, it exercised oversight by approving designs for federal-aid projects on a project-by-project basis. It did not develop or seek to apply nationwide standa¡ds but did begin sponsoring early research on geometric design, the results of which were later incorporated in state standards and American Association of State Highway and Transportation Officials (AASIITO) design policies. For example, n 1925 the Bu¡eau of Public Roads, predecessor of FFIWA, reported that 18 ft was the minimum pavement width for trucks and automobiles to pass safely; later in 1944 a federal study recommended a lane widthof12ft(1). 2RRR "ccot¡nted for 47 and 26 percent, respectivel¡ of capiøl outlay for søte-funded projects on mral and urban collectors. These percentages were applied to toøl capital outlays by counties and municipalities ($2.4 and $3.6 billion, respectively) to estimate their RRR expenditures, yielding $1.1 billicn for cormties and $1.0 billion for municipalities.

BACKGROUND AND ISSUES 25 Until the late 1920s, state agencies generally adopted ståndards indepen- dently of one another, leading to design inconsistencies between adjacent states as well as duplications of effort. To add¡ess these problems, the American Association of State Highway Officials (AASHO) (now the Ameri- can Association of State Highway and Transportation Officials), which ini- tially confined itself to disseminating information, began to adopt design policies in1928 (12).Thesr- policies, though not obligatory, were intended by AASHO to guide its members on technical matters in which state-to-stâte uniformity was needed. By 1944 it had adopæd seven design policies that were incorporated into the design standards of many states (13). Over the years, AASIITO has revised these design policies a number of times and issued many additional policies (14) thar not only recommend minimum design values but also procedures to be used in the planning and design of highways. For example, AASIITO has recommended procedures for citaen participation, envi¡onmental studies, and project evaluation (15) and has issued guidance maferials for designing pavements (16), úaffrc barriers (17), and lighting (18). The federal government adopæd many of AASITTO's design policies and guides as stândards for federal-aid projects; approximately 20 ne incorpo. rated into federal regulations of highway design û9). Although FFIWA has supplemented AASHTO design policies in some areas, it has generally pre- ferred to adopt them rather than develop design standa¡ds independently and has participaæd in AASHTO committees responsible for developing design policies. In practice, each state highway agency continues to incorporate AASHTO policies in its design standards, often with some alterations or extensions. FHWA reviews and must approve these standards for use on federal-aid projects. For road geometry, AASI{TO policies recommend minimum (or max- imum) design values for features such as ¡ Lane widths, o Shoulder widths, r Horizontal and vertical curves, r Superelevation at curves, . Sight dis[ance, . Bridge widths, o Sideslopes and ditch drainage, and r Pavement cross slopes. To increase flexibility and adaptability to a variefy of nationwide condi- tions, AASHTO recommends different design values for variations in terrain, setting (urban versus rural), traffic volume, traffic characteristics (e.g., per-

26 DFSIGNING SAFERROADS centage of heavy rucks), and function (local, collector, arûerial, etc.). The recommended values also vary with speed. For roads intended for high-speed driving, the values specify wider lanes and shoulden, longer sight disances, and more gentle curves. AASIIIO policies recommend design speeds based on function, setting, tenain, and traffrc characteristics. Not all highway features are Eeated with firm, numerical design standards. Roadside featu¡es such as obstacle-free clea¡ zones and protective ba¡riers generally have been covered by guidelines rather than recommended mini- mum design values. As a result, new construction design is more uniform on the roadway than on the adjacent roadside. Geometric Standards for RRR Projects Geometric standards for RRR projects specify whether particular geometric feahues must be upgraded as part of the project. Features that do not meet minimum søndards must be upgraded unless a design exception is sought and approved. When Congress authorized federal aid for RRR projects inI976, AASHT0 had not developed design policies specifically for RRR work, nor had FTIWA adopted minimum geometric standards for RRR projects. Existing stândårds and policies were geared to new construction or reconstruction. As an interim measure, the FTIWA applied its new construction standârds to RRR projects while it considered separate RRR standards for nonfreeway highways. For freeways, mostþ Inters[ate highways constructed in the past 30 years, FHWA concluded that no special RRR standards were v/arranted because these highways generally met the most stringent new construction standards. Design exceptions, permitted on a case-by-case basis for any federal-aid project quickly became commonplace for nonfreeway, federal-aid RRR pro- jects because upgrading to the geometric srandards for new highways is often extraordinarily expensive. In the northeastern ståtes, where highway systems are relatively old and the topography is severe, the FTIWA reported that 75 to 90 percent of RRR projects were granted design exceptions. In those middle and far west states where highways are newer and the topography is relatively flat, the percentåge of RRR projects granæd design exceptions was far lower (0 to 30 percent) (20).Differences in highway widths contribute to the higher percentåge of projecß in the northeast gËnted design exceptions. For exam- ple, in all but one strate west of the Mississippi more than 60 percent of the primary highways now have lanes 12 ft wide, whereas few primary highways in northeastern states have 60 percent of primary highways with lanes of this width (4). FHWA noted that not all of the differences among states in the percentåges of RRR projects granted design exceptions can be explained by differences in

BACKGROUND AND ISSAES 27 topography and the age of highway systems (20). As discussed in more deøil in Chapter 2, other factors that affect these percentages include the amount of state funds available for RRR, the design standards in effect, and the pro- cedures followed by FHWA division offrces in approving design exceptions. By November 1976, AASIIIO had developed and adopæd a policy on geometric design for RRR work, which was published in the RRR geometric design guide the following year (21). Commonly refened to as the "purple book," this guide contains recommended minimum design values for lane and shoulder widths, cross slope.s, superelevation, and bridge widths, as well as advisory information on grades, curvature, sight disønce, and clear zones. Overall, it is considerably less stringent than AASHTO policies for new construction (Appendix A). The purple book was opposed by safety organiza- tions, and within FIIWA it was opposed by the Office of Highway Safety. The FtrwAs inærim measure of requiring new construction standa¡ds for RRR projecs, with lenient exceptions, lasted far longer than expected because selecting separate RRR geometric design standards for nonfreeway highways proved to be complex and controversial. safety and design issues were raised that involved a large number of geometric features affecting different types of highways in rural and urban settings. Underlying the question of standards was the need to use federal RRR funds in the most cost-effective manner ûo enhance safety while preserving and restoring federal-aid highways for the nation as a whole. controversy was inFoduced by the different perspectives of safety organizations and state highway agencies. It was not until June l9S2thatthe FtrwA issued new regulations addressing RRR standards. During the intervening 6 years, the FFIWA considered a number of alternative policies and reversed itself with respect to preferred action. The following alærnatives were considered: l. Continue to use nerv construction geometric design standards with exceptions permitæd on a case-by-case basis. 2. Adopt guidelines conúained in the AASHTO RRR design guide. 3. Adopt RRR søndards developed by the FHWA. In August 1978, after opposition to the AASHTO RRR guidelines arose, the FF{WA proposed RRR standards developed internally (22). In general, the FTIWA standards are somewhat more stringent than the standards in ttre AASHTO RRR guidelines, but are similar in terms of scope and format (Appendix A, Table A-1). 4. Adopt a flexible approach under which states could develop and use tJreir own RRR standards subject to FFIWA approval. This protracted rulemaking process attracted comments from a variety of institutions and individuals, including safety-oriented organizations such as the Center for Auto Safet¡ the Insurance Institute for Highway Safety, and the

28 DESIGNING SAFER ROADS National Transportation Safety Board" Safety organizations generally opposed any regulation that might lead to special standards for RRR projecs and favored the first altemative as least objectionable. Although the FTIWA had granæd a large number of design exceptions under the first alternative, safety organizations believed the process of explicitly considering design exceptions on a project-by-project basis will occasionally result in substantial geometric improvements. Safety organiza- tions acknowledged the need for design exceptions, but they viewed the AASIITO RRR guidelines and FFIWA proposed standards as too lax, permit- rinn tha ÞÞÞ nrn¡nm fn fmrrc almncf exellrsive.lv on roa¡l surface imnrnve-ru¡6 r¡¡v ¡vu\ l,rvõre¡ --r --- ---- ments and discouraging a safety-conscious design process. In addition, it was argued that reductions in standa¡ds for federally assisted RRR projects would violate legislative mandates concerning safety. The fourth altemative, permit- ting states to develop their own ståndards, also was unpopular with safety organizations, which feared that states would choose, and the FFIWA would approve, standards similar to the AASHTO RRR guidelines (23). State highway agencies initially supported the AASHTO RRR guidelines, but later indicated a general willingness to accept the more stringent RRR stândards proposed by the FIIWA. New construction standards rilere viewed by many states as inappropriate for RRR projects. State officials generally believed that the new standards, if followed rigorously, would greatly increase project cosls, thereby concentrating available funds on a small number of improvement projects. Such a policy, it was argued, would leave unattended many miles of federal-aid highways in need of pavement repair and would meet neither safety nor repair objectives. It was also contended that if wide- spread exceptions were permitted, needless adminisEative costs and delays would be inctrred (23). In June 1982 the FHWA selected the fourth approach, permitting states to develop their own RRR ståndffds subject to FTIWA approval. By this time, some stâtes had grown accustomed to using new construction søndards, with case-by-case exceptions, and under the June 1982 rule, ståtes were permitted to continue this practice (9). However, the issue remained unsettled because of congressional reservations concerning this approach. These reservations ini- tially surfaced during the fall of 1981 when FIIWA officials, representatives of safety organizations, state highway offrcials, and others debated RRR issues extensively in hearings held by the House Subcommittee on Investigations and Oversight of the Committee on Public Works and Transportarion (24). Safety Considerations In the congressional hearings on RRR standards, debate focused on (¿) the practical impact of RRR standards on safety and (b) the role of safety in the

BACKGROUND AND TSSAES 29 RRR program. A number of questions were raised with respect to the impact of the sønda¡ds on safety. V/ill accident rates increase if highways with existing geometric deficiencies are resurfaced and no other improvements are made? What changes in accident rates can be expected if different types of geometric improvements are made? What would be the nationwide con- sequences on safety and highway condition of alternative RRR standards when budgetâry resources a¡e ûxed? What would be the biggest safety payoffl Although such questions could not be answered fully, the FIIWA prepared the RRR Technical Analysis report to address them. This analysis, which concluded that standa¡ds less stringent than those for new construction would be appropriate for RRR projecs (1), was crittcizeà by the National Transporøtion Safety Board for methodological shortcominSs (25). Overall, many issues concerning the desired level of safety o be included in the program, and the type of standards and policies needed to balance safety and pavement preservation, were left unresolved during the RRR rulemaking process. With respect to the second issue, the role of safety in the RRR program, the FHWA adopted the position that safety ,¡/as an essential consideration of the RRR program, but secondary to preserving and extending the service life of highways. On the other hand, safety organizations expressed the concern that effofs to upgnde the safety of federal-aid roads were being relaxed in the RRR program. Debaæ over the relative priority of safety and repair arose repeatedly in the testimony during the congressional RRR hearings (2J). Ultimately, this debate led to a provision in the Surface Transportation Assistance Act of 1982, which stated that the objective of the RRR pro$am is ". . . to preserve and extend the service life of highways and enhance highway safety." Congressional deliberations were unclear about how much of a change, if any, was required by this provision (2ó). Subsequently, the FI{WA modified its June 1982 rule on RRR standards to echo this restatement of program objectives. Reflecting the legislative ambiguity, the FFIWA changed the policy statement in the preamble to the rule, but made no changes to the procedures (27). These changes to statutory language and rules have not resolved the problem of the cost-safety trade-off within the RRR program. To address "safety cost-effectiveness," an additional provision was included in the Sur- face Transportation Assist¿nce Act of 1982 that called for the National Academy of Sciences to study the safety cost-effectiveness of highway geo- metric design søndards and recommend minimum stândards for resurfacing, restoration, and rehabilitation projects on existing federal-aid highways, except freeways.

30 DESIGNING SAFER ROADS KEY ISSUES Many questions and issues that bear on minimum RRR sønda¡ds were left unresolved during the RRR rulemaking process and relaæd debate. In organiz- ing a study that would respond to the congressional request, the study identi- fled six key areas of inquiry that add¡ess these issues and, taken together, provide the technical foundâtion necessary for specitc recommendations. State and Local Procedures for Selection, Design' and Construction of Highway fmprovement Projects (Chapter 2) To analyze altemative design standa¡ds, it is necessary úo understand the relationship between standards and other parts of the RRR process: What types of RRR projects are funded with federal aid? How are these projects selected? What design stândards are currently used? Are stringent design stânda¡ds frequently circumvented? How are safety needs taken into account? All of these questions are directed to state and local highway agencies because they have ttre primary responsibility for selecting and performing RRR work. During the congressional hearings on RRR st¿ndards (24), wiøesses relied heavily on either nationwide statistical data or personal observations about specific stâte and local practices. Although useful, nationwide ståtistical datå or studies may mask signiûcant variations among the states with respect úo RRR needs and the ways by which they are met. Personal observations may be accurate but are difficult to compare systematically. Further, they may portray an atypical situation reflecting the practices of a single highway agency. To ob¡ain a more balanced picture, reviews of prior studies and analyses of nationwide data bases were supplemented with in-depth case studies of highway practices in 15 states and interviews with local public works officials throughout the country. Relationship Between Safety and Geometric Design (Chapter 3) What are ttre safety payoffs (i.e., reductions in the number and severity of accidents) from geometric changes such as increasing lane and shoulder widths, straightening sharp curves, or removing roadside obstacles on existing highways? The trade-off befween the cost of such improvements and their safety payoff is fundamental to the issue of minimum RRR design stândards. To make this trade-off requires quantitative knowledge of the relationships between safety and different roadway features. Despite numerous statistical

BACKGROUNDÁilDISSUES 3I studies of accident data, these relationships are not well known, and divergent relationships are suggested by different analyses (28). Isolating the effects of a specifrc geometric feature from other conditions of the roadway environment, vehicle cha¡acteristics, and driver cha¡acæristics has proved to be â formid- able research fask. This task is often complicated by the lack of comprehen- sive and consistent accident and exposure (usage) daø. Highway resea¡chers will probably never develop deflnitive safety relation- ships that cover the full range of highway design features. The complexity of the causes of accidents, the infrequency of accidents, and the continuing evolution of highway vehicles, naffic regulations, and enforcement policies work against this. Nevertheless, wherever possible, the study committee made judgmens about the most probable relationships between safety and key highway features using what was considered to be the most credible data available. Tìo help provide a basis for these judgments, the commiftee spon- sored critical reviews ofprior resea¡ch and ¡vo special studies that addressed major gaps in existing knowledge. Relationship Between Cost and Geometric Design (Chapter 4) Like the relationships between safety and geometric design, the relationships between cost and geometric design a¡e critical in determining how safety and road repair needs can most effectively be balanced. How much does it cost, for example, to widen lanes and shoulders in addition to resurfacing as opposed to simply resurfacing? Nationwide statistics compiled by FFIWA for federal-aid projects indicate that typical resurfacing projects on rural arærials cost approximately $150,000 to $200,000/mi. When minor widening is included, the cost more than doubles. Full reconst¡uction, with wide lanes, costs approximately $1.25 million/mi, more than six times the unit cost of simple resrrfacing (29 ). Although these rough estimates provide a sense of the magnitude of costs involved, they mask the large variability tlnt exists from region úo region, state to stâte, and from project to project. To develop cost relationships for use in the study, the committee examined published cost daúa, cost records, and estimating procedures for a sample of søte highway agencies throughout the country. This work supplemented reviews of existing nationwide cost records and data sets. Safety Cost-Effectiveness of Geometric Design Standards (Chapær 5) The principal questions that underlie ea¡lier debate over RRR st¿ndards are

32 DESIGNING SAFER ROADS . What are the cost and safety ûade-offs of making incremental geometric improvements to existing highways? . How do minimum RRR sønda¡ds affect the balance between preserving highways and improving safety on a systemwide basis, when available funds are limited? These questions were add¡essed from project-level and system-level per- spectives using the safety and cost relationships identiûed. The added cost per accident eliminaæd that can typically be expected for improvements to exist- ing iúghway geometry was estimatecÍ at the project ievei. How much rioes if. cost, for example, to eliminate an accident by lane widening, and how does this compare with shoulder widening or straightening a sharp horizonal curve? Where system data were available for existing highway conditions, on a nationwide basis and for selected súates, the study estimated the effect of alternative RRR minimum standa¡ds on systemwide safety and the total expendinrre needed to meet the sfandards. Also considered was the likely impact that such stândards would have on the frequency of major pavement repairs and operational benefis in the form of reduced user costs that may result from geometric improvements along with improved safety. Tort Liability and Geometric Design (Chapter 6) Highway agencies have become increasingly concerned about the number of tort claims ûled against them and the resulting costs of settlements, awards, and legal defense. These claims allege negligence in rhe design or operation of public highways. Some highway agencies have feared that special geometric design stan- dards for RRR projects, less stringent than new construction standards, might make them more susceptible to tort claims; others have concluded just ttre opposite. The limited data available on tort claims against highway agencies were analyzed to determine how frequently geometric design is at issue as opposed ûo maintenance practices, signing, or other aspects of highway management. In addition to standards, other ways were considered in which the design and construction of RRR projects may reduce a highway agency's exposure to tort claims. REFERENCES L America's Highway's 1776-1976: A History of tle Federal-Aid Program. FFIWA, U.S. Department of Trarsportatioa 1976. 2. P. K. Wheeler. Highway Assistatae Prograrns: A Hßtorical Perspective. Con. gressional Budget Office, Washington, D.C., Jan. 1978.

BACKGROT]NDáNDISSUËS 33 3. Federal-Aid Road Act of July 11, L916, Ch.Z4l,30 Stat. 355. 4. Highway Statístìcs /985. FHWA, U.S. Deparunent of Transportarion, 1986. 5. Híghway Safety Performance-L995, Fatal and Injury Accident Rates on Public Roads in the United S¡ø¿s. FIIWA, U.S. Department of Transporøtion, (forth- coming). 6. Status of the Nation's Highways: Conditiotts ard Performa¡¡c¿' FÍIWA, U'S. Depafment of TransporøtiorL Jrme 1983. 7. Fifth Anwal Report to Congress: Higlway Bridge Replacemøtf and Rehabílita' tion Progran. FHWA, U.S. Department of Transportatior¡ May 1983. 8. Financing Federal-Aíd Highways. FHIVA, U.S. Departnent of Transportation, Sept. 1983. 9. The 1986 Awual Report on Híghway Safety Improvemenl Prograns. FHWA. U.S' Department of Transporøtioa April 1986. L0. Resu$acing, Restoratíon, and Rehabilitation (R-R-R) I4'ort' FTIWA Notice N 5040.19. U.S. Departrnent of Transportation, June 28, 1976. 11. "Design Standards for Highways," final rule. Federal Regìster, Vol. 4?, No. 112, June 10, 1982, pp. 25268-25n5 (FHWA project coding deûnition)' 12. F. W. Cron. "Highway Design for Mo¡or Vehicles-A Historical Review." Part 8: "The Evolution of Highway Standa¡ds." Public Roads, Vol. 40, No. 3, Dec. 1976, pp. 93-100. 13. A Polìcy onGeometric Design of Rural Highways: /9ó5. American Association of State Highway Offrcials, Washingtor¡ D.C., 1966. 14. A Policy on Geometric Desígn of Highways and Streets,1984. American Associa- tion of State Highway and Transportation Officials, IVashington, D.C. 15. A Policy on Design of [Jrban Highways and Arterial Streets, 1973. American Association of State Highway and Transportation Officials, Washingtor¡ D.C. 16. Interim Guide for Design of Pavem¿n! Structutes. American Association of State Highway and Transportation Officials, Washingtor¡ D.C., 1972. 17. Guide for Selecting, htcating, and Designing Trafic Barriers. American Associa- tion of State Highway and Transportation Officials, Washington, D.C., 1977. 18. An Infornøtion Guide for Roadway Lighting. American Association of Søæ Highway and Transportation Ofûcials, Washingtor¡ D.C., 1976. 19. Design Standards for Highways, Code of Federal Regulations, Title 23, Part 625. 20. U.S. Congress. House of Representatives. Testimony of L. l,amm. Subcommittee on Investigation and Oversight, Committee on Public Works a¡rd Transportation, Resurfacing, Restoration, and Rehabilitatíon of Roads Other Than Freeways. Hearings . . . . , 97th Congress, Sept. 17, 1981 (Serial No. 97-75, p.45). 21. Geometric Design Guide for Resurfacing, Restoration, and Rehabílítation (R-R-R) of Highways and Streets. American Association of State Highway and Transporta- tion Officials, Washington, D.C., 1977. 22. "Design Standards for Highways." Notice of proposed rulemaking. Federal Register, Vol.43, No. 164, Aug. 23, 1978, pp. 37556-37568. 23. Unpublished summaries of comments to rulemaking Dockets No. 77-4 and No. 77-10, FFIWA, U.S. Department of Transportation (undated).

34 DESIGNING SAFER ROADS 24. U.S. Congress. House of Representatives. Subcommittee on Investigations and Oversight, Commitæe on Public Works. Reswfacíng, Resforation, and Rehabilita- tion of Roads Other Than Highways, Hearings . . .97fh Congress, Sept. 17, Oct. 27,28; Dec.IS,1981. 25. Federal Highway Adtninistration Non-Inferstate Resurfacing, Restoratìon, and Rehabílitation Program: Safety Effectiven¿ss Evaluation. National Transportation Safety Board, \Yashingøn, D.C. Sepr. 22, 198L. 26. House (p. H10717, December 21,1982) and Senate (p. 516067, December 23, 1982) colloquies to the effect that the "enhance safety" provision would not require the application of full new design standards. ?7 "llpcim Sfqnãa¡¡lc fnr Eliahrwa¡rc-P...'.f"^i-- Dao+nrotiaa ^-á Þ-L^!.ili+^+:^-@ru ¡\w¡¡cu¡u4uvu of Streets and Highways other than Freeways." Final rule. Federal Regísrer, Vol. 48, No. 63, March 31, 1983, pp.1347O-13412. 28. Synthesis of Safety Research Related toTrSfic Cortrol and Roadway Elemenfs. FHWA, U.S. Departrnent of Transportation, Yol. l, Dec. 1982 (see Chapær 4, "Roadway Cross Section and Alignment"). 29. State of the Nation's Highways: Condìtions and Perþrmance. (Table A-4, p. A-15). FIIWA, U.S. Department of Transporrarion, June 1983.

2 State and Local Procedures for Selection, Design, and Construction of Highway Improvement Projects State and local (county and municipal) highway agencies are responsible for constructing and maintaining the nation's federal-aid highways. These agen- cies select and design resurfacing, restoration, and rehabilitation (RRR) proj- ects as part of this responsibility. The federal role is limited to providing funding assistance and exercising oversight authority, largely by setting design standards and guidelines, determining project funding eligibility, and approving individual project designs. Because of this division of respon- sibilities, ttre effectiveness of any federally imposed requirement, such as RRR design st¿ndards, cannot be easily judged or evaluated without also examining parts of the overall process controlled by state and local highway agencies. Nor can changes in federal RRR requirements be reasonably assessed wittrout flrst understanding the context in which they operate and how they interact wittr other steps in the process to promote preservation and enhance highway safety-the statutory objectives of RRR conducted with federal aid. This chapter contains the results of a review by the study committee of resurfacing, restoration, and rehabilitation activities in state highway agencies and local governments. The review was conducted to o Determine the nature of state and local RRR projects-their characteris- tics and objectives in the context of state and local agencies' overall con- struction and maintenance programs; 35

36 DESIGNING SAFER ROADS . Develop an understanding of the process in state and local agencies that determines the characteristics of RRR projects. This process includes not only project design (application of RRR standards, but also programming (select- ing and scheduling RRR projects) and finance (determining the overall level of ¡esources to be devoted to RRR and the sources of funding); and o Focus the study committee's evaluation of RRR design standards and practices on the critical areas of uncertainty or disagreement, and identify possible alternatives to current practices that may be valuable in meeting federal-aid RRR program objectives. The characteristics of RRR projecß were considered relevant to the com- mittee's review because l¿) RRR standards and guidelines must be appropri- ate for the projecs to which they are applied, (b) project characteristics determine which søndards are likely to be binding, and (c) the nature of ttre projects reflects the objectives of state and local highway agencies in the use of federal aid for RRR. The following specific questions about RRR project cha¡acæristics are addressed in this chapter: What are the immediaæ purposes of ståte and local federal-aid RRR projects-ile they mainly resurfacing and other pavement repairs, safety enhancement, a general upgrading of the level of service, or a combination of these? What type of geometric or other safety- motivated improvements are commonly made as part of resurfacing projects? How do federal-aid projects differ from non-federal-aid projects? Project selection, programming, and ûnancing are important because deci- sions in tlpse areas critically affect the success of the federal-aid RRR program in meeting its objectives. In the discussion of RRR programming and ûnance the following questions are addressed: How do state and local agen- cies select RRR projects? What factors (pavement condition, geometric fea- tures, accident history trafûc, or others) do they consider? Do federal RRR ståndards influence project selections? How do financial constraints such as a shortage of state capiøl improvement funds or funding allocation procedures imposed by a søte legislature affect these choices? How does the state choose between federal and state funding for RRR? Design standa¡ds and practices determine the nature and quality of resulting projects. The following questions relating to project design a¡e addressed in this chapter: What design stândards are used? How are safety needs and objectives considered? What is the Federal Highway Administration @tfWA) role in influencing t¡e shtes' design practices? In addition [o RRR practices, the committee also examined the relationship between RRR activities and other highway programs with similar objectives, including the special federal-aid programs supporting bridge replacement and rehabilitation, hazar d elimination, and state-funded pavement maintenance activities.

STATE AIVD I,OCAL PROCEDURES 37 REVIEW OF RRR PRACTICES: INFORMATION SOURCES To answer the preceding questions, fhe committee first reviewed existing studies and daø that address procedures for developing federal-aid RRR projects or describe resulting projects. Relevant previous studies are prin- cipally an FIIWA review of RRR projects in 19 states (l) and two General Accounting Office (GAO) projecs---one directed at the nature of RRR design exceptions granted by FrI'ü/A (2), and another aimed at determining the type of federal-aid RRR and reconstruction work performed (3). The committee also examined nationwide data maintained by F[{WA to record the ståtus, work classiflcation, and basic financial information on all federal-aid projects. Information available from these sources was expanded and updated through case studies of RRR programs in 15 søte highway agencies and interviews with more than 40 local highway ofûcials familiar with RRR activities. The case study states were selected for the diversity of factors that may influence RRR needs and programs: highway age and condition, terrain, mix of urban and rural conditions in the ståte, agency organization, size of federal-aid apportionments, availability of søæ and local funds to supplement federal aid, and RRR design standards used. (The case study states, along with local highway agencies and their RRR activities, are described in Appendix B, Tables B-1-B-15.) The 15 states selected account for nearly 50 percent of federal-aid apportionments and include most of the states with the largest highway programs. Committee søff visited the highway agency and the FI{WA office in each of the 15 states and reported to the commiuee RRR activities and the way these activities fit into the overall søte highway progrrim. Of the 15 state highway agencies visited, 9 were using special RRR design standards approved by FFIWA and 6 had special agreements with FTIWA exempting them from many routine FHWA reviews of project designs (an arrangement known as certification acceptance) (Appendix B, Table B-1). Telephone interviews were conducted with local highway officials repre- senting 16 counties, 20 cities, and 3 metropolitan planning organizations to discuss RRR practices. Most of the ofûcials were selected from jurisdictions in the case study ståtes through state highway agency referrals or independent contacts with state or national associations of local officials. Interviews with state and local highway agency ofñciirls were conducted during 1984. Follow-up interviews conducted in the fall of 1986 indicated that few important changes occured in the ståtes between 1984 and 1986. The conditions that existed in 1984 are reported in this chapær and the øbles in Appendix B; the summaries of RRR standards in Thbles B-9 and B-10 have

38 DEsrcNrNc sAFER RoADS been updated to reflect conditions in 1986. Other changes summarized in ttre introduction to Ap'pendix B. STATE RRR PROGRAMS 1984 are This section contains background information on the organization of state highway agencies and an examination of typical state RRR project cha¡ac- r--io¡i¡¡ -*-^-.-:-- ñ'Mô'l'rÂõ Gnon¡i-n on¡{ ¡locim ctonrforrlc an¡l nro¡-lv¡¡ùuvù, PrvË¡4r¡¡rrr¡rré P¡wvuurvùr ¡¡rrq¡¡wrr¡ér d¡u uvù¡ér¡ tices. State Highway Agency Organization Functions The principal functions of state highway agencies are o Construction; . Project development and design; . lvlaintenance; o Traffic operations; . Programming and budgeúng; . Local government financial assistance; and . Support, including general administration, research and testing, and long- range planning and data collection. Søte highway agencies are responsible for the major intercity routes in the sûate, including federal-aid Interståtes and most federal-aid primaries, but share responsibility for administering federal-aid secondary and urban roads with local governments. The extent of this shared responsibility varies widely, paÍiculary for federal-aid secondary roads (Appendix B, Table B-2). In Missouri the state administers all federal-aid secondary roads; however, in New Jersey the state administers only 5 percent of such roads. Among the case study ståtes, ttre share of toøl federal-aid mileage administered by the state ranges from 94 percent in Missouri to 26 percent in New Jersey. In most of the case study states, cities and counties maintain nearly all local rural roads and city streets, which are ineligible for most federal-aid programs, but a few st¿te highway agencies (Virginia, New Hampshire, Texas, and Missouri) administer substantial non-federal-aid rural mileage. srnce

STATE AND LOCALPROCEDURES 39 O r g anizational Str uc twe In most state highway agencies, responsibilities are divided between a central office and several disrict offices. The cenEal office is organizeÅ into func- tionally defined divisions (e.g., design, construction, maintenance) whereas disrict offices have geographically deflned responsibilities for maintenance and other functions as well. Larger state highway agencies tend to be more decentralized (e.g., New York, Califomia, Texas) with significant respon- sibility for project programming and design handled by the districts and policy and review functions performed by the central offices. Smaller states are more likely to have highly cenralized organizations (e.g., South Dakota, New Hampshire) with the central offlce responsible for all functions except mainte- nance and supervision of construction. Highway agency organizational charts differ from state to state, but there are usually parallel divisions for the major functional areas ofproject develop- ment and design, construction, maintenance and traffic operations, planning, and administrative services (Figure 2-1). Assignment of project programming responsibilities varies most in the organizational structure of highway agencies; ttre programming office may be under planning, project development, or adminisEation, or it may be a sepa- rate division. The programming oftce is responsible for assembling a multi- year construction program (listing projects, costs, and funding sources) based on progam proposals from the functional divisions and the district offices, matching projects to funding sources, ensuring tiat the program meets all departmental and legislative requirements regarding funding allocations to program categories and geographic aleas, and ensuring that all federal-aid funds for which the state is eligible are used. The programming office may also have substantial authority for ûnal project selection and scheduling decisions, especially for smaller projects. Top management also is active in the selection of all major projects and in final approval of the program. In decentralized highway departments, ttre major activity of a central office functional division is to oversee district activities in its area of responsibility. For example, the cenFal office design division will develop ståndards that district designers must follow, approve project designs for the districts, and periodically review district performance. RRR Responsibilities In the past, most state highway agencies performed resurfacing as a part of thei¡ maintenance programs. Historically, maintenance activities were funded almost entirely by the state (without federal assistance), but this situation

FIGURE 2-1 Typical state highway agency organization.

SIATEAAIDLOCALPROCEDURES 4I changed in 1976 when Congress made RRR eligible for federal aid. Some states responded by using federal aid for projecs conducted under their maintenance programs. However, this practice created problems-mainte- nance staffs were not familia¡ with the adminisnative aspects of federal aid, and, moreover, were oriented toward pavement repair projects that did not require extensive design plans or involve the concurrent geometric or roadside improvements that came to be expected in federal-aid RRR. These problems sometimes led to misunderstandings between federal and state personnel, and to delays in project implementation. As a consequence, some ståtes chose to shift some resurfacing projects from maintenance to construction programs and assign RRR project development responsibilities to their design divisions, which were more familiar with federal procedures and better equipped to prepare more complex designs. Currently, resurfacing responsibilities are divided in many states--design divisions develop the plans for federal-aid RRR and sometimes for major state-funded resurfacing projects while maintenance divisions handle most spot pavement repairs, light resurfacing, and other pavement Eeatments that the FTIWA continues to view as maintenance. Central office and district maintenance departments are also responsible in many states for developing the budget and program for both state-funded and federal-aid resurfacing and for conducting pavement monitoring progmms that guide project selection. Safe ty Pro gram Responsibílities All states collect records of accidents on the søte highway system as a requirement for federal-aid eligibility. The data are used (ûo varying degrees among the states) in RRR and reconstruction project design and to identify high-hazard locations for corrective treatment. Also, all states have a special category of safety improvement projects in their construction programs. These projects are usually funded by federal-aid categorical hazard elimination and rail¡oad grade crossing apportionments, but are sometimes funded (e.g., Cal- ifomia, Michigan, Ohio, Washington) with additional federal or state funds. Maintenance funds are also used for minor safety improvements in all søæs. Typically, collection and analysis of accident data are the responsibility of an office in the traffic operations division. In some state highway agencies the functions of this office may be limited to data processing and reporting accident statistics to the design division and the districts (e.g,, Arizona, Florida). In others the safety office may øke a leading role in developing projects and assembling the statewide safety improvement program (e.g., Califomia, Michigan), a funcúon analogous to the maintenance division's role in the resurfacing program.

42 DESIGNING SAFER ROADS In most states, district ftaffrc or maintenance engineers receive reports of high-accident locations from the accident data system and other sources and are expected to investigate the sites and recommend solutions to problems. These investigations lead either to traffic engineering improvements (e.9., signals, signs, pavement marking), no action (if the problem appears unrelaæd to road conditions), or, in a small number of cases, to capital improvement projects. Federal, State, and Local Relationships To oversee the federal-aid program, FHWA operates a division office in each state capital. Because these offices maintain day-to-day contâct with their respective state highway agencies and have first-hand experience with state highway needs and issues, FHWA delegates considerable authority to them. FHWA division administrators approve all federal-aid projects, design stan- dards, and design exceptions. To each division offrce, FIIWA assigns area engineers and technical specialists (e.g., safety engineers) who monitor federal-aid projects in different parts of the staæ. Under an adminisFative anangement known as certification acceptance, FHWA may delegate some of its federal-aid program oversight respon- sibilities to the state agency. Under full certiflcation acceptance procedures, the FTTWA division office's only responsibilities for an individual project are to approve it for inclusion in the state's federal-aid program, approve excep- tions to federal standârds in the design, and conduct a final inspection after the project is completed. The sarc certifles that federal requirements have been met regarding design standa¡ds, public participation, competitive bidding, and other aspects. FHWA conducts periodic audits of the state's performance under certiûcation acceptance. Certification acceptance may be limited to certain phases of project review (e.g., it may cover construction but not design) or to certain categories of projects, for example, federal-aid secondary road projects (in which case the certification acceptånce arrangement is often referred to as a secondary road plan), or local government federal-aid projects. Five of the case study stâtes (Illinois, Missouri, New York, Virginia, and Washington) have certification acceptânce covering design of federal-aid primary projects (other than Inters- tate projects, which are never covered by certif,cation acceptance). Applica- tion for certification acceptance is at the discretion of the st¿te. Although the procedure can reduce federal paperwork and speed up project development, many states have not adopted it because of the increased staff burden it entails. Søte highway agencies a¡e the direct recipients of federal aid, but as noted previously, they do not administer all federal-aid highways; consequently,

STATE AIID I.OCAL PROCEDARES 43 federal funds a¡e passed on to local governments. The state agency usually makes funds for federal secondary and urban highways available o local governments by either a formula distribution, a discretionary process, or a combination of the two. Regardless of how federal funds are passed on to local governments, ståte highway agencies remain accountåble to the FHWA for ttre use of all federal-aid funds. Søte highway agencies act as the liaison on all communications between FTIWA and local govemments and approve project designs (but not design exceptions) before submitting them to FI{WA (or on behalf of F[{WA under certification acceptance agleements). In addi- tion, highway agencies sometimes let construction conEacts and supervise construction for local government federal-aid projects-services that smaller jurisdictions are often not equipped to handle. RRR Project Characteristics To provide an indication of the relative scale of RRR work in state highway programs, differences in spending patterns for federal-aid and non-federal-aid projects, and general types of RRR work undertaken, highway project lists obtained from the case study stat€s were used to classify highway improve- ment projects inûo seven câtegories: new construction urd reconstruction, resurfacing and minor widening (including pavement rehabilitation short of full reconstruction), bridge work, intersection improvements, safety improve- ments, thin overlays and seal coats, and other (Appendix B, Tables B-3 and B-4). l. New construction and reconstruction: construction of a new road; relocating an existing route on a new alignment; major widening (i.e., adding lanes) on an existing road; or reconstruction of an existing route on approx- imately the old alignment where ttre old pavement structure is removed and. replaced; in general, any project intended primarily to increase the traffic- carrying capacity of the highway system. Occasionally a state may undertake a full pavement reconstruction project, primarily because of failure of the old pavement structure, without any substantial realignment or capacity increase, but such projects are r:ìre. 2. Resurfacing ønd minor widening:a project to overlay an existing pave- ment with new material, usually asphalt or concrete, when the overlay is the principal activify of the project lbut excluding thin overlays and seal coats (see deûnition of ttrin overlays and seal coats)l; recycling existing surface material and reapplying the surface; substantial shoulder improvements such as surfacing or regrading; and resurfacing in combination with widening to increase lane or paved shoulder width (but excluding widening to add a new I

44 DESTcNINcsAFERRoADS lane). This category would also include the resúoration and rehabilitation categories of RRR work. The stâtes appear to use these terms synonymously with resurfacing as defined here, and also to describe major concrete pave- ment repairs. hojects may include incidental improvements related to safety or úaffic operations. Pavement preservation and improved ride qualiry are the primary objectives of all projects in this category. 3. Bridge work: a project in which the principal activity is building a new or replacement bridge, bridge rehabilitation (e.g., widening, deck replacement or major repairs for structural soundness); minor construction improvements SuCh as insfellinq railinss anr! decL rehahilitatinn lctnr¡r¡rnl ¡anoirc rn o Írr.i¡rca_---___--Þ ÉÉv¡¡ \ùÉüvíu¡êi iw¡jeüõ ùu é ijiü¡¡;v deck short of replacement). capacity increase, safet¡ or structural preserva- tion may be the primary objective of these projects. 4- Intersection improvementJ.' any construction project in which the pri- mary activity is improvement of the operation of an intersecton. Includes projects ranging from installation of traffic signals to construction of new grade-separated interchanges. other common activities are addition of turning lanes, channelization, and realignment of intersections. The objectives of these projects are to improve capacity or safety. 5. safety improvements; any project in which the principal objective is to correct a specific problem and improve safety. These projects may involve the same types of work as the preceding categories, for example, installing signals at an intersection, upgrading bridge railings, resurfacing to improve skid resistance, or reconstructing a short road segment to straighten a dangerous curve. other common projects are guardrail installation and removal of roadside obstacles. A project is placed in this category if the state identifies it as motivated primarily by safet¡ ratler than by capacity or preservation concerns. In all of the case study states these projects are selected through a special safety improvement programming procedure, usually involving anal- ysis of accident statistics to identify hazardous locations on stât€ roaãs and progmms to install a specific category of safety improvement (e.g., søndard bridge railings) throughout fle state highway system. Because of differences in the ways in which the states reported information, in some instances, a project that could be defined as a safety improvement may have been classi- fred under the functional type of work [e.g., a bridge improvement intended ûo correcf an identified haza¡dous condition may be classified as bridge work rather than a safety improvement (Appendix B, Tables B-3 and 84 )1, but trris misclassification is not frequent enough to substantially affect the percentage disrributions of spending given in the tables. 6. Thin overlays and seql coats: any pavement overlay not substantial enough to meet the FFrwA pavement reatment criterion for work qualifying for federal aid G'HWA requires federal-aid overlays !o be more than % in. thick). Excludes thin overlays that are a stage of a larger pavement project,

STATE Al,lD LOCAL PROCEDURES 45 patching, and overlays on short road segments. These projects have the same general objectives as the heavier resurfacing projects classified in the resurfac- ing and minor widening category-retarding structural deterioration of the road and improving ride quality-and therefore complement the states' major resurfacing activities. However, they have shorter lifespans than heavier overlays, and many ståtes regard them as unsuitable on high-volume roads except as a stop-gap measure. States that have systematic pavement monitoring programs often select thin overlay and seal coat projecs in the same programming procedure that they select heavier ovedays (e.g., Arizona, California, Washington). The states may administer and finance these projects either as capiøl improvements or as part of the maintenance department's operating budget. The work is most commonly performed by conEact, but may be done by súate personnel. Projects usually involve no improvement beyond ttre pavement treâtment. 7. Other: any projects not included in the preceding categories. Typical projects are rest area improvements, lighting, painting, and building con- struction. FTIWA defines RRR to include all "resurfacing and minor widening" projects. In addition, bridge rehabilit¿tion (structural repairs or improvements to a bridge deck other than replacing the deck), a part of the bridge work category and some projects in the safety improvement category (e.g., resur- facing to improve skid resistance, or shoulder improvements for safety pur- poses), also fall under the federal RRR definition (4). The purpose of the FF{WA RRR definition is to identify the types of work eligible for funding from the states' federal-aid primary, secondary, and urban apportionments but ineligible before the 1976 congressional authorization of RRR as a legitimate federal-aid improvement. All projecs in the construction and reconstruction category have been eligible for federal aid since the inception of the federal highway program and are beyond the scope of RRR as defined by FFIWA. Nearly all projects in the intersection improvements category are eligible for federal-aid primary, secondary, or urban funding (regular federal aid) as reconstruction or as raffic operations improvements and therefore are excluded from the federal RRR definition. Thin overlays and seal coats, although they have objecúves similar to RRR project objectives, are excluded from the federal definition and from federal-aid eligibility. Nationwide, about one-tfth of the stâtes' totâl non-Interstate capital expen- ditures is for RRR projects (Table 2-l). The majority of the states' federal-aid RRR projecs are in the resurfacing and minor widening category. In several of the case study states some safety improvement projecrs (e.g., Ohio and Michigan) and bridge rehabilitation projects (e.g., New York, South Dakota) are approved for federal-aid primary funding as RRR projects. However,

lIABLE 2-l State Highway Agency Capital Outlays on Non-Interstate Flderal-Aid Projects, by Type of Improvement and HighwayFunctional Class, 1985 Improvement Type Right-oÊWay New Engineering Construction RÞconstruction lr4ajor Widèning RRR Bridge Sr'ork siafety/Other TotalHighwav $mit- gmr- $-il- s*it.- smit- smir- $imir-¡rrË'wêj Þ rþ S rl- mil- $mil $ il- $ il- fi il- $mil- Functional Class lions Vo lipns olo lions olo lions % lions a/o lions o/0 lions % lions %, Rtral iArterials 537 t6.2 549 16.5 sL4 i5.5 2ûS 6.3 813 24.5 5g5 17.6 tt4 3.4 3,320 100.0Çollecors 136 13.6 105 10.5 t32 13..1 30 3.0 294 29.3 264 26.i 41 4.2 1,OOZ 100.0 UrbanArte¡ials 733 zLI 697 20.1 449. 12.9 2t4 6.2 505 t4.6 550 t 5.9 1,22 9.2 3,470 100.0collectors 45 14.8 43 l4,L 55 18. I 30 9.9 52 t7.1 47 15.5 32 10.5 304 100.0Totar ir4-51 r7.9 T3ç4 ti.2 T;150 t4.2 T{2 6.0 T364 zo.a T746 17.8 i09 6.3 I,Oç¡ 100.0 Sou¡C-e : H ighwa.v. Sta.tistic.s J985. Täble SF-12lr'.

STATE AI'ID TNCAL PROCEDURES 47 nationwide, federal-aid projects whose main objective is safety improvement are usually funded by the federal categorical safety programs, and projects for which bridge work is the main activity are usually funded by the federal categorical bridge replacement and rehabilitation program. On most federal-aid RRR projects resurfacing for pavement preservation is the principal activit¡ but many of these projects involve improvements to bridges and intersections within ttre project limits, roadside safety improve- ments, and even reconstruction of short segments to improve alignment or replace failed pavement. The extent of these ancillary improvements-lane and bridge widening, spot realignments, and roadside safety upgrading-that should be required in the course of a federal-aid resurfacing project, is the major issue in setting RRR geomenic design standards or requirements for safety enhancemenß. C haracteristic s of F ederal-Aid Re surfacing and Minor Widening Proiects The typical federal-aid project in this category involves pavement resurfacing or rehabiliøtion, often with minor cross-secúon or roadside improvements within existing rights-of-way. Lanes and shoulders may be widened, shoulder construction upgraded (e.g., replacing a gravel shoulder with asphalt)' selected roadside obstacles (e.g. culvert headwalls) removed or shielded, and obsolete guardrails removed or replaced. Changes to vertical and horizontal alignment are infrequent. Narrow bridges are occasionally widened or replaced as part of these projects, but the more common treatment is to install approach guardrail and upgrade existing bridge rails. Both FTIWA RRR fleld reviews (2) and the ståte case studies indicate that federal-aid resurfacing projects almost always include one or more geometric or roadside improvements, such as minor widening or removal of ¡oadside obstacles, fhat enhance safety. Exceptions are most likely gentle topography where roadside haza¡ds are infrequent and existing road geometry meets applicable design standards. However, the extent and nature of safety enhan- cements is not consistent from state to state, particularly roadside improve- ments beyond the shoulder edge. For example, policies vary with respect to upgrading guardrail, moving versus protecting culvert headwalls, and relocat- ing utility poles. The case studies revealed that identical situations on RRR projects are treated differently in different states. In addition, resurfacing and other pâvement repair projects do not always take advantage of relatively simple, low-cost opportunities to improve safety. After examining completed RRR projects in 19 states, FHWA engineers reported a number of missed opportunities for such improvemenß even though safety had been enhanced ûo some degree on all projects. Opportunities

48 DESIGNING SAFER ROADS were missed because low+ost and relatively simple safety improvements were not considered during design. These missed opportunities principally involved roadside improvements and safety hardware. Simple improvements often overlooked by the states include o Improving roadside traversability through slope flattening, ditch regrading and relocation, or removal of unnecessary gua¡d¡ail; o Removing or shielding roadside obstacles such as sign supports, utility poles, and rigid mailbox supports; . Upgrading obsoieie guardraü and bridge raü sysrems; o Improving Eaffic operations by cutting trees and brush to restore ade- quate sight distånce and insølling new or upgraded signing and pavement markings (1). Most of the projects examined by FFIWA engineers were designed before Congress enacted the requirement that all federal-aid RRR projects enhance highway safety. Review of the 15 case study stâtes indicated that low-cost safety improvements are receiving greater attention in RRR project design as a result of FTIWA rules implementing the new congressional provision. Most officials in state highway agencies and FFI'WA division offices indicated that roadside safety improvements are more frequently included in RRR projecs now than before 1983. However, substantial variability still exists in state practice. This variability appears related in part to differences among the states in typical existing geometric and roadside conditions (frequent occurrence of roadside obstacles may result in protection or signing rather than more costly removal) and traffic volumes (at very low volumes, higher cost roadside improvements are less likely). However, not all of ttre variability can be explained by these factors; it also arises from ¡ The practice in some state highway agencies of not using federal-aid on RRR projects when its use would require improvements that they would regard as unreasonably expensive; r The severe restriction in some ståtes on the funds available for pavement preservation, which necessitates streæhing resurfacing dollars as far as possi- ble; o Differences in the stringency of FFIWA division office policies regarding requirements for RRR project design and procedures for project review because of minimal guidance (before 1983) from FFIWA headquarters on RRR design policy; and

STATE AND LOCAL PROCEDT]RES 49 o Differences in general philosophy among the state highway agencies regarding the pdority and value of safety improvements, reflected in the states' programming decisions and design procedures. In fhe sections on RRR project programming, ûnance, and design, the influ- ence of the preceding factors on the outcomes of state RRR activities will be described. Chøracteristics of Special S$ety Improvement Projects The primary intent of special safety improvement projects is to correct specific haza¡ds. These projects are usually selected through a special pro- gramming procedure t}tat involves screening state accident records to identify high-hazard locations and also targeting speciûc categories of potential haz- ards (e.g., substandard guardrail) for statewide treatmenl They are usually funded by the federal-aid hazañ elimination and railroad grade crossing progÍìms, but additional federal or state funds are used in some states. The stâtes usually do not regard these projects as RRR, but they a¡e related to RRR because they involve the same types of safety improvements that many states' RRR standards require in the course of federal-aid resurfacing projects. Safety improvement projects are generally spot treatrnents such as installa- tion of new guardrail and traffic conFol devices, pavement marking, and skid- resistant pavement treatments. Occasionally, more costly improvements, such as intersection reconstruction or straightening a hazardous curve are made. Project Programming In directing their highway progtams, state agencies must balance a number of competing objectives, the principal ones being the preservation of roads, improved service levels, and enhancement of safety. Success in meeting these objectives depends on the quality of individual project designs and project programming decisions--+he allocation of resources to categories of con- struction and maintenance activities and the selection and scheduling of projects. Although the states rely heavily on federal aid for financing con- struction, they reøin ttre major responsibility for programming, with the federal role limited to setting funding levels for broad program categories and establishing standards and other requirements governing design and con- struction of individual federal-aid projects. Thus, while the federal govemment may set RRR design standards inænded to encourage preserving existing highways and enhancing highway safety, the

50 DESIGNING SAFER ROADS extent to which these objectives are actually met in a state's highway program as a whole depends on how they are considered in fhe state's programming decisions. As described in this section, the state case studies revealed that RRR programming decisions a¡e influenced by federal RRR søndards, and an undersfanding of these influences is necessary üo evaluate the effect of RRR standards on safety and preservation. D e termini n g Overall P ro gr am O bi e c tiv es Although ttre professional staffs of highway agencies have latitude to select the construction projects that, in theirjudgment, best further the objectives of the stâte highway program, they must also comply with many restrictions, imposed by the state legislatures and tl¡e federal govemment, on how funds may be spent. For instance, legislative designations of state funds for pafticu- lar uses, or legislative requirements for geographic allocations of funds, frequently limit the programming flexibility exercised by the agency's profes- sional staff. Similarly, categorical spending apportionments for federal aid constrain state programming decisions. Therefore, in practice, søte highway agency programming decisions reflect a compromise among the judgments of agency professionals, the state legislature, and the federal govemment as ûo how best ûo meet the priorities of the public at large. Review of staæ highway agencies for this study generally indicated that public perceptions regarding highway needs, as evidenced by letters and conversations, public hearings, and communications with elected officials, have focused on preserving the road system (e.g., pothole repair), capacity and service-level deficiencies (e.g., congestion), and special local concerns (e.g., access to new development). Although these concerns vary from sta¡e !o state, system preservation has consistently been cited as a high priority public concern. This concern has been mirrored by state legislatures that have recently increased state motor fuel taxes to fund infrastructure preservation, or have enacted stâtutes that set spending minimums for preservation work (e.g., Michigan, Mississippi). Safety is also a public concern, but state officials reported that often the public's perception of safety needs does not coincide with objective indicators of safety problems, and public demands tend to focus on a few highly visible problems rather than on the types of modest, low-visibility improvements (e.g., roadside obstacle Featments and lane widening) that enhance system- wide safety. Søte highway agencies reflect local public and legislative concerns in setting priorities for their construction and maintenance programs. The con- cern for preservation is most clearly reflected in the project programming

STATE AND I.OCAL PROCEDURES 51 decisions of the highway agencies in the case studies. preservation-oriented work, from seal coats to reconstruction, is rapidly becoming ttre dominant component in most states' capital improvement programs. Søte highway agencies attempt as best they can to adapt the federal-aid program, with its various restrictions on how funds may be used, to match their own objectives. Federal aid for RRR is no exception; in spite of the program's dual-st¿ted goals of preservation and safety, states have used federal RRR funds primarily to help achieve their road preservation objectives. RRR P ro gramming P rocedures As noted earlier, resurfacing or pâvement rehabilit¿tion is the primary activity of most federal-aid RRR projects, which may also include minor geometric and roadside improvements. In all 15 case studies, highway agencies selected projects of this type primarily on the basis of pavement condition regardless of funding source. Safety needs play an insignificant role in RRR project selec- tion, and are fi¡st considered during the design process, after project selections have been made. Only occasionally when federal-aid primary or secondary funds are used for safety improvement projects such as guardrail replacement and pavement marking is safety a key factor in selecting a project that qualifies as federal-aid RRR work-bur only for projects intended primarily as safety improvements, not for resurfacing and minor widening. Although all of the state highway agencies in the 15 case srudy states considered pavement condition, their approaches to measuring pavement needs and selecting speciûc projects varied widely (Appendix B, Table B-5). Six of the stafÊs use mechanical means to measure systemwide pavement roughness, deflection, or skid resist¿nce; eight rely on systemwide visual surveys of pavement conditions; and six do not use any type of systemwide pavement survey. In seven states the central office has the primary respon_ sibility for selecting RRR projects, disrrict or regional offices have this responsibility in four states, and the responsibility is divided in four states. Although there is a trend ioward more systematic and objective methods of selecting resurfacing projects, professional judgment and knowledge of local conditions continue to be major factors in project selection and a¡e unques- tionably the primary means of determining the type of surface treatment or repair. Choosing Between Federal and State Funding for RRR An important step in developing a state's highway construction program is assigning a funding source---one or more of the federal-aid program catego-

52 DESIGNINcSAFERRoADS ries or full staæ funding-to each scheduled project. In this procedure the ståte must ensure that none of its apportioned federal funds are lost because of failure to use them in the required time period. At the same time the ståte attempts to minimize the extent to which the restrictions on the various federal-aid categories-regarding the kinds of projects eligible and standards that must be applied---cause it to deviate from carrying out the projecs it judges to have highest priority. Marked differences were observed in the criteria the states have adopted in choosing between state funding and federal aid for the RRR projects in ttreir programs. The important factors the states consider in making this decision are o Avsilability of state nnney for resurfacing: Some states rely so heavily on federal aid for all resurfacing that their actual choice for most projects is either to fund RRR work wittr federal aid or to defer it. o Cost of complying with federal RRR standards: Required geometric or roadside improvemenß can be very costly on substandard roads, or may be regarded as infeasible on roads with narrow rightof-way and extensive adjacent development, but on roads built to high standards the costs of federal RRR requirements are often small. . Benertß of meetingfederal RRR standards:States are often more willing to underûake expensive geometric or roadside improvements on high-volume arterials and roads with poor safety records than on roads that have no history of operational problems. . Size (total cost) of the project: Because of the fixed costs of federal paperwork and review procedures, stâtes prefer to fund large projects with federal aid and small ones with state funds. c (Jrgency of the projec¡; Several ståtos reported that the time required from identification of a resurfacing need to completion of the work can be less than a year if the project is state funded rather than federally funded. All the case study ståtes consider the second factor, cost of complying with federal RRR standards, to some extent in choosing a funding source for a RRR project, but the weight placed on this factor varies considerably. Most of the 15 case study states follow, in general, one of the following three models in choosing funding for resurfacing: ¡ The state has substantial stâte funds available for resurfacing and selects for federal aid those projects ttrat can meet applicable RRR design standards without costly geometric or roadside improvements (Arizona, Florida, Ohio, Virginia). This strategy maximizes the miles of pavement repair within the totat federal and staæ resurfacing budget. In developing its federal-aid RRR pro$am, the state agency employs some means of informally screening

STATE AND T,OCAL PROCEDURES 53 projects, comparing existing conditions to RRR standards to identify projects already in or close to compliance. Søæ funds are used for projects for which compliance with federal ståndards would be costly. The FHWA division office often concurs in this practice or even assists in the screening to avoid federal or siate conflicts over project design. . State funds for resurfacing are limited, thus only a few state-funded RRR projects may be undertaken per year, or the only state supplement to federal- aid RRR is a small maintenance resurfacing program suitable only for stop- gap measures or low-volume roads (California, Michigan). A state in this category can be much less selective in choosing which RRR projects receive federal aid because the only altemative for most projects is to defer adequate pavement repair indefinitely. These staæs still attempt to stretch their federal- aid resurfacing expenditures as fa¡ as possible by defening ttrose projects for which compliance with standards would be the most costly and pavement repair needs are less than critical. In exceptional cases in which pavement repair cannot be delayed and the state is unable to reach agteement with the FHWA regarding geometric and roadside upgrades it frnds acceptable, full state funding may be found for a major RRR project. o Substantial state funds are available for resurfacing, but the cost of compliance with federal standards is not the predominant consideration in choosing between federal and state funding for resurfacing in Mississippi, New Jersey, and Washington. These søtes explicitly consider not only pave- ment preservation, but also the moderate capacity increases and service enhancements obrained from minor widening, spot realignment, or intersec- tion improvements in projects constructed to RRR søndards in selecting resurfacing projects to fund with federal aid. They therefore tend to select for federal aid the more complex projects on high-volume roads that have as objectives service improvement, as well as pavement repair, and use state funds for simpler projects on roads already in compliance with sønda¡ds or on low-volume roads where the benefits of full RRR treâtment would be small. Thus, states in this category also screen resurfacing projects for federal-aid suitabilit¡ but the objective of rhe screening is not simply to minimize the nonresurfacing costs of federal-aid RRR. Federal-aid RRR is only a small fraction of toøl annual mileage resurfaced in most of these states. Although the strategies differ, in every case study ståte the federal require- ments for RRR geometric or roadside improvements affect project selection. This influence may work against obtaining the best combination of system- wide safety improvement and road preservation wittr the funds available. For example, in several case study søtes, fully state-funded resurfacing projects involved pavement work only, regardless of geometric or roadside conditions (although the design of state-funded projects approached federal standards in i ;

54 DESIGNING SAFER ROADS other states). In states dependent on federal aid for RRR, resurfacing of some roads may be defened such that pavement cost are higher than they otherwise would have been had the project not been delayed. Financing RRR Work Alttrough resurfacing and minor widening projects are the predominant types of federal-aid RRR conducted in ttre case study states, they do not dominate total federal-aid spending on the states' f-ederai-aid primary, secondary, anri urban systems. In 7 of the 15 stâtes these projects accounted for less than 20 percent of federal-aid expenditures in the year for which the data were tabulated, and in all states except Ohio, South Dakoh, and Washington the percent¿ge is less than 40 percent (Appendix B, Table B-3). Evidently the shift in the shtes' priorities from new construction to preservation does not necessarily imply that RRR has become the major federal-aid work category. Much of the new preservation work is reconstruction that entails significant capacity increases ând bridge replacement or improvement projects funded largely through the federal-aid categorical bridge program. Many stâtes continue to rely on their own funds rather than federal aid for a substantial share of their resurfacing activities (Appendix B, Thble 8-6). Eight of the 15 case study stâtes spend mofe on st¿te-funded resurfacing than on federal-aid resurfacing; of the 12 ståtes for which resurfacing mileage data were obtained, 10 resurface more miles using ståte funds than federal funds. The portion of federal-aid system miles receiving federal-aid resurfacing in a year was less than 2 percent in at least nine case study stâtes. Mileage comparisons somewhat underståte the importance of federal-aid resurfacing because on stâte-funded projects the tendency is to use thinner overlays on lower volume roads, but it is clear that mosf states have some flexibility in choosing between federal-aid RRR and state-funded work to address their pavement preservation problems. Nonetheless, reliance on federal aid for resurfacing is substantial in many stâtes and has been growing in recent years. Use of federal aid for RRR work largely follows from the increasing emphasis on preservation needs, but it also follows from shortages or constraints on sûate funding. As state motor fuel revenues declined and construction costs escalated during the mid- to late 1970s, some state highway agencies that had initially resisted using federal aid for RRR work felt compelled to begin doing so. Now that state highway tax increases have been enacted in many states in recent years, some states have reduced, or plan to reduce, their use of federal aid for resurfacing (e.g., Florida and Missouri). The state of Virginia has consistently resisted using federal aid for resurfacing work, but has resorted to

STATE AAID I,OCAL PROCEDURES 55 using it recently, not because of shortages in total capiøl resources but to release state funds needed to undertake urban system reconstuction projects for which federal-aid urban funds were not available. Despiæ these examples, the Eend toward increasing reliance on federal aid for resurfacing and other RRR work appears likely to continue. The federal RRR program is responsive úo the mounting preservation concerns of state highway agencies, and increased federal-aid authorizations since the 19g3 increase in federal road user taxes will tend to encourage use of federal aid for all eligible work including RRR. Increased federal aid means increased maæhing fund requirements so that more state funds become tied up in federal-aid work and are unavailable for full state funding of other projects. In the early years of federal RRR funding, some states experienced diffi- culty in meeting the requirement that at least 20 percent of their primary and secondary federal aid be used for RRR work because they were not prepared to use federal aid for resurfacing. Apparently some FtrwA division offices were willing úo classify work bordering on reconstruction as RRR to help meet this requirement, and little, if an¡ federal aid authorization was allowed to þse. currently the requirement is a[ least 40 percent for RRR and reconstruction. None of the st¿te highway agencies in the 15 case study ståtes reporæd a problem meeting this requirement, a result consistent with GAo ûndings l3). F inancing S pe cial S ofety I mprovenlent p roj e ct s In most of the highway agencies in the 15 case study states, special safety improvement projects (the primary p.opose of which is to impróve safety by correcting specific hazards) are funded only up to the limii of the state's federal-aid hazard elimination and rair¡oad grade crossing apportionments. In the past a few states have accumulated large unobligateã balances of hazard elimination apportionments and avoided the lapse of their federal safety funds only by developing occasional large safety projects that could absorb their safety apportionment balances. Several states, however, (California, Michigan, Ohio, and Washington) routinely use regular federal-aid primary, secondary, or urban funds or 100 percent sf¿te construction funds to substantially supplement categorical federal aid for safery improvemenrs (Appendix B, TabËB_7). california and ohio have agreements with their FFIWA division offices linking state spending for special safety improvements to requirements for upgrading safety in RRR projects. In catifornia, rhe s[ate highwãy agency has earma¡ked funding for a statewide priority bridge rail upgrading piogram in lieu of any requiremenr in rhe stare's federal-aid RRR Jan¿ar¿i ttraiurioge

56 DESIGNINGSAFERROADS rail automatically be upgraded in the course of RRR work. The rationale for this arrangement is that priority selection of bridge rail improvements based on accident potential will be substantially more cost-effective ttran the hap- hazard selection ttrat would result if bridge rail improvements were conducted solely when the bridge happened to fall within tl¡e limits of a resurfacing project. ohio has a similar arrangement with its FHWA division office for guardrail improvements. Design Practices and Standards In each of ttre case study stâtes, the design process for RRR projects, FFIWA role in the process, design standards or guidelines used, and procedures for considering design exceptions were examined. State Design Process Federal-aid RRR projects are designed either in district offlces or in cenEal office design divisions. Of the 15 case study statßs, 9 performed most RRR design in district offices and 6 performed it in the cent¡al offlce (Appendix B, faUte A-a¡. When the cenEal oftce staff do not design individual projects, they usually perform review functions and are responsible for all direct "ontu"t with ttre FHWA regarding project approvals and design exceptions. In states with certiûcaúon acceptance agleements covering design, ttre central office takes responsibility for the reviews and interim approvals the FTIWA would otherwise Perform. once a project is programmed, the responsible state design gfoup prepares the plans, speciflcations, and estimates necessary to request conEact bids. (Sometimes, final cost estimates may be prepared by â separate group within the agency.) For federal-aid RRR projects, designers and the various reviewers use a number of techniques to identify and respond to opportunities for geometric and safety improvements during this process. The techniques used vary from state to state (Appendix B, Table B-8) and include . Field reviews. Project designers almost always perform predesign field reviews of existing conditions. In some states, cenüal office staff or FHWA søff may participate in the predesign field reviews or in follow-up reviews related to specific design exception requests. Photologs are sometimes used in lieu of site visits. c Predesign reports. In some state highway agencies, designers prepare predesign reports that cover items such as existing road geometry, existing

STATE AND LOCAL PROCEDURES 57 roadside features, accident history, proposed improvements, anticipated design exceptions, and expecæd project costs. In some states (e.g., New Jersey and souúr Dakoø) these reports are routinely reviewed by FFIWA division offices and commonly contain ttre justification for any requested exceptions to applicable design sønda¡ds. o Accident data analysis. In 13 of the 15 case study states, highway agencies routinely review accident data as paÍ of the RRR design process. This practice probably has become universal as a result of an April 19g4 FHWA directive requiring analysis of accident history for RRR projects 15,). Such reviews are usually intended to help identify haza¡dous locations or assess the reasonableness of a possible design exception. The accident ana- lyses may be confained in a predesign report, or treated as a separate step in the process. c Reviews by trffic and sofety specialists. All 15 states had central traffic and safety specialists, who in most cases oversee the federal hazardelimina- tion program. In two of the states, these specialists routinely review or critique designs for RRR projects. o cost-effectiveness analyses. Formal cost-effectiveness analyses of indi- vidual projects are rare. South Dakota and Michigan use such analyses to consider nade-offs between cost and accident reductions for removing, pro- tecting (with guardrail), or not treating roadside obstacles. The analyses consider the location and nature of the roadside obstacle, average traffic volumes, and frequency of run-off-the-road accidents. c Post-construction evaluations. Some slates, such as Washington, rou- tinely review a sample of all recent construction projects, including RRR, to evaluafe operational and safety characteristics. The results of such reviews can be used to encourage betær designs on future projects. The timing of these techniques within the overall design process also varies among the stâtes. For example, FHWA division offices may review a RRR project before it is programmed, after programming but before a formal exception request, or after a formal exception reques[ @igure 2-2). Schematic representations of the design process in Illinois and ohio (Figure 2-3) illus- trate the variability in the use and timing of techniques to treat safety improvement opportunities. Although any of these techniques may promote safefy-conscious design, and it is helpful to know how commonly they are used, the case studies suggest that it would be misleading to assume that the presence of any single technique, or even group of techniques, will automatically lead to effective, safety-conscious designs. Final designs are influenced not only by design and review practices but also by the standards applied, the rigor of the exception process, and the attitudes of state designers and FF{V/A reviewers.

58 DESIGNINCSAFERROADS FHwA-approved RRR d€3lgn standards FHWA llêld revlew lnlormal FHWA guldance State lleld rovle Fleld revlsw bY deslgne¡s - RevtetsJ o! acclden! dala by deslgnefs DEslgners pr€pare predeslgn report FHWA aPProves Programmlng prolect lor tederal ald lnformal FHWA Euldance FHWA revlews Predeslgn r€port lnlormal FHWA guldance FHWA lleld revlew Formal m€ellng between FHWA and slate FHWA revlew ol acaldent hlstory Deslgners Peñorm salety cost- êflectlveness analYsl9 Rovl€w þy state tralllc and saloiy spåclallste State lustlflos Posslblo deslgn exceptlons FHWA revlews and approves llnal plans, sp€clttcatlons, and estlmateg Pogt-constructlon safety evaluatlon bY state FIGURE 2-2 Generalized RRR design process þoints where safety- enhancement may be considered). Most of the highway agencies in the case study stâtes have a two-tiered approach to RRR design---one for federal-aid RRR projects and a second for projects funded with state funds. In states where søte-funded RRR projects are treated as maintenance activities, the design process may be geatly simplified, focusing almost entfuely on pavement repairs. However, Illinois and New York use the same process regardless of funding, but apply less stringent stândåfds for state-funded projects. As a result, the typical state- funded RRR project is more difficult to characterize than federally funded projects. Søte-funded RRR projects almost always include resurfacing, but

Ohio Department of Transportalion (nondivided highways) Stato lleld revlew lllinols Department of Transportation AASHTO nèw conslructlon standards FHWA tle¡d revlew; ldentlflcatlon ol nec€ssary roadglde lmprov€ments; lnlorm¿l assossmenl ot posslble d€slgñ êxceptlons Fleld revlew by deslgners Deslgners prepar. predeslqn report lncludlng revlew of accldent dala and justlflcatlon for proposed deslEn €xcêptlong Stâte prepares brlef report documentlng lustltlcatlon f or proposed deslgn excepllons ând relevant accldenl data Mêetlng between FHWA ànd state; ¡nfomal declslons on deslgn exceptlons Flnal plans clte how safety wlll be enhanced by prolect lmprovemenls FIGURE 2-3 RRR design process in Ohio and lllinois.

60 DESIGNINGSAFERROADS geometric and roadside improvements may range from no change to improve- ments approaching those made on federal-aid projects' F ederal Highway Adninistation Role For federal-aid construction projects, FIIWA normally performs initial reviews of project scope; intermediate reviews of designs; and reviews of frnal plans, specifications, and estimates before authorizing construction. Respon- .iUitity ior these reviews is deiegatui io FFi't{A division offices, and division administraúors have ttre latitude to specify format and frequency of review activities. In general, the extent and detail of Ft{wA reviews depend on project complexity, and RRR projects are regarded as being less complex than many highway construction projects. However, an April 1984 notice Ûo divi- sion adminisEators provides gleater emphasis for RRR projects by suggesting that freld reviews may be particularly appropriate on RRR projects ". . . to maximize opportunities to enhance highway safety" (6). There are many stages at whiitr a FHWA division office may intervene, formally or infor- mall¡ in the RRR design process figure 2-2). IrL a Ståte where certification aCceptance agleements cover design and construction, the division office's formal actions may be limited to a final design inspection. However, divisions may still perform various project and periódic program reviews in accordance with the specific certification accep- ^tuna" ugtaarentS. Moreover, even when these agreements exist, the FHWA division office must approve all design exceptions on a project-by-project basis. The role of FIIWA division offices in approving design exceptions and special RRR standards varies among the states, as does the nature and extent o1 other project review activities. FFIWA division offices in lllinois, New Jersey, and Texas have worked closely wittr the state highway agencies to deveiop special RRR standards; division offices in Ohio and New York discourage the development of special RRR standards, preferring Úo use new construciion standards; still other division offices in A¡izona and Virginia have neither encouraged nor discouraged special standards' Some division offrces (e.g., New Jersey and South Dakota) require and review predesign reports for possible design exceptions; others (e.g., Texas) wait until a specific exieption iequest is submitted. The Ohio division office conducts field reviews of all projects before programming and makes preliminary determina- tions about design exceptions and necessary roadside improvements for federal aid. A number of other division offices conduct occasional ûeld reviews, often in connection with a speciflc exception request. Also, many division offices participate in regular meetings with highway agencies to

STATE AAID LOCALPROCEDURES 6T review the sfatus of all federal-aid projecs. In some cases (e.g., Illinois) FHWA staff may make decisions regarding design exceptions at these meet- ings. Most of the state highway agencies used informal contacts with FTIWA to handle design issues or problems. For RRR projects, possible design excep- tions are often addressed during such informal contacts, and a number of states reported that they were unlikely to pursue a formal request unless FHWA staff reacted positively to such informal inquiries. Division office procedu¡es and requirements for RRR projects were in hansition in several søtes during the period in which the visits to case study states took place (spring and summer of 1984). Changes were occurring because special RRR standards had been recently proposed or approved in some strates, division offices were tightening the documentation of the formal process for reviewing design excepfion requests tpartly in response to GAO ûndings andrecommendations (2)l,and many division ofûces were strengrh- ening their review of RRR projects to reflect úe greater emphasis on safety required in the surface Transportation Assistance Act of 1982. Moreover, after passage of this act, FIIWA headquarters office provided more formal guidance to division ofûces conceming RRR policies. This guidance included memo- randa to regional adminisEators on safety analysis (5) and design exceptions(7) for RRR projects. As a result, there is a trend toward grcater federal influence in the RRR design process. Standards Nine of the 15 case study states had special RRR standards in effect for federal-aid projects in 1984; the remainder applied st¿te new construction stândards, which are usually based on AASI{T'O guidelines (d, 9). Regardless of whether or not special RRR design standa¡ds are in effect for federal-aid projects, state highway agencies and their respective division offices always have some mutual underst¿nding concerning the geometric design values to appl¡ roadside improvements required, and the process for seeking and documenting design exceptions. About one-third of the case study states had separate RRR standards for søte-funded RRR projects with less stringent requirements than those applied on federal-aid projects. other sråtes applied the same standards for all projecs but allowed more frequent design exceptions on state-funded projects. States that treat resurfacing work as maintenance did not routinely apply geomeric standards or consider geometric improvements during state-funded resurfac- ing. In the case study states, all special RRR standards for federal-aid projects contain minimum design values for key geometric features such as lane and

62 DESIcNING sAFER RoADS shoulder widths and horizontal alignment-values generally less stringent than new construcfion standards (Appendix B, Table B-9). These values are compared with existing conditions at candidate RRR project sites to determine which geometric features are substandard and must be upgraded to be eligible for federal-aid. The values also may be used as a basis for selecting a particular design value when upgrading is required; however, once a súate ãgency makes the decision to upgrade a feature fo qualify for federal aid, it will in many circumstances design to new construction standads rather than to the minimum value in the RRR ståndards. For example, if an existing lane is i0 tt wirie an<i the RRR. standard requires ii it, the siaie oiten wiii wi,jen 'uo 12 ft (new construction stândârd) because it believes the additional cost (for a 12-ft insæad of the minimum ll-ft lane) is low and consistent with a long-term goal of designing highways to full new construction standa¡ds. The state-to-slate variation of minimum values for geometric features (Appendix B, Table B-9) falls within a relatively narrow range, somewhere between AASHTO new construction guidelines and the special RRR søn- dards proposed by FIIWA in 1978. This uniformity is particularly consistent for lane and shoulder width, but less so for horizontal and vertical alignment, reflecting mutual state and federal willingness to alter cross-section charac- teristics more often than to undertåke more costly alignment changes. Although the range in minimum design values is narrow' the highway agencies perceive ttrat the existing differences are necessâry and accommo- date unique circumstances in their stâtes. Most offlcials in the case study stâtes expressed doubts about the practicality and faimess of a single set of RRR minimum design values for nationwide application. In any event, they had largely resolved or put aside their contention with FHWA over the speciflc values selected for roadway geometric features and were not anxious to reopen this issue. For roadside features beyond the shoulder edge, however, special RRR standards varied considerably (Appendix B, Table B-10); this was an issue of some contention between the state highway agencies and FIIWA division offices. In some cases special RRR guidelines require the removal or protec- tion of roadside obstacles (e.g., utility poles, headwalls, or trees), flatlening of sideslopes, and replacement of outdated guardrail. States that apply new construction standards to RRR differ in their handling of the roadside. New construction standards in New York do not address roadside obstacles beyond the shoulder edge, whereas Texas new construction standårds specify roadside clear zones for all highways. Also, regardless of the new construction standards, some FFIWA division offices have established guidelines for roadside Eeatments, even in states where special RRR standa¡ds have not been formally promulgated. The FHWA division offices in the case study ståtes cited a number of factors considered in establishing (or approving) design requirements for RRR pro_iects:

STATEAND LOCAL PROCEDURES 63 . Age, design, and right-of-way of existing highways; o Topogaphic and geologic conditions; ¡ Extent and nature of RRR project proposals; o Extent of safety-oriented RRR activities; and o Financial resources of the state. Officials in Arizona, Califomia, and New York cited increasing concerns over tort liability related to highway design. This concem appeared greatest in states that had lost sovereign immunity and had no statutory ceilings (or high ceilings) on awards. However, none of these states cited examples of ¡ort claim concerns influencing selection of speciûc RRR design standards. Because FHWA division offices differ in the extent of safety enhancements required for RRR projects, particularly on the roadside, the proportion of RRR spending on safety-related improvements varies from state to scate, and RRR project designs approved for federal aid in one state may not be approved in another. Design Exceptions Design exceptions on federal-aid RRR projects must be reviewed and approved by the FHWA division office. The circumstånces under which exceptions are required vary from stâte to state depending on the applicable standards for each RRR project. Usually, formal exception requests are lim- ited to geometric features, whereas policies on roadside improvements are not regarded as standards and therefore deviating from them does not require a formal exception. However, tl¡is distinction normally is only procedural. The division offices visited simply enforced their policies on roadside improve- ments through interim or final approval of design plans, or other means separate from the exception review. State highway agencies generally lnow what to expect before they make a formal request for design exceptions because of informal contacts with divi- sion offices and previous experience. Many state highway agencies reported that they only submit formal exception requests when they are confident. of approval. Most reported that formal exception requests were usually approved, provided adequate justification was given. A GAO review of design exceptions on RRR projects in six ståtes revealed that vertical and horizontal alignment accounted for 44 percent of all excep- tions, shoulder width accounted for 2l percent, and bridge width accounted for 13 percent (2).The case studies, like the GAO review, revealed that high cost and absence of prior accidents are the most commonly cited justifications for I I

& DESICNING SAFER ROADS design exception requests submitted by state highway agencies and approved by FÉIWA. Procedures for evaluating exception requests differ among states, although division administrators must approve formal design exceptions. In many states, FIIWA field engineers make these decisions at site reviews, meetings with state engineers, or after reviewing predesign reports or formal exception requests. Among the case study states, Texæ has the most elaborate review procedure, whereby a three-member panel from tïe FHWA division office reviews exception requests and recommends approval or disapproval to the ¡lir¡i cinn arlminicfrqfnrc Summary State RRR Programs ¡ The most common federal-aid RRR project includes pavement resurfac- ing or rehabilitation, often with minor cross-section or roadside improvements within existing rights-of-way. The more extensive RRR projects constitute a kind of søte highway project uncommon before federal-aid RRR-one with service improvements substantially beyond the scope of simple resurfacing but wifh a much lower cost than full reconstruction. ¡ In all 15 case study stâtes, the primary consideration for selecúng poten- tial federal-aid RRR projects is pavement condition. Safety needs and accident experience are generally not considered in project programming. o Divergent strategies have been adopted by the søtes in selecting the RRR projects for which federal aid will be used. Some state highway agencies avoid using federal aid on projects that would require costly geometric and roadside improvements, seeking to maximize the miles of pavement repair per federal-aid dollar. Others tend to reserve federal aid for more complex pave- ment repair projects that generally include geometric improvements.¡ About one-half of the state highway agencies in the United States apply special design guidelines approved by FFIWA in the design of RRR projects. These guidelines are usually less stringent than new construction sfandards, and the design values applied by different staæs fall within a relaúvely narrow range. o Special minimum design values for RRR projects are used to determine which geometric featu¡es on candidate projects must be upgraded !o qualify for federal aid, but once the decision to improve a feature is made, many states make the improvement to the more stringent standards for new consEuction. r Even in states where new construction design standards are applied to RRR projects, stâte highway agencies and their respective FHWA division

STATE AND LOCALPROCEDT]RES 65 offices have reached informal agreements about when these geomeric design values must be applied rigorously, the roadside improvements that will be required, and the circumstances under which design exceptions will be approved. ¡ For road featu¡es beyond ttre shoulder edge, special RRR standards and RRR design practices vary $eatly and are of continuing contention between some state highway agencies and the FIIWA. Role of Safety in the RRR Process o State highway agencies reflect local public and legislative concerns in setting priorities for their overall construction and maintenance progïams. They report that such concerns are focused more on preservation and capacity needs than on safet¡ and that the emphasis on pavement repair in the use of RRR federal aid is consistent with this focus. o The impact of the federal-aid RRR program on safety rests primarily on the extent of geometric and roadside improvements appended to a resurfacing or pavement rehabilit¿tion project. Because FtrwA division offices differ in the extent of safefy enhancements, particularly on roadsides, required for RRR improvemenfs and because existing conditions differ from ståte to ståte, the proportion of RRR spending on safety-related improvements varies from state to ståte. currently, RRR project designs approved for federal aid in one state may not be approved in another. uncertainty over the safety benefis ttrat result from incremental design improvements underlies ttris va¡iability in ståte and FFIWA division office practices. . The typical RRR project does include one or more improvements that enhance safet¡ but in recent reviews of completed federal-aid RRR projects, the FFrwA has found many missed opportunities for low-cost safety improve- ments; most involve roadside and safety hardware. It is in the area of low-cost safety improvements, however, that FFrwA division offices are evolving toward more stringent requirements. o Nevertheless, because safety problems are not explicitly considered at the project selection stage, the impact of the federal-aid RRR progmm on safety is limited to modifying the design of selected projects. The federal progam is unable to ensufe that the selected projects will address the most urgent safety needs on a systemwide basis. Noteworthy RRR Practices . In general, the federally sponsored hazañ elimination program is the only construction program in which state highway agencies explicitly con-

6 DESIcNINc sAFER RoADs sider safety in project selection and programming. This program is not popular wittr some highway agencies because it requires a substantial amount of paperwork in comparison to other progfams and its funding is comparatively small. Nevertheless, some states, California and New York, for example, supplement federal hazard elimination funds with additional state funds or regular federal aid. In doing so, they can progam safety-motivaæd projects that might not meet the federal benefit-cost guidelines for hazard elimination projects. r In Califomia and Ohio, FTIWA division offices link the required design practices for resurfacing and minor widening projects io ihe siaies' wiiìittg- ness to spend regular federal aid for safety-motivated improvements elsewhere. For example, the Ohio division offtce requires less guardrail replacement on resurfacing projects because the state is using regular primary funds to replace its most obsolete guardrail as a separâûe activity. ¡ The mechanisms by which state agencies and FFIWA division oftces interact to improve the safety of specific projects vary widely and reflect unique adaptations to the specific circumstances and organizations involved. c special RRR design guidelines in Illinois and Texas define specific Eeatments for certain roadside obstacles. o The Ohio FHWA division office conducts field reviews of prospective federal-aid resurfacing projects before programming to idenúfy neces- sary roadside improvements and to make tenüative decisions about design exceptions. o Illinois and New York stâte highway agencies prepare predesign reports for most RRR projects, regardless of funding source, that contain a review of accident daø and justifications for any proposed exceptions to applicable design standards. r In Michigan, stâte fafûc and sâfety speciatists review all design plans for federal-aid RRR projects. . South Dakota and Michigan state highway agencies perform formal cost- effectiveness analyses to determine when barrier protection is warranted for roadside hazards. . The Washington Ståte highway agency routinely evaluates a sample of completed RRR projects to assess safety and operational effects. LOCAL RRR PROGRAMS I-ocai governments administer about one-half the road miles of tie federal-aid secondary system and approximately three-fourths of the federal-aid urban system. County governments are responsible for locally administered second-

STATE AI,ID I.OCAL PROCEDTJRES 67 ary highways, and incorporated cities and towns are responsible for urban system highways. However, this division of responsibility can valy, par- ticularly in u¡banized areas where county governments may also administer some urban system highways. Further, some states allow local governments to use federal-aid bridge and haza¡d elimination funds. compa¡ed wittr state governments, local governments are usually less reliant on federal funding. This difference occurs because local governments administer about 80 percent of the nation's highway mileage off the federal-aid system, for which federal aid is generally not provided, and because per-mile secondary and urban system federal aid is subsøntially lower than primary system aid. Federal RRR requirements have less influence on local progïams than on state prognms because local governments are less dependent on federal aid and federal-aid highways are a smaller share of their responsibility. RRR Project Characteristics l¡cal offlcials interviewed suggested that local govemments in urban areas tend to use federal aid for major reconstruction projects whereas those in rural areas tend to use federal aid for resurfacing and minor widening projects (Appendix B, Täble B-11). These tendencies are supported by daø for all federal-aid construction projects that show that the dominant improvement type (excluding new construction) is reconstruction and major widening (55 percent) for urban system funds and resurfacing and minor widening (50 percent) for secondary funds (Table 2-2). on the urban system, resurfacing and minor widening account for about one-fourth of federal-aid expenditures (excluding new construction). Like state highway agencies, local governments generally view resu¡facing and minor widening projects as the principal federally funded RRR activity even though other RRR work, such as bridge rehabilitation, can be funded with federal assistance. Local officials also indicated that many urban areas currently use federal aid for intenection improvements in addition to major reconstruction projects. Such projects are not separately identified in federal project data and are usually included in the safety category (as minor taffic management improve- ments such as signals or left-tum lanes) or in the reconstruction category (as major intersection improvements fo increase overall capacity). Intersection projects are usually not considered RRR work by either local or ståte agencies unless they are part of a resurfacing and minor wideninþ project andäe not covered by special RRR design standa¡ds. Local and state officials interviewed reported that tle typical federal-aid resurfacing and minor widening project undertaken by a tocal government is similar to a state project in this category-resurfacing or other pavement

68 DESIGNINGSAFERROADS TABLE2-2 Comparison olExpenditures by Type of Improvement (Excluding New Construction) in Federal-Aid Urban, Secondary, and Primary Programs Federal-Aid Program Category Type of Improvement Secondary Urban(ok) (qo) Primary Other Total(70) (Y.) (Vo) Reconstruction and major widening Resurfacing and rehabilitation Minor widening Bridge work Safety improvements (includes traffic control and minor intersection improvements) Total Total cost ($millions) 39. l 50.3 10.4 0.2 100.0 850 54.5 25.7 t'7.5 2.3 r00.0 959 44.0 39.9 15.5 0.6 r 00.0 2,409 12.6 62.5 28.0 34.5 22.9 36. r 2.0 1.4 100.0 100.02,944 7,16l Norr: Compiled from FHWA Fiscal Management Information System, Form FHWA-37 (Project Status Record) data for projects completed between April 1983 and March 1984. Expenditures include state match. Expenditures for a project funded by more than one program category (e'g., a project using both lederal-aid primary and other funds) are allocated accordingly among the categories. rehabilitâtion, possibly with minor cross section and roadside improvements. Local officials reported that compared with a ståte project, however, a local project is less likely to include as many geometric improvements because the road more often meets applicable geometric standatds before the improve- ment is made (a common situation, for example, on low-volume secondary roads, where design standards are less stringent than on high-volume roads), and cross-section design exceptions may be granted more frequently in urban areas because of environmental and community considerations. Project Programming Responsibility for selection of local federal-aid projects usually rests with local govemments or metropolitan planning organizâtions that represent local governments in larger urban areas. The state highway agencies in tum review these selections, primarily to verify program eligibility, and complete some (if not all) of the necessary paperwork for securing federal funds. Project selection and programming is more complicated in larger urban areas, where metropoliøn planning organizations exist, than in small u¡ban and rural areas. lMetropolitan planning organizations are regional organiza- tions established in response to federal law and are responsible for planning,

STATE AND T.OCAL PROCEDURES 69 coordinating, and approving federal Eansportation investments in urbanized a¡eas with populations greater than 50,000. Local elected officials are respon- sible for decision making in the metropolitan planning organization (10)1. Annually, the metropolitan planning organization must approve federal-aid urban projects undertaken within its jurisdiction. In some areas, constituent governments submit potential projects to the metropoliøn planning organiza- tion technical staff who rank the projects using benefrt-cost procedures. Projects that have large road-user benefits related to capacity increases or safety improvement (e.g., major reconstruction and intersection projects) may be assigned priority, sometimes making it difficult for resurfacing projects to be programmed wittr federal aid. In other large urban areas, meFopoliøn planning organizations a¡e less involved in project selection but decide how much each of their member jurisdictions will receive in annual allocations. Ofæn informal agreements exist among members so that one jurisdiction may receive a larger share for a specific project if it is willing úo forego allocations in a future year. In smaller u¡ban and rural areas, project selection is usually a simpler matter-the state highway agency informs the local government of is alloca- tion of secondary or urban system aid and ttre local government selects the projects. In many instances the annual amounts of federal aid are small and the city or county may allow it to accrue over several years so that a major project can be undertaken. There are exceptions úo these practices. In ståtes fhat maintain most second- ary highways (e.g., Texas and Virginia), the highway agency retains the same key role in project selection for secondary roads that it has for other state programs. In other cases (e.g., New York) the state may not allocate its secondary and nonattributable urban system funds (urban funds not earmarked for urbanized areas with populations greater than 200,000) to local govern- ments, instead it may make them available through a discretionary process in which local governments request project funding from the state. In such cases the ståte highway agency plays a more active role in the programming of local federal-aid projects. In general, however, local agencies select federal-aid projects without a great deal of directon or pressure from ståte highway agencies. In the case of resurfacing and minor widening projects, local agencies often have more latitude in project selection because unlike states, they are less dependent on federal aid for this type of work. Because of this financial independence, local officials, particularly in larger jurisdictions, reported that their govemments a¡e often able to select for federal aid only those projects that do not require geometric and roadside improvements beyond those they consider worth- while. If officials view required geometric improvements as excessive, local funds will often be used for improvements on federal-aid highways, and using

70 DESIGNINGSAFERROADS local funds to resurface federal-aid highways is a common occurrence. Only in financially pressed rural counties where federal aid is a significant share of total highway funding do stringent design stândards appear to cause diffl- culties for local offrcials. Financing RRR Work As described in the preceding section, financing responsibility is often linked to progïamming re.sponsibility. When fe<leral aid is directly allocaæd ¡o local governments, they are primarily responsible for project selection; when the stâte makes this aid available through a discretionary process (i.e., distributing funds on a project-by-project basis), the state highway agency has some responsibility in project programming. In either case, state highway agencies closely conFol the distribution of federal aid to local governments. The state highway agency most often reserves federal aid for use by a specifrc local government or for a specific local project; later it lets a construction conEact and spends these funds on behalf of the local government. In some states, the state highway agency provides the necessary matching funds along wittr the federal aid it disributes to local governments. Of the highway agencies in the 15 case study states, 7 routinely provide matching funds for local urban system projects and 8 either provide matching funds or are directly responsible for local secondary proj- ects (Appendix B, Table B-14). In cases in which matching funds are not provided, the ståte may allocate gasoline tax or other state revenues to local governments for highway prrrposes, which can be used to match federal aid. Under federal law, urban system funds are apportioned to ståtes based on urban population, and the portion of these funds attributable to urbanized areas with populations greater than 200,000 in each sfate is earmarked for use in those a¡eas. State highway agencies may make the remaining urban system funds (about 30 percent of all urban system funds) available to local govern- ments through a variety of means. Eleven of the case study states use population-based formulas to allocate "nonattributable" urban system funds; the remainder use a discretionary process (Appendix B, Table B-14). Federal law formerly required that one-half of secondary funds be allocated to counties. Although this requirement no longer exists, many states continue this practice, often using the same allocation formula that the federal govem- ment uses to apportion secondary funds to the states. Eight of the ståte highway agencies in the 15 case study states allocate secondary funds by formula; 7 do not; of those that do not, 4 administer more than 90 percent of their state's secondary highways.

STATEAND LOCALPROCEDURES 7I Design Practices and Standards state high\¡/ay agencies frequently share (or assume) responsibility for design and consfuction oversight on local federal-aid projects. Six of the stâte highway agencies in the case study usually assume responsibitity for the design of local federal-aid projects (Appendix B, Table B-14), and all perform design reviews and coordinate wittr the FHWA on behalf of the local govern- ments. Moreover, most state highway agencies generally let the consrucdon contracts and perform construction oversight (10 of ttre highway agencies in the case study follow this procedure). The federal role is less in local federal-aid projects than in state federal-aid projects. Many state highway agencies have secondary road plan agreements (14 of the case study states) or certification acceptance agreements (8 of the case study states) covering urban system projects below a certain cost, through which the state highway agency is responsible for the design reviews the FHWA would otherwise perform. consequently, most F[rwA division offices will not formally review a local secondary RRR project until the flnal inspec- tion unless a design exception is required. with regard to design standards for federal-aid RRR projects, practices for local government projects vary more than those for stâte projects (Appendix B, Table B-15). If a stâte has special RRR standards, it may or may not apply these standards for local secondary or urban system projects. secondary projects may be covered by separate RRR standards or guidelines included in the secondary road plan agreement; new construction søndards may be applied for urban projects even if special RRR standards have been developed for rural projects. Interviews with local ofñcials revealed mixed feelings about RRR stan- dards. Many representatives of urban areas believed that the standards had not adversely affected their road progams for several reasons. First, major recon- struction and intersection projecs were the predominant forms of work undertaken with federal aid because of local capacity and safety needs, technical ranking of projects by metropolitan planning organization staffs, or the desire to concentrate on a few large projecs because of the delay inherent in using federal aid. (In the last case, local officiats indicated they would rather have one or lwo large projects rather than tve or six smaller ones delayed because of federat procedures.) Second, some urban areas could use federal aid for RRR-type work because thei¡ roads already had good design characteristics. Finally, a number of local ofÍìcials indicated that design exceptions were often $anted when standards presented significant problems. However, even though design standards pose little problem, most local ofûcials interviewed would like to have greater flexibility when using federal aid, especially given that future needs may change. Further, a significant

72 DESIGNINGSAFERROADS number of local officials in both urban and rural areas indicated that RRR design standards were having negative effects on thei¡ road progfams. In some cases, the stândards precluded needed preservation work, forcing the locality to spend federal aid in less productive \ryays. In mral areas, local officials often endorsed upgrading road geometry in principle but quesúoned the cost- effectiveness of rigorous design standards, echoing concerns raised by many state highway agencies. In urban areas witfi older sfreets and narrow rights-of- way that cannot be widened without signiûcant costs, some offrcials expressed concem that the AASIilO design standatds were ûoo stringent not only for RRR projects but aiso ior reconstruciion projects as weii. Summary o Like súate highway agency projects, the most common local federal-aid RRR project involves resurfacing with minor geometric or roadside improve- ments. In comparison with ståte programs, however, local government RRR activities are less reliant on federal aid, and as a result local governments often have greater programming flexibility. Local govemments can often select for federal-aid only those projects that already meet applicable federal design ståndards or that would require improvements believed to be clearly cosf effective. r In urbanized areas, geometric improvements raise environmental and community issues seldom encountered in rural areas. . Because of special agreements under which state highway agencies assume many of the project review functions otherwise performed by FFIWA for smaller projects, the federal oversight role is more limited on local federal- aid RRR projects, often consisting only of approving design exceptions and conducting ûnal, post-construction inspections. STjMMARY OF FINDINGS Ståte highway agency federal-aid RRR projects are, in general, selected and designed with the primary objective of preserving pavement and extending the life of the road. Typical federal-aid RRR projects involve pavement resurfac- ing and roadside improvements or minor cross-section improvements (lane or shoulder widening) within existing rights-of-way. Most non-Interståte federal- aid RRR projects are carried out on two-lane ru¡al roads because most federal- aid road miles are of this type, and federal aid designated for urban areas is extremely limited and usually devoted to capacity improvements rather than RRR.

STATE AND I.OCAL PROCEDURES 73 In the last few years, state highway agencies have paid increasing attention to safety on federal-aid RRR projects, and most projects inciude some improvements that enhance safety. Nevertheless, highway agencies still miss opportunities for cost-effecfive safety improvements on Rnn projects. Missed safeny opportunities sometimes result from the standards that guide RRR project design. About one-half of the states have special standards, less rigorous than sønda¡ds for new construction, that govern design offederal-aid RRR projects. The remaining states must seek approval from the FFrwA to deviate from new construction standa¡ds. The specìal RRR standards specify the minimum values of some geometric dimensions of the roadway 1e.g., tanä and shoulder widttr) that a project qualifying for federal aid-may leave unaltered, and also often identify other road conditions for which special consideration by the project designer is mandatory. RRR standards often fail to specify firm requirements for several key design features that affect safety. In particular, the standards in use in many states set no frrm requirements dictating whether ttre alignment of a horizontal c_urve, 1!e stopping sight distance at a hill crest, the contour of sideslopes, or the width of the clear zone on a RRR project must be improved. when RRR standards do not provide definite guidance for treatment of a feature, new construction standards apply in principle, but in practice the features specifi- cally addressed in RRR ståndards usually constitute the possible upgradesgiven serious consideration during project design. Also, new cons^truction standa¡ds do not address all imporønt aspects of RRR project design. For example, new construction standards do not specifically require the upgrading of obsolete guardrails. Special RRR sønda¡ds are sometimes ambiguous and subject to va¡iable interpretation and application. For example, in many state RRR standards, minimum acceptable geometric conditions depend on the design trafûc vol- ume and design speed of the road, but the method of determining ttrese design values is not specified. The geometric improvements dicøæd ty ttre standard can differ greatly depending on whether current Eafûc volume or volume projected at the end of the expected life of the project is used as the design value, or wlether design speed is selected on the basis of running speeds or the road's functional classification. stândards, however, are only one factor influencing the characteristics of RRR project designs. Re¡¿iews of state highway agenciès revealed many casesin which two states with similar standa¡ds produced very differeni nnnproject designs. These discrepancies result in part from differences in the physical conditions of road systems: standards cannot provide specific guid- ance to ût all possible ci¡cumstances encountered on the highway system. Also, federally required RRR standards can be ci¡cumvented*by uring ,t t" funds for pavement repairs on roads whose existing geomeû-y falls short of the standards.

74 DESIGNINGSAFERROADS The nature of compleæd RRR projects depends on the process the highway agency follows in project planning, selection, and design. This process ofæn rans o produce a fully safety-conscious design because of a lack of (¿) emphasii by top management, (å) required resources, and (c) necessary tools to identify safety improvement opportunities and evaluate options for address- ing them. Little information is available about the payoffs of the kinds of salety enhancements most readily incorporated in RRR projects, and existing information usually is not accessible to designers. Project design routines nlece little emnhasis on i{entifvins sâfetv needS or seeking Opportunities tO imptoue safety. The highway agency professionals most likely to appreciate the connections between geometric design and safety (e.g., faffrc and safety engineers) are seldom involved in the RRR design process' The process can also fail to produce the safest practical design because the scope õf RRR projects is often too narrowly conceived, unnecessarily excluding types of improvements that could enhance safety. In m¿ìny states, alignment adjustments, roadside improvements, and some types of bridge improvements are seldom considered as appropriate or feasible components of RRR designs. In most stâtes, any improvement requiring more than a minor addition to the right-of-way is beyond the scope of a RRR project, largely because of the amount of time needed for right-of-way acquisition. Highway agencies that produce the most safefy-conscious RRR projects actively seek opportunities for safety improvements. These agencies follow systematic procedures for evaluating existing condiúons to detect safety needs and analyzing a range of options to meet these needs. To evaluate existing conditions, they analyze accident records at project sites, conduct site inspec- tions that emphasize safety opportunities and often involve safety specialists, verify existing geometric and t¡affic conditions for comparison with design ståndards or guidelines, and prepare written design reports with prescribed contents that document design procedures and safety analyses. These agencies also tend to consider a broad range of possible safety-motivated improve- ments as options in RRR projects, including spot alignment adjustments; attention to signs, markings, and signals; improvements at bridges and inter- sections; and establishment of safer roadside conditions. They look for mod- est, low-cost soluúons in situaúons where upgrading to full standards would be impractical, for example, spot widening at curves and other locations prone to edge-drop problems. Although, for the greatest impact on safety, this safety-conscious approach would extend to the planning and project selection ståges of a state's highway improvement programs, serious consideration of safety in RRR generally has been limited to the design stage. In a few stâtes, however, needs for safety improvements of the types appropriate for RRR projects are being addressed on a systemwide basis by screening the highway system for substanda¡d

STATEAND LOCALPROCEDURES 75 geometric or roadside conditions and setting st¿tewide priorities for their correction. None of the cæe study states that produce the most safety-conscious RRR designs has comprehensively analyzed the benefrts and costs of this policy. The safety emphæis usually implies higher average cost per mile for RRR projects and greater demands on the resources of the design søff. The søtes have not estimated the overall accident savings achieved by their policies, and these would be diffrcult to measure. REFERENCES I' RRR Field Revíews4irøl Report. FHWA, u.s. Deparunent of rransportation, Jan.25,1984. 2' The Departmcnt of rransportation's progrant to preserve the Higlutays: sdety Remaíra an Issue. U.S. General Accounting Office, Dec. 23, lgga. 3. Types $ work Perþrmed using Resurfacing, Restoration, Rehabiritation, and Reconstruction Federal Highway Funds. lJ.s. General Accounting office, Feb. 29, t984. 4. FHWA Notice N 5M0.19, FHWA, u.S. Department of rransporrarion, June 2g, 1976. 5. safety Analysis--Nor{reeway 3R prograrn. Memorantlum from Director, FHWA office of Highway safery ro FHV/A Regionar Adminisrrators, FttwA, u.s, Department of TransportatiorL April 6, 1984. 6. "Monitoring of Federal-Aid Highway Design projects." Federal-Aid Higttway Progrant Manual. FHWA, U.S. Department of Transportation, Vol. 6, Ch. 2, Sec. 1, Subsection 2, April 2,1994. 7. 4R Program: Design Exceptiora. Memorandum from Director, FHWA office of Engineering to FHWA Regional Administrators, u.s. Deparrnent of rransporta- tion, July 9, 1984. 8' Geometric Desígn Guides for Local Roads and streets, American Association of State Highway and Transportation Officials, Washington, D.C., 19g1. 9. Geometric Design standards for Highways other than Freeways. American Asso- ciation of state Highway Transportation officials, washington, D.c., 1969. 10. "Transportation Planning in cerrain urban Areas." code if Federal Regulations. Title 23. Section 134.

4J Relationships Between SafetY and Geometric Design Relationships between safety and highway design features routinely improved on resurfacing, restoration, and rehabilitation S.RR) projects are described in this chapter. These relationships were used for the safety cost-effectiveness comparisons presented in Chapter 5, and can also be used by engineers making design decisions on individual RRR projects. The relaúonships per- øin primarily to two-lane rural roads, which account for about 75 percent of all fôderal-aid highway mileage, 25 percent of vehicle miles traveled (vMT) throughout the United States, and 35 percent of U.S. highway faølities. Background information on the natufe and frequency of accidents on federal- aid highways is presented in Appendix H. APPLICATION OF SAFETY RELATIONSHIPS TO DESIGN STANDARDS The following questions about the safety effects of highway improvements were the focus of earlier debates over RRR design standards: o What changes in accident rates can be expected if different types of geometric improvements a¡e made? r Will accident rates increase if highways are resurfaced without correcting existing geometric defi ciencies? . Wiãt afe the safety benefìts of low-cost alternatives, such as warning signs and markings, compared with more expensive geomeEic improvements? Despite the widely acknowledged importance of safety in highway design, the scientific and engineering research necessafy to answer these quesüons is 16

SAFETY AND GEOMETRIC DESIGN 77 quite limited, sometimes contradictory, and often insuffrcient to establish ûrm and scientifically defensible numerical relationships. Further, in ttrose cases in which relationships can be established with substantial conf,dence, the results are often not known or applied by highway designers. In general, relationships between safety and highway features are not well understood quantitatively, and the linkage between these relationships and highway design søndards has been neither straightforward nor explicit. The American Association of State Highway and Transportation Officials (AASI{TO), which has historically assumed primary responsibility for setring design standards, relies on committees of experienced highway designers to do this work. The committees use a participatory process that relies heavily on professional judgment. Quantiøtive estimates of the overall safety or cost implications of recornmended design policies are not usually developed, although the process takes into account not only safety but also cost and other facûors (such as the effect of design on trafflc operations and capacify, maintenance implications, and design consistency for similar traffrc condi- tions). Although relationships between safety and highway features should be important for assessing the cost-safety trade-offs that underlie any highway standards, they are even more importânt for RRR standa¡ds than for new construction stândards. In RRR work the costs of making incremental geo- metric improvements are often large relative to other project costs. For new construction and reconstruction projects, where the entire highway is being constructed from the bottom up, often on newly acquired rights-of-wa¡ the added costs of building to higher standards are comparatively low. As a result, stringent stândards (that require wider lanes, flatter curves, etc.) can be much more costly for RRR work than for new construction, and development of RRR ståndards requires more ca¡eful estimation of the safety payoffs expected from incremental geometric improvements. In addition to geometric features, a variety of other factors affect highway safet¡ including other elements of the overall road environment (e.g., pave- ment condition, weather and lighting, traffic, and t¡affìc regulations), driver characteristics (intoxication, age), and vehicle characteristics (size, weight, braking capability). The effect of highway design is obscu¡ed by the presence of these factors. Indeed, most. accidents result from a combination of factors interacting in ways that preclude determining a single accident cause. Even when a vehicle runs off the road because of driver error or equipment failure, the design of ttre roadside still may affect accident severity. This interacúon between road, driver, and vehicle characteristics complicates attempts to estimate the acci- dent reduction that can be expected from a particular safety improvement.

78 DESIGNING SAFER ROADS RELATIONSHIPS BETWEEN SAFETY AND KEY ROAD FEATURES Highway features affect safety by o Influencing the ability of the driver to maintain vehicle conFol and identify hazards. Significant features include lane width, alignment, sight distance, superelevation, and pavement surface characteristics; o Influencing the number and types of opportunities that exist for conflicts between vehicles. Signifrcant features include access control, intersection design, number of lanes, and medians; o Affecting fhe consequences of an out-of-control vehicle leaving the travel lanes. Signifrcant features include shoulder width and type, edge drop, roadside conditions, sideslopes, and gua¡drail; and o Affecting the behavior and attentiveness of the driver, particularly, the choice of travel speed. Driver behavior is affected by virtually all elements of the roadway environment. For nearly 50 years, resea¡chers have ried þ measure the effects of various road features on safety. Generally, accident rates associated with different roadway designs have been estimated by using actual accident records and travel daø. The latter, usually expressed as vehicle miles of travel, is needed to express accidents relative to the number of opportunities for their occur- rence. Despiæ these long-term efforts, surprisingly little is known about the decrease in accident rates that, result from improvements in road design. Explicit, widely accepted, quantiøtive relationships have not emerged. At times, researchers have been unable !o a$e€ on even the most fundamental of findings. In part, this unfortunate situation can be attributed to inherent difficulties in accident research: ¡ Accidents are relatively infrequent so that sound statistical studies require consistent data collected over long periods of time for many miles of highway. . Many factors-some related to fhe road environment, the driver, and the vehicle--interactively contribute to the occurrence and severity of accidents. Information describing the plethora of related factors is seldom included in the accident data base. Even with reasonably compleæ data bases, however, researchers a¡e often unable to sort out effects attributable to the specific roadway feature of interest. Controlled experiments a¡e difficult to design and conduct. . Reporting practices for nonfatal accidents differ among slates and, in some cases, within states. Thus, estimates of accident rates developed using data from one area might not be appropriate elsewhere.

SAFETY AND GEOMETRIC DESIGN 79 . Some factors, such as vehicle performance and crashworthiness, that unde¡lie relationships between safety and road design, change over time so that relationships developed at one time may no longer be representative in later years. shortcomings in the way research has been organized and conducted also have hindered the development of reliable relationships between safety and road design. Knowledge about the safety effects ofroad design hæ often been uncoordinated and has lacked rigorous søtistical controls. with the exception of a modestresearch program sponsored by the Federal Highway Administra- tion (FI{WA), few opportunities exist for coordinated, purposeful research by experienced researchers using adequafe data that would over time provide the missing information about the safety effects of alternative road designs. other countries sponsor simila¡ research, but the results are diffrcult to apply in the united states because of critical differences in vehicle characteristics, traffic rules, or accident reporting practices among countries. In any event, wise investments in safety improvements are difûcult to formulate in the absence of ¡eliable safety-road design relationships acquired through a soundly managed and adequately financed research prognm. with these difficulties in mind, the study committee commissioned two special research projects and several critical reviews of the existing highway safety literature in order to assess the most likely relationships between-safety and the following highway design features: ¡ I¿ne and shoulder width and shoulder type, o Roadsides and sideslopes, . Bridge width, o Horizontal alignment, . Sight disÍance, o Intersections, o Pavement surface condition, and¡ Pavement edge drops. In the committee's judgment, improvements to ttrese design features on RRR projects are most likely to have significant and measurable safety effects. some geometric features such as cross slopes (the transverse pavement slope from the centerline on straight sections) and vefical alignment (except as a sight distance consideration) have been excluded because they do noì meet these criteria. Pavement surface condition and edge drops, nongeometric features, have been included because they bear directly on the overall safety effectiveness of RRR work. The study committee made is best judgments about the most probable relationships between safety and each of the highway design features. For each feature, the study assessed

80 DESIGNING SAFER ROADS ¡ Whether a relationship between safety and the design feanres exists (e.g., is shoulder width related to safety?); o Direction of any relationship (e.g., whether increasing shoulder width improves or degrades safety); and . Where possible, the magnitude of the safety impact most likely over the range of improvements considered in RRR projects (e.g., the reduction in accidents expected if shoulders are widened from 2 to 4 ft). For several of the more important features, such as lane width, horizontal curvature, and bridge widttr, evidence was judged to be sufficient to generalize quantitatively about the safety effects of design improvements. In the case of each feature, the generalized relationship is principally applicable to two-lane rwal highways. The effects in urban settings could not be clearly documented because a substantial portion of the prior resea¡ch focused on rural highways. For features such as pavement edge drops and sideslopes, development of quantitative models proved to be impossible even though considerable safety- related information was collected. The quantiøtive relationships were used to estimate incremental safety benefits expected from adoption of nationwide RRR standards for the geo- metric design of two-lane rural roads and to assess the cost-safety trade-offs involved (Chapter 5). Also, the relationships, summarized in Appendix C, should prôvide guidance no highway designers who make daily decisions about trade-offs between safety and cost. Clearly much remains unknown about safety and geometric design relationships. Better understanding of these relationships should continually evolve over time as new research results become available. Cunent knowledge about each design feature is sum- marized next. Lane and Shoulder Width and Shoulder Type 'Wide lanes and shoulders provide motorists increased opportunity for safe recovery when their vehicles run off the road (an important factor in single- vehicle accidents) and increased lateral separation between overtaking and meeting vehicles (an imporønt factor in sideswipe and head-on accidents). Additional safety beneflts include reduced interrupúon from both emergency stopping and road marntenance activities, less wear at the lane edge, improved sight distance at critical horizontal curves, and improved roadway su¡face drainage. Prior research reviewed as part of this study (/) indicates that

SAFETY AND GEOMETRIC DESIGN 81 o Accident rates decrease with increases in lane and shoulder width; . In terms of accidents eliminated per foot of added width, widening lanes has a bigger payoff than widening shoulders; and¡ Roads with søbilized shoulder surfaces, such as asphalt or portland cement concrete, have lower accident rates than nearly identical roads with unstabilized earth, turf, or gravel shoulden. r---- - -- -¡ Fltl hatl sectlon-hlghway II constructed on com-oactéd I I earth (embankment) transported II lo sectlon IL*----- ----- -----J r------------ --- --*lI Cut haf soctton-hlghway II construclod aller oxcavailon of II earth or rock,lL-------- --------J Sldeslope (also foreslope) Clear zone wldth (lo obslacle or nontraversable slope) FIGIJRE 3-1 Cross section design features and terms. Although the effects of individual cross-section features were estimated in several prior studies, the literature did not provide a single, internally consis- tent model of the simultaneous effects of lane width, shoulder width, and shoulder type on accidents. (see Figure 3-1 for an illustration of typical cross- section features.) consequently, reseârch was commissioned, in conjunction with FtrwA, to study the combined effects of cross-section features and other va¡iables (roadside clear zones, terrain, and naffic levels) that affect accident rates. The research (2) produced relationships between cross-section feanrres and accident rates consistent with the ûndings outlined previously (Figure 3-2). The relationships, described in Appendix c, cover single-vehicle, head-on, and sideswipe accidents. Such accidents are directly affected by lane and shoulder conditions and account for roughly 60 percent of highway fatalities on non-Interstâte federal-aid highways in rural areas. The relationships predict that widening lanes from 9 to 12 ft without shoulder improvement reduces accidents by 32 percent. v/idening shoulders is less effective than widening lanes: adding a 3-ft unstabilized shoulder where

Shquldff wldth (fi) 91011 LANE WIDTH (ft) s1011 LANEWDJH(lt) NOTES: Aô¿ldênr rêlaliÞnship qoverg singlo-vèhiclg,,éidesÍPo' and optrjosité- diÞction a@idenß ôt two.lam rural .highwâys. Relatv€ acq¡dênt råt9.ig d€lned aS a mutiplé oflhå aôcidl:ns per miÌl¡on vshiole mileÞ tor 12-fi lanes qnd 1O*ft stáblliæd thouldeÉ! FIGIIRE 3-2 Normalized relationship between accidents and lane and shoulder conditions l2).

SAFEW AND GEOMETRIC DESIGN 83 none existed reduces accidents by 19 percent. If the 3-ft shoulder addition were paved, tlre expected reduction would be somewhat greater-about 22 percent. The greatest gains result from a combination of improvements. For example, widening a highway with 9-ft lanes and no shoulders to 12-ft lanes and 6-ft paved shoulders reduces accidents by about 60 percent. However, the accident reduction as a result of improving a specific feature will be less when ottrer features also are improved. For example, adding 2-ft shoulders to â highway with ll-ft lanes will eliminate fewer accidents than adding 2-ft shoulders to a highway with 9-ft lanes. Roadsides and Sideslopes Roadside encroachments begin when the vehicle inadvertently leaves the travel lanes, veering toward the roadside. Most encroachments are quiæ harmless: the driver is able to regain control of the vehicle on the shoulder and safely return to the travel lanes. When coupled with nearby roadside hazards, however, encroachments can result in roadside accidents (Figures 3-3,3-4, and 3-5). Such accidents comprise a significant number of the accidents that occur: on two-lane rural roads, more than 30 percent of all accidents involve single vehicles running off the road (3). FIGURE 3-3 Large úees near roadway.

FIGURE 34 Steep sideslope. FIGURE 3-5 Rigid drainage structure.

SAFETY AND GEOMMRIC DESIGN 85 Past research on the safety of the roadside environment has produced important improvements to roadside hardware, including, for example, the development of barriers ttrat better contain and more safely redirect erant vehicles and sign and luminaire supports that break away on impact, causing little damage to ttre striking vehicle and its occupants. In addition, design stândards occasionally provide for clear recovery areas-borders beginning at the edge of the fravel lanes with Eaversable slopes and free of hazardous obstacles. Improved designs for drainage structures such as culvert headwalls reduce the hazard posed by unforgiving roadside obstacles. Also, speciûca- tions for sideslope and diæh confrguration now recognize safety benefits, as well as the more conventional objectives of construction economy, main- tainability, and slope stability. Entry of an errant vehicle onto the roadside border does not in itself mean that an accident is inevitable. Although some danger always exists, tie chances of a safe recovery are excellent if the border is reasonably smooth, flat, and clear of fixed objects and other nontraversable hazards. The chances of successful recovery diminish as the gtound slope within the border increases a¡rd the width decreases. Although there are no clear breakpoints, safety researchers generally agree that at speeds of approximately 55 mph, "safe" clear zones should have sideslopes no steeper than about 6:1 and should extend outward at least 30 ft from the edge of the t¡avel lanes. When the border is flat, unintended encroachments on tangent alignments seldom extend beyond the 30-ft range. Despiæ the noteworthy research described, much of what is lnown about roadside safety relationships remains qualitative in nature, and only tentative steps have been taken úo develop comprehensive accident models. previous studies have found signiflcant relationships between accident rates and com- posite measures of roadside condition (4-B). Research commissioned for this study revealed a significant relationship befween the roadside recovery dis- tance and accident rates on two-lane rural roads (Figure 3-6) (2). Increasing the clear recovery area from 5 to 20 ft, for example, is estimated to reduce the number of single vehicle, head-on, and sideswipe accidents by about 35 percent. Roadside encroachment models have been used to examine the safety effects of specific roadside features (9-13). These models take into account the size and shape ofa roadside feature, its distance from the travel lanes, and the probability that a collision with the roadside feature will resulr in an accident. For purposes of this study, a roadside encroachment model was calibrated using ttre daø base from a recent study of utility pole accidents (Appendix F). This new calibration was able to effectively replicate the effect of traffic volume and pole offset on utility pole accident rates. It also performed better

86 DESIGNING SAFER ROADS 024681012141618 CLEAR RECOVERYAREA BEYOND OUTSIDE SHOULDER EDGE (ft) NOTES: Accident relat¡onship covsrs s¡ngl€-veh¡cle, s¡desw¡pe, and oppos¡te- direction accidents on two-lane rural highways. Clear ræovery area is reasured from the outs¡de should€r edge lo the nesest roadside obstacle or hæard. Relalive accidenl rate is defined as a multiple of the acc¡dents per million vehicle mìles lor a clear recovery area ol 20 11. FIGURE 3-6 Normalized relationship between accidents and width of clear recovery uea (2). than previously calibrated models in reproducing observed accident rates for different roadside environments. Accordingly, it is recommended for interim use in examining the safety effect of specific roadside features. Under this model, the probability of an object being struck by an enant vehicle decreases with its distance from the edge of the travel lanes such that, for example, an object located 10 ft from the travel lanes is about twice as likely to be struck as an object located 20 ft from the travel lanes. Bridge Width Hazards associated with bridges can be significant. Roadway constriction at n¿urow bridges reduces the opportunity for safe recovery by out-of-connol vehicles and can result in end-of-bridge collisions. Furthermore, bridge approaches a¡e often on a downward grade, a factor responsible for increases in speed, and, particularly in ttre case of older spans, are often sharply curved (Figure 3-7). When coupled with other factors such as premature icing in H r.z fE l- 1.6z tuI 1.5oo< 1.4 tu k r'rI t¡JÉ 1.2 1.9 1.8 1.1 1.0

SAFETY AND GEOMETRIC DESIGN 87 7,t&.. FIGURE 3-7 Narrow bridge on curve. winter and substandard bridge rail, the special hazards associated with bridges are readily understood. Investigations of bridge safety (14) have revealed that bridge width is the principal factor affecting bridge safety: fewer accidents occut on wide bridges than on narrow bridges. More precisely, these studies have found the dif- ference between the clear bridge width and the width of the approach lanes (referred to as ttre relative bridge width) to be a betær indicator of hazard than bridge widttr itself (Figure 3-8). As this difference increases, observed bridge accident rates, commonly expressed in terms of total accidents per million vehicles, markedly decrease. From several quandrative relationships developed in ea¡lier research stud- ies, a single one was selected for use in this study as the most likely relationship for a two-lane bridge (75) @igure 3-9). This relationship, described in Appendix C, predicts rhat ¡ Increasing the difference between the width of the bridge and the width of the approach lanes from 0 to 4 fr will decrease bridge accidents by about 40 percent, with the fìrst foot of widening accounting for nearly one-third of this reduclion. e The incremental safety gains of widening bridges decrease as bridge width increases-the first foot of bridge width beyond the travel lanes has th¡ee times the effect on accident rates ¿ìs the tenth foot. I þ

88 DESIGNING SAFER ROÀDS wher€ A. wldth oltravel lanes, I = brldg€ wldthl ç = bridse (struaiæ) iengihj atri B - A = rqlatlvè brldge wldth, FIGURE 3-8 Bridge width terms and dimensions-Plan view- BRIDGE WIDTH - WIDTH OF TRAVEL LANES (ff) NOTE: Rêlawe a@ident rqte js a multiple ol lhe nqmbs ql brìdge apc¡deRts. Þermillion vehlcles whêñ th6 brldge w¡dth minuslhè w¡dth Òl lhe travel lanes FIGURE 3-9 Normalized relationship between accidents and bridge width (/5J. Safety at nanow bridges can also be improved by transition guardrails at bridge approaches, new or rehabilitated bridge rails, and warning sþs. However, rcse¿r.chers have not been able ûo develop reliable quântitative estimates of safety benefits for these improvements. I I

SAFE(Y AND GEOMETRIC DESIGN 89 Horizontal Alignment Accidents ate more likely to occur on horizontal curves than on strâight segments of roadway because of increased demands placed on the driver and the vehicle and because of friction between tires and pavements @igure 3-10). The safety effect of an individual curve is influenced not only by the curve,s geometric characteristics, but also by the geometry of adjacent highway segments. The hazard is particularly intense when the curve is unexpected, such as when it follows a long straight approach or when it is hidden from view by a hill crest. ,; FIGURE 3-10 Sharp horizontal curve. The safety effect of flattening sharp horizontal curves is of particular interest on RRR projects. Vy'hen a sharp curve is improved, transitions from the straight to curved portions of the highway are smoother; the length of the curved portion of the roadway is increased; and the overall length of the highway is slightly reduced. (See Figure 3-11 for horizontal curve geometry.) Neither the combined nor individual effects of these changes on accidents are well understood. Numerous researchers have attempted to relate changes in accident rates úo speciûc characteristics ofcurve geometry, usually concentrating on degree (or radius) of curve. Past studies differ considerably in estimates of accidents per I I I I I I I

90 DESIGNING SAFER ROADS (a) Plan vl€w of slmPle curv€ L where R = radius of curve ln f€€t, I = central angle of curve ln degr€es, D = degree of curve = 5730/R' and L = lêngth of curve ln feet = 100 (l/D) (b) Cross section ø = Superelevation angle ¡n degrees The rate of superelevatlon of a roadway ls the roadway slope (tan ø ). FIGURE 3-11 Horizontal curve geomeEy and terms. vehicle mile as a function of degree of curve, partly because of differences in techniques used for calculating the amount of travel and identifying accidents considered to be curve related. Also, some of the accident datâ bases were limited (encompassing only a single yeår in some cases), and influences of other geometric and traffic cha¡acteristics on curve-relaæd accidents were not properly Eeated in some of the analyses (16). A recent study sponsored by the FFIWA (17) srcceeded in eliminating many of these problems and, in so doing, assembled the most reliable accident datâ base cufrently available for horizontat curves. Like the data in earlier studies, the FI{WA data indicate a snong link between degree of curve and accidents. The link between accidents and other measures of curve geometry, including curve length and cenEal angle, is much

SAFETT AND GEOMETRIC DESIGN 9I weaker. For cost-effectiveness analyses of horizontal curve improvements, the study used a numerical relationship based on the FTIWA data (Appendix D). In this relationship, the expected change in accidents resulting from a horizon- tal curve improvement is based on the change in degree of curvature, taking into account the minor reduction in road length that accompanies cuwe flattening (Figure 3-12). 051015202530 DEGREE OF CURVATURE NOIES: Accidentretationship is 'or h¡ghway segments 0.6 m¡ long. Relative rate is a mulliple of acc¡dents per million vehicle; on langent secÍõns. FIGLIRE 3-12 Cresr curve with restricted sight distance. The relationship between accidents and degree ofcurve must be regarded as rough in nature because horizontal curves are considered in isolation- without regard to the alignment of adjacenr highway segments-and because the relationship does not fully correct for inærrelated effects of other geo- metric features (e.g., sharp curves occur more frequently on roads with narrow lanes and dangerous roadsides). As degree of curvature decreases, this relationship predicts that the number of accidents at the curve also decreases, on average by about 3 fewer accidents per degree of curvature for each 100 million vehicles passing through the curve. Flattening a sharp curve on a road carrying 2,000 vehicles per day eliminates about one accident every 8 years for each reduction in curvàture of 5 degrees. Although researchers have not been able to estimate the benefits quan- titativel¡ a number of other design elements also affect sâfety at curves:

92 DEsrcNINc SAFER RoADs . Adequate superelevatio¿.'Horizontal curves on high-speed highways are usually superelevated, or banked, for safety and passenger comfort. AASI{TO design policy specifies superelevation requirements based on degree of curve and design speed (/S). On cuwes where superelevation is less than that speciûed by AASHT0, improving superelevation as part of a RRR project is a relatively inexpensive way of increasing design speed. o Clear roadsides and mild sideslopes: As noted earlier, single vehicle, run-off-road accidents are particularly common at horizontal curves. Conse- quently, roadsides with mild (4:1 or flatær) sideslopes would be expected to yield greater beneûts in terms of reduced accicient severity at curves than on tangent sections, especially on the outside of curves where more than two- thi¡ds of the fatal, curve-related, run-off-road accidents æcw (19). . Spiral Ûansirtons: Increased accident frequencies at horizontal curves appear to be related more to enü.y and exit effects than to steady-ståte travel on the curved roadway (Appendix D). Spiral Eansitions, which help drivers make smoother entries and exits, reduce the hazard at these locations. c Pavement surface: Because of the vehicle dynamics involved, pavement condition on curves is particularly important for safety. Rough pavement, wittt potholes or bumps can contribute to loss of vehicle control, and surface friction must resist lateral forces in wet weather. o Striping and other trafic contol devices: Striping and reflectorized markings on pavement edges and centerlines, curve waming signs, and post delineators may help drivers successfully negotiate curves at night. . No-passing zones: Hoilzontal curves can exacerbate difficulties in carry- ing out passing maneuvers. The marking of no-passing zones may reduce accidents at such curves. Sight Distance Sight distance is the length of road ahe¿d visible to the driver. To enhance safety on highways, designers must provide sight distances of sufñcient length that drivers can avoid striking unexpected objects in the highway lanes (18). Sight distance requirements vary sharply with vehicle speeds. For example, according to AASHTO design procedures, the required stopping sight dis- tance is 400 ft at 45 mph, 550 ft at 55 mph, and725 ft at 65 mph (18). If the increase in Interstate speed limiS n 65 mph results in higher driving speeds on non-Interstate highways, sight distance improvements will become more important. Sight-distance restrictions result from obstructions on the inside of horizon- tal curves, at intersections, or at sharp hill crests (Figure 3-13). Although obstructions at horizontal ctuves and intersections can sometimes be elimi-

SAFET! AND GEOMEIRIC DESIGN 93 nated without changes to highway geometry (e.g., by cutting brush or Eees), obstructions at hill crests can only be corrected by changes in vertical align- ment-lengthening the existing vertical crest curve (see Figure 3-14). A recent Nadonal Cooperative Highway Resea¡ch Program study for which accident data were collected for carefully matched sites with and without sight-distance restrictions due to verúcal curvature, found accident frequen- cies to be 52 percent geater overall at sites with sight restrictions than at control siæs (20). The safety effect of a sight-distance restriction is influenced not only by the sight-distance restriction itself, but also by the nature and location of any potential hazards hidden from view. For example, a heavily used but hidden intersection greatly increases the likelihood of accidents at crest curves. Without the heavy use, the necessity for rapid stopping would be $eatly diminished and, as a result, so would the likelihood of stopping-related accidents. However, with sight distance problems at intersections, the hidden object is another vehicle. Thus, tie 6-in. object height, which is recommended FIGURE 3-13 Crest curve with restricted sight distance.

94 DESIGNING SAFER ROADS Profllê of Vertical Crest Curve Slght dlstance LIne of slghl ln, Parabollc curve L wh€tE L = length of curve, g1 = percentgradeofapproachtangent, 92 = percent grads of exlt tangent, a = ls2-s11, hl = heightotdriver'seYe,and hz = helghtofoblect. FIGURE 3-14 Vertical curve geometry, sight distance, and terms. by AASHT0 to represent small animals, rocks, or other debris in the roadway, is inappropriate for intersection problems. Rather, the object height should be the height of passenger cars (AASHTO recommends 4.25 ft in passing sight distance calculations) or, if nighttime conditions are the worst c¿ìse, the height of passenger car headlamps. The critical review of restricted-sight effects revealed no empirically based, quantitative relationship describing the influence of sight restrictions at hill crests on highway accidents (21). In the absence of such a relationship, a theoretical model, based largely on professional judgment (accident data for crest curves are not available), was used in the safety cost-effectiveness analyses discussed in Chapter 5. This model, detailed in Appendix E, can be used by highway designers investigating potential improvements at specifrc sites. Careful attention must be given, however, to fully understanding the type and nature of hazard at the site, and thorough analysis of available accident daø is mandatory. The model assumes that accident rates at crest curves depend on the severity of sight resricdon, gradients on the approaches to the curve, and type ofhazard located in the sight-restricted area. The severity ofsightrestriction is measured by the difference between the speed at which vehicles operate on the curve and the speed at which, according to AASTITO procedures, they could safely stop before striking an unexpected object. As an example of expected improvement ât a crest having a 9 percent gradientchange andcontaining a "significant" hazard,as described in Appen-

SAFETY AND GEOMETRIC DESIGN 95 dix E, increasing ttre design speed from 35 ûo 45 mph by lowering the crest is expected to reduce the frequency of accidents about 15 percent on the 0.6-mi segment containing the curve. Climbing lanes for slow-moving vehicles can improve safety on crest cr¡rves with inadequate sight distance for passing. Marking no-passing zones with bottr pâvement markings and signs also improves safety at such loca- tions. Intersections On two-lane rural highways, intersections are ranked together with horizontal curves and bridges as the most likely locations for accident concenftation (22). According to National Safefy Council estimates, 56 percent of all urban accidents and 32 percent of all rural accidents occur at intersections (23). Although the average accident occurring at an intersection is not as severe as one occurring on the open road, there is nonetheless a concentration of severe accidents at intersections. Of all ttre fatal accidents in the United States, 29 percent of those that occur on urban highways and 16 percent of those that occur on rural highways are intersection-related. It is therefore logical for safety improvement programs to place special emphasis on these natural locations of accident concentration. Intersection improvements include changes to the physical elements of the intersecting roadways and operational measures for the control of raffic. These improvements generally focus on reducing conflict and improving driver decision making. Reducing approach speed and improving skid resis- tance can also be important. To achieve these objectives, intersection improvemenfs are tailored to each individual situation, due recognition being given to üafûc volumes on each of the intersecting roadways, prior accident pattern, physical cha¡acteristics of the site, and so forth. Useful procedures for selecting safety improvments at intersections include the following: ¡ Collision diagrams showing vehicle paths, time of occurrence, and weather conditions for individual accidents; o Condition diagrams showing important physical features that affect traf- frc movement at the intersection; and¡ Field review of the intersection to detect hazards not apparent from the collision and condition diagrams. Modeling the accident effects of specific design parameters is difflcult because of ttre large number of physical and operational features that affect highway safety at an intersection and regression-to-the-mean inaccuracies.

96 DESIGNING SAFER ROADS Also, improvements commonly add¡ess a number of intersection deficiencies simultaneously. Although simple relationships to predict the effects of specif,c intersection improvements are generally unavailable, a substantial body of information exists that designers use in remedying defrciencies at hazardous siæs (see Appendix G for a summary of physical and operational features affecting intersection safety). One researcher, for example, has concluded that accident reductions of 30 percent or more are possible at intersections wittr correc[able deûciencies such as poor sight disfance, inadequate signs and markings, and no channeüzauon, (24). Pavement Surface Condition Almost all RRR projects involve resurfacing or other pavement repairs. In addition to preserving the pavement structure and improving ride quality, resurfacing also has safety implications. Indeed, substantial controversy has surrounded this point with highway organizations arguing that routine resur- facing (without geometric improvements) enhances safety and safety organi- zations arguing the opposite (25). The potential effect of resurfacing on safety is a result of two factors working in opposite directions. First, resurfacing reduces surface roughness and improves ride quality, generally leading to increased average speeds. Second, resurfacing often increases pavement skid resistance, which reduces stopping distance and improves vehicle controllability when the pavement surface is wet. As part of this study, a reviey of available resea¡ch on the safety effects of resurfacing was conducted (26). This review supports the following tentative findings: o Routine resurfacing of rural roads generally increases dry-weather acci- dent rates by an initial amount of about l0 percent, probably because of increased speeds. Dry-weather skid resistance and stopping are unaffected by resurfacing unless the original pavement was extremely rough, so that tires did not maintain contåct with the paved surface. o Routine resurfacing of ru¡al roads generally reduces wet-weather acci- dent rates by an initial amount of about 15 percent. Apparently, this follows from improvements in wet-weather stopping distances and vehicle control- lability that more than compensate for any effects of somewhat higher speeds following resurfacing. o For most rural roads, the net effect of resurfacing on accident rates is small and gfadually diminishes with ûme. Initially, the total accident rate

SAFETY AND GEOMETRTC DESIGN 9'] typically increases following resurfacing, likely by an amount less ttran 5 percent. when averaged over the project life, the effect of resurfacing is much less. o Resurfacing improves the safety performance of roads that experience an abnormally high frequency of accidents in wet weather. Resurfacing projects provide the opportunity to correct deficient pavement cross slopes at little or no exFa cosfs. Correcting cross slopes allows better drainage of ttre pavement surface and improves vehicle control in wet weather. on individual resurfacing projects, careful attention to the removal of surface defects and necessary improvements to skid resistance, surface drainage, and superelevation may help offset the potentially adverse effects of increased speeds. Pavement Edge Drops Pavement edge drops (vertical discontinuities at the edge of the paved surface) result either from resurfacing activity unaccompanied by desirable shoulder improvement or wear or erosion of weak shoulder materials (Figu¡e 3-15). A particularly susceptible location of edge drops is the inside of horizontal curves, owing in part to the inward off-tracking of ttre trailing wheels of turning vehicles. RGIIRE 3-15 Pavemenr edge drop.

98 DESIGNINc sAFER RoADS Deøiled investigations of specifrc accidents have revealed the poæntially deleterious effects ofedge drops on vehicle safety. Although the daø needed to make reliable estimates of ttre frequency with which edge drop problems contribute to highway accidents are not available, some safefy resea¡chers believe that vertical discontinuities, particularly at the edge of a narrow üavel lane, pose a serious haza¡d to driverS who make Otherwise minor encroach- ments onto the shoulder. RRR projects provide an opportunity ûo correct pavement edge-drop prob- lems, and indeed FHWA procedures require that all such problems be cor- rected on iederai-aid RRR projects. However, rosuriacirrg raises the elevation of the roadway unless existing pavement materials are recycled. Thus, resur- facing can increase the likelihood that edge-drop problems will reoccur, particularly where shoulders afe constructed of earth, turf, or unbound gravel. Track tests, as well as theoretical studies of vehicle dynamics, have investi- gated the likelihood that drivers can safely recover once they have traversed edge drops of varying heights and shapes. As vehicle speed increases, the Vertical Shape Rounded Shape ø= Bevel ang¡e FIGURE 3-16 Pavement edge drops (cross-section views). Beveled Shape

SAFETY AND GEOMETRIC DESIGN 99 difficulty of successful recovery clearly increases. However, there are no generally accepted standards or guidellnes on the degree of edge drop, as chanctenzeà by height and shape (Figure 3-16), that constitutes an unaccept- able risk. To bener understand the hazard posed by edge drops, trre stuây commissioned additional experimental track testing (27) and a critical review of previous rcsearch (28). unlike much of ttre earlier track testing, the new tests used ordinary as well as professional drivers, testing vertical and beveled edge-drop shapes with nominal heights of 3 and 4.5 in. for ttre vertical shape and 4.5 in. for ttre beveled shape. The tests measu¡ed the frequency with which drivers, whose vehicles had their right tires off the pavement and scrubbing against the edge, could return to a l}-ft lane without intruding into the adjacent lane. This scrubbing situation is generally the most difficult for a driver to handle. Tþst results showed that ordinary drivers had more difficulty recovering from vertical drops than had been expected based on earlier tests using professional drivers. Edge drops on the order of 3 in. were generally negoti- ated adequately at speeds of about 30 mph in a large pâssenger car. However, tests on the same edge drop with small cars driven by a professional driver suggest that the safe speed would have to be lower, probably between 20 and, 25 mph. Successful recovery at the 4.5-in. vertical edge was impossible at almost any speed. Test results also showed that ftre shape of an edge drop affects the problems that many drivers experience in attempting to negotiate it. The use of a beveled edge, with a bevel angle of 45 degrees, was found to greatly reduce the control problems atnibuøble to edge drops. Both simulations and track testing require somewhat arbitrary assumptions about the nature of the inadvertent encroachments and the concept of a successful recovery, and neither addresses the frequency with which minor inadvertent encroachments can be expected. As a result, curent understanding of the edge-drop hazard is incomplete. In the interim, edge drops of any heighi or type must be considered potentially hazardous, and should not be built into the cross section as a result of either pavement surfacing or resurfacing. Combined Effects Estimation of the combined effects on safety of simultaneous improvements to fwo or more highway features is difficult in the context of RRR work because of problems in experimental design and model calibration @igure 3-17). The accident relationships discussed in this chapter address individual highway features and do not explicitly account for the possibility of mulriple alignment changes or changes in alignment coupled with changes in cross

100 DESIcNINc sAFER RoADs section. However, Appendix C contains a procedure that can be used to deveþ reasonable estimates of the combined effect of multiple improve- ments. LOW.COST SAFETY MEASURES Highway design practice provides a broad range of low-cost safety measures that can be used to ameliorate the geomegic defrciencies of existing highways.Ë-.^--r^^:-^1,.¡ôDÀ4rrrPrçù ¡¡¡v¡uuw Geometric Deftciency Low-Cost Safety Measure Narrow lanes and shoulders o Pavement edge lines o Raised pavement markers r Post delineators Steep sideslopes; o Roadside hazard markings roadside obstacles r Round ditches o Guardrail Narrow bridge r Traffrc control devices o ApProach guardrail o Hazard and pavement markings Sharp horizontal curve o Trafûc control devices o Shoulder widening ' Appropriate suPerelevation r Gradual sideslopes o Pavement antiskid treatrnent c Obstacle removal or shielding Poor sight distance at hill o Trafûc conüol devices crest o Fixed-hazard removal o Shoulder widening Hazardous intersection r Traffic control devices o Trafñc signalization r Fixed lighting o Pavement antiskid treatment o Speed controls Estimates of accident reduction factors (the percenøge decrease in acci- denß or in accident rates expected if the measure is applied) afe common for these measures. Many state highway agencies maintain lists of accident reducúon facüors for use in connecúon with haza¡d elimination programs, and a number of compilations from different sources have been published (29-34)' State highway agencies use accident reduction factors in selecting and

SAFETY AND GEOMETRIC DESIGN 101 FIGURE 3-17 Pavement edge drop, large Eee, and steep sideslope. programming safety improvements for locations where high accident frequen- cies have been observed. When federal aid from the hazard elimination program is involved, the total projected accident cost savings plus any opera- tional benefits, are expected to equal or exceed the required construction costs. Therefore, state highway agencies must estimate expected accident reducúions for each federal-aid hazard elimination project. The accuracy of the published accident reduction factors has been increasingly challenged (35,36). These challenges have arisen trecause most of these factors are derived from simple comparisons of accident da¿a for periods immediately before and after a measure is applied. The comparisons often a¡e biased toward overestimating accident reductions att¡ibuøble to applied measures because the sites at which measures are applied are not selected at random (sites with recent histories of abnormally high accident frequencies are usually selected) and control sites a¡e not used. From the standpoint of RRR work and safety cost-effectiveness, a further problem with published accident reduction factors is that they are generally insensitive to the degee of hazard. For example, a typical factor might indicate that the installation of warning signs at curves reduces accidents by 30 percent regardless of curve geometry. Although this factor mighe be appropriate for very sharp curves, it will overstate the safety effects of insølling warning signs at all curves with design speeds less than the speed limit.

LOz DESIGNING SAFER ROADS Recognizing the need for caution in the use of published accident-reduction factors, the study committee nonetheless concluded ¡trat the low-cost safety measures described can provide significant reductions in the frequency and severity of accidents. The safety benefits of these measures, coupled with their low costs, are such that the measures can be highly cost-effective on RRR projects. EFF'ECT OF CHANGING VEHICLE FLEET To assure safe use, highways must be designed and built to accommodate the types and volumes of vehicles in operation. Because important vehicle charac- teristics, such as average size, weight, performance, and crashworfhiness evolve over time, the compatibility between highway geometry and vehicle characæristics may be altered in ways that affect the severity and frequency of accidents. Because of these possibilities, the committee commissioned a review of new vehicle ftends and forecasts and an âssessment of whether fleet changes over the next 15 years are likely to cause fundamental changes in the current relationships between safety and highway design elements (37).Find- ings of this assessment include the following: o Automobile and light-truck characteristics. Changes in safety-related automobile and light-truck characteristics in the recent past were precipitated primarily by the fuel shortages and price increases of the 1970s. These forces have now diminished and are not likely to precipitate further change of signifrcant magnitude during the next 15 years. Gross vehicle weight. Sparked by gasoline shortages and rapid price increases in the 1970s (and resulting federal fuel economy standards), U.S. manufacturers inroduced substantially lighter automobiles and lighrduty trucks. Between 1978 and 1980, for example, average gross vehicle weight for new automobiles dropped from approximately 3,600 to 3,200 lb. Since then, average weight first søbilized and then increased slightly as a result of size increases in the smallest imported vehicles. Because fuel prices are expected to rise only gradually and further improvements in fuel economy are possible without weight reductions, no significant change in average weight is expecæd through 1990 and probably through 2000. Dimensions. Accompanying reductions in gross vehicle weight were reductions in average vehicle length and width. No further reductions in vehicle length and width are expected ttrrough 1990 and probably through 2000.

SAFETY AND GEOMETRIC DESIGN 103 Brøkíng. Possibly increased use of antilock braking systems in new automobiles could ultimately reduce average stopping distånces. Driver eye height. Average driver eye height decreased as vehicles were downsized. For some vehicles, the actual height is now below the va].oe (42 in.) assumed in AASHTO stopping sight disønce standards. However, no further reduction in average driver eye height is expected through 1990 and probably through 2000. underclearance. Reduced underclearances have also accompanied reductions in automobile size. underclearances as low as 4 in. a.ì no* common. This underclearance is less than the 6-in. object height assumed in AASHTO stopping sight distance policies. t Heavy-truck characteristics. The history of motor freight transportation reflecfs a continuing increase in ruck size and weight ovei time. Although continued upward adjustmenß can be expected as a result of st¿te legislative activities, no new regulatory initiatives are anticipated at the natioial level during the next 5 years. Dimensions. The surface Transportation Assist¿nce Act of 19g2 required states to permit longer, wider trucks on Interstate highways and major primary routes than were previously permitted by many iøæs. Now allowed nationwide are tractor semitrailers wittr semitrailei lengths of at least 48 ft (usually replacing 45-ft semitrailers), twin trailer trucks with two 28-fr t¡ailers (usually replacing the 45-ft semitrailer), and 102- in. wide rucks (replacing 96-in. wide trucks). Gross vehicle weight. Although the change in federal law did not generally increase truck weight limits, the increased volumetric capacity will increase average gross weights. Braking. New trucks are not expected to have impaired braking capability, provided that brakes are properly designed and adjusted. Handling- New trucks will perform differently from the trucks they replace, particularly with respect to off-nacking where tle rear wheels of a turning vehicle follow a different parh than its front wheels. The longer semitrailers experience more low-speed, inward off-tracking, and twins experience less inward off-tracking than the semirrailers* they replace. under high-speed conditions, twins exhibit greater outward off- tracking, and 48-ft semitrailers exhibit less outwa¡d off-tracking than the semitrailers they replace. o Patterns of use vehicle mix. on major highways during at least the past 20 years and possibly longer, trucks have comprised an increasing pioportion of total

104 DESIcNINc sAFER RoADS t¡affic. Although this trend may well continue into the future, the rate of increase is likely o diminish. Dffision of large-truck operations' The Surface Transportation Assistance Act of 1982 addressed the use of longer and wider trucks on a nationwide "system" of limited mileage, but it is expected úrat these larger trucks will be increasingly used on other highways as well. Growth of li7ht trucks. Sales of light rucks (pickups, panels, vans, and utility trucks) have grown relative to automobile sales. These light trucks have a higher center of gravity than automobiles, which may have implicaûons for barrier and guardraii designs' Although anticipated changes in the vehicle fleet, as previously sum- manzú, are likely to affect highway safety, the effects will be small and almost certainly mixed. For example, continued vehicular improvements, such as increased use of antilock braking systems, are expected to have positive effects on safety. At the same time, increasing disparity between the sizes and weights of automobiles and heavy trucks and the increasing proportion of heavy trucks on major highways may result in safety being degraded. These anticipaæd fleet changes may be sufficiently large to warrant reexamination of selected new design standards (e.g., stopping sight distance, which is influ- enced by both braking performance and driver eye height) and roadside hardware design (e.g., breakaway signs whose design is affected by weight of the wayward vehicle). However, the changes do not appear to be sufflciently large to measurably influence the relationships between highway safety and the roadway and roadside features affected by typical RRR improvements. Such relationships appear to be relatively insensitive to evolutionary changes in vehicle characteristics of the type experienced in the past and expected in the future. ROADWAY CONSISTENCY Unfortunately, safety relationships presented elsewhere in this chapter fail to capture situational influences present in the roadway environment that con- tribute gfeatly to roadway hazards (38-40). Illustrative of these particular hazards are high-volume intersections in isolated rural settings, sharp horizon- tal curves following long segments of generally straight alignment, and compound curves-contiguous horizontal curves tuming the same way-in which a flat curve precedes a much sharper one. Common to such situations is the violation of driver expectâncy: the unfamiliar or inattentive driver, lulled to complacency by the gentleness of the approaching roadway, is surprised by the sudden appearance of a potenúal hazard. The response is uncertain and

SAFETY AND GEOMEÎRIC DESIGN 105 slow, leading possibly to inappropriate maneuvering and an increase in acci- dent potential. In addition fo inconsistencies that occur at point or spot locations, such as illustrated previously, other situations are found in which clues from the physical environment belie the nature of the roadway hazard. perhaps the most important arises from possible incompatibilities between the roadway cross section and its horizont¿l and vertical alignment. In the case of a rgadway improvement, for example, upgtading cross-sectional elements with- oi.rt corresponding upgrading of alignment can result in an erroneous and potentially hazardous illusion of safety and the selection of operating speeds excessive for the critical alignment conditions. In general, the degree of hazard inherent in a specific feature, such as a narrow bridge, sharp curve, or a roadway without shoulders, depends not only on the feature itself but also on the nature of the nearby roadway environment. Although the safety effects of such interactions have not been quantified, it is quite likely ttrat expected safety gains, estimated by using the relationships described in this chapter, will understate achievable gains if roadway inconsis- tencies are eliminated as a part of the RRR improvement. useful techniques for eliminating spot inconsistencies or compensating for their potentially adverse effects include ¡ Provision of gradual geometric transitions appropriate to the anticipated vehicle operating speed; . Improvement of sight distance for early detection of the presence of the critical feature; . Provision of gentle sideslopes with few roadside obst¿cles at critical locations; and ' Installation of traffic control devices appropriate for the situation. SUMMARY Motor vehicle safety is critically influenced by the way highways are built. Improvements to the following design features on RRR projects are most likely to have significant and measurable effects on safety: o Lane and shoulder width and shoulder type, o Roadsides and sideslopes, o Bridge width, o Horizontal alignment, . Sight distance,¡ Intersections,

106 DESIcNINcSAFERRoADS o Pavement edge drops, and ¡ Pavement surface condition. For the first ûve features listed, evidence is sufficient to support quantitative estimates of the reduction in accident rat€s as a result of design improvements on two-lane rural highways. For intersections and pavement edge drops, the safety effects of design improvements are well-established, although quantitative relationships between accident rates and design improvements are not available. Design improvements on RRR projects can affect not only accident rates but also accident severity. Past research on the sâfety of the roadside environ- menf, for example, has produced roadside hardware that reduces the likeli hood that an accident will result in a fatality or a serious injury. However, most of what is known about the effects of design improvements on accident severity remains qualitative in nature. Pavement resurfacing without ottrer highway improvements will have a small negative effect on safety on most roads. On roads with unusually high percentages of wet-weather accidents, however, resurfacing may slightly reduce the total number of accidents. In both cases, the effects on safety are small initially and decrease with pavement wear. Although the safety effects of low-cost operational measures are frequently overståted in the literature, these measures can provide significant reductions in the frequency and severity of accidents. The safety benefits of these measüres, coupled wittr the low cost to implement them, a¡e such that the measures can be highly cost-effective on RRR projects. REFERENCES 1. C. V. Zegeer artd J. A. Deacon. "Effect of Lane Width, Shoulder Width, and Shoulder Type on Highway Safety: A Synthesis of Prior Literature." h TRB State-of-the-A¡t Report. TRB, National Resea¡ch Council, Washington, D.C. (forthcoming). 2. C. Y. Zegeer, J. Hummer, D. Reinfurt, L. Herf, and W. Hunter. Safety Effects of Cross-Section Design for Two-Lane Roads-Vol. I and II. Report FÉIWA-RD- 87/008 and 009. FHWA, U.S. Department of Transporøtion, 1986. 3. J. L. Graham and D. W. Ha¡wood. NCHRP Report 247: Effectiverwss of Clear Recovery 7-ones.'lRB, National Research Council, Washington, D.C., 1982, 68 pp. 4. P. H. Wright and K. K. Mak. "Single Vehicle Accident Relationships." Traffic Engíneering, Vol. 46, No. l, Jan. 1976, pp. L6-21. 5. D. E. Cleveland and R. Kitamrna. "Macroscopic Modeling of Two-Lane Rural Roadside Accidents." In Trarxportation Research Record ó81. TRB, National Research Council, Washingtor¡ D.C., 1978, pp.53-62.

SAFETY AND GEOMEIRIC DESIGN IO7 6. T. J. Foody and M. D. Long. The ldentiftcøion of Relationships Between Safety and Roadway Obstructiotts. Columbus Br¡reau of Trafûc, Ohio Departrnent of Transportation, J an. 197 4. 7. C. V. Zegeer urd M. J. Cynecki. "Deærmination of Cost-Effective Roadway Treatments for utility Pole Accidents." rnTransportation Research Record 920. TRB, National Research Cormcil, Washingron, D.C., 1984, pp.52-&. 8. C. V. Zngær and M. R. Parker, Jr. "Effect of Trafñc and Roadway Features on utility Pole Accidents." rnTransportation Research Record /7\.TRB, National Research Council, Washingtor¡ D.C., 1984, pp.65-76. 9. J. C. Glennon. NCHRP Report 148: Roadsíde Safety Improvernent prograrns on Freeways: A Cost-Effectiveness Priority Approach. TRB, National Research Council, Washington, D.C, 1974, & pp. 10. J. C. Glennon and C. J. Wilton. Effectiveness of Roadside Safety Intprovements: Vol. I, A Methodology for Determíning tle Safety Effectivercss of Improvemenfs on All Classes of Higlu,ays. Report FHWA-RD-75123. Midwest Research Instin¡te, Kansas City, Mo., Nov. 1974. 11. J. C. Glennon. NCHRP Report 214: Design and Traffic Corxrol Guidelines for Low-Volume Rural Roads. TRB, National Research Council, Washington, D.C., 1979,41 pp. L2. Guíde for Selecting, Locating, and Designíng Trafic Barríers. American Associa- tion of Søte Highway and Transporrarion Officials, Washington, D.C., L977. i3. T.C.Edwa¡dsetal.NCllRPAeportTT:DevelopmentofDesignCriteriaforSafer Lutninaìre Supports. HRB, National Research Council, Washington, D.C,, 1969, 82 pp. 14. K. K. Mak. "Effect of Bridge Width on Highway Safery: A Synthesis of prior Research." h TRB State-of-the-Art Report. TRB, National Resea¡ch Council, Washingûon, D.C. (forthcoming). 15. D. s. Turner. "Prediction of Bridge Accident Rates." rourrnl of rrarsportation Engineering, Vol. 110, No. l, American Society of Civil Engineers, New york, Jan. 1984. 16. J. C. Glennon. "Effect of Alinemenr on Highway Safery: A Synthesis of prior Research." In TRB State-of-the-Art Report. TRB, National Resea¡ch Council, Vy'ashington, D.C. (forthcoming). 17. J. C. Gle¡non, T. R. Neuman, and J. E. I-eisch. SSety and Operational Considera- tbns for Design of Rural Higlu,ay Curves. Report FHWA-RD-86/035. FHWA, U.S, Department of TrarLsportatior¡ Aug. 1986. 18. A Policy on Geom¿tric Design of Higltways and streets. Americar¡ Association of State Highway and Transportarion Ofñcials, Washington, D.C., 19g4. 19. K. Perchonak, T. A. Ranney, A. Baum, M. Srephen, and F. Dominic. Methodotogy for Reducing the Hazardous Effects of Highway Features and Roadside Objects. Report FHWA-RD-78?02. FHWA, U.S. Department of Transporrarion, May r978. 20. P.L. Olson et al. NCHRP Reporr 270: Paratræters Affecting Stopping Sight Distance. TRB, National Resea¡ch Council, Washington, D.C., June 1984,169 pp.

108 DESIGNING SAFER ROADS Zl. L C. Glennon. "Effect of Sight Distance on Highway Safety." In ÏRB State-of- the-Art Report. TRB, National Research Council, Washington, D.C' (forthcom- i"g)' 22. C.P. Brinkman and S. A. Smith. "Two-Lane Rural Highway Safety'" Public Roads,Yol.48, No. 2, Sept. 1984, pp. 48-53. 23. Accident Facts-1985 Edition. National Safety Council, Chicago, Il1. 24. D. Cleveland. "Effect of Intersection Safety Improvements on Highway Acci- dents: A Synthesis of Prior Research." In TRB State-of-the-A¡t Report. TRB, Nationat Research Council, Washington, D.C. (fortltcoming). 25. U.S. Congress. House of Represenlatives. Subcommittee on Investigations and Oversight, House Committee on Public Works and Transportation, Resurþcing, Restoration, Rehabilitation (3R) of Roads Other Than Freewals. Hear- ings ...gTth Congress, Sept. 17, Oct.27,28, Dec. 15, 1981 (SerialNo' 97-75). 26. D. Cleveland, "Effect of Resurfacing on Highway Safety: A Synthesis of Prior Research." In TRB State-of-the-Art Report. TRB, National Research Council, 'Washington, D.C. (forthcoming). 27 . P. L. Olson, R. Zimmer, and V. Pezoldt . Pavetnenf Edge Drop ' The University of Michigan Transportation Institute, Ann Arbor, 1986. 28. J. Glermon. "Pavemenlshoulder Drop-offs As They Affect Highway Safety." In TRB State-of-the-Art Report.TRB, National Research Council, Washington, D.C. (forthcoming). 29. J. A. Smithet øT.Identiftcatíon,QuanfiftcationandStructuríng ofTwo-LaneRural Highwq Safety Problems and Solutions, Vol. I. Report FHWA-RD-83/022. FHWA, U.S. Department of Transportation, June 1983. 30. Highway Safety Engirvering Studies. FHWA, U.S. Department of Transportatior¡ 1980. 31. T, Creasey and K. R. Ãgent. Development of Acciderú Reduction Facfors. Report UKTRP-85-6. University of Kentucky Transportation Research Program, Louisville, March 1985, 82 pp. 32. Accident Reduction Levels Wh'tch May be Attainable From Various Safety Improvenænfs FHWA, U.S. Department of Trarsportatior¡ Aug. 1982. 33. Roy Jorgensen Associates. Evaluation ofCriteriafor Safety Improvemenfs on th¿ Highway. Gaithersburg, Md., 1966. 34. Roy Jorgensen Associates. NCHRP Report 197: Cost and Safety Effectiveness oJ Highway Design Elemenrs. TRB, National Resea¡ch Corurcil, Washington, D.C.. 1978, 46 pp. 35. E. Hauer and J. [,ovell. "New Directions for Learning about the Safety Effect ol Measures." ln Tratuportation Research Record 1075. TRB, National Researct Council, Washington, D.C. 1986, pp.96-102. 36. K. S. Opiela. "Evaluating the Effectiveness of Highway Safety Improvements." Presented at the ASCE Nashville Conference on Highway Safery, Nashville. Tenn., March 1986. 37. W D. Glauz. "Effect of Possible Future Change to the Vehicle Fleet on Highwal Safety." In TRB State-of-the-Art Repor. TRB, National Resea¡ch Council, Wash' ingtorL D.C. (forthcoming). 38. C. J. Alexander and H. Lunenfeld. Positive Guídance inTraffic ConJrol. FHWA U.S. Department ol Trarsportatioa April 1975.

SAFETY AIID GEQMEÎRTC DESIGN 109 39. T. J. Post et al. A Users' Guìfu to Positfue Guidarce. FHWA' U.S. Deparrnent of Transportation, June 1977. 40. C. J. Messer, J. M. Mounce, and R. Q. Bracken. Hìghway Geometríc Desígn Coraistenry Retøted to Driver Expectansy, Yols. I and II. Report FHWA-RD- 81/035 and FÌ{WA-RD-81/036. FHWA, U.S. Deparünenr of Transporratior¡ April 1981.

4 Relationships Between Highway Costs and Geometric Design This chapter contains a discussion of the relationships between cost and key highway design features. More specifically reported are the added costs highway agencies incur on resurfacing, Iestoration, and rehabiliøtion (RRR) projects when existing road geometry is improved (e.g', through lane and shoulder widening or horizontal curve reconstruction). Along with the safety- geometric design relationships described in Chapter 3, these cost data are the key elements of the safety cost-effectiveness analyses presented in Chapter 5. In the sections that follow, the inherent problems of generalizing about highway cost ¿ìre reviewed, and typical RRR project costs obtained from selected states (Chapter 2) are reported. The relationships between RRR project cost and incremental improvements to existing road geometry are presenred followed by discussions of the right-of-way requirements for RRR projects and the longer term maintenance implications of improved highway geometry. COST RELATIONSHIPS-PROBLEMS AND LIMITATIONS The scope of geometric improvements included in a RRR project affects highway agency costs in two ways. First, it can increase or decrease the initial capital investment required for the design and consEuction of the RRR project, including right-of-way acquisition. These costs are referred to in this report as RRR project costs. Second, changes to highway geometry may affect longer term maintenance requirements. Initial RRR project costs tend to be the dominant consideration for highway agencies making decisions about geo- mefiic design improvements. In some cases, however, maintenance costs also 110

HIGTTWAY COS(S AND GEOMETRTC DESIGN 111 can be important, and neglecting to tâke them into account can result in poor design decisions. For example, placing a guardrail may appear to be an attractive alternative fo flattening sideslopes because the initial project cost is less; however, a different conclusion may be reached when the long-term cost of maintaining the guardrail is taken into account. Also, in snowbelt states the placement of quardrail might lead to higher costs for plowing and sanding because of problems with drifting snow. Although maintenance costs can be an important consideration in decisions about geometric design improvements for individual projects, ttre establish- ment of generalized relationships between maintenance costs and specific types of design improvements is very difficult. There are many different components of maintenance costs, some of which may be increased and others of which may be decreased by a specifrc type of improvement. For example, paving shoulders increases the amount of paved surface to be mainøined but may reduce pavement raveling by moving the pavement edge further from traffrc. Because of these difficulties, the cost relationships presented in this chapter do not include maintenance costs. Unfortunatel¡ a variety of factors make it diffrcult to relate RRR project costs to incremental geometric improvements in ways that could be used to calculate the potential cost impact of particular design standards. These factors include o Vqriuble site conditions. Costs for specific geometric improvements, such as widening lanes from 10 to 12 ft or increasing a horizontal curve radius from 500 to 1,000 ft, can vary g¡eatly depending on site-specifrc factors such as topography, right-of-way availability, and drainage requirements. o Variable labor and nwterial cos¡s. Costs for identical geometric improve- ments on similar sites can vary within a given state, as well as betwe€n states because of different construction unit costs. Labor costs in the San Francisco atea, for example, are nearly double those in Jackson, Mississippi (1). Even within a given state, cost variation can be substantial; a resurfacing or widening project in south Florida (Monroe County) may cost 40 percent more than an identical project in the Florida panhandle (2). t Variable design practlces. Highway agencies differ in the way they design pavements, shoulders, drainage structures and other items not directly related to geometry. For example, New York State routinely paves highway shoulders, whereas Virginia constructs gravel or turf shoulders; Ohio often applies l-in. overlays; Washington Søte applies 1.5- to 2.0-in. overlays. Such differences cause variations in cost even when the geometric improvements are the same, the sites are similar, and unit prices are identical. . Variable project scale. Unitprices for construction can vary depending on the quantity involved so that cost estimates based on average unit cost can

tt2 DESIGNING SAFER ROADS mask the effects of economies of scale related to geometric improvements. For example, review of sample RRR projects in \Mashingron S¡ate revealed excavation unit costs 10 to 15 percent less for projects where more than 10,000 yd2 of ea¡th were excavated compared with 3 projects where between 1,000 and 10,000 yd2 were excavated l3). Expticitly incorporating economies of scale inûo RRR cost estimates, however, is complicated by the general lack of quantity-sensitive unit cost datå and the reality that the scopes of RRR projects are often arbinary. For example, the length of a project may be based on the limits of the last resurfacing project or the level of funding set aside. D^^^.,oa mora.inl n,,ontitiao fnr recr¡rFe¡ins an¡l lqne nnd chnrrl¿le¡ wi¿leninsuwquùv r¡¡slv¡¡q Yse.!¡!¡vr " ^--"___Þ are related to project length, unit costs will increase or decrease depending on the length selected. o Complementary consÛuction requirements. In addition to promoting economies of scale, multiple geomeFic improvements on a RRR project may entail complementary consFuction requfuements that can further reduce incre- mentâl cost. For example, lengthening a crest vertical curve to improve sight disønce involves excavating earth that must be tlansported off-site if there a¡e no fill requirements for other improvements. The material could, in some cases, be used for flattening sideslopes or widening shoulders. . Scale of historical cost data. Highway agency cost datâ are often avail- able at levels that are either too coarse or too deøiled to be directly useful for estimating the effect of specifrc geometric improvements on RRR projects. At one exEeme, costs may be available on a per-mile basis for different project categories; for example, resurfacing and widening projects on fwo-lane rural arterials in a given state might cost $250,000/mi. Such estimates are useful for early budgeting and programming, but shed no light on incremenfal costs for widening 10-ft lanes to 12 ft. At the other extreme, most state highway agencies maintain records on unit. costs for hundreds of construction items such as excavation, subbase, or guardrail installation. These data are used in conjunction wittr quantity estimates to develop detailed project cost estimates. However, estimating the effect of specific geometric improvements on quan- tities for each construction item can be a difficult and time-consuming task. o Project costs bias. Cost estimates for incremental geometric improve- ments based on typical project costs can be biased because they exclude projects rejected because of high cost. Hence, if flattening horizontal curves typically costs $150,000 each on a completed RRR project, úre cost of flattening all similar curves would probably be higher. Because of these factors, no single set of universally applicable cost relationships can be developed for incremenøl geometric improvements. At best cost relationships can be developed that indicaæ the added costs for geometric improvements on RRR projects typical of a particulâr ståte or

HIGTIWAY C)STS AND GEoMETRIC DESIGN 113 region. such relationships can be used in conjunction with accident relation- ships to examine the safety cost-effectiveness of particular geometric improvements in a state or region. TYPICAL RRR PROJECT COSTS As described in Chapter 2, the typical federal-aid RRR project involves pavement resurfacing, often with minor lane and shoulder widening or road- side improvements. occasionally, spot improvements to vertical or horizontal alignment are undertaken as paÍ of a RRR project. Similar projects funded without federal aid tend to be less complex, constructed to lower geometric design standa¡ds, and more often involve resurfacing without widening or other geometric improvements. In a number of states, however, state-funded RRR projects are generally of the same scope as those funded using federal aid. Generally, RRR projects entail costs in the following work categories: c Site preparation and earthwork: clearing, tree removal, excavation, placement of new embankment, and grading; o Drainage: construction of ditches, drains, culverts, and other minor structües required for drainage; c Pavem¿nt: all pavement consFuction-subbase, base, and top course- on lanes and shoulders; ¡ Structures: rehabiliøtion or replacement of bridges and larger culverts; . Traffi.c and safety: placement of permanent traffic control devices, light- ing, signing, fencing, striping, markers, and similar driver aids; c Trafrc control: temporary Eaffic control measures during construction; o Miscellaneous: other items such âs utility pole relocation or curb and sidewalk construction; . Right-of-way: purchase of land or use e¿ìsemenfs; and o Engineering, mobilization, and other: engineering design and oversight, allowances for contractor front-end costs to assemble necessary equipment and personnel on-site (mobilization), and ståte sales taxes (if applicable). For minor widening and resurfacing projects in washington State, for exam- ple, the pavement category accounts for 45 percent of all costs exclusive of righrof-way (Table 4-1). Engineering, mobilization, sales tax, and con- tingency allowances account for an additional 24 percent, with the remaining costs divided equally among the other categories. In Florida, however, where the terrain is much flatter, the pavement category accounts for an even greater share of project costs (76 percent).

II4 DESIGNINGSAFERROADS TABLE 4-1 Percent of RRR Costs by Category for Typical Resurfacing and Minor Widening Projects Category Washington Florida Illinois Site preparation and earthwork Drainage Pavement Structures Tiaffic and safety; traffic control Miscellaneous Engineering, mobilization, and other Tota! Shoulder type l0 5 65 I 2 76 9 9 45 ll 2 24 i00- Paved 8 l3 T00 Paved 5 6 9 r00- Unpaved NorEs: Right,oÊway is excluded. Category includes highway with I I -lt lanes and 3-ft shoulders widened 1o l2-ft lanes and 4-ft shoulders. Federal-Aid Project Costs In the 15 case study states, the cost of resurfacing and minor widening projects (including resurfacing without widening) using federal aid ranged from $z:,OOO io $557,000/mi, almost an eightfold difference (Table 4-2). This TABLE 4-2 Resurlacing and Minor Widening Construction Costs for Different Funding Sources State Funds Only (cost/mile, $thousands) Federal Aid (cost/mile, $ thousands)State Medium and Thick Overlaysu Thin Overlays and Seal Coatsá Arrzona California Florida Illinois Michigan Mississippi Missouri New Hampshire New Jersey Ohio Tèxas Virginia Washington 222 202 156 225, 196 122 oo 411 430 '73 551 2'76 89 198 r30 206 148 l9 lt il ll 47 35 ).o-48 Norr: Dashes indicate data not availabìe souncn: compiledfromstatehighwayagencyñscalyearconstructionprograms;fiscalyearsvaryfrom 1983 to 1985ì còsts generally exclude right-of-way acquisition and engineering' aOverlay thickness usually greater than 3/4 in. 1'Overlay thickness usually less than 3/4 in. cExcludes lnterstate transfer project in Chicago area. l I ì

HIGTIWAY COSTS AI,ID GEOMEÎRIC DESIGN 115 variation results from many of the factors noted previously-including dif- ferences in topography, scope of geometric and roadside improvements, and usual pavement and shoulder design practices. Urban system RRR projects are generally the most costly on a per-mile basis, followed by primary syst€m projects, with secondary system projects costing less than the other two. This sequence holds in flve of the six case study ståtes for which comparable data are available for all three systems (Table 4-3). The exception is New Jerse¡ where costs for primary system projects exceed those for urban systems, possibly because the primary system is heavily urban (nearly 50 percent) in New Jersey. TABLE 4-3 Resurfacing and Minor widening construction costs for Different Federal-Aid Systems Average Cost per Mile (g thousands) State Primary Secondary Urban Combined Calilo¡nia Michigan Missouri New Jersey Ohio Texas t73 138 81 243 54 328 203 195 95 603 83 521 237 229 150 431 92 |,255 202 t96 99 430 73 557 Norr: Federal-aid projects only. SouncE: Compiled from state highway agency frscal year construction programs; fìscal years vary from 1983 to 1985; c-osts generally exclude right-oÊway acquisition and enginèering; for cáse study slates reporting data for all three systems. Includes some nonfreewa¡ multilãne projðcts. Non-Federal-Aid Project Costs Resurfacing and minor widening projects funded without federal aid are less costly on average than those funded with federal aid. Average costs reported for such projects in five of the case study states range from g89,000 to $206,000/mi, or between 40 and 98 percent of the average per-mile cosrs of projects funded with federal aid (Table 4-2). Some states also repair pavement surfaces by using thin overlays (usually less than % in. thick) or seal coâts, both of which are ineligible for federal aid and therefore must be funded exclusively from ståte or local sources. Such

116 DESIGNINGSAFERROADS projects usually involve little work and cost only about one-tenth as much as federal aid resurfacing and minor widening projects. ADDED PROJECT COSTS FOR GEOMETRIC IMPROVEMENTS The frgures reported in the preceding section illustrate average costs of RRR projects and the variation of costs by staæ and highway system. These data do not indicate the proportion of RRR spending related úo geometric improve- ments, nor do they indicate how RRR costs might be aifecte<i by design standards that require changes to existing highway geometry' To explore the safety cost-effectiveness of design standards, such informa- tion is required, ideally as relationships between incremental project costs and incremental changes in highway geometry that can be generally applied in all RRR project situations. However, for reasons summarized earlier in this chapter no such generally applicable relationships exist. Therefore, studies of cost and safefy trade-offs must rely on more limited approaches to cost estimation. Usually cost. relationships are tailored ûo a particular state's experi- ence; the relationships may not be directly applicable to any given project but nevertheless illusftate the added project costs that are incurred when specific geometric improvements a¡e made. Resea¡chers use two general approaches for developing such relationships. The fust, an engineering approach, assumes hypothetical project conditions and required improvements, estimates the construction quantities (e.g., cubic yards of excavation, tons of asphalt) needed to make the improvement, and translates these quantity estimates into costs using typical unit prices from earlier contract experience. The second approach, a more statistical one, relies on fecords from a sample of actual projects and attempts to relate the variation in costs between projects to differences in project scope, including geometric improvements. The engineering approach clearly links costs to speciûc design differences, but the hypothetical conditions selected may or may not be representative. The statisúcal approach, on the other hand, is more broadly represenfåtive of actual project experience, but does not have such a clear, reproducible link between specifrc design features and cost. To estimate relationships between RRR project costs and geometric design, both approaches were used in this study. Where reliable data on project costs and scopes were available for certain features, statistical relationships were used' Elsewhere the study relied on illustrative engineering estimates. RRR project costs are summarized next for resurfacing-only projects and for the added costs of lane and shoulder widening, flattening sideslopes, removal of roadside obstacles, bridge widening, horizontal curve reconsEuc- tion, and lengthening crest vertical curves.

HIGTTWAY COSTS AI,ID GEOME-TRIC DESIGN LI1 Resurfacing Only For projects that involve only resurfacing and no geometric improvements, suffrcient daa a¡e generally available ro relare typical project costs within a state to the repaving width. For two-lane rural roads, such relationships are available for the three case study stâtes (24) in the sysúemwide cost-effective- ness analyses presented ln Chapter 5. Simila¡ relationships have been developed in Kentucky (5) and, Wisconsin (6); data from Kentucky are presented here along with data from the three case study states (Table 4-4). Because topography and tenain have littlê or no effect, resurfacing-only project costs are reasonably consistent for the four states, ranging from a low of $102,000/mi in lllinois to a high of $134,000/mi in Florida for a rwo_lane highway wittr ll-ft lanes and no shoulders (Table a-a). The 30 percent spread between the states is related primarily to diffe¡ences in pavement unit costs TABLE 4-4 Resurfacing and Incremental widening costs - Two-Lane Rural Hiehways (2-6) Added Cost per Mile ($thousands) Resurface Two I I -ft Lanes, No Shoulders Florida Illinois Kentucky Washington Widen Lanes I ft in Each Direction Florida Illinois Kentucky Washington Widen Paved Shoulders 1 ft in Each Direction Florida Illinois Kentucky Washington Widen Unpaved Shoulders I ft in Each Direction Florida (sodding) Illinois (crushed stone) Kentucky (earth or turf) Washington (crushed stone) State t34 102 105 r22 20 32 38 44 t4 l0 I 8 (includes some dense graded) 2t 4 4 l4 8 (new construction) NorEs: Costs include allowances lor engineering (7 percent) and mobilizatio¡ and minor roadside,ìntersection,andsåfetyimprovementstypicallyapaitófaresuilacingproje"t.corti^ãuJ¡riìJto tgg3 levels using FHWA highway construction cost indices. Right-oÊwãy aóquisition costs äre'exclude¿.

118 DESIGNING SAFER ROADS and pavement overlay thicknesses (always greater than % in. and usually in the 1.5- to 3.0-in. range). Lane and Shoulder Widening The added costs for lane and shoulder widening on resurfacing prcjects vary more than the costs for resurfacing only because of ttre effects of terrain and topography. In the four stâtes given in Table 4-4,Íhe added costs for widening ianes by i fi in each ciireciion rangss irom $20,ûÛûin¡i in Fioriria o $44,000i mi in Washington Statô, more than a twofold difference. Widening paved shoulders costs about one-half as much as widening lanes because widening at the shoulder edge requires less complex construction methods and thinner pavements are used. If shoulders are widened but left unpaved, shoulder widening costs are reduced by about 50 percent (Table 4-4). Flattening Sideslopes Flattening sideslopes at spot locations on RRR projects is relatively common, but the costs vary widely depending on the scope of the sideslope improve- ment and site conditions. Earthwork is ttre principal cost component of sideslope flattening and involves excavation (on or off the site), ûansport, placement, and compaction of earth to build up slopes on fill sections (see Figure 4-1). Flattening sideslopes on cut sections involves excavating the existing sideslopes and removing the excavated soil or rock. Earthwork unit prices for excavation, transport, placement, and compaction range from $2 to $12iyd3 depending on the total quantity involved (larger quantities have lower unit prices), source of the material (on or off-site), amount of transport involved, and ttre nature of the material (rock, clay, loam, FIGURE 4-1 Flattening sideslopes on fill cross sections. Should€r edge

HIGÍMAY COSTS AI,ID GEOMETRIC DESIGN 119 etc.). To illustrate ttre sensitivity of sideslope flattening costs to the slope geometry of fill sections, the sideslope flattening costs were compared (i.e., added RRR project costs in terms of cost per mile) assuming earthwork at $Z yd3. These added costs are moderately sensitive to trre ãriginal and new slopes, but are very sensitive to the fill heighr (Table 4-5). For example, flattening a2:l onginal slope to 4:1 costs $9,000/mi with a frll height of 2 ft and $34,000/mi with a fill height of 4 ft. TABLE 4-5 lllustrative costs for Flattening Sideslopes on Fill Sections Construction Cost per Mile (gthousands) for One Side of Highway Original Slope New Slope Height = 2lt Height : 4fr Height : 6fr Heighr : gfr 5 9 l3 5 9 l8 5 t4 2:I 3:l 2:l 3:l 4:l 3:l 4:l 6:l 4:l 6: I t'7 2? 50 l8 34 67 l8 5l 37 74 110 38 '74 t47 39 TT2 66 r30 t94 61 l3l 2s9 68 t96 NorEs: Based on an ea¡thwo¡k-unit price of $/ydl; seeding at $ 1,000/acre; drainage at l5 percenf of earthwork costs; pÌus 25 percent for engineering, mobilization, and miscellaneous expinses. Ni r.ight-oÊ way costs are included. Removal of Roadside Obstacles Extending the width of the recovery zone beyond the roadway surface can lead to improved highway safety. To accomplish this, objects can be removed, relocated, or protected. The associated costs of these actions have been well- documented in past studies and by several state highway agencies (Table a-6). Guardrail installation costs range from about $10 to g2O/inear ft with no apparent economies of scale. Tree removal, however, does exhibit rather high fixed costs; the unit cost of removing only a few hees is quite high. The costs ofrelocating signs and utility poles are highly variable, in part because right- of-way requirements differ from site to site. Bridge Widening As for many other types of improvements, the added costs for bridge widening can be estimated from statistical studies of past construction costs or from

TABLE 4-6 Unit Costs for Selected Roadside Obstacle Removal and Protection Strategies (7-10l Unit CostTypeofAction (1985 $) Sourçe Remarks Guardrail Removal 1.65/lincar ft (8) 2.86/linear ft Vr'ashington State DePartment of TiansPortation Installâtion I l.O0/linearft (8) i r ^^ ¡r:---^- r. f,Í^-¡^6^ nâhõrrñâñt ^.1áìtinnal (7'11 t t '¿vt.tltca. " *';'¡ir';;öï;t'ü"' "Ëäh'."d " treatment 10.00/linearft (7) 2 t. l6llinear ft Washington State DePartment of TiansPortation Replacement 25.00/linear ft New York State Includes removai Department of olold rail and Tiansportation some bridge rail Bridge rail end trõatment 6,300/bridge Montana Department New rail costs oftansportation same as guardrail 5,280/bridge (7) Tiee removal 238/tree NewYork state Based on removal DePartment of of 184 trees on a TiansPo¡tation single Project 2641Íree (7) 2\Oltree (8) 660/tree (9) Less than 100 no.nmarketabie trees removed 85/tree Ø More than 100 nonmarketable trees removed Utility pole relocation Wood/telephone 405/pole U0) Rrtral WoodTlow'power 1,490/po1e (10) Rural Wood/high power 2,66}lpole (10) Rural Non*ooð z,oEO/pote Qq Rural All types 1,585/Pole (7)2,580/Pole (9) Sign relocation 195/sign (S) "Small" sizes Breaka.way sign 210/sign (9) installation Impact atteûuator - :i'riiãriãii". 4,400/unit (S) Sand-ñlled type Nore: 1985 costs câleulated using FHWA composite construct-ion indices'

HIGTIWAY COSTS AND GEOMETRIC DESIGN I21. engineering analyses. Statistical data a¡e commonly used for planning studies, and widening costs are usually expressed in terms of dollars per square foot. The influences of site conditions on the costs of many widening projecs are minimal and are not considered in ttre analysis. Because widening of existing bridges has not been a common feature of RRR projects, cost estimates must rely on daø collected for other types of project funding. Typical costs (in 1985 dollars) for widening existing structures and install- ing new bridge rail in Washington St¿te ranged from $125Æt2 for "minimal" widening (up ø 5 ft) to $87Æt for "moderate" widening (10 or more fr). This range suggests the existence of a fixed-cost component not influenced by the amount of widening, such as the costs of installing bridge rail, and possibly economies of scale as well. An inverse relationship, captüing these effects and showing reduced sensitivity of unit construction costs to each foot of widening as the totâl amount increases, was fit to these two points and yielded the following cost model: Ç= where c equals the unit cost ofbridge widening in dolla¡s per square foot and w equals the width to be added to rhe bridge (in feet). In calibraring this expression, minimal widening was assumed to be 4 ft and moderate widening, 12 ft. A project that involved widening a bridge 8 ft would thus be expecred ro cost about $97/f&. other cost estimates obtained from the literature (7, It) ue somewhat lower but reflect the variability expected among the stâtes. These per-square-foot esúmates, when factored to 1985 dollars, were $45 for Colorado, $145 for Illinois, $70 for New York, and $85 for virginia (II).Later,other researchers estimated bridge widening costs to range from about $66 to $B2lftz, again expressed in 1985 dollars (7). Alrhough on the high side, the washington State model as expressed by Equation 1 is not substantially different from other estimates and is favored because of the sensitivity it expresses between unit construction costs and the amount of widening. Many existing highway bridges cannot be widened because of the bridge type (e.g., through trusses) or structural inadequacy. If additional bridge width is required, the only remedy is to remove the existing structure and replace it with a new, wider one. using data from washington State (expressed in 19g5 dollars), estimated costs for such ân altemative are as follows: øo +ff (Ð i

r22 DESIGNING SAFER ROADS Costtf? (8) Remove existing sfructure Construct new sEucfure P¡esEessed concrete girder Concrete posttensioned box girder Steel plate girder Short span concrete slab Horizontal Curve Reconstruction Flattening a horizontal curve, which increases its length, usually requires full reconsmr;úon between the beginning and ending points of the curve. Thus, while RRR projects generally do not entail reconstruction, it will be required for a portion of the project length if a curve is to be flattened. t-ittle has been published about the costs of reconstructing horizontal curves. Because curve reconstruCtion is relatively uncommon and costs a¡e highly dependent on site conditions, state highway agencies generally do not accumulate enough experience to søtistically relate average curve reconstruc- tion costs úo changes in curve geometry. Hence, data on typical relationships between cost andiurve geometry based on completed projects ale not avail- able. Nevertheless, the cost to flatæn a horizonøl curve is related to the original curve geomegy (central angle and degree of curvature) and the change in that geometry [i.e., degree of curvature of the reconsEucted curve (see Figure 3-11, Òhaptet 3)1. In general, feconstruction costs increase as either the curve cental angle, the original degree of curvature, or the change in degree of curvature increases. At a parúcular curve on an existing roadway (with a given cengal angle and initial degree of curvature), the cost to reconstruct the curve will increase as the change in degree of curvature increases. To illustrate the order of magnitude of reconstruction costs and ttre sen- sitivity ¡o curve geometry, a cost relationship was developed for hypothetical conditions using Washingfon State unit costs (Appendix I). For a central angle of 50 degrees and initial degfee of curvature of 15 degrees, this relationship predicts ttrat it will cost about $150,000 ûo reduce the degree ofcurvature to 10 ã"gt"".. To reduce it to 6 degrees, the cost increases to about $225,000 (Figure 4-2).In this case, ttre initial l5-degree curve would accommodate a maximum design speed of 36 mph, whereas the lO-degree curve would be suiøble for a design speed of about 45 mph, and the 6-degree curve would accommodate a design speed of 55 mph. These calculations ale based on American Association nf Søte Highway and Transpofation Officials' (AASIITO) new construction standards for a superelevation rate of 0.08 (11). 6 56 69 106 56

HIGTIWAY COSTS AND GEOMETRIC DESIGN I23 qooog 3ooF U' oor- 200(J u¡ oÍ r00 v:¿46910 NEW DEGREE OF CURVATURE ll¡r 70 60 NEW DESIGN SPEED, mph g;*pE5r**i;$¿,g5;*','''ruff'r'}d'"l*":" FIGURE 4-2 lllusradve horizontal curve reconstruction costs. Lengthening Crest Vertical Curves As is tle case with reconstructing horizontal curves, little empirical informa- tion is available about costs for improvements to vertical curves. The most common safety-motivated improvement to vertical curves is lengthening crest curves to increase sight disønce. Even though such improve*ãnts ar" ,ore common on RRR projects than horizon¿al curve reconstruction, the practice is still not so common that highway agencies have accumulated enough experi- ence to relate average costs to changes in geometry. At a given location, however, fhe cost of lengthening a crest curve will be strongly influenced by existing curve geometry and tle changes to that geometry. In general, costs for lengthening a crest curve increase as the change in curve length increases, as the difference between approach and exit grades increases, and as the initial length of the curve ¿ecreaies (see Figure 3-14, Chapter 3). To illustrate the order of magnitude of costs involved and the sensitivity to curve geometry, a cost relationship was developed for hypothetical conditions based on washington State cost daø (Appenãix I). As an example of cosr effects, consider a 2b-ft curve at a location where the differenie berween Central angle : 50 degrees D = orlglnal degree ol curye DS = orlglnal deslgn speed, mph l\";:r_s- 0

r24 DESIGNING SAFER ROADS approach and exit grades is 4 percent. The maximum design speed for such a curve is 35 mph (Table 4-7). According to the cost relationship, lengthening the curve to 400 ft with a maximum design speed of 43 mph would cost about $68,000. længthening it ¡o 800 ft, which would allow a maximum design speed of 53 mph, would roughly double the cost to $136,000' If the grade difference was 6 percent instead of 4 percent the costs would be about 15 percent higher. TABLE 4-7 Illustrative Costs for Lengthening Crest Vertical Curves to Increase Stopping Sight Distance lnitial Curve Improved Curve AASHTO Design Speed mpho AASHTO Design Project Length, Speed Cost ft mpho ($thousands) 200 200 400 400 600 200 200 400 600 600 4 4 4 4 4 6 6 6 8 8 Difference in Grades Length, ok lt 400 800 800 r,000 800 800 1,400 800 1,000 r,800 NorËs: Costs are based on unit prices lrom Washington State for two-lane rural roads (see Appendix l)l schematic ofvertical cu¡ve geometry is shown in Figure 3-14 Chapter 3. 'AASHTO maximum design speed based on stopping sight distance criteria (12 p' 308)' Intersection Improvements Costs for intersection improvements are highly variable, depending on the physical and operational features to be improved and other site-specific conditions. Representative costs for widening and new channelization of an existing intersection are $100,000 to $150,000 (Table 4-8). Improvements such as the construction of new turning lanes and realignment of curbs Íue leSS costly-typically $10,000 to $20,000 per intersection. Also, some intersection improvements, such as rechannelization using pavement markings and upgaded traffic control devices, can be implemented for much less than $10,000 per intersection as part of a RRR project. 35 35 43 43 48 3r 3l 38 39 39 43 53 53 57 53 47 56 47 46 55 68 136 t26 165 109 t51 28 l 144 198 399

HIGHWAY COSIS AI'ID GEOMEÍRIC EESIGN TABLE 4-8 Representative Costs of Intersection Improvements (14) Type olProject Construction Cost (1983 $) Widening and new channelization Installation of new traffic signals Reconstruction of one approach Construction ofnew turning lanes Realignment olcurb l 00,000- 1 50,000 60,000- 100,000 50,000-75,000 r 0,000-20,000 10,000 RIGHT-OF.WAY REQTIIREMENTS As noted in Chapter 2, most federal-aid RRR projects do not involve right-of- way acquisition. Existing rights-of-way are often large enough to accommo- date minor lane and shoulder widening and some roadside improvements. Of the 15 søte highway agencies visiæd for this stud¡ only Virginia routinely acquired additional right-of-way for RRR work. In Virginia, right-of-way acquisition is necessary on some secondary highways because narrow existing rights-of-way are common, sometimes as n¿¡rrow as 30 ft. Although many geometric improvements can be completed within existing right-of-way, others cannot. I¿ne and shoulder improvements occasionally require additional right-of-way as do roadside improvements such as slope flattening or removal of ftees and other ûxed objecs. Reconstruction of horizontal curves almost always requires additional righrof-way. Søte highway agencies generally resist acquiring right-of-way for RRR projects because right-of-way acquisition will be costly and time consuming, or it will adversely affect aesthetics and the community. Financial Costs The ûnancial costs of right-of-way acquisition include not only land purchase prices, but also adminisnative costs associated with negotiation and condem- nation. As a result, right-of-way costs can include a fixed per-parcel compo- nent and a variable component based on parcel size and location. Washington State, for example, reported a fixed cost of approximately $10,000/parcel and a variable component of approximately $5,000/acre in rural areas and $50,000 to 100,000/acre in urban areas. Time Costs Righrof-way acquisition often requires considerable time, in some cases 2 years or more. In Florida the state highway agency allows 18 months for land t25

126 DESIGNING SAFER ROADS acquisition. Such a requirement conflicts with RRR project schedules, which are usually geâred toward providing urgent pavement repair within one year. In many stâtes the delay in project scheduling is as important as ûnancial costs for avoiding right-of-way purchases for RRR improvements. Aesthetic and Community Impacts Removing Fees or taking portions of homeowners' front yards is unpopular ^-r --- .--i+., ^--^^i+:^- +^ DDì) ìr¡¡¡n"ama¡fc À l+hn¡¡ah thacaalru ç¿1Ir aluuJç lurrrr¡lu¡lrLJ uPlruùruvrr w ruu\ u¡rPrvYv¡¡¡vr¡Lù. tuu¡vuór¡ q¡vùv possibilities arise more often in wban and suburban settings, they occur in rural areas as well. Some state highway agencies have issued specific policies on circumstances warranting tree removal. Geometic improvements on RRR projects are rarely foregone because of negative aesthetic and community impacts alone; however, in conjunction wittr financial and time costs they form an additional barrier to rightof-way purchases. MAINTENANCE COST IMPLICATIONS The effect of geometric improvements on maintenance costs and requirements has been considered in only a general way in a few prior studies. Ståte maintenance records generally have been underutilized, and it is only since tle advent of pavement management systems that they have been used to relate maintenance activities and costs to pavement condition. Future use of these records may lead to a better underst¿nding of the effects of geometric design on maintenance costs, but in the meantime no reliable quantitative relation- ships are available that can estimate added maintenance costs that are a result of geometric improvements. Nonetheless, improvements that directly affect the amount of roadway surface can be expected !o increase maintenance costs. Other improvements, such as sideslope flattening and widening roadside clear zones, will not have a substantial effect on maintenance requirements. Maintenance requirements can be expected to increase as highways are widened and as more surface areâ must be maintained. Although some economies of scale can be real:øed, in maintaining the additional atea, they are probably small, so that as an approximation, lane and shoulder maintenance costs âre roughly proportional to surface area. The added maintenance costs that result from cross-section improvements are small relative to either the total maintenance costs or the capiøl costs of widening. For example, consider the routine maintenance costs on the follow- ing hypothetical RRR project:

HIGTTWN COSTS AND GEOMETRTC DESIGN I27 Exístíng Conditíons Requíred Improvemcrús Length = 2.0 mi Lane width = 11 ft Lane width = 12 ft Shoulder width = 3 ft Shoulder width = 4 ft Rolling ærrain Paved shoulders Consfuction costs for hypothetical RRR project ile as follows: Cost andImprovemenÍ Costlft-mi ($) Lane widening 89,200 (22,300) Shoulder widening 42,400 (10,600)Resurfacing 280,000 (5,000) Total 411,600 A¡mual construction costs (7 percent discount rate over a 3O-year project life) 33,200 Annual routine pavement maintenance costs at $250Æt-mi are as follows: Increase as Percenf ofNo With Arnualízed Improvemcrt Widening Increase 4o Consfruc-(8) (8) (8) Increase tion cost 14,000 16,000 2,000 14.3 6.0 Norr: Construction unit costs were derived empirically from case study state project data. Pavement maintenânce costs are from ¡.he Washingtør Søte Pavement Management System, Alttrough maintenance costs on this segment will increase 14 percent because of additional surface area, this increase represents less than 10 percent of annualized project construction costs. Data from the Highway Performance Monitoring System (IIPMS) were used to obúain a rough estimatê of the nationwide effect of cross-section improvemenfs on maintenance costs. These estimates are given for two-lane rural federal-aid highways:

r28 DESIGNING SAFER ROADS Lanes Shoulders Total State Local Total Existing Surface fft-mì)" 11,681,000 5,146,000 16,827,000 Added Surface Area If Improued Percent(fr-mi)o lrrcrease 432,0æ 3.7 150,000 2.9 582,000 3.5 1984 Maíntenance Cost (6thousands)" r,329,773 629,063 i,958,836 Cost Increase Resultíng From Improvernent ($thousands) 46,542 22,017 68,559 Source: Analysis of 1983 Highway Performance Monitoring Sys- tem data base and Highway Statistics 1984. "Average lane width = 10.9 fq average shoulder width = 4.8 ft; total mileage - 536,415. \¡ tSlS FIIWA RRR standards are ap,plied. "Includes costs for facilities, slructurcs, and snow and ice control only. State and local governments spend nearly $2 billion on maintenance activities on two-lane rural federal-aid highways Q3).If hnes and shoulders of these highways were to be widened to the special RRR ståndards proposed by FHWA in 1978, an increase of 3.5 percent would have to be maint¿ined in the surface area. Assuming that maintenance costs rise proportionally with sur- face area, an extra $68 million would have tn be expended nationwide. However, because cross-section improvements would be made over a period of time, maintenance costs would increase only slightly each year; mainte- nance costs would increase by approximately $3.4 million/year over 20 yeats. Maintenance costs can be an important consideration in choosing among alternative roadside improvements. For example, flattening sideslopes becomes more attractive as an alternative to placing guardrail when the costs of maintaining the guardrail a¡e taken into account. Alignment improvements can also affect highway maintenance costs. For example, alignment improvemenfs that require pavement reconstruction at curves will reduce maintenance costs in the short term because newer pave- ments cost less to maintain.

HIGTTWAT COSTS AND GEOMETRIC DESIGN r29 SIJMMARY Geometric design improvements on RRR projects affect both the initial capital investment required for the project and future maintenance require- ments. Although initial project costs tend to be the dominant consideration for highway agencies making decisions about geometric design improvements, maintenance costs can also be important, and neglecting to tâke them into account can result in poor design decisions. _ The added cost for specific design improvements was found to vary widely from project to project because of variations in site conditions, ùbor and materials costs, design practices, and project scale. In the case study states, federal-aid resurfacing and minor widening proJect costs ranged from $73,000 to $557,000/mi. RRR projects on urban systems are generally the most costly, followed by primary and seconda¡y system RRR projects. Resurfacing and minor widening projects funded without federal aid are, on average, less costly than those funded with federal aid, although in some states the difference is small. unit costs were assembled for resurfacing-only projects and for la¡\.e and shoulder widening. Differences in resurfacing costs rìmong states were the result of differences in unit costs for materials and the thickness of typical overlays. Larger state-to-sfate variations were found in the unit cost for lane and shoulder widening because of the effects of terrain and topography. unit costs were assembled for flattening sideslopes and removing ioadside obstacles. sideslope costs were found to be highly sensitive to nil heights. costs for both types of improvements vary considerably depending on whether acquisition of righrof-way is required. Rough, quantitaúve relationships were developed for estimating costs for reconstructing horizontal and vertical curves and widening bridges. The rela- tionships account for economies of scale in these improvemenis and for fhe considerable effects of central angle (horizontal curves) and grade (crest curves) on costs, REFERENCES 7. Dodge Guide to Public works and Heavy construction cosfs. McGraw-Hill, New Yo¡k, 1985. 2. unpublished cost estimates from [,ong Range cost Estimation procedure, Florida Department of Transportation, Tallahassee, 1985. 3. unpublished data from a sample of RRR projects provided by washington State Department of TransportatiorL Olympia 1985. 4. unpublished cost estimates provided by Illinois Department of rransportatior¡ Springfield, 1986.

130 DESIGNINGSAFERROADS 5. C. V. Zegeer,R. C. Deen, and J. G. Mayes. "Effect of Lane and Shoulder Widths on Accident Reduction on Rural Two-Lane Roads." lnTrarcportation Research Record 806,TR8, National Research Cotmcil, Washington' D.C" 1981' pp. 33-43. 6. An Evaluation of Alternate Shoulder Wídth Standards for New Cotstructìon on Wisconsin's Two-Lan¿ State Truck Highways. Summary Report. Wisconsin Department of Transporøtioru Madisoru Feb. 19M. 7. S. A. Smitlu et al. Identification, Quantífication ard Stucturing of Two-Lane Rural Híghway Safety Problems and Solutions. Report FHWA/RD-83/022. FHWA, U.S. Department of Transportation, June 1983. 8. Cost-Effective Cross-Section Design for Two-Lane Roads' FI{WA' U.S. Depart- ment of Transportation, (tbrthcoming)' g. LL. Graham, and D. W. Harwood. NCHRP Report 247: Effectiveness of Clear Recovery Zones. "lB,National Research Council, Washington, D.C.' May 1982' 10. C. V. Zegeer and M. R. Pa¡k, Jr. Cost-Effectivencss of Counfermeasutes for Utility Pole Accidents. FFIWA, U.S. Department of TransportatiorL Jan. 1983. 11. J. C. Glennon. NCHRP Report 148: Roadside Safety Improvemerú Prograrns on Freeways: ACost-Effectivercss Priority Appro¿cå. TRB, National Research Coun- cil, Washington, D.C., 1974. 12. A Poticy on Geometric Design of Higltways and Streets. American Association of State Highway and Transportation Ofûcials, Washington, D.C.' 1984. 13. H iglw ay St atistícs,1984. FHWA, U'S. Department of Transportation. 14. T. R. Neuman. NCHRP Report 279: Intersection Channelization Design Guide. 3. TRB, National Resea¡ch Council, Washington, D.C.' 1985.

5 Safety Cost-Effectiveness of Geometric Design Standards The safety cost-effectiveness of design standards was analyzed for resurfac- ing, restoration, and rehabiliøtion @RR) projects on the basis of relationships between safety and geometric design (chapter 3) and the relarionshþs between cost and geometric design (chapær 4). The princþl emphasis in this chapter is on the safety effects of design improvements. Howevei operational effects, such as user travel time and operating cost savings as a resuliof design improvements, are also considered. cost-effectiveness analyses were conducted at both the project and system level. The objective of the project-level analyses was to establish the circum- stances under which a given design improvement is cost-effective by examin- ing cost versus safety trade-offs for a series of representative proþts. This information can be used in setting minimum standards or esøbiishing design practices so that opportunities for cost-effective improvements are carefully considered by designers. systemwide (or statewide) safety gains and costs of geometric improve- ments were also considered in order to add¡ess the overall financial and budgeary consequences of design standards and the rade-off between pave- ment condition and safety tfrat is cenEal to the debate over RRR standards. In essence, the systemwide approach addresses two questions: From a safety standpoint, are certain kinds of geometric improvements to existing highways worth making throughout a highway system? At current funding levðls, can state highway agencies afford these improvements while mainøining pave- ments in reasonable condition? 131

I3Z DESIGNINGSAFERROADS EARLIER STI.JDIES OF SAFETY COST.EFFECTIVENESS IN HIGHWAY DESIGN Existing highway design standards-both for new construction and for RRR work-generally are not linked to explicit assessments of cost-effectiveness. Although the standards reflect judgments about the effects of changes in design variables on safet¡ traffic operations, consfuction cost, and other considerations, they afe not the products of formal cost-effectiveness studies in which these effecS are quantified and rade-offs are analyzed. State high- ----- ----^:^^ ¿L^ A-^-:^^ñ ,{¡¡,-iatinn af Qrare lfichw¡v qn¿l Transnnrfationway a8,frrrçlcùr Lllç rul¡çrlvdr ôùùw¡sLw¡¡ offrcials (AASHT0), the Federal Highway Administration (FIIWA), and others who adopt søndards or recommend design guidelines have not relied on cost-effecúveness analyses for several reasons: o Surprisingly little is known about how much accident rates will decline with improvements in road design. Researchers have often reached conflicting conclusions, and explicit, widely accepted quantitative relationships ¿ìre not available from the literature (Chapter 3). o The cost effects of design stândards vary widely because of differences in site conditions, unit costs, design practices, and project scale (Chapter 4). o Conclusions ftom a cost-effectiveness analysis can be very sensitive to discount rates, imputed accident costs, imputed value of time, and other assumptions made to facilitate comparisons of benefits and costs. o Even if the costæffectiveness analyses produce the results intended, questions remain about how these results should be used in se$ing standards. What other factors should be considered in addition to cost-effectiveness, and what are the relative weights of these factors? How should the underlying uncertainties of cost-effectiveness analysis be øken into account? Because of the preceding problems, researchers who have analyzed the safety cost-effectiveness of highway improvements generally have sought to provide designers with methods that can be used to produce more cost- effective designs, rather than a better basis for setting standards (1-7). I-eisch and Neuman (8) explicitly examined the cost-effectiveness of new construction ståndards for locally adminisæred highways in Minnesota. They argued that design standa¡ds afe ". . . one of the most important tools the trigtrway engineer hâS . . ." and ". . . limited construction and maintenance budgets and increased public pressure on the engineer to justify the expendi ture of public highway funds demand these standards ¡o be cost-effective." Other researchers have presented the development of cost-effective designs for individual projects as an altemative to designing to standards. For exam- ple, Jorgensen Associates û) describe the objective of their study as develop-

C O ST.EF F ECTN EN ESS O F DESIGN STAN D ARD S 133 ing a methodology for "tailoring the designs of individual projects rather than developing designs.through rigid application of design standards." Similarly, Graham and Harwood 13) state "the cost-effectiveness of roadside design improvements can vary widely between highway sections based on accident rates, traffic volumes, tenain, required construction quantities, unit con- struction costs, and right-of-way requirements. There is a clear need for a roadside design process bæed on cost-effectiveness considerations rather ttran a single, fixed roadside design policy." The question of customized design (versus reliance on stândârds) depends on the extent to which safet¡ cost, and other impacts of a given design improvement vary from site to site. Customized designs are appropriate when site-to-site variations are large and the factors influencing tiese variations a¡e not easily represenúed in design stândards. Standa¡ds are appropriate when little site-to-site variation exists or when facúors influencing these variations (e.g., traffic levels) are easily taken into account when applying the søndard. In any case, it is likely that design standa¡ds will remain in some form and it is reasonable to question the cost-effectiveness implications for highway sys- tems as a whole. To the extent that the authors of the preceding studies examined or com- mented on standards, they did so with respect to new construction standards rather than RRR sfanda¡ds. Both Jorgensen Associates (1) and, Leisch and Neuman (8) noted the strong sensitivity of cost-effectiveness to traffic volume and implied that new consfuction design standards might be too stringent for low-volume rural highways and perhaps not stringent enough for nonfreeway highways with high traffic volumes. SCOPE AND FRAMEWORK OF COST.EFFECTIVENESS ANALYSES The principal measure of cost-effectivness used in this study is added RRR project cost per accident eliminated. Improvements to geometric design for a given RRR project generally reduce the number of accidents and increase the cost of the project. cost per accident eliminated shows the added cost required to eliminate one accident. Cost per accident eliminated is calculated as follows: o Estimate ttre change in accident rate for the design improvement under consideration and, given average daily traffic (ADT), the change in the number of accidents per year. o Estimate the added cost required to implement the design improvement as part of a RRR project.

r34 DESIGNING SAFER ROADS . Annualize the added cost, based on an assumed project life and discount rate. . Calculate cost per accident eliminaæd as added cost divided by accidents eliminated. For simplicity, cost-effectiveness analyses presented in this chapter assume constant trafûc volumes expressed in terms of average daily traffic. As calculated in ttre preceding steps, cost per accident eliminated does not include the benefits to highway users associated with travel time and operating cosr sawinss Theqe. lp.nefifs c¡n he accounted for in the cost Der accident eliminated framework by subracting them from the cost required CI imple- ment a design improvement and calculating net (consruction minus user) cost per accident eliminated. However, a shortcoming of this method is that if user savings exceed the cost to implement the improvement, the net cost per accident eliminaæd is negative and therefore not meaningful. The advantage of using cost per accident eliminaæd as the measure of safety cost-effectiveness is that it focuses directly on the safety versus cost trade-off and allows designers or policymakers to impute their own values úo accidents eliminated. To use this information, however, users must understand the underlying disribution of accidents by severity (e.g., for the type of accident under consideration, how many faølities and injuries are there per thousand accidents?). A disadvantage of this approach is that it understates safety beneflts when design improvements act principally to reduce the sever- ity of accidents, rather than the number of accidents, and overstates safety benefits when design improvements result in higher speeds and increased accident severity. In both c¿ìses, cost per fatal or injury accident eliminated may be a more appropriate measure of safety cost-effectiveness. Past studies of the cost-effectiveness of highway safety improvements differ considerably in the value imputed to eliminating different types of accidents, particularly fatalities (Table 5-l). The National Highway Traffrc Safety Administ¡ation (NHTSA) uses the future economic production of individuals (the amount of compensaúon individuals would have received had the fatality not occurred) in assigning cost ûo a faølity, whereas the National Safety Council (NSC) uses production minus consumption. The "willingness- to-pay" approach uses estimates of the amounts individuals would pay for small reductions in the probability of death. When the values for different severity classes are weighted together using a typical disribution of accidents by severity for two-lane rural highways, the resulting average value per accident eliminated ranges from $10,000 to $50,000 in 1985 dollars. A speciñc value is neither recommended nor applied in this study but it is concluded that improvements witfr a cost per accident eliminated less than $10,000 are clearly justifled on safety cost-effectiveness

CO ST.EF F ECTN EA¡ ESS O F DESIGN STAN DARD S TABLE 5-l Alternative Estimates of Accident Costs by Severity Cost per Accident ($thousands 1985), Severity ofAccidents NHTSA NSC "Vy'illingness- Approachó Approach. to-Pay"Approachd Fatal Injury Property damage only All accidents on twolane rural highways" 394.6 lt.t 1.4 t6.9 256.5 13.2 t.2 13.3 |,348.7 r0. r t.9 45.3 aCosts are updated to 1985 using the implicit price delìator for gross national product.óCosts are based on the 1983 NHTSA ieport'(9). costs are p.o,-uid"d by incident-fatality, injury, or property damage involvement. The unit costs perincidentwere converted to unit costs per acòident using national level summary data on accidents and incidents provided in the report. 'NSC costs are based on Estimating the Cost ofAccidents I9S4 (il1).dThe willingness-to-pay estimates aie from Kiagh et al. (t l). eCosts were calculated using a distribution ofaccidents by severity lrom Smith et al. (4/. The distribu- tion-3 percent fatal, 37 percent injury, and 60 percent property damage only - is for twolane rural highways with 400 ro 2,000 ADT grounds whereas improvements with a cost per accident eliminated greater than $50,000 are not justified solely on safety cost-effectiveness grounds. Improvements with cost per accident eliminated in the $10,000 to $50,000 range may or may not be wafÏantÊd depending on (a) the specifrc value impuæd to accidents eliminated; (å) uncerøinties sunounding estimates of the added cost for an irnprovement and the number of accidents it will eliminaæ; and (c) other factors (e.g., environmentâl effects) that are not accounted for in the calculation of cost per accident eliminated. The relationship between cost-effectiveness analyses using cost per acci- dent eliminated and the benefit-cost ratio approach employed in other studies is illustrated in Appendix J. In a benefit-cost ratio approach, the analyst imputes dollar values to accidents eliminated. The numerator of the benefit- cost ratio is the sum of annual safety and operational benefrts to highway users for the design improvement under consideration. The denominator is annualized construction cost. for the improvement. The benefit-cost approach permits a direct comparison of safety benefits with costs and other consequences of design improvements that can be valued in dollar terms. Further, benefit-cost can account for changes in accident severity because separate unit costs are used for accidents by severity class. Benefit-cost results, however, will be of limited value to users who disagree subsøntially with the dollar values imputed for accident costs. cost per accident eliminaæd (rather than the benefit-cost ratio) was selected as the principal measure of cost-effectiveness to avoid the arbitrary impuøtion of dolla¡ values to accidents eliminated. Had a benefit-cost approach been used instead, ûndings about cost-effectiveness would not change, provided 135 i Ì i

t36 DESIGNING SAFER ROADS that the dollar value impuæd to accidents eliminated falls within the $10,000 to $50,000 range. Economic Assumptions For the calculations of cost per accident eliminated presented in this chapter, a discount rate of 7 percent and a project life of 30 years ale assumed. The 7 percent discount raæ splits the difference between the 4 percent rate recom- mended by AASHTO (i2) ior iow-risk investmenß and the i0 percent rate recommended by the Office of Management and Budget in Circula¡ A-94. The fixed project life of 30 years was assumed for simplicity. The useful lives of safety improvements vary considerabl¡ and in many cases will be less than 30 ye¿fs. For a hypothetical safety improvement with a cost per accident eliminated of $10,000 at a discount rate of 7 percent and a project life of 30 years . Increasing the discount rate to 10 percent would increase cost per accident eliminated to $13,200 (about 30 percent); ¡ Decreasing the discount rate to 4 percent would reduce cost per accident eliminated to $7,200 (about 30 percent); o Increasing project life oo 40 years would reduce cost per accident elimi- nated to $9,300 (about 7 percent); and . Decreasing project life ¡o 20 years would increase cost per accident eliminated to $11,700 (about 15 percent). SAFETY.COST TRADE-OFFS This section contains an examination of safety-cost trade-offs for lane and shoulder widths, horizontal curvature, roadside obstacles, sight disønce on vertical curves, and bridge width-the key design features for which quantitå- tive safety relationships have been developed for two-lane rural highways. Formal analyses of safety-cost rade-offs "vere not conducted for traffic controls and other low-cost safety measures because of the lack of reliable quantitative relationships for estimating the safety effects of these measures. However, the committee found sufficient information to conclude that such measures can provide significant reductions in the frequency and severity of accidents (Chapter 3). The safery benefrts of the measures, coupled with their low costs, a¡e such that the measures can be highly cost-effective on RRR projects.

COST.EFFECTNP¡,IESS OF DESIGN ST}WDARDS 137 Lane and Shoulder lVidths of all the highway geometric features considered in this stud¡ lane and shoulder widths are the most amenable to quantiøtive analysis. Although the caveats presented about uncertainties in accident relationships and costs apply to lane and shoulder widths, more is known about the safety effects and costs of these features. Also, costs for lane and shoulder widening vary less from siæ to site than do costs for other feârures so that findings ãboui th.i, .ort- effectiveness are more easily generalized. Better system-level data are available for existing lane and shoulder designs. Thus, the number of highway miles above and below a given lane and shoulder width ståndard could be determined because øø on these design features are reported in ståte highway inventories. such systemwide com- parisons between standa¡ds and actual conditions are generally not possible for horizontal curves and sight distance on vertical cu*es becåuse standards for these features are usually expressed in terms of design speeds, and information on design speods for individual curves is not available in most søte highway inventory data sets. P roj e ct-Level C o st-Effe ctive ne s s Important determinants of the cost-effectiveness of lane and shoulder widen- ing are o Traffrc volumes, o Roadside environment, ¡ Terrain and highway alignment, and o Lane and shoulder widths before improvement. Traffic volumes are a prime consideration in the cost-effectiveness of lane and shoulder widening (Figure 5-l) because the number of accidents elimi- nated by lane and shoulder widening increases almost in proportion to ADT, whereas costs are not signitcantly affected by ADT. current RRR standards frequently place little weight on ADT. In some súates, lane width standards for RRR projects are based solely on design speed. This may result, for example, in widening 10-ft lanes to ll fr on a highway with 300 ADT but not widening 1l-ft lanes ro 12 ft on a highway with 5,000 ADr even rhough rane widening on highways with high ä¡r'n* u much lower cost per accident eliminated.

138 DESIGNING SAFER ROADS " o 1,000 2,000 3,000 4,000 5,000 6,000 7,000 AVERAGE DAILYTRAFFIC NOTES: Exmple assures rolling t€rain, 1o-ft lanes wjth 2-ft shoulders before improverent and 1l-ft lanes w¡th 4-tt shoulders aiter improverent Costs are ¡n 19d5 dollars and wer€ calculated using a dis@unt rate of 7 percenl and a prol'ect l¡fe of 30 Years. FIGURE 5-1 Cost-effectiveness of lane and shoulder widening by ADT. The roadside environment is an important consideration in the cost-effec- tiveness of lane and shoulder improvements because roadside haza¡ds such as steep slopes and fixed objects affect both the likelihood that an accident will occuf and the severity of the accident. In the relationship befween accident rates and lane and shoulder width described in Appendix C, the degree of hazardassociated with the roadside is represented by a roadside rating ranging from 1 (least hazardous) to 7 (most hazardous). A roadside hazud rating of 1 might involve a flat roadside with no fixed objects wittrin 30 ft of the oußide edæ of the shoulder. A roadside râting of 7 might involve a steep downward slope or a sheer rock wall along the edge of the shoulder. For highway ,.gr"ntr with roadside ratings of 6 or 7, the cost per accident eliminated is less than one-half ttrat for sections with roadside ratings of 1 or 2 (Figue 5-2). Tenarn and highway alignment are also important considerations in the cost-effectiveness of lane and shoulder improvements' Costs for lane and shoulder widening on more rugged terrain generally increase, because of the greater amount of earthwork required. The accident reduction associated with iane and shoulder widening is also affected by terrain and alignment because more rugged terrain and poor alignment means higher accident rates and thus ø8sooe oH40 = =teo l-z lrJ9zooo fE Ë10F UI o()^

COST-EFFECTNENESS OF DESIGN STANDARDS 139 6oo &40 o t¡lF Êrof lr¡ Fz ot 2o õ o ffro(\ Fa o o Least Most Hazârdous ROADSIDE HAZARD RATING Hazardous NOTES: Exarìpte assures 2,OOO ADT, rolting tera¡n, 1O_ft lanqs with 2_ft shoutders before improv€rent and 1f _fl lane¡ with 4_ft shoulders afterjmproverent. Costs ar€ in lggs doltars and w€rs calcularea usin! À ãìscount rale of 7 percent and a projæl lile of 30 years. FiGURE 5-2 Cosreffectiveness of lane and shoulder widening by roadside hazud rating. the reduction in accidents as a result of a given lane and shoulder width improvement is greater.t The study examined the cost-effectiveness of adding lane or shoulder width using typical unit costs and accident rates for flat, rotting, and mountainous terrain (Table 5-2). on flat terrain, shoulder wideni'g is l'ore cost-effective than lane widening because of the relatively low cosl of widening shoulders on flat terrain. on mountainous terrain, however, the cost-effectiveiess of lane and shoulder widening are neæly equal. The cost-effectiveness of lane and shoulder widening diminishes as widths approach the levels mandated by new consrucdon standards (Figure 5-3). Although the cost for widening lanes from l0 to 1l ft is about the sÀne as the cost for widening lanes from ll to 12 ft, the number of accidents eliminated is less in the latter case. Lane and shoulder width improvements result in time and cost savings to lAccording to the accident relationship for lane and shoulder width, the percenøge improvement in accident rate is constant for given lane and shoulder width irnprovements. aaãing i f, oilun" width, for example, always causes a r2.l percent decrease in accidents. For highway Jections withpoor alignment, the base accident rate is higher so that a given percentage reduction in accidents results in a higher absolute reduction.

TABLE5-2IllustrativeCost-EffectivenessoflaneandshoulderWidening, 2,000 ADT Tèrrain Lanes Paved Shoulders Unpaved Shoulders Added Cost pe¡ Mile of Widening (l ft each direction) Flat Rolling Mountainous 31.6 44.6 68.2 14.0 21.2 44.6 4.6 I 1.8 35,2 ¡ -^:r^-+- El:*:-^r-Á ^ar h¡lilc ner VearâuuluçlrL5 Lr¡¡l¡¡r¡urvÚ Pvr 1r^rrv P-' ^-*- FIat Rolling Mountainous 0.086 0. r09 0.172 0.058 0.073 0.r15 0.048 0.061 0.097 Cost per Accident ElimiÍated lor Widening Flat Rolling Mountainous 35.3 33.0 3 1.9 19.6 23.4 31.2 7.7 15.5 29.3 Norns: costs are in thousands of 1985 dollars. unit costs for widening are basg! on unìt costs fori;ãi"iäñ;;ã;i.'*t.o in chãøl'i, i"ui" ¿-+; t" calcuratecostper"*"lg:i1:ITl::I*^1i1t-*f .å,i, *.o un"""iize¿ using a z peicent áiscount rate and a project life of30 years; theexample assumes l0-lt lanes and 2-ft shoulders belore imprÔvement' 9-0to 10-2 10-2to 11-4 11-4to'12-6 LANE AND SHOULDER WIDTH IMPROVEMENT(fi) NôTFq F,âmnle assmes 2.ooo ADT and rolllng to(a¡n Cosls are in thousands oitããã ¿oir"i"'un¿ *ere calculaled us¡ng a d¡scount rate of 7 percent and a pro¡ect life of 30 },ears. FIGURE 5-3 Cost-effectiveness of successive lane and shoulder width improvements. øoooI Aso¡- = =!! zoF tuo õoÍto ulÈ t-tt o

CO ST.EF F ECTN EN ESS O F D ES IGN STAN DARD S r4t highway users, as well as improved highway safety. Narrow lanes and shoul- ders on two-lane rural roads cause motorists to drive closer to vehicles in the opposing lane. Motorists must compensate for driving closer to opposing traffic by slowing down and allowing larger headways between vehicles in the same lane. Thus, motorists driving on roads with narrow lanes and shoulders will experience more delay and will d¡ive at lower speeds than on roads with wider lanes and shoulders. Speed increases as a result of lane and shoulder width improvements offset part of the safety benefit of these improvements because, other things being equal, accident rates increase with speed. This speed-relaæd effect is already accounted for in the accident relationship for lane and shoulder width because the relationship provides estimates of the net effect on safety of width improvements and related speed increases. T\e Highway Capøcity Mønual (13) provides factors for adjusting the capacity of two-lane rural highways to account for lane and shoulder width. This study estimated the effects on fravel times of improvements to lane and shoulder widths based on the Highway Capacity Manuql methodology (Appendix K). At ADT levels less than 2,000, the time savings associated with lane and shoulder width improvements are minimal @igure 5-4). These savings increase sharply with increasing ADT, however, because the effects of narrow lanes and shoulders are exacerbated when traffic volumes are greater. The AVERAGE DAILYTRAFFIC NOTE: Example assumes lo-tt lanes wìth 2-tt shoulders before ¡mproverenl and 11-ft lanes wilh 4-ft shoulders atter ¡mprov€mnt. FIGIIRE 5-4 Travel time savings for lane and shoulder widening. tr ¡¡l> 1.600É. l¡l o. t¡JJ r.zoo =.cc ulÀ o 800 IJJ U)g 400 f o I 0

142 DESIGNING SAFER ROADS effect of a given lane and shoulder width improvement on travel time will also vary depending on terrain because curves and grades add to the adverse effects of na¡row lanes and shoulders on vehicle operating speeds. Travel time savings for the first few feet of widening are usually substan- tially greater than those for further widening (Figure 5-5). For highways with 12-ft lanes and Gft shoulders, the time savings for further widening are minimal. tt;t-l" troi.loto trtfu" LANE AND SHOULDER IMPROVEMENT (ft) NOTE: Example assumes 2,o0o ADT and rolling terrain, FIGURE 5-5 Travel time savings for successive lane and shoulder width improvements. Part of the travel time benefit to users associaæd with wider lanes and shoulders will be offset by ttre higher operating costs associâted with higher travel speeds. This point is illustrated in the following table, which gives the travel time and operating cost effects of an increase in speed from 50 to 55 mph. User Costs per 1,00A Vehicle Míles Travel Operating Total Titttc Cost User (at $750thr) (in 1985 $) Cost ($) 50 mph 150.00 55 mph L36.36 Difference 13.64 t90.72 340.72195.15 331.514.43 9.21.

COST.EF FECTMENESS OF DES¡GN STáNDII.RDS In this example, about one-third of ttre travel time savings are offset by greater vehicle operating costs. At greater ADT levels, the net user savings-time savings less added vehicle operating costs---can be substantial in relation to the cost of lane and shoulder widening (Figure 5-6). At 5,000 ADT, for example, the user savings are about 40 percent of the widening cost. At lower ADT levels, however, the user savings are small; for example, less than 2 percent at 1,000 ADT. AVERAGE DAILY TRAFFIC i "li:;":iillf ,î:äiå13;T,l:üiäilå;Xi:iii:fi "i*"",'il'i:Iliff ''lime savings are valued at $7.50 per vehicle hour. FIGURE 5-6 User savings as a percent of the cost for lane and shoulder widening. Sy s tem-Lev el C o s t-Effe c tive ne s s At the system level, the study examined the cost-effectiveness of alternative minimum lane and shoulder width standards using data on two-lane, rural federal-aid highways from the FtIWAs Highway Performance Monitoring System (HPMS), as well as more deøiled data from the roadway inventories of Florida,Illinois, andWashington State. The standards are minimum accept- able lane and shoulder widths for a highway segment, given the segment's traffic volume, design speed, percentåge of trucks, and highway functional classification. Designers compare the standards with existing lane and shoul- der widths to determine whether the existing widths can remain in place or whether they must be upgraded. t43 lt o t- ul() ,- u=o.z <qø=<= thæ eç5r-<Ø@OEOt¡¡ uJu, Í.lt- 0

144 DESIGNING SAFER ROADS Alternative standards examined at the national level include . Application of AASHTO new construction standards to RRR projects; ¡ RRR standards developed by AASHTO in 1977 [Geornetric Design Guide for Resurfacing, Restoration, and Rehabilitation of Highways and Streets (14),rcfened to as the AASHTO Purple Bookl; . Standards proposed by F[{WA in 1978 (but not adopted); and r A modifred version of the 1978 FTIWA søndards in which the ADT ranges for riggering lane and shoulder width improvements a¡e shifted and wide¡ shoulde¡s æe required at tle. highest ADT levels (Table 5-3). The three alternative standards examined at the national level are described in Appendix L. TABLE 5-3 FHWA Proposed Standards for Lane and Shoulder Widths With Modifications l0 Percent or More Trucksr' Less Than l0 Percent Tiucks Design Year Running Volume Speed'(ADT) (mph) Combined Lane Lane and Shoulder width width. Combined Lane Lane and Shoulder Width Widrh. l -750 7sr 2,000 Over 2,000 Under 50 50 and over Under 50 50 and over All II 1l "Highway segments should be classified as "under 50" only ifmost vehicles have an average speed olless than 50 mph over the length ofthe segment. /'For this comparison, trucks are defrned as heavy vehicles with six or more tires. 'One loot less for highways on mounlainous terrain. The lane and shoulder width standards currently used by states generally fall between the special RRR standards proposed by FTIWA in 1978 and AASI{TO new construction standards (Chapter 2). Once it has been determined that a given highway segment must be upgraded because it does not meet minimum standards, state highway agen- cies will frequently improve the segment beyond ttre minimum standards up to the levels suggested by AASFITO for new construction (Chaprer 2). To estimate accidents eliminated and cosfs for the four sets of minimum stan- dards less stringent than AASHTO new construction st¿nda¡ds, assumptions about the extent of this practice are needed. Because highway agencies are more likely to make improvements beyond the minimum standards when t0 t2 l0 t2 I I 13 t2 15t2 l8 It t2 t2 t4 T] 9 l0 10

CO ST.EFFECTNENESS OF DESIGN STANDARDS 145 traffic levels are greater, this study assumed for simplicity thât segments not meeting minimum standards would be upgraded to la) minimum standa¡ds if ADT is 2,000 or less and (b) new construction standards if ADT is greater than 2,000. In practice, site-specific considerations such as congestion levels and the roadside environment will govern whether and by how much improved cross sections will exceed minimum standards. At the national level, the cost for lane and shoulder widening under AASHTO RRR sr¿ndards is about 20 percent of the cost under AASrnro new construction standards (Table 54). The number of accidents eliminaæd is less than 20 percent, however, so that AASHTO RRR standards perform poorly in terms of cost per accident eliminated. This occurs because, unlike other standards, AASIITO RRR standa¡ds do not vary with traffic volume. As shown in the project-level analysis, cost per accident eliminated decreases .lgptv with increasing ADT. AAsIrro RRR standards are not as safery cost- effective overall because they do not distinguish between roads with low and high ADT. The cost for lane and shoulder widening under the l97g FHWA proposed stândards is about 45 percent of the cost under new construction standards. The number of accidents eliminated is about 40 percent, however, so that improvements under the FHWA proposed søndardi are somewhat less cost- effective than under new construction standards. Three modiûcations were made to the 1978 FHWA proposed standards in order úo improve their cost-effectiveness. First, t}te ADT ianges for specific lane and shoulder widths were shifæd. The original FHWA proposal defined certain lane and shoulder widths for 1 to 400 ADT, 401 to ¿,ooo eor, -o TABLE 5-4 cost-Effectiveness of Alternative Lane and Shoulder width Standards, Nationai Level Analysis Using HpMS Data Land and Shoulder Width Standarcl Accidents Cost per Year Eliminated($millions), per Year Cost per Accident Eliminated ($thousands) AASHTO new construction standards AASHTO RRR standards 1978 FHWA proposed standards 40.6 56.0 43.4 31.5 ¿Costs are in 1985 dolÌars and were calculated using a 7 percent discount râte and a project life of30 years.álhe t9z8 r'Hwe proposed standards were modifiãd by l¿) shirtine the eI)r ureatpoiiìiio ..qri.. n o..improvements at higher traffic levels and fewer imfrovements at lower traffic levels, (bj inc..asing minimum shoulder widrhs lor highways wirh greare r ihan 2,000 ADT levels, and (c/ rJü.iíg ,'inlru- shoulder widths for highways on mountainous terrain. 2,360 480 r,040 58, l 00 8,600 23,800 33,900 FHWA proposed standards with modifrcations, 1,069

146 DESIGNINGSAFERROADS gJearer than 4,000 ADT. The breakpoint at 400 ADT was shifted to 750 ADT' and the breakpoint at 4,000 ADT was shifted to 2,000 ADT. The shift from 400 to 750 ADT has the effect of making the standards less stringent-ttre least stringent lane and shoulder widths þreviously applicable to highways with I o 400 ADÐ are no\il applicable ûo highways \,/ith 1 Ûo 750 eor. ftre shift from 4,000 to 2,000 ADT has ttre opposite effect-the most stringent lane and shoulder widths þreviously applicable o highways with ADT greater tfran 4,000) are now applicable to highways with ADT gfeater than 2,000. The net effect of the two shifts is to decrease the cost of lane and -r.^,,l,t^- .,,i¡{aninc innreqsp rhe. nnmher of ac.c.idents eliminated. and deCfeaSgù¡¡vuruv¡ w ¡uvrr¡¡¡6, the cost per accident eliminated. The second modiûcation t0 ttìe 1978 FFIWA proposed standalds was t0 increase the shoulder widths required at higher ADT levels from 4 to 6 ft on each side. As noted in the project-level analysis, shoulder widening is more cost-effective than lane widening on most of the nation's highways' The third modifrcation to the 1978 FI{WA proposed standafds was to decrease minimum shoulder widths by I ft in each direction on mountainous terrain. As shown in the project-level analysis, shoulder widening is less cost effective on mountainous than on flat or rolling terrain' The combined effect of the three modifications is to increase the cost for lane and shoulder widening by about 3 percent, increase the number of accidents eliminated by about 45 percent, and reduce the cost per accident eliminated by 26 percent. The FTIWA proposed ståndards with modifications could save approximately 1,000 lives and prevent nearly 30,000 injuries each year.2 As of January 1986,26 staæs had special RRR standards and 24 states used new construction standards for RRR projects. The modified FHWA proposed standards are generally less stringent than special RRR standards at low trafûc levels, more stringent than special RRR standa¡ds at high Eaffic levels, and less stringent than new construction stândards at all ADT levels. The cost-effectiveness of standards proposed at the national level were compared with special RRR standards in Ftorida, Illinois, and Washington state (Table 5-5). In these three states, the modifred FFIWA proposed standards were about as costæffective as special RRR standards, but involved higher spending for lane and shoulder widening and eliminated more accidents. 2These ertimates are based on 33,900 accidents eliminated (lable 5-4), distributions of accidents by severity from Smith et al. (4), and information on the number of fatalities per fatal accident and injuries per injury accident from NHTSA (lJ).

COST.EFFECTMENESS OF DESIGN STANDARDS I47 TABLE 5-5 Cost-Effectiveness of Alternative Lane and Shoulder Width Standards: State-Level Analysis for Florida, Illinois, and Washington Cost per Accidents Accident Lane and Shoulder Cost per Year Eliminated EliminatedState Width Standard ($millions) per year ($thousands) Washington 1978 FHWA proposed standards 13.33 359 37.t FHWA proposed standards with moditcations 16.43 532 30.9 State RRR srandards I t.l3 321 34.7Florida 1978 FHWA proposed standards 7.70 224 34.4 FHWA proposed standards with modifrcations 8.60 364 23.6 State RRR srandards 6.33 316 20.0Illinois 1978 FHWA proposed standards 10.73 198 54.2 FHWA proposed standards with modifrcations 15.89 404 39.3 State RRR srandards 13.62 357 38.2 NorEs: Costs are in 1985 dollars and were calculated using a 7 percent discount rate and a project life of 30 years. The 1978 FHWA proposed standards were modified by (a/ shifting the ADT breakpoints to require more improvements at higher traffic levels and lewer improvements at lower traffic levels, (ó/ increasing minimum shoulder widths for highways with ADT greater than 2,000 and (c) reducing minimum shoulder widths for highways on mountainous terrain. Summary of Findings on Lane and Shoulder Widening ¡ Trafflc volumes are an importânt consideration in the cost-effectiveness of lane and shoulder widening. Standards that are not sensitive to traffic volumes, such as the RRR ståndards presented in the AASHTO RRR guidelines, ¿¡re not cost-effective. . Lane and shoulder widening not only eliminate accidents but also result in time savings to highway users. These time savings are relâtively small at lower traffrc levels but can be an importånt consideration for highways with ADT greater than 2,M. o The cost-effectiveness of the lane and shoulder width standards proposed by FTIWA in 1978 can be improved by la) shifting rhe ADT breakpoints to require more improvements at higher traffic levels and fewer improvements at lower levels, (á) increasing minimum shoulder widths for highways with ADT gfeater than 2,000, and (c) reducing minimum shoulder widths for highways on mountainous tenain. Modified standards reflecting these changes are given in Table 5-3.

148 DESIGNING SAFER ROADS Horizontal Curves The cost-effectiveness of reconstructing horizontal curves by decreasing their degtee of curvature was examined at ttre project and system levels using (ø) the safety relationship presented in Chapter 3 and discussed in more detail in Appendix D, (b) the cost relationship discussed in Chapter 4, and (c) AASIITO methods for estimating highway user travel time and operating cost savings from flattening curves 03). Søndards for horizontal curves are usually expressed in terms of design speeds. As describcd in AASHTO's Pclicy a* the Geometric Design o! Highways and Streets,1984 (15), the design speed of a curve is determined by radius, superelevation, and the friction generated by the pavement surface' For new construction, design speed is used as a concept for unifying the various aspects of the new highway. AASIIT0 defrnes design speed as "the maximum safe speed that can be mainøined over a specified section of highway when conditions are so favorable that ttre design featu¡es of the highway govem" (15), All pertinent features of the highway, including hori- zontal curves, should be related to the design speed to obtain a balanced design. The selection of an ideal design speed using a rationale similar to that employed for new construction is generally not practical for RRR projects. It is more practical to compare the as-built design speeds of existing curves with some measure of the speeds of approaching vehicles before they slow down for the cuwe. The 85th percentile speed is useful for this pulpose because it exceeds the speed of most approaching vehicles and has traditionally been used by traffic engineers for setting speed limits. P roj ec t Lev el C o st -Effe ctiv ene s s Important determinants of the cost-effectiveness of flattening horizontal curves afe ¡ Design speed of the curve before and after improvement; . Other factors influencing accident rate on the curve, such as the distribu- tion of actual operating speeds, whether the curve is isolated on a long tangent section (versus one of many ctrrves on a highway segment), lane width, sight distance, vertical alignment, shoulder width, and roadside environment; ¡ Cost of flattening the curve, which in tum is affected by the cennal angle of the curve, cost of additional right-of-way, and unit costs for construcúon quantities; and

CO ST.EFFECTNENESS OF DESIGN STANDARDS r49 ¡ Trafûc levels; the number of accidents eliminated depends not only on the improvement in the accident rat€, but also on the volume of traffic using the curve. using a hypotfietical curve flattening project, the study examined the effect of design speed before and after improvement, cental angle, and ADT on cost per accident eliminated. other factors influencing accident rates are not explicitly accounted for. This is an important limitation of the analysis; because of other factors that must be add¡essed on a sife-specific basis, the accident fttþ at a curve could be much grcater or much less than assumed. The cost-effectiveness of cuwe flanening varies sharply with design speed before improvement (Figure 5-7). The cost per accident eliminated for flatten- ing a 30-mph curve is about one-fourth the cost for flatæning a 45-mph curve. DESIGN SPEEO BEFORE IMPROVEMENT Ï"Îffi ;'""i:'ü:f ''S[:::åTåi:ï:'J,ÍX""'"13åïï'åiîi." l'#,i"]ii5*",, fJ.i:ï3"XtrJüJ.':,Sff '::ì:triÍ"àåi',iï;":i:f.":ïålXi'..,:"jf åî FIGURE 5-7 Cost-effectiveness of flattening horizontal curves by design speed before improvement. As discussed in chapter 4, a relatively large share of the cost for flattening a horizontal curve is ûxed and is not sensitive to the amount by which ttre curve is flattened. For this reåson, it is not cost-effective to flatten a curve only slightly because the cost will be out of proportion to the improvement in the accident rate ffigure 5-8). cost-effectiveness considerations are not just limited to the question of whether or not a given curve should be improved, 6ooo@I zoo ¡r¡ 1- =E lsoJ l¡¡ F.z l¡lg 100 oo cc H50 F. U) oo 0L 20

150 DESIGNING SAFER ROADS DESIGN SPEED AFTER IMPROVEMENT i*:ffi ,r:i:ï:S''f [:"'3å Tå iåi"o'i":gï:Jg!'ï'å'.",iin" 3i'],1'i,lJll'***, f;J."il.Iå:Xl"'J","i#:ffi 'fr"fJf"Í"àAi'Ï"',*i:Î."liltlå1'-':"¿''"f" FIGURE 5-8 Cost-effectiveness of flattening horizontal ctrrves by design speed after improvement. designers should also consider cost-effectiveness in deciding the degree to which the curve should be upgraded. The amount of earthwork and right-of-way required to flatten a horizontal curve, and consequently the cost per accident eliminated, is strongly affected by the cenfial angle (Figure 5-9) because the curve length increases roughly in proportion to its central angle. As with other design features examined in this chapter, cost per accident eliminated varies inversely with ADT (Figure 5-10) because the number of accidents eliminated is directly proportional to ADT. User savings realized from flattening horizontal curves are appreciable in relation to the cost of these improvements, particulffly at ADT gfeater than 2,000. Taking user savings into account along with safety beneûts sEengthens the case for flattening curves. Drivers reduce speed when they approach a curve and then accelerate after they enter or pass the curve. This speed change cycle adds Ûo travel time, fuel consumption, and ofher components of operating cost. Improving horizontal curves reduces the size of the speed change, which in turn reduces travel times and operating costs for highway users. Curves also affect user costs in two other ways, both of which are insignificant relative to the speed-change effect. First, vehicles expend more energy and experience mofe tire wear on curves oooo6 o UJF =J u¡ t-z lrJo õo fr tuÀ t--Ø oo l nêrldn sóeed bêfore imprãverìrent = so mph

oô @o6 ;Bo l¡¡ k =EsoJ lr.tF z, ¡¡¡o40 oo EHeo F c, oo CENTRALANGLE NOTES: Charactedsìiçs 01 the hypothelicat horÞontal curye on which this æ¡silivlty analys¡9 is based are (a) dsign speed of 30 mph before imlrrovorent,(b) 55 mph aft€r lmprowmnt, aM (c) ã,ooö ADT, Cosls are in 19gs àollars and werê calculated us¡ng a dlsæuú ¡atâ of 7 Þq@nt and a projecl life.of 30yeare. FiGIIRE 5-9 Cosr-effectiveness of flattening horizont¿l curves by cennal angle. 160 úg 1Ao 6 fil tzoF Ët* =u¡¡80z l¡¡ Euo o c¡4o lr¡ o.âzoo o 0 1,000 2,000 4000 ADT NOIESi çbârætêri$tiG ef thô hypolhofca¡ horilontal cuNè on which this s€nsiliv¡ty analysis ¡s basod are (a) dslgn spæd ol30 mph bofor€ improverent,(b) 55 mph aft€r ¡rÞrovmnl, and (c) æntrâl anot€ of 3ó deare€s. Cbsre are tn 1985 doltars and were calculatêd using a d¡sæunt rate ol Z pércenl and aprolèct lile òf 30 ys¿¡rs. FIGURE 5-10 Cost-effectiveness of flattening horizontal curve$ by ADT. 0

r52 DESIGNING SAFER ROADS because of the tire side friction required to keep the vehicle on the curve. Second, flaúening a curve provides a small reduction in Eavel disønce. Using the methodology prcsented in AASHTO's Manual on User Benefit Analysis of Highway and Bus-Transit Improvemcnts (12), the study estimated the effects of horizonøl curves on user costs. The manual provides estimates of the extent to which drivers decrease their operating speeds on curves with low design speeds. The estimates of speed decreases are consistent with more recent estimates (,¡ó). Flatæning a horizontal curve to increase its design speed from 30 to 55 mph will reduce user costs by $27 per i,00Û vehicies Gigure 5-ii). For a highwav with 1,000 ADT this savings amounts to $10,000 per year' The $10,000 per year in user savings is a considerable portion of the annualized cost for curve flattening (Iable 5-6). At higher ADT levels, user savings for curve flattening might be well in excess of the annualized cost. Sys tem-Lev el C o s t-Effec tivene s s The cost-effectiveness of cuwe flattening was examined by using curve inventory data from state highways in Florida, Illinois, and Washington State. DESIGN SPEED BEFORE IMPROVEMENT NOTE: Travel tire savings aro valued at $7.5O por vohicl€ hour. FIGURE 5-11 User savings for improving the design speed of a horizontal curve to 55 mph. t,5soI !9so oooJ40 ct UJ i30 oz =20u, ffiroØl 20

COST.EFFECTNENESS OF DESIGN ST¿NDINOS I53 TABLE 5-6 Illustrative cost and user Savings for Horizontal curve Flattening Annualized Cost lor Curve Flattening Cost ($) Added construction cost User savings per year I,OOO ADT 3,OOO ADT 5,OOO ADT Net cost (construction cost less user savings) l;000 ADT 3,000 ADT 5,OOO ADT l 3,300 9,900 29,700 49,500 3,400 - 16,400 -36,200 Nor¡s: The example assumes design speed before improvement of 30 mph, a design speed afterimprovement ol 55 mph, and a central angle of 30 degreei. Time savingr ur. ,åru.ã uiliio fier vetrictehour. Consl¡uction cost is annualized using a 7 percent discounl raté and a project life oi 30 y.u.r. These states maintain deailed roadway invenúories that contain the daø needed (degree of curve, central angle, and traffic volume) to estimate the cost-effectiveness of crlrve flattening for individual curves. These daø are not available in many other states and are not compiled ât the national level. curve flattening costs were estimated using the cost relationship from washington søte (Appendix I) with adjusments based on construction price indices for Florida and Illinois. cost per accident eliminated was calculated witl and without accounting for user travel fime and cost savings because there is no generally accepted practice for weighting these savings in relation to construction costs. Although all three states have about the same mileage for two-lane rural highways, washington St¿te has many more curves with low design speeds reflecting difference in ferrain (Figure 5-12). Most of the tenain in wáshington is rolling or mountainous whereas most of the terrain in Illinois and Florida is flat. cost per accident eliminated for flattening horizontal curves in the úrree states was estimated by using a design speed of 40 mph as the minimum standard. For highways with 85th percentile speeds of 55 mph or more on tângent sections, under this minimum stândard curves will be flattened only when the design speed is more rhan 15 mph below the g5th percentile speed of vehicles approaching the curve. v/ithout accounting for user time and cost savings, system-level cost per accident eliminated for flatæning horizontal curves with design speeds below the minimum sr¿ndard of 40 mph ranged from g15,000 in Rori¿ato $100,000 in Illinois (Table 5-7). cost per accident eliminated by flattening horizont¿l curves is higher in Illinois than in washington and Florida for two reasons. First, mosiof the

154 DESIGNING SAFER ROADS 3,000 2,500 2,000 1,500 1,000 s00 0 DESIGN SPEED (mph) NOTES: Des¡gn speeds of ind¡v¡dual curu€s al€ est¡mated from their radi¡ of curuatur€ using the equation pres€nted ¡n AASHTO'S PolÆv o" the Geometric Des¡gn of Highways and Streett, 1984 (l 1, p. 174) A combined side frìction factor and super€l€vat¡on rale of 0.20 js assured. FIGURE 5-12 Number of curves by design speed on two-lane rural stâte highwaYs. substandard curves in Illinois are just below the 40 mph minimum, whereas in Washington and Florida, many of the substandard curves are below 30 mph. As shown in ttre project-level analysis, cost per accident eliminated is very sensitive to design speed before improvement. Second, Illinois has a number of curves with central angles of 90 degrees because roads in Illinois are frequently laid out along the boundaries of rectangular parcels of land. The TABLE 5t Cost-Effectiveness of Flattening Horizontal Curves on Two-Lane Rural State Highways Washington lllinois Florida l,690 29,2'7 | 481 60.9 28,s82 689 1.4 Nores: The assumed minimum standard is an as-built design speed ol40 mph. Substandard curves a¡e assumed to be reconstructed with a design speed ol55 mph. When the net cost is negative. net cost per accident elminated is not meaningful (NM). Costs are in thousands of 1985 dollars. Added construction costs are annualized using a 7 percent discouÌlt rate and a project iile ol 30 ,vears |J' t¡J É. :) o lt o cÉ ¡¡l d¡ =fz Substandard curves Added construction cost per year Accidents eliminated per year Cost per accident eliminated User time and cost savings Per Year Net (construction less user) cost per year Net cost per accident eliminated 145 693,095 t,0923t 7499.8 14.8 1,7 45 4,39 I r,350 -3,29943.s NM 1 I I l I i I I I

C O ST.EF F ECTNENES S O F DESI GN S?WDARDS 155 project-level analysis showed that cost per accident eliminated for recon- structing curves is very sensitive to the cental angles of the curves. cost per accident eliminated in Florida is much lower than in Illinois and washington primarily because average ftaffic volumes are much higher. ADT for the reconstructed curves in Florida is 5,000 versus r,600 in washington and 1,700 in Illinois. 'when user travel time and operating cost savings are taken into account, curve flattening is substantially more cost-effective. cost per accident elimi- nated drops drastically from $61,000 to $1,400 in Washington and from $100,000 to $43,000 in lllinois. In Florida, the user rime and cost savings for curve flattening exceed the added cost so that these improvements can bejustified at the system level based on user savings alone. In washington state, a more detailed sraüfication of curves by ADT was used to establish the circumstances under which curve flattening is frequently cost-effective (Table 5-8). Ti*ing user savings into account, the average cost per accident eliminated for flattening curves in the 750- to 1,500-ADT range with design speeds greater than 15 mph below 85th percentile speeds is about $20,000. For the 500- to 750-ADT range, however, average cost per accident eliminated is about $100,000, double the $50,000 figure used in this study as an upper limit for improvements justited purely on the basis of safety cost- effectiveness. These findings support the use of 750 ADT and a difference of TABLE 5-8 Cost-Effectiveness of Flattening Horizontal Curves by ADT Washington State System-Level Analysis ADT Under 500 1,500 500-149 750-1A99 and Over Torat Substandard curves Added construction cost per year Accidents eliminated per year Cost per accident eliminated User time and cost savings per year Net (construction less user) cost per year Net cost per accident eliminated 246 4,6r4 l4 329.6 782 3,832 212.9 t64 2,982 l8 t65.1 1,t l0 t,8'72 104.0 754 t3,334 t66 80.3 9,855 3,419 21.0 526 1,690 8,341 29,27 t 283 48 I 29.s 60.9 16,835 28,582 -8,494 689 NM NorEs: The assumed minimum standard is an as-built design speed ol40 mph. Substandard curves are assumed to be reconstructed with a-design speed of55 mph. When the net còst is negative, net cosr per accident eliminated is not meaninglul (NM). Costs are in ihousands of 1985 dollars. A?ded ôonrtru"tion costs are annualized using a 7 percent discount rate and a project lile of30 years.

156 DESIGNING SAFER ROADS 15 mph between the as-built design speed of curves and 85th percentile speeds as breakpoints t0 sepafate situations in which curve flattening is frequently cost-effective from those in which it is not cost effective. Summary of Findings on Horizontal Curves . Curve flattening is frequently cost-effective when ADT is greater ttnn 750 and the design speed of the existing curve is more than 15 mph below the 85th percentile speaÍ oi vehicies approaching iire curve. . Cu.ue flatæning can result in subsøntial travel time and operating cost savings to highway users. Taking these savings into account strengthens the case for curve flattening. . A firm, nationwide standard for horizontal curves-such that curves falling below the standard are routinely upgfaded-is inappropriate because of the high degree of site-to-site variation in the cost-effectiveness of curve flattening. Sight Distance at Crest Curves Søndards for sight dishnce at crest c¡rves a¡e defined in terms of design speed. According to AASHTO's Policy on the Geometic Design of Highways and streets (15), "ttre minimum sight disønce available on a roadway should be sufficiently long to enable a vehicle Eaveling at or near the design speed to stop before reaching a stationary object in its path." In calculating sight distance, AASHTO's policy on design assumes the driver's eye is 3.5 ft above the road surface and the object is 6 in. above the roadway. The policy also assumes a brake reaction time (the time interval from the instant the driver recognizes the existence of a haza¡d on the roadway to the instant the driver actually applies the brakes) of 2.5 sec, which is $eatef than that required by most drivers under normal highway conditions. The safety cost-effectiveness of reconstructing crest curves to improve sight distances was examined by using hypothetical projects. Cost per acci- dent eliminated was calculated using the safety relationship presented in Chapter 3 (and discussed in more detail in Appendix E) and the cost relation- ship discussed in Chapær 4. This analysis probably overstates the cost- effectiveness of flattening a typical crest curve because, as noted in Chapter 3, the safety relationship provides an upper tround to accident reductions result- ing from increased sight disønce. Crest curves were not analyzed at the System level because data on cost and accident relationships-length of curve and degree of hazard in the sight- restricted area-âÍe not available'

COST-ffi FECTNEI'IESS O F DESIGN STAIIDARDS t57 P roj e c t-Lev e I C o st -Effe c tiv e ne s s Important determinants of the cost-effectiveness of improving sight disønce at crest curves are ¡ Difference between the design speed of the curve and the speeds of vehicles as ttrey Eavel through the sight-restricted area, . Degree of hazard in the sight-restricted area, and o Traffic volumes. The cost-effectiveness of sight distance improvements at crest curves varies sharply with the difference between design speeds and operating speeds in the sight-restricted area (Figure 5-13). The cost per accident eliminated for improving a curve with a speed difference of 20 mph is about 60 percent of the cost per accident eliminated for improving a curve with a speetl difference of 15 mph. The degree of hazud in the sight-restricted a¡ea is also an importânt consideration. A major hazard, (such as a high-volume intersection or sharp horizonøl curve) in the sight-restricted area might reduce the cost per accident eliminated by one-half. As with other design features, cost per accident eliminaæd varies inversely 0 r,000 2,000 3,000 4,000 s,000 AVERAGE DAILYTRAFFIC NOTES: Example assures operating speed of 55 mph and a major hæard ¡n the sight-restr¡cted a¡ea. Cosls are ¡n 1985 dollars and were calculal€d us¡nq a d¡scount rale of 7 p€rc€nl and a project lif€ of 30 y€ars. FIGURE 5-13 Cost-effectiveness of reconstructing crest curves by ADT and design speed before improvement. a 100oooc o80 !JF zE60f uJ , ä40õ o É20U o- t- u)$o

158 DESIGNINGSAFERRoADS wittr ADT because the number of accidents eliminated is directly proportional to ADT. længthening vertical crest curves reduces mo¡or vehicle operating costs by , reducing the length of vertical tangents. This effect is small relative to the construction cost of lengthening curves, however. Consider, as an example, a 600-ft-long vertical curve connecting a 6 percent upgrade and a 6 percent downgrade on a highway with 4,000 ADT. On the basis of estimafes of the effect of grades on operating costs in AASHTO's Mønual on User Benefit Analysis of Highway and Bus-Transit Improvements /!ô! !--:aL_:!__.L:_u¿), tc.rrgLrrörlrng uus curve oy l,\ ^, Il. wul reoucc moior venrc¡e operaûng costs by $2,û0lyear. The annualized construction cost for this improvement is about $35,000/year. Thus, even at 4,000 ADT, user savings will cover less than 10 percent of this improvement. Summary of Findings on Sight Distance at Crest Curves ¡ Reconstructing crest curves to improve sight disønce can be cost-effec- tive when a major hazud, (such as a high-volume intersection or narrow bridge) exists in the sight-restricted area, the design speed of the existing curve is more than 20 mph below operating speeds in the sight-restricted area, and ADT is greater than 1,500. i . Highway user travel time and cost savings associated with reconstructing crest curves are generally small relative to ttre added cost of construction for highways with less than 4,000 ADT. ¡ A firm, nationwide ståndard for sight distance at crest curves-such tlat curves falling below the standard are routinely upgraded-is inappropriate because of the high degree of site-to-site variation in the cost-effectiveness of sight distance improvements. Bridge Width The safety cost-effectiveness of bridge width improvements was examined on two-lane rural highways using the relationship bet'ween bridge width and accidents presenûed in Chapter 3 and tt¡e relationship between bridge width and cost presented in Chapter 4. P roj e c t-Leve I C o st-Effe c tiv e ne s s Important determinants of the cost-effectiveness of bridge width improve- ments are

C O ST.EF FECTN EN ESS O F DESI GN S?WDARDS 159 o The width of the bridge relative to the widrh of the navel lanes on bridge approaches; o Whether widening is practical or, alternatively, whether it is necessary to demolish and reconstruct the bridge to provide more width; o Whether the bridge requires major rehabiliøtion or reconstruction for other reasons (e.g., inadequate load-bearing capacity); o The lengfh of the bridge, which affects the cost of bridge width improve- ments; and o Traffic volumes. The accidentrate for bridge-related accidents depends on the relative bridge width-the difference between the widrh of the bridge and the width of the travel lanes of the highway on which the bridge is located (chapter 3). If tanes a¡e úo be widened as paÍ of a RRR project, the new travel lanes (rather than the Eavel lanes before improvement) should be used in calculating the rerative widths of bridges. Another important consideration is whether it is practical to widen the existing bridge, or alærnativel¡ whether it is necessary to demolish the existing bridge and construct a new one to provide additional width. Based on the unit costs presented in Chapær 4, widening a 50-ft-long bridge from 20 to 28 ft would cost about $40,000. If widening is not practical, the cost of demolishing the 2O-fr-wide bridge and consrrucring a 28-ft-wide bridge would be about $100,000. Thus, cost per accident eliminaæd would be more than two times greater if widening is not practical (Figure 5-14). Bridge widening is less cost-effective for longer bridges. Bridge widening costs increase roughly in proportion to ttre length of the bridge, but. the number of accidents eliminated by widening a narow bridge is only marginally affected by bridge length. Thus, the cost per accident eliminaæd is greater for longer bridges. Bridge width improvements are highly cost-effective when bridges require reconstruction for other reasons (e.g., inadequate load-bearing capacity) because the added cost for providing a wider bridge is substantially less if the bridge is going to be reconstructed anyway. Designers should consider cost-effectiveness not only in deciding whether to widen a bridge, but also in deciding the extent to which it should be widened. A substantial part of the costs for bridge widening are fixed and must be paid regardless ofhow much rhe bridge is widened (Figure 5-15); therefore it is generally not cost-effective to widen bridges by only a few feet. Also, large improvements to bridge width may not be cost-effective because there is little payoff in terms of accident reduction in increasing bridge widths beyond levels mandated by AASHTO new construction standards.

70øÞoott 60 o uJäso =E =40l¡ll-ñsoÕõ3zo &Hro F 1t8o 1,000 2,000 3,000 4,000 AVERAGE DAILYTRAFFIC NOfES: Example assures a bridge lenglh of 50it, reìatìvsw¡dth before improverent of o, and relative width after ¡mproverent of I fl. Costs âre in 1985 dollats a¡d were calculated using a discount rate of 7 percent and a projêct lile of 3o Years. FIGIIRE 5-14 Cost per accident eliminated for bridge width improvements. ADDED WrDTH (ft) NQTES: Example assure$ 2,OOO ADT, br¡dge length of 50 ft, and a relalive w¡dlh of 0 belors improverent. Cosls arè in 1986 dollars and were calculated using a discgunt rale of 7 pqcent and a project lite ol 30 yèars. FIGURE 5-15 Effect of ¿dded width on cost per accident eliminated for widening an existing bridge. G16ooÞt¿14 o l¡l e12 =EroJ t¡¡ ,8 l¡¡o- õb o$c ut .Lt-2 an oo0

COST-EFFECTNENESS OF DESIGN STAI\IDARDS 161 ADDED WTOTH (ft) i{*iìr*:rff *äft"ri:ijr i:,,;ff ii}i:':f,i:åå ïrÎrå :" FIGURE 5-16 Effect of added width on cosr per accidenr eliminated for demolishing a bridge and constructing a wider bridge. The cost per accident eliminated decreases sharply with added width when an existing bridge is demolished and a wider bridge is constructed in its place Gigure 5-16). unless there are other considerations in the decision to replace the bridge, such as inadequate load-bearing capacity, this improvement makes sense only for relatively large improvements to bridge width. user costs are seldom an important consideration in bridge widening on two-lane rural highways. The effect of a narrow bridge on user costs is similar to the effect of a horizontal curve on user costs: drivers reduce ttreir speed when they approach a narrow bridge and accelerate after they cross the uriage. However, field observations of driver behavior at bridges on fwo-lane rural highways indicate that the amount by which drivers reduce speed is slight(about 2 mph) and not süongly relared üo the widrtr of the bridge. Instead, drivers compensate for narrow and hazardous bridges by repositioning their vehicles closer to tlte centerline, even crossing it in exremecases (17). A second possible consideration is the role of bridge width as a prime determinant of the bridge's vehicle-carrying capacity. on high-volume high- ways, a narrow bridge may be a bottleneck at which congestion occurs, which in turn increases havel time and cost. such congestion seldom occurs on two- lane rural highways. a870oI o60u kz50 = Ju40 t-z ul e30ooÍro IJJÀ b10 oo 0

162 DFJIGNING SÀFER ROADS Sy stem-Level C os t'Effec rtvene s s The cost-effectiveness of bridge widttr improvements was examined at the system level using data from the FTIWA bridge inventory. Cost per accident e-liminated was calculated under the assumption that widening ttre existing bridge is not practical; therefore, it must be demolished and replaced with a widãr smcture. As suggested by the project-level analysis presented earlier, this assumpúon will overstate cost per accident eliminated by 50 to 75 percent in those cases in which widening is practical. This system-ievei anaiysis does not arjtiress siiuaúons in which ii is neces- sary to demolish and replace the bridge for other reasons' such as inadequate loaâ-carrying capacity. In those situations, the cost per accident eliminated for bridge width improvements will be subståntially less because the bridge *ouid b" demolished in any case and the added cost is only that required tn provide a wider structure. Also, the system-level analysis addressed only bridge width improvements, not lower cost improvements such as improved signing and lanes that gfadu- ally narrow as they approach a nalrow bridge. The cost-effectiveness of these lower cost improvements could differ greatly from the cost-effectiveness of bridge width improvements. At the system level, the cost-effectiveness of bridge width improvements varies depending on bridge lengh, width, and ADT (Table 5-9). For bridges less than 100 ft long with ADT gfeater than 4,000, the average cost per accident eliminated ranges downward from $30,000 for relative widths less TABLE 5-9 System-Level cost-Effectiveness of Bridge width Improvements Cost per Accident Eliminated by Relative Width Bridge Length ADT 0-1.9 2.0-3.9 4.0-5.9 0-r00 Over 100 0-750 75 I -2,000 2,001-4,000 Over 4,000 0-750 75 I -2,000 2,0001-4,000 Over 4.000 224.1 57.6 29.2 I 5.0 803.9 258.0 r62.8 9'1.9 40 1.6 98. I 40.0 2t.9 1,6 1 9.6 420.0 221.9 139.2 429.4 t 37.0 59.9 31.3 I ,875.9 600.5 28 5.0 t64.1 No¡rs: The analysis assumes that exisling bridges are demolished and replaced by wider bridges designed to AASHîO new construction stanãards. The analysis is based on data lrom the FHWA bridge inu"itory. Bridges with relative widths less than zero are not included because they are outside the range tã *nic¡itre acõi¿ent relarionship applies. It is likely that cost per accident eliminated lor replacing such ùiiág"i *árf¿ be considerably bèloï the ualues shown in the column for a relative width of0 to 1.9 ft. Cosis are in thousands ol t98j dollars and were calculated using a discount rate ol7 percent and a project life ol 30 years.

COST.EFFECTNENESS OF DESIGN STANDARDS L63 than 6 ft. For bridges greater than 100 ft long with less than 2,000 ADT, the average cost per accident eliminated is more ttran $200,000. Summtry of Findings on Brídge Width ¡ Bridge width improvements are frequently cost-effective when the length of the bridge is less than 100 ft and the widtl¡ is less than the following values: Usable Bridge Wídth (ft)" Width of appoach lanes Widtlt of approach lanes plus 2 ft Widttr of approach lanes plus 4 ft Width of approach lanes plus 6 ft alf lane widening is planned as part of the RRR project the usable bridge width should be compared with the planned width of the approaches after they are widened. . Highway user Favel time and cost savings associated with bridge widttr improvements are generally small relative to the added cost of making these improvements for highways with less than 4,000 ADT. ¡ A firm, nationwide stândard for bridge width on RRR projects-such that bridges falling below the stândard are routinely upgraded-is inappropriate because of tle high degree of site-to-site variation in the cost-effectiveness of bridge width improvements. Roadside Obstacles The cost-effectiveness of removing roadside obstacles was examined using several hypothetical projects. Cost per accident eliminated was estimated for roadside improvements by using accident relationships presented in Chapter 3 and discussed in more deøil in Appendix F. No system-level analyses were conducted because data on roadside obstacles are not compiled by stâtes at the system level. P roj ec t-Lev el C o st-Effe c tiv ene s s Important determinants of the cost-effectiveness of removing a roadside obstacle are Design Year Volurne (ADT) 0-750 751-2,000 2,001-4,000 Over 4,000

I& DESIGNING SAFER ROADS . Degree of hazard posed by the obstacle, for example, as measured by the probability that a collision with the obstacle will result in an injury or fatality; . Degree of hazard remaining after the obstacle is removed; . Distance of the obstacle from the edge of the travel lanes; o Horizontal alignment, as vehicle encroachments on the roadside are more likely to occur at curves; and o Traffrc volumes. Removal of isolated roadside obs0acles can be highly cost-effective, even ^- r^.., ,,^1,,..ô - ^á. En¡ o-amnla ramm¡al nf ¡n icnlatarl free. 1O ft from thev¡l lvw-Yvrulltw rvquù. ¡ vr v ø¡¡P¡vt road can result in a cost per accident eliminated of less than $15,000 at 1,000 ADT and less than $10,000 at 2,000 ADT (Figure 5-17). 0 1,000 2,000 3,000 4,000 5'000 6,000 AVERAGE DAILY TRAFFIC NOTES: Example assures thê obstacle is located 10 fl from lhe edg€ of the lÍavel lanes and the area b€h¡nd the obstacle ¡s a 4:l fill slope. Cosls for rercving isolalêd trees and ut¡lity poles are assured to be $660 and $2'580, respectively (Chapter 4, Table 4-6). Costs are in 1985 dollars and are calculated using a discount ratê of 7 pe.cent ând a pro¡oct life ol 30 yêârs, FIGURE 5-17 Cost-effectiveness of removing isolated trees and utility poles by ADT. In analyzing the cost-effectiveness of removing a roadside obstacle, impor- tant considerations include not only the hazard posed by the obstacle, but also the remaining hazards once the obstacle is removed. For example, removing a tree might be highly cost-effective if it is an isolated obstacle on a relatively flat slope. However, if the tree is one of many obstacles or if it is on the edge of a steep slope, then removing it is less cost-effective because drivers who otherwise would have struck the tree ale still likely to have a serious accident. ?]s0ooog ffcoF =ã.0 t¡J Fz utq-20 o o ffro(L F(r, oOg

CO ST.EFFECTNENESS OF DESIGN STAI,IDARDS 165 The cost per accident eliminated of removing obstacles increases with distance because there is less chance that errant vehicles will srike obs¿acles farther from the edge of the road (Figure 5-18). For example, the cost per accident eliminated for removing a roadside hazud located l0 ft from the roadway edge is about 50 percent less than that for removing ahazardlocated 20 ft from the roadway edge. DISTANCE FROM EDGE OF TRAVEL LANES NOTES: Example assumes 2,OOO ADT and lhe area behind the obslacle is a 4:1 f¡llslope. Costs for remv¡ng ¡solatsd trees and ut¡lity poles ar€ assured to be $660 and gA,SgO, respælively (Chapter 4, Tabte 4-6). Costs are in 19gs do¡lars and were calculated using a d¡scount rate ol 7 percent and a proiêct l¡fe of 30 yeaß. FIGURE 5-18 Cost-effectiveness of removing isolated trees and utility poles by distance from edge of travel lanes. Although the factors involved in examining the cost-effectiveness of removing roadside obstacles are complex-involving subtleties such as the hazards that remain after an obstacle is removed-it is nonetheless clear that this type of improvement can be highly cost-effective, even on low-volume roads. Summary of Findings on Roadside Obstacles ¡ Removal or relocation of isolaæd roadside obst¿cles such as Eees and utility poles can be highly cost-effective, even on low-volume roads. G50oooI fiì 40l- =EgoJl¡l t-z o' 20õo firo(L FU' oo0

166 DESIGNING SAFER ROADS . The safety cost-effectiveness of removing a roadside obstacle depends on the distance of the obstacle from the roadway edge, the presence of other obstacles nearby, sideslopes on which ttre obstacle are located, and traffic volume. SAFETY.PRESERVATION TRADE-OFFS Because both pavement resurfacing and geometric improvements for RRR --^i^^¿^ ^-^ Â--l^l G-^* +k^ -^ñó ô^¡t' Â tha tn¡{a-nff }rof¡¡¡cen ccfefv enflPfUJq/tù dç tUl¡UW uV¡¡¡ urv ú.¡¡v ùvuwv, pavement condition is cenhal o the debate over RRR standârds. Changes in design standards for RRR projects could have a substantial effect on pavement condition if the amount of money available for pavement resurfacing is increased or decreased. About 70 percent of the 540,000 mi of two-lane rural federal-aid highways have lane or shoulder widths less than the widths recommended by AASHTO new construction standâfds (Iable 5-10). The one-time lane and shoulder widening cost ûo upgade these highways to new construction standards is about $30 billion. This is about one-half the one-time cost to resurface all 540,000 mi of two-lane rural federat-aid highways. Thus, if new construction stândards for lane and shoulder widttrs were applied to all resurfacing projects, the number of miles that can be resurfaced in a given year would drop sharply without additional funding. The cost per mile for resurfacing and minor widening of two-lane rural federal-aid highways will vary depending on the width sfândards used (Table 5-11). Under a fixed budget, about 30 percent more mileage can be completed in a year under AASHTO RRR standards than under AASHTO new con- struction standalds because the per-mile costs for resurfacing and minor widening under AASHTO RRR sønda¡ds is about 30 percent less. Under the 1978 FHWA proposed ståndards, and the two variations on these standa¡ds, TABLE 5-10 Deficient Mileage and Widening Costs Under Alternative Standards: National-Level Analysis Using HPMS Data Alternative Minimum Lane and Shoulder Width Standards Deficient Mileage One-Time Cost lor (as a Percent ofAll Lane and Shoulder Two-Lane Rural Widening (in billions Federal-Aid Highways) of 1985 $) AASHTO new construction standards AASHTO RRR standards FHWA 1978 proposed standards FHWA proposed standards with modiñcations 11.5 19.1 37.9 35.9 29.3 6.0 12.9 r 3.3

C O ST-EF FECTN EN ESS O F DES IGN STAN D ARDS TABLE 5- I 1 Cost per Mile Under Alternarive Width Standards 167 Alternative Minimum Lane and Shoulder Width Standard Resurfacing and Minor Widening Cost per Highway Mile (thousands) AASHTO new construction standards AASHTO RRR standards FHWA 1978 proposed standards FHWA proposed standards with modiÍÌcations t74 l3l t44 t45 NorEs: Costs are for all two-lane rural federal-aid highways, as estimated lrom HpMS data. The cost differences under alternative standards are one-time ðosts. about 20 percent more mileage can be completed in a year than under AASI{TO new consFuction ståndards. In considering these trade-offs, it should be noæd that service lives for geometric design improvements such as lane and shoulder widening are considerably longer than service lives for pavement resurfacing. once a highway segment is widened to meet standards it does not hâve to bi widened again when the pavement is resurfaced. In contrast to resurfacing, the safety benefits of widening continue over the life of the highway. Also investigated were the year-by-year safety and pavement preservation consequences of applying alærnative lane and shoulder width søndards to washington state's two-lane rural highways over a period of 30 years. Road- way inventory daø and pavement deterioration relationships from wash- ington's Pavement Management Sysæm were combined with accident and cost relationships developed in this study. The analysis assumed the follow- ing: ' Highway segments will be selecæd for resurfacing on the basis of pave- ment condition. . Projects will be scheduled for improvement until a specified budget is expended. ¡ Pavements will be improved to new condition if resurfacing is performed; otherwise pavement condition is downgraded to account for deterioration over time. Four alternatives were examined: la) AASHTO new construction stan- dañs, (b) modifred FTIWA srandards, (c) V/ashington Staæ RRR søndards, and (d) resu¡face only (no lane and shoulder widening). If standa¡ds were used, the cross section of the selected project was improved, if necessary, along with resurfacing. The alternatives varied substantially in ttreir effect on systemwide pave- ment condition, as measured by the highway mileage with pavements in need

168 DFSIGNING SAFER ROADS 2,500 2,000 1,500 1,000 500 0 l¡lt5 u.lJ =Fz l¡l ÞL lllo 1999 YEAR IHJ:3;.',,ä""-,'3H¿iiL'i[îå, #ÅXå?:y"ffi""¿';.i'ï"ff iffL".i:["Ii" Washington Stat€ paverent managerent pract¡ces (PSl of approx¡mately 2 5) Total syst€m m¡leage is 5,6f0 FIGURE 5-19 Miles of deficient pâvement in Washington State under alternative lane and shoulder width standards' of resurfacing at the end of 1 year (Figure 5-19). All altemaÍives exhibited a common pattern over time-highway mileage in need of resurfacing peaks about 10 years in the future (because a large number of pavements are due for resurfacing after 10 years), followed by a few years of improvement and eventual st¿bilization. AASIITO new construction standards have the most adverse effect on pavements: deficient mileage increases sharply and then stabilizes at a level twice that observed in the base year of 1984. Under the washington state RRR and modified FHWA standafds, miles of deficient pavement stâbilizes at levels 20 and 40 percent worse than current conditions. Under the resurface-only alternative, miles of deficient pavement stabilizes at a level slightly below that in the base year. Lane and shoulder widening not only shifts construction funds from current pavement repair, but also influences the process in futüe years by adding to the surface area that must be maintained and repaved when a segment is resurfaced in the future. If all deficient lanes and shoulders are widened in accordance with Washington State RRR standards, an additional $46 million in systemwide resurfacing costs (3 percenQ would result. Under AASHTO new construction ståndârds, systemwide resurfaCing costs r¡iould increase by nearly $190 million (18 percent) as a result of the extra surface area. Although stricter standards lead to more deficient pavements, their use also enhances safety (Figure 5-20). Under each of the three sets of standards, the number of accidents eliminaæd (relative to the resurface-only alæmative)

1,000 COST.EFFECTNENESS OF DESIGN STANDARDS 169 1999 YEAR G y 800 É u¡ fL fil 600F =5o* t¡J rnFz ä 200 õ o oprion, 3,841 ;iffiålTJi::Í:'S;i:Jlnliååil,:'';,,,T,ffJ.'i?jiî'ff illìg;""' over lhe 30-yed span and reasured relaùve lo the resurfac¡ng_only opt¡on. For exampte, tn the year 2000 application of AASHTO now constru¿t¡on;taildards will be elimlnating acc¡dents at a rate ol 7SO per yer ov€r resulac¡ng only. FIGURE 5-20 Accidents eliminated from Washington State highways under alternative lane and shoulder widttr standards. increases over time and then strbilizes after most of the segments with deûcient lane and shoulder widths have been upgraded. under AASHTO new construction standards, the number of accidents eliminated per year approaches 1,000. under less stringent standards, fewer accidents are elimi- nated-530 per year for modified FHWA and 460 per yeâr for washington state RRR standards, once all deficient highway segments have been upgraded. In addition to lane and shoulder widening, other improvements also might involve important safety-preservation trade-offs. The iost of a major effort directed toward curve flattening could be considerable in stâtes that, because of terrain, have many deficient curves. In washington state, for example, flattening all curves with a design speed less than 40 mph on two-lane rural federal-aid highways would have a one-time cost of $:50 mittion. This is about 40 percent of the cost to resurface all of these highways. The results from the washington State analyses illustrate the safety versus pavement preservation trade-off faced by all staæs; increased spending for safety improvements will reduce the number of miles that can be resurfaced with ûxed budgets. However, exrapolating speciflc results from the analyses to other states must be done with caution because the results depend on existing geometric and pavement conditions. v/ashington State has far fewer t//¿¿¿¿¿ ." ---

i70 DESIGNINC SAFER ROADS highways with lane widths of 12 ft tlan do other states, and has somewhat lower shoulder widths than the national average. However, the pavements in Washington State are in better condition than pavements in most states. These facûors will deærmine how many miles can be upgraded on a ûxed budget and the safety versus pavement preservation issues faced by each statÊ. SIJMMARY OF FINDINGS This section contains a summaly of general findings on the safety cost- effectiveness of design standards, as well as key findings on specific geo- metric features. General . Traffrc levels greatly affect the safety cost+ffectiveness of geometric design improvements. Some special RRR standards place little weight on ADT. Fewer accidents and I4ore overall user beneûts would result if expendi- tures for design improvements were shifted ûo higher volume roads. o States differ widely on the levels of investment required to achieve a given standard because of factors such as ænain and past design practices. A ståndad easily achieved in one state might require a considerable portion of the RRR budget in another. . Changes in geometric design sfânda¡ds for RRR projects can have a substantial effect on pavement condition by increasing or decreasing the amount of money available for pavement resurfacing' For example, the one- time cost of upgrading lane and shoulder widths on all ¡ryo-lane rural federal- aid highways ûo AASIITO new consEuction standards is $30 billion, which is about one-half of the cost to resurface all of these highways. . Cost per accident eliminated for geometric design improvements made as part of RRR projects frequently falls in the $10,000 to $50,000 range' Improvements with cost per accident eliminated in this range may or may not be warranted depending on the specific value impuæd ûo accidents eliminaæd; uncertainties surrounding estimates of the added cost for an improvement and the number of accidents it will eliminate; and other factors (e.g', environmen- tal effects) not accounted for in the calculation of cost per accident eliminated. o Geometric design improvements produced under a given set of design standards will vary in cost-effectiveness because of variations in accident rates, unit costs, design practices, project scale, and other siæ conditions. Thus, standards must be regarded as tools for ider'tifying situations in which improvements are more likely to be cost-effective. On the one hand, there will

CO ST.EF FECTNEI,IESS OF DESTGN STAI,IDARDS 17t be opportunities for cost-effective improvements beyond those called for by standards. On the other hand, there will sometimes be specific improvements called for by standards that are unusually costly (or that have other undesir- able consequences) such that design exceptions are justifiable. Lane and Shoulder Width o The cost-effectiveness ofthe lane and shoulder width standards proposed by FIIWA in 1978 can be improved by (a) shifting the ADT breakpoints ro require more improvements at higher traffic levels and fewer improvements at lower ADT levels, (å) increasing minimum shoulder widttrs for highways with ADT greater than 2,000, and (c) reducing minimum shoulder widttrs for highways on mountainous terrain. o Highway user travel time savings for widening lanes and shoulders can be signiûcant at higher ADT levels. Taking these savings into account, along wittr safety benefits, sEengthens the case for wider lanes and shoulders on high-volume highways. Horizontal Curves . Flattening horizonøl curves can be cost-effective, particularly when ADT is greater than 750 and the design speed before improvement is more than 15 mph below the speed of vehicles approaching the curve. However, the consid- erable site-to-site variation in the cost for alignment improvements indicates that firm nationwide sfândards for horizontal curves are inappropriate. ¡ Highway user travel time and operating cost savings for flattening hori- zontal curves can be considerable. Taking these savings into account, along with safety beneûts, sffengthens the case for tìese improvements. Sight Distance at Crest Curves r Reconstructing crest curves to improve sight distance can be cost-effec- tive when a major hazard exists in the sight-restricted areâ, ttre design speed of the existing curve is more than 20 mph below operating speeds in the sight- restricted area, and ADT is geater than 1,500. However, fhe considerable siæ- to-site variation in the cost of these improvements indicates that firm nation- wide standards for sight distânce at crest curves are inappropriate. . Sight distance improvements at crest curves provide user time and operat- ing cost savings, but the savings a¡e small in relation to the cost of these improvements and can usually be ignored in cost-effectiveness estimates.

r72 DESIGNING SAFER ROADS Bridge Width o Bridge width improvements can be cost-effective, particularly when ADT levels are high; the bridge is short (e.g., less than 100 ft); and its width is close to or less than that of the travel lanes on bridge approaches. However, the considerable siæ-to-site variation in the cost-effectiveness of bridge width improvements indicaæs that frrm nationwide standards for bridge width are inappropriate. ¡ Bridge widening results in user time and operating cost savings, but the savings a¡e smaii in reiation to Íhe cost oi iiiese improveÍÌren'rs and can usriaiiy be ignored in cost-effectiveness estimates. Roadside Obstacles ¡ The removal or relocation of isolated roadside obstacles such as trees and utility poles is frequently cost-effective, even on relatively low-volume roads. However, if an obstacle is close to other haza¡ds, then its removal or reloca- tion is substantially less cost-effective. REFERENCES 1. Jorgensen Associates. NCHRP Report 197: Cost and Safety Effectíveness of High- way Design Elemenfs. TRB, National Resea¡ch Council, Washington, D.C., 1978, 46 pp' 2. C.y. Zege,er and M. J. Cynecki. "Dete¡mination of CosçEffective Roadway Treatrnents for Utility Poles." In Transportation Research Record 970. TRB, National Research Council, Washington, D.C., 1984, pp.52-&. 3. J. L. Graham a¡rd D. W. Harwood. NCHRP Report 247: Effectiverwss of Clear Recovery 7,orcs.TRB, National Research Council, Washington D.C., May 1982, 68 pp. 4. J. A. Smith et al. Ideníftcation, Quantiftcation and Structuring of Two-Lane Rural Highway Safety Problems and Solutíons, Vol. 1. Report p¡1ry67RD-831022. FHWA. U.S. Department of Transportation, June 1983. 5. W F. MacFarla¡rd and J. B. Rollins. Cost-Effectiveness Techniques for Highway Safety: Resource Allocatíon. Report FHWAiRD-M/011. FHWA, U.S. Department of Transportation, June 1985. 6. Guide for Selecting, Locaing, and Designing Trafic Barriers. American Associa- rion of State Highway and Transportation Oftcials, Washingûon, D.C., 1977. 7. L C. Glermon. NCHRP Report 148: Roadside SSety Improvem¿nts on Freeways: A Cost-Effectiveness Priority Approach. TRB, National Research Cor¡ncil, Wash- ingtor¡ D.C., 1974, Ø pp. 8. J. R. t eisch a¡rd T. R. Neuman. Study of Width Standards for State Aid Streets and

COST-EFFECTNENESS OF DESIGN STAI\IDARDS I73 Highways. Minnesota Department of Transportation, Report FV/HA/IvIN-79-04. St. Paul, July 1979. 9. The Economíc Cost to Socíety of Motor Vehicle Accidets. NHTSA, U.S. Deparr ment of Transportation, 1983. 70. Estimating tle Cost of Accidents 1984. Nationat Safety Councit Annual Bulletin, Chicago, Ill. 11. B. C. Kragh, T. R. Miller, and K. A. Reinert. "Accident Costs for Highway Safety Decisionmaking." Public Ro¿ds, Vol.50, No. l, June 1986. 12. A Manual of User Bencfrt Analysis of Highway and Bus-Transit Improvemøtt* American Association of State Highway and Transportation Offlcials, Washingtor¡ D.C.,1977. 73. Specíal Report209: Higlway Capacþ Man¡¿l. TRB, National Research Cormcil, Washington, D.C., 1985. L4. Geotnetric Design Guide for Resurfacing, Restoration, and Rehabilitøion (RRR) of Híghways and Streets. American Association of State Highway and Transporta- tion Officials, Washington, D.C, 1977. 15. A Polícy on thc Geometric Design of Highways and Streets. Lmsrican Association of State Highway and transportation Officials, Washington, D.C., 1984. 16. J. C. Glennon, T. R. Neuman, and J. E. læisch. S$ety and Operarional Considera- tionsfor Desígrs of Rural Higlway Curves. FHWA, U.S. Departrnent of Transpor- tation, Aug. 1983. 17. D.L. Ivey, R. M. Olson, N. E. Walton, G. D. Weaver, and L. \Y. Fr¡n. NC/r'RP Report 203: Safety at Na¡row Bridge Sites. TRB, National Research Cormcil, Washington, D.C., 1979, 63 pp.

6 Tort Liability and Geometric Design In recent years highway agency administra¡ors have become increasingly concerned about the growth of ¡ort claims. Such claims allege that highway agencies have committed a legal wrong by improper or negligent highway design, operation, or maintenance that became a cause or partial cause of a highway accident. Claims against highway agencies are part of a nationwide problem of rising liability insurance premiums and incre¿sing costs of tort actions. As a result of tort claim concerns, several important questions have arisen about resurfacing, restoration, and rehabilitation ßRR) ståndards and design practices: . Are RRR geometric standards a frequent issue in ûort claims? ¡ If a state adopts and applies special RRR standards less stringent than new construction st¿ndards, will this action increase or decrease the chances of subsequent tort claims? ¡ In what ways can RRR projects affect a highway agency's chances of tort claims? ¡ What design practices will reduce a highway agency's exposure to suc- cessful tort claims? These questions are explored in light of available ståtistical evidence, case study findings, and published research. Analysis of information on tort claims in four states (Rorida, Loursiana, New York, and Pennsylvania) indicates that the geometric design features covered in RRR standards are usually not the t74

TORT T.TABILITY I75 cenEal focus of ort claims. Pavement features, üafûc control devices, and roadside bar¡iers account for the large majority of ¡ort claims in Pennsylvania and Louisiana. By conecting hazards related to fhese conditions as part of RRR projects, highway agencies have an opportunity to reduce the chances of tort claims. Furthermore, by documenting key design decisions on RRR projects, highway agencies improve their defense against tort claims. BACKGROIJND ON TORT LIABILITY Tort is deûned as a civil wrong or injury, and a tort action seeks repayment for damages to property and injuries to an individual. If a defendant is found negligent in his actions or lack of action, he is liable for a tort claim and must compensate the plaintiff. Staæ laws and rulings differ regarding tort claims against a govemmental entity. In most states the courts or state legislatures have eliminated sovereign immunity whereby an individual cannot sue the state or its agents for negligence. Only three states continue to maintain sovereign immunity against tort claims, and even in these states the doctine has been weakened. A series of court decisions between 1959 and 1961 est¿blished the legal foundation for abolishing the doct¡ine of sovereign immunity (1). In these decisions courfs ruled that immunity of states or other govemment entities was inherently unfair and illogical. Also, past judicial rulings contained numerous exceptions to the doctrine, producing incongruous results. Courts concluded that the states wero capable of assuming financial loss from tort judgments, especially with the availability of liability insurance. Increasingl¡ courts have held government entities responsible for negligent conduct. while state legislatures have revised their statutes to allow tort suits against the state for designated activities. Nevertheless, 47 states have retained a limited form of immunity known as discretionary immunity under which planning and design activities are exempt from liability. Because tort laws vary and local courts rely more heavily on precedents set by their stâte courts than on those set in other states, guidelines developed for inærpreting discretionary immunity and handling tort claims in one state may not be applicable to another. Concern over the costs of tort claims and liability insurance has grown sharply during the 1980s. An increasing number of claims have received considerable media exposure (2) and have led many business and municipal government leaders to advocate some type of tort reform l3). The unavailability of insurance coverage, or large premium increases, have stimu- lated nationwide discussion of an impending insurance crisis. The insu¡ance industry contends ttrat the increase in the number and unpredictability of multimillion dollar jury awards has caused large

176 DESICNING SAFER ROADS unforeseeable operâting losses. Lending support to this, the U.S. Deparrnent of Justice Tort Policy Working Group issued a report citing a sevenfold increase in product liability cases handled in federal courts over the past decade (4). Arecent survey of 145 cities by the U.S. Conference of Mayors revealed that 15 percent of U.S. cities have canceled or cut back services because of rising tort claims and insurance costs (5). In conEast, a study by the National Center for State Courts termed the so-called "litigation explosion" of the 1980s a myth (6). T\e study indicated that the increase in total tort claims filed in ståte courß r,vas no greater than the population increase during tha IOROq Claims Against Híghway Agencies Highway agencies are spending substantial sums as a result of tort claims. The costs of handling tort claims include not only the direct costs of judgment awards, settlements, and insurance but also attomeys' fees and the cost of engineers' and other staff time. A series of surveys conducted by the American Association of Søte Highway and Transportation Officials (AASIITO) provides the best informa- tion available on the extent and costs of tort claims against state highway agencies. The latest survey includes information through 1982 (7). The AASHTO survey indicates that the average amount paid by a state in judgments and settlements during fiscal year 1982 was $894,000, based on 26 stâtes reporting. In some siates, however, the costs of tort claim awards and settlements are much greater than these average figures. In fiscal year 1982, for example, California spent more than $5 million on tort claims (7).Penn' sylvania spent $72 million on tort claims alone between 1979 and 1986 and currently budgets $25 million per year for awards and settlements l8). From 1982 to 1984 Louisiana was assessed more than $32 million in judgments from tort claims (9). Large cities are also susceptible to large claims-New York City paid more than $6 million in claims in 1985 for pothole-related accidents alone (10). AASFilO surveys indicaæ that the average insurance premium for state highway agency liability coverage initially escalated and then declined in the period between 1975 and l98I (7). The average premium peaked in tscal year L976-1977 at about $900,000 per state, but dropped to less than $500,000 by fiscal year 1980-1981. This variation in average premium cost can be attributed to the changes made by søte highway agencies in the management of their tort liability exposure and the competition in the insurance industry in the late 1970s and early 1980s. Some ståtes iniúally increased their insurance coverage, but later dropped coverage completely, became self-insured, and

TORT UABILIT-r T77 developed in-house expertise to litigaæ thek tort claims. The high interest rates in the late 1970s and early 1980s created favorable investment oppor- tunities for insurance companies causing them to lower rates in an attempt to increase volume. Basis for Tort Claims Against Highway Agencies Negligence can be alleged on two grounds particularly relevant to highway agencies: . Agency (or person) improperly performs its duties (misfeasance). o Agency (or person) fails to perform its duty (nonfeasance). state courts use different procedures to estâblish negligence in an accident. Forty-two states use some form of comparative negligence by which the judge orjury assesses the degree to which each party is responsible for fhe accident. The initial sæp in this process is to establish the proximate cause of the accident; that is, the primary cause without which the accident would not have occurred. Proximate cause is often at issue if the plaintiff is in part responsible for the accident; for example, if he was intoxicated and attempts úo prove that the roadway featu¡e in question would have caused the accident even if intoxication had not been a facüor. under comparative negligence, if an award is made ûo the plaintiff each defendant is responsible for paying ttre assigned percentage of negligence. However, in approximately 80 percent of the states in which the doctrine ofjoint and several liability prevails, a defendant's liability is not limited to his percentage of fault-he is potentially liable for all of the award if other defendants cannot pay their share (11). Theoretically, the defendant may bear only I percent of the fault but be forced to pay all of the damages when the doctrine ofjoint and several liability applies. This principle has led plaintiffs to include in their claims defendants with "deep pockets," such as govem- ment agencies and large corporations, in order to be assured of receiving a payment if damages a¡e awa¡ded. In response fo this action, the legislatures of 11 states have recently passed laws limiting or abolishing joint and several liability (12-14). Followed in eight states and the District of columbia, contributory negli- gence is another procedure courts use to apportion fault. under this procedure, if the plaintiff is judged at fault in an accident, even if only by a small degree, he is not entitled to claim damages from any other party involved. As state legislatures or state supreme courts have eliminated or weakened sovereign immunity or adopted comparative negligence as the basis for

178 DESIGNINGSAFERROADS apportioning fault, opportunities for successful negligence claims against the state have increased. I-awsuits have alleged negligence in virtually every activity of state highway agencies, but maintenance activiúes are more vulner- able to tort suits û). Because maintenance divisions of highway agencies have standard operat- ing procedures, their work can often be easily tested for negligence. Divisions that plan and design roadways, however, must rely on personal judgments, and assessing negligence is not so straightforwafd. Recognizing this distinc- tion, some st¿tes retain discretionary immunity to cover duties of the state that -^-,,ia c,,^L i¡',{c-a-ro i-^1"¡{ina nlannìnc nnd ¡lesisn nctivities Provirle¡l arçquuv ùuvr¡ Juué¡¡¡v¡rÐr ¡¡¡vru\A¡¡ó e¡¡s evv¡Þ¡- reasonable process is followed, planning and design work may be immune from fort claims. As states abandoned sovereign immunit¡ municipaliúes that had immunity under the umbrella of state law also became vulnerable to tort suits. However, many towns and cities, formed by incorporating, were already susceptible to tort suits because they are considered private corporations. IMPLICATIONS FOR RRR DESIGN STANDARDS AND PRACTICES RRR Improvements and Tort Claims Little is known about how frequently the geometric features addressed by RRR design stândards a¡e cited in tort claims against highway agencies- Few states maintain data on toft claims by alleged defect. Further, classifying tort lawsuits is difficult because most involve several defects that differ in impor- tånce. Four states-Florida, Louisiana, New York, and Pennsylvania-maintain summary information on tort claims that indicates alteged defects in roadway condition. As discussed in the following paragraphs, the geometric features addressed by RRR standalds are seldom cited in tort claims in these four Stâtes. However, tlte experience of these states may not be representative of other st¿tes. In California, for example, state officials indicaæ that geometric features ¿lre more frequently cited in tort cases than experience from the other four states would suggest. Pennsylvania maintains a summâry list of tort claims, beginning at the time the state eliminated its sovereign immunity in 1978, and has compiled infor- mation on the number of claims, amount of settlements, and the number of fatalities for the various conditions alleged in tort suits (8). Similar informa- tion is available for Louisiana covering the period 1982 tD 1984. Louisiana reported the amount of judgmens awarded by the court and the number of

TORT UABILITT I79 claims alleging a defect in roadway condition (g). rn addition, New york reported the otal number of claims by atleged defect in roadway condition for 1983 rhrough 1984 (15),andFlorida provided summary data on the number of tort claims filed against the søte highway agency by alleged defect in roadway condition (Jó). Geometric features usually covered by RRR standa¡ds account for a small percentage of the tot¿l claims filed against highway agencies in pennsylvania, Louisiana, New York, and Florida (see Tables Gl and 6-2). Daø from Pennsylvania and Louisiana are based on awa¡ds and settlements, whereas TABLE 6-1 Costs of Settlements or Judgments in Tort Cases Pennsylvania (1979-March 1986) Louisiana (1982-t984) Condition Percent Amount Paid($) of Total Judgments(g) PercentolTotal Geometric featureso Pavementó Tiaffic control devices" Roadside barriersr' Work area protection, Other./ Totals 2,6s6,094 9,434,973 4,s64,015 5,477,t15 1,235,109 7,9t3,942 3 1,28 1,308s 44t,928 1.2427,523,066 84.88 3,17 3,rt3 9.7946,266 0.14001,282,103 3.95 32,426,476 100.00 8.49 30.16 14.s9 t7 .51 3.9s 2s.30 100.00 aCross section, alignment, intersections. 'Påvement edge drops, slippery pavements, potholes, surface deterioration. 'Signs, signals, pavement markings.4Guardrails, median barriers. ?Tèmporary conditions during construction. /E_mployee operations, weather-related conditions, other roadway features.gClaims frled.afterPennsylvania's Sovereign Immunity Act of Sêptember 1978. The total paid in tort claims was$72 million, which includes the listed claims, plusapproiimately 450 claims submitteã beforeSeptember 1978 and claims that were not coded with theiisted cìaims. Cases fileo after Sept..¡"i rpu s -the date of the act - are subject to a payment limit of $250,000 per person and $ I millión per accidenr. daø from New York and Florida are based on tort claims filed against the highway agency. cross section, alignment, and intersections account for g.5 percent of costs in Pennsylvania and 1.2 percent ofjudgment costs in l-oui- siana. of the cases in which a geometric feature is at issue, horizonlal and vertical curves (4 percent of totål costs in pennsylvania and !.2 percent ofjudgment costs in Louisiana) are the most often cited. The impòrtance of geometric features in tort claims may have been obscured in l-ouisiana because of the unusually large percentâge of claims relaæd to pavement, condition. Improper design of geometric features represents 8.0 percent of toøl ¡ort claims filed in New York from 1983 to 1985. Data from Florida indicate generally similar results; between Júy 1972 and November 19g6 design of geometric featu¡es accounted for 7.1 percent of total claims ûled.

T8O DESIGNINGSAFERROADS TABLE 6-2 Tort Claims Filed in New York and Florida New York (1983- 1985) Florida (1972- 1986) Condition No. olClaims Percent No. olClaims olTotalPercentolTotal Geometric featureso Pavementå Tiaffic control devices" Roadside barriers/ Work area protection' Otherr Total 467 1,0r6 t,238 397 426 2,295 5,839 8.0 r7.4 2t.2 6.8 't.3 39.3 i0û.00 556 2,0t9 876 '7 .t 25.8 1t.2 4,3',14 55.9 l,ózJ luu.uu aCross section, alignment, intersections. åPavement edge drops, slippery pavements, potholes, surface deterioration csigns, signals, pavement markings. dGuardrails, median barriers. "Tèmporary conditions during construction./Empioyee operations, weather-related conditions' other roadway features' In Pennsylvania the Office of Attorney General settles most ûort claims out of court, whereas in Louisiana tort claims are more often decided by a judge't Neither Pennsylvania nor Louisiana reports cross-section features, which include lane widths, shoulders, sideslopes, and ditches, as the basis for any tort claims. Curve alignment accounts for nearly 50 percent of the tort claim settlements in the geometric category in Pennsylvania. Intersections account for a smaller percent of total settlements. In 1982 five claims in Louisiana a[eged inade4uate sight distånce. Pavement features, including edge drops, potholes, surface deterioration, and slippery pavements, account for nearly one-third of the s€ttlement costs in Pennsylvania and more than 80 percent of the judgment costs in Louisiana. Of these features, pavement edge drops account for more than 20 percent of seülement costs in Pennsylvania, whereas shoulder condition (which includes edge drops) accounts for 66 percent in Louisiana. Although geometric features account for slightly less than 2 percent of the total number of claims in Pennsylvania, they account for more than I percent of the associated fatalities. ll,ouisiarra laws on tort liability are unique in that they are based on the Napoleonic Code of Civil l-aw, whereas lhe laws of other states arc based on English Commor I:w.

TOKI UABILITT 181 Susceptibility of RRR Projects and Standards to Tort Ctaims The standards selected for RRR projects, the design process followed, and the scope of the improvements may influence ttre litigation of future tort claims. The issues that might a¡ise in a fort action a¡e o Did the project meet the appropriate design standards? . Are the standards reasonable? o W'as the desþ prccess reasonable? r Did the improvements correct existing dangers? r should unimproved roads be judged by standards used for roads that are improved? The resolution of tort claims alleging an inadequate geometric design is contingent on deærmining the appropriate set of design standards used ûo assess negligence. New construction design sønda¡ds are continuously updated. In most states, the stândards in effect af the time a roadway was constructed or reconstructed are generally used to assess the adequacy of the road. For example, in the 1972 case of Hampton v. state Highway convnis- sion, Kansas, the court held that liability could not be predicated on design defects alone because the design was adequate when tre highway was built and must be judged by standards prevailing at rhe time (1). Determining whether a highway improvement project is sufficiently exten- sive to qualify ¿ìs reconstruction can be a key issue in a tort claim'because reconstruction projecs usually must meet curent new construction standards. For example, the defrnition of reconstruction was cenEal to the 19g5 case of Brendv.Iowain which the judge ruled that resurfacing and seal coating do not constitute reconstruction and therefore the søndards of the period during which a highway was originally built continue to apply e7). Even when a highway has not been reconstructed, the adequacy of ttre existing design may be challenged when conditions have changed iubstan- tially such that a clear danger exists. For example, the state of california lost a tort c¿¡se (Baldwinv. state) in which the omission of lefçtum lanes was known to be dangerous. Alfhough the highway w¿rs designed in accordance with applicable standards, raffrc conditions had changed over time. The court ruled that "once ttre entity has notice that the plan or design, under changed physical conditions, has produced a dangerous condition of public property, it must act reasonably üo colÏect or alleviate the hazard,, (I). The RRR project design process provides ân opportunity for highway agencies to review safety and raffic conditions. An accident analysis and site inspection should reveal any hazardous conditions, which, if leflunattended, could become the basis of a future tort action.

I8Z DESIGNINGSAFERROADS Deficient roadside signs or pavement markings and pavement edge-drop problems, which are ofæn the bases of tort claims, can be routinely conected on RRR projects. Defense of a RRR Project Design Although planning and design activities are exempt from liability in most states, this immunity has been held not ø apply to decisions made without nrior studv or conscious deliberation. I¡ King- v. State, for example, the court held that discretionary immunity was not available to a state highway agency that failed to exercise "due care" in planning a traffic light system (1). ht such cases, documentation of the planning process should be a part of the state highway agency's defense. For RRR projects, documentation should demonsFate that safety aspects of the roadway design were properly considered. Reports that identify deficien- cies in existing roadways are poæntially threatening to the public agency preparing the report ifthe defrciencies are not addressed. Thus, ifan exception to an applicable design standard was granted, documen[ation should explain the re¿sons for the exception and show that logical and orderly procedures were followed in obøining it. When a highway agency contemplates a design exception for a geometric or roadside feature, it should be prepared to prove why the feature need not meet the same standards as other facets of the roadway design. Often, the best defense in ttris situation is to demonstrate that the safety cost-effectiveness of further upgfading the fean¡re does not meet any reasonable criteria. Part of this defense is evidence that special care was taken in determining that an excep- tion was appropriate. Courts seldom rule that the unavailability of funds is justiûcation for not correcting an alleged defect, but the issue of availability of funds can be part of the defense in relation to the agency's programming procedures. The following points are important to such a defense: the agency is aware of the condition of its facilities; defrciencies have been ranked on a logical basis; and, given the existing funding, are being corrected in the order of priority. Appropriate warnings or other temporary measures should be used to alert the public that deficiencies have not been corrected (18); the highway agency can then afñrm that it has performed its duties in the best way possible with the available resources. However, in some stâtes such as Pennsylvania, courts allow neither inadequate funding nor more critical safety priorities elsewhere as a justification for inaction. Sometimes segments of a highway may be designed to new construction stândards and other segments may simply be upgraded according to less

TORT TIABIUTY 183 stringent RRR standards. A highway agency involved in a tort case under such circumstances should use a defense similar to tle one it offers for a design exception, and, again, costs of construction alone are not a suffrcient criterion. If the process by which portions of a roadway are selected for rehabilitation instead of reconstruction is not a¡biEary, courts will usually hold that RRR stândards are applicable. If RRR design srândards are directly implicated in a tort case (e.g., the stândffds are alleged fo be ûoo lenient), the søæ highway agency may have immunity from liability. As noted previousl¡ many sf¿te legislatures have ganted states immunity from liability for planning and design activiries, which include setting søndards. such activities a¡e considered discretionary and require the special expertise of trained personnel. states without sovereign immunity are clearly liable for operational activities such as day-to-day maintenance. As søæd in the landmark case W¿jss v. Fote (New york, 1960), "Lawfully authorized planning by govemmental bodies has a unique charac- ter deserving of special treatment as regards the extent to which it may give rise ø tort liability. To accept a jury's verdict as to the reasonableness and safety of a plan of governmentral services and prefer it over ttre judgment of the governmental body which originally considered and passed on the matter would be to obstruct normal governmental operations and to place in inexpert hands what the legislature has seen fit to entrust to experts,' (1). In order to receive immunity for planning and design activities, a state must thoroughly document the design process in order to defend challenges. The ruling in weiss v. Fote set certain conditions under which states should not be granted immunity from liability for planning and design activities: (a) rhe plan or design was not duly considered, (b) there is no evidence that due care was used in preparation of the plan or design, (c) no reasonable offlcial could have accepted the plan, and (d) approval of the plan was arbitrary. Simply stated, a rational and orderly process must be followed if a plan or design is to be considered immune from claims of negligence. If a feanre built during construction \ryas not called for in the plans or was alæred from the specifica- tions, it is open to a claim of negligence in a tort action. STJMMARY Available data on the characteristics of tort claims against highway agencies indicate the following: ¡ Geometric featu¡es account for a small percentåge (about 10 percent or less) of ¡ort claims in the four states with available data. The experience of these states may be too atypical to generalize, but it is probable thai geometric features account for less than one-fourth of all tort claims.

184 DESICNINGSAFERROADS . RRR projects routinely correct or upgrade fe¿tures such as paYement edge drops, signing, guardrails, and median barriers, often the targets of tort claims. o When roadway geometrics are an issue in a tort claim related to a RRR project, application of RRR desþ standards, less stringent than new con- struction standards, is unlikely to be a basis for the tort claim. Most states have some type of design immunity that may cover the use of design standards as long æ rcasonable procedures are followed. Highway agencies can minimize their susceptibility to tort claims by o Using the RRR design process as an opporfirnity to identify and correct hazardous conditions, with special emphasis on nongeometric features prone to tort actions; . Documenting the reasons for making a RRR improvement on a highway segment rather than reconstructing it, particularly if nearby segments are being reconstructed; and o Documenting the entire design process with particular emphasis on the rationale for seeking design exceptions. REFERENCES 1. L. W Thomas, ed. Selected Studies in Highway Law' Yol. IV TRB' National Research Council, Washington D.C', 1982. 2. "Cowting Disaster: Lawsuit Costs Keep Rising, Affecting Economy, The Wall Street fournal, May 1ó, 1986. 3. "Finger-Pointing Distinguishes Anempts to Fix Blame for Liability Crisis." Natíonal loumal, Feb. 15, 1986. 4. Report of the Tort Policy Working Group on the Causes, Erteú, and Policy Implicatiotts of tle Current Crisis inlnsurance Availability and Affordaåiltfr. U.S' De,partment of Justice, Feb. 1986. 5. Municípal Liability Concerns: A 145-Citt Sumey. U.S' Conference of Mayors, Washingüon, D.C., July 1986. 6. R. T. Roper.The State Court Case I'oad Statistics, l9M Anrunl Report. National Center for State Courts, Williamsburg, Va., 1986. 7. Survey on the Status of Sovereign Immunity in the States,1983. Administrative Subcommittee on lægal Atrairs, American Association of State Highway and Transportation Officials, Washington, D,C. 8. Unpublished data. Pennsylvania Department of Transportation, Harrisburg, March 1986. 9. Unpublished data- Louisiana Department of Trarsportation and Development, Bafon Rouge, Jan. 1986. 10. "New @onding) Agent in the War on Potholes." The New YorkTírncs, Nov. 12, 1986.

11. L2. 13. 14. 15. 16. 17. 18. TORT T]ABTLITY 185 The Need for Legislative Reform of the Tort System. American Tort Reform Associatior\ Washingtor¡ D.C., 1986. "Insurance Woes Spur Many States ø Amend Law on Liability Suits.,, The New YorkTimes, March 31, 1986. "Limits on Lawsuit Awards Are Voted in Califomia." The Walt Street fournal, Iune 5, 1986. "Insurance Firms Proût From Crisis." The Washington Post, De¡,. Zl, 1986. Unpublished data. New York Department of Transporøtion, Alban¡ Jan. 1987. Unpublished data. Florida Deparunent of Transportation, Tallahassee, Nov. 1986. "State Wins: Road Was Not 'Reconstn¡cæd'," TranSafety Reporter, Aug. 1985. NCHRP Synthesis of Highway Practice 106: Practical Guidelines for Minimizing Tort Lîabílity. TRB, National Research Council, Washingron, D.C., 1983, 40 pp.

7 Findings and Recommended Design Practices for Resurfacitg, Restoration, and Rehabilitation Projects Summarized in this final chapter are findings on tlle selection and design of resurfacing, restoration, and rehabiliøtion (RRR) projects; the cost and safety trade-offs involved in improving geometric features on these projects; and the influence of design standards. Recommended practices are presented that will promotô more safefy-conscious design and, in turn, more safety cost-effective RRR projects. FINDINGS Resurfacing, restoration, and rehabiliøtion projecs enable highway agencies to improve highway safety by selectively upgrading existing highway and roadside features without the cost of full reconstruction. For example, widen- ing lanes and shoulders on two-lane rural highways on the federal-aid systems alone could save about 1,000 lives and prevent nearly 30,000 injuries each year (Chapter 5). In the last few years many highway agencies have paid increasing attention to safety on federal-aid RRR projects, Nevertheless, further opporn¡nities for safety cost-effective improvements on RRR projects often exist. Highway agencies can take additional steps to improve highway safety on RRR projects while performing pavement repairs or other highway preservation activities. In practice, however, most highway agencies have had difficulty in striking a balance between these two objectives. This happens in part because the effect of RRR projects on highway service life is relatively predicøble 186

FINDINGS AND RECOMMENDATIONS 187 whereas the effect on safety appears less certain. Inadequate information about the safety payoff of improvements to existing highways underlies much of fhe confusion and difference of opinion about the appropriate level of safety-motivated improvements on RRR projects. Although relationships between safety and highway geometry are not clearly understood, available data demonstrate that highway design strongly influences highway safety. Moreover, for selected features on two.lane rural highways, the study committee found sufûcient data and research to make judgments about the most probable relationships so that it could analyze fhe safety and cost Eade-offs involved in improving these features. These ana- lyses show that improving existing highways to match new construction design standards is generally unwarranted. By permitting more highway miles to be improved earlier, less stringent RRR design Standards can better enhance systemwide safety. The review of current RRR design practices conducted for this study concluded that federal-aid RRR projects usually do enhance safety and that highway agencies generally have paid gleater attention to safety since Con- gress added safety enhancement as a RRR objective. But the review, coupled with the study's cost-effectiveness analyses, also concluded that many oppor- tunities for low-cost safety improvements are neglected and that RRR funds currently spent for safety improvements could be redirected for greater sys- temwide safety gains. A number of factors are responsible for this situation: o RRR dcsign practices vary widely from agency to agency. Some highway agencies follow exemplary practices to address safety needs, some of which are incorporated in later recommendations. Others do not place enough emphasis on evaluation of existing conditions and accident histories to detect safety needs and analysis of opportunities for meeting these needs. o RRR projects are initiated primarily to address paveïtent repair and rehabilitation needs. safety needs are often not addressed until a project has been programmed and preliminary design has been initiated. By then, little schedule flexibility remains to accommodate geometric improvements that require additional time for design or right-of-way acquisition. o Federal-aid RRR projects frequently widen lanes and should¿rs but seldom reconstruct sharp curves or replace bridges with narrow decks. Because tlere is a higher concentration of accidents at curves and bridges, improvements at these locations, despite the high costs, can sometimes be more cost-effective with respect to safety than routine cross-section improve- ments. . Not enough is known about the s$ety gains that will occur after the geometry of existing highways is improved or other s$ety-oriented improve- ments are nade. Avulable information is not always in the hands of designers, or in a form that can be applied without ambiguity.

188 DESIGNING SAFER ROADS o Engineers who administer state traffic and safety programs seldom partícipate in tlæ design of RRR proiects. They are usually the agency staff members most knowledgeable about accident daø and special safety mea- sures, but they have other duties and assignmenfs. RRR project designers often have not received suffrcient education in safety engineering. Design standffds alone cannot address these factors fhat collectively limit the safety gains of federally funded RRR projects. Within the overall process of planning, selecting, and designing RRR projects, the influence of safety standards is small. RRR søndards, which can affectronly a few key design features, cannot be tailored to fit all possible, or even most, circumstances and consEaints encountered in a given stâte or a speciûc siæ. Both the special RRR søndards proposed by the Federal Highway Administration (FIIWA) and the American Association of Staæ Highway and Transporøtion Offrcials (AASIITO) for nationwide use on federal-aid RRR projects and those adopted for use in individual states seldom contain firm minimum standards for features other than lane and shoulder widths. Furthermore, where ûrm lane and shoulder width standards are applicable for federal-aid projects, such standards can be circumvented. For instance, some highway agencies reserve federal aid for pavement repairs on projects in which lane and shoulder widttrs already meet the standards. Pavement repairs on highways for which cross-section improvements are most needed are deferred or undertaken using state or local funds so that widening either is not required or is determined by less stringent standards. A variety of practices are recommended ttrat encompass the entire RRR process but with special focus on design practice. In selected instânces, federal, state, and local highway agencies cân use the recommendations, along with design aids, published manuals, and local experience to develop or modify minimum design ståndards for RRR projects. The Secretary of Trans- portåúon is required by statue to ensure that, for federal-aid RRR work, projects are designed and constructed in accordance with standards that extend the service life of highways and enhance highway safety. To accom- plish this, the Secretary, acting through the Federal Highway Administration, must either set nationwide RRR standa¡ds or approve ståndards adopted by individual states. For either approach, the committee's recommendations provide guidance. The recommended design practices provide for a safety impact evaluation in which the safety consequences of existing conditions and potential improvements are evaluated to develop more safety+onscious designs. This process should improve the procedures for selecting the most safety cost- effective improvements. If these recommendations are followed for federal-aid projects on nonfree- way highways, project spending for lane and shoulder widening will generally

FTNDINGS AI,I D RECO M MEN DATI oNS TABLE 7-l Organization of Study Recommendations Safet y Cons cious D es ign Proces s r89 [. Assessment of Site Conditions Affectine Safetv2. Determination of project Scope3. Documentation of the Designprocess _ 4. ReviewbyTiafiìc andSafetyEngineers Design Practicesfor Key Highway Features 5. Minimum Lane and Shoulde¡Widths 6, 7. Horizontal Cu¡vature and Superelevation 8. Vertical Curvature and Stopping Sight Distance9. Bridge Width 10. Sideslopes and Clear Zones I 1. Pavement Edge Drop and Shoulder Type 12. Intersections 13. Normal Pavement Crown Other Design Procedures and Assumptions 14. Traffic Volume Estimates for Evaluating Geometric Improvements 15. Speed Estimates for Evaluating Geometric Improvements 16. Design Values lor Geometric Improvementsi7. Design Exceptions Planning and Programming RRR projecß 18. Screening of Highways Programmed for RRR projects ^19. Assessment of the Systemwide potential for Improving Salety Safety Research and TÌaining 20. Special Task Force to Assess Highway Safety Needs and priorities 2 I . compendium of Information on safety Effects of Design Improvements22. Increased Research on theRelationships Between Safeiy and nesigrr-- - 23. Salety Tiaining Activities for Design Engineers decline and spending for alignment, bridge, roadside, and intersection improvements, as well as project design, should increase. In some states these shifts may decrease typical RRR project costs; in others they will increase typical costs. Nationwide, the typical project cost will probably increase slightly but not enough to measurably interfere with RRR pauemenirepair and preservation activities. Although the study provides guidance in many areas (such as the relation- ships between accidents and specific geometric features), more detâiled guidelines, fu¡ther research, and increased safety education for designers will be needed to address many of the questions that remain about how !o best enhance safety through highway design and operations. Nevertheless, the study committee concluded that a design process that emphasizes safety, even wittr some parts that will be handled differently by different agencies and designers, will ultimately produce better RRR projects. T!.e 23 study recommendations are organized into ûve categories (Table 7-r): 1. safety conscious design process-recommended general practices that comprise a systematic process for considering safety durinf RR project design.

190 DESIGNINGSAFERROADS 2. Design practices for key highway fe¿tu¡es-recommendations that spec- ify the existing highway conditions that wdnant a geomeEic improvement ouright or a serious evaluation of a geometric improvement, as well as design practices that should be followed routinely for key features. 3. Ottrer design procedu¡es and assumptions-recommendations that cover design factors such as speed and traftc volume, as well as guides for selecting design values for geometric improvements and seeking design exceptions. 4. Planning and programming RRR projects-recommendations for including safety considerations in ttre planning and programming of RRR n¡nianfç 5. Safety research and faining-recommendâtions for increased attention to highway safety research, training, and education. These recommendations generally apply to nonfreeway RRR projects whether or not t¡ey are funded with federal aid. For federal-aid RRR projects in particular, the Secretary of Transportation, through the Federal Highway Adminisfation, should take ttre necessary steps to implement the recommen- dations in the first three categories-safety-conscious design process, design practices for key highway features, and other design procedures and assump- tions. Taken together, these recommendations @ecommendations I through 17) comprise a practical national policy on RRR project design that will be more safety cost€ffective and comprehensive than an extensive set of rigid minimum stândârds. The fourth category of recommendations, planning and programming RRR projects, is direcæd to state and local highway agencies that have the authority to perform these functions for federal-aid projects wittrout federal oversight. The flnal category safety research and training, is directed to the lalger highway community with specific recommendations intended for the Con- gress, FFIWA, AASI{T0, and state and local highway agencies. SAFETY.CONSCIOUS DESIGN PROCESS Significant improvements in safety are not automatic by-products of RRR projects; safety must be systematically engineered into each project. To do this, highway designers must deliberately seek safety oppornrnities specific to each project and apply sound safety and traffic engineering principles' High- way agencies must sEengthen safety considerations at each major step in the design process, neating safety as an inægral part of design and not as a secondary objective. These actions require that highway agencies devote greater resources to RRR project design, and to be successful, the added design labor and effort must be complemented with safety-oriented training.

FINDINGS AAID RECOMMENDATIONS 191 Assess Current Conditions Recommendation l: At the begínning of RRR project design, highway designers should assess existing physical and operatíonal conditions fficting safery: o Analyze accident and tavel dats tn identify specific safety problems that might te corrected and to determine if the site has been unduly hazardous compared with the systemwide performance of similar highways. o conduct a tlørough site inspection using penonnel trained to identify features that pose safety hazards under common operating conditions and recognize opportunities for safety improvements. c Determine and verify eisting geometry, including lane and shoulder widths; degree, lengfh, and superelevation of horizontal curves; stopping- sight-distance ¡estrictions; locations and design of intersections; sideslòpes; and clea¡ recovery distances. . Determine prevailing speeds at approaches to horizontal curves and at curves or hill crests with possible stopping-sight-distance restrictions. The designers of RRR projects can draw on substantial information that bears directly and indirectly on safety. unlike designers of new highways, designers of RRR projects work with a highway in operation, one with an established safety record, and one for which design and operational charac- teristics can be observed and measured. Not all state highway agencies take advantages of these favorable circumstances; few perform all of the activities recommended. Many failures by highway agencies to correct safety hazards or to make low-cost safety improvements on RRR projects can be traced to a breakdown in the design process at this early søge. Determine Project Scope Recotnrnendation 2: In addition to pavement repairs and geomztric improve- mcnts, designers of RRR projects should consider ønd, where appropriate, incorporate other intersection, roadside, and trafftc control improvemenx tlat may enhance safety. such improvements should be routinely considered by designers and should include improvements at intersections and driveway entrances to increase sight distance and reduce vehicle conflicts; replacement or rehabiliøtion of obsolete bridge rails and guardrails; removal of roadside obstacles and unnecessary guardrails; slope flatæning; ditch relocation and regrading; and new or improved signing, pavement markings, and other traffic control devices.

I92 DESIGNINGSAFERROADS These improvements can provide significant reductions in the frequency and severity of accidents. The safety benefrts of these improvements, coupled wittr ttreir low costs, are such ttnt they can be higtily cost-effective on RRR projects. In its review of completed federal-aid RRR projects, the FIIWA òoncluded that many highway agencies missed opportunities ûo enhance safet¡ ofæn at low cost, ttrough improvements such as those described (Chapær 2). Such omissions occur when designers conceive their missions too nanowl¡ for example, concenEating on pavement repairs or confining the safety review to a few key highway features. Document the Design Process Recomm.endation 3: Beþre developing construction plans ønd specífications, designers should. prepøre a søfery and design report. Safety components of this report should describe tlu following: o Existing geometric and roadside features, taffic volwræs and speeds, and accident history; . Applicable minimum design standards; o Specifrc søfery problems or concerns raised by accid¿nt data, fteld inspection, or concerns expressed by the publíc; . Design optiôns for correcting safety problems and the cost, safety, and other relevant impacts of these options; . Proposed exceptions to applicable design standards and the rationøle to support the excePtions; and . Recomm¿nded design proposal and its cost and safety impact' Approximately one-third of the state highway agencies reviewed for this study currently use some form of design report to document the design process (Chapter 2). Documentation of the design process improves design decisions and generally increases the accountability of those involved in the design, allowing more meaningful design reviews. Moreover, good docu- mentation is a sound defense against tort claims alleging improper design (Chapter 6). The length and complexity of the recommended safety and design report will vary depending on existing conditions and the extent of any safety problems.

F I N DTNGS A1'I D REC O M M EN DATT ON S t93 Review the Design Recommendation 4: Trffic and safety engineers should rourinely review safery and design reports, as well as proposed RRR designs before final approval. Direct participation by traffrc and safety engineers in the design of RRR projects contributes to a more safety-conscious design process. However, such participation is often not feasible. Design divisions or units within søte highway agencies are generally responsible for the design of RRR projects. Because naffic and safety programs are often adminisæred elsewhere (fre- quently in maintenance divisions), traffrc and safety specialists rarely partici- pate in the design of RRR projecs. They have other responsibilities and are not always based in disrict or regional offices where they can work directly with design teams. Nevertheless, it is feasible to have Eaffic and safety qpecialists routinely review and critique RRR design teports. Such a critique could be accom- plished ai part of design reviews common in many state highway agencies and could be coordinaæd with raining activities aimed at improving the safety engineering skills of designers (see Recommendation 23). DESIGN PRACTICES FOR KEY HIGHWAY FEATURES Recommended design practices include minimum values that can be used by FHWA and state and local highway agencies in setting minimum RRR geometric design søndards for selected featu¡es. Designers use such st¿ndards to determine whether a particulff geometric feature must be upgraded as part of a RRR project. A fe¿ture not meeting the minimum standard must be upgraded unless a design exception is sought and approved. Numerical minimum RRR sønda¡ds a¡e warranted for nationwide use when tle following conditions are met: ¡ Trade-offs between safety and performance against cost can be evaluated quantitatively, and conclusions can be d¡awn about the safety cost-effective- ness of different stândards generally applicable regardless of the stâte or project; ¡ Sønda¡ds would help refocus RRR expenditures on more safety cost- effective geometric improvements; and o Standffds would simplify parts of the design process and FIIWA ap'proval procedures, freeing design resources for the analysis of site improve- ments that cannot be covered by numerical standards.

194 DFSIGNINC SAFER ROÁ,DS Lane and shoulder widths on two-lane rural highways meet these condi- tions, and minimum values ¿¡re recommended. Cross-section features a¡e particularly importrnt because they can affect highway safety and cost over the length of a highway. Twolane rural highways account for about three- fou¡ths of all nonfreewa¡ federal-aid highway mileage and serve roughly one- fourth of all vehicle miles traveled throughout the United States (1). Where these conditions are not met, for other key features or categories of highways, other design practices are recommended that will help achieve ttre same safety objectives as minimum ståndards. In some cases, the recommen- dations specify threshold conditions that warrant deøiled evaluation of par- ticular improvements @ecommendations 5,7, and 8). In others, they specify improvements that should routinely be made or evaluated on RRR projects (Recommendations 6, 11, and 12), or design policies that should be developed on a state-by-state basis (Recommendations 9 and 10). Minimum Lane and Shoulder Widths Recomm¿ndation 5: The following minimum lane and shoulder width values are recommended for two-lane rural highways: Design Year Rurning Volum* Spee&(ADT) ("'ph) 1-750 Under 50 50 and over 751-2,000 Under 50 50 and over Over 2,000 All Combined Larc and Lane Shoulder Lane Width Widtha Width 10t29 l0 Percent or More Truclc,s LessThan I0 Percent Truclcs Combined Lane a¡ú Shoulder Widthð 11 1.2 12 L4 l7 10 10 11 11 i0 12 11 13 L2 15 12 l8 "See discussion of design traffic volume later in this chapter @ecommendation l4).öHighway segments should be classified as "under 50" only if most vehicles have an average speed of less than 50 mph over the lengrh of the segrnent. .For this mmpariscn, trucks are deñned as heavy vehicles with six or mo¡e tires. ¿One foot less for highways qr mountainous tenain.

FINDINGS AIVD RECOMMENDATIONS 195 The FFIWA and søæ highway agencies can use fhese recommended mini- mum lane and shoulder width values ûo set minimum RRR design standa¡ds. These recommended values are similar to the minimum lane ánd shoulder width values proposed by the FIIWA in 1978 but include several modificarions to improve safety cost€ffectiveness: o The average daily üaffic (ADÐ ranges are adjusted so that a larger number of roads with high ADT and fewer roads with low ADT would be improved. Lane and shoulder width improvements a¡e more cost-effective on high-volume roads than on low-volume roads. o Minimum roadway widths on roads with high ADT are increased by 2 ft in each direction. The cost per accident eliminated of adding shoulder width on high-volume roads is particularly low. o Minimum roadway widths on mountainous terrain are reduced by I ft in each direction because shoulder width improvements are, on the average, less cost-effective on moun[ainous terrain. In terms of cost per accident eliminated, the recommended values are more safety cost-effective than other standards proposed for nationwide use-about $32,000 for each accident eliminated on average compared with $43,@0 for the RRR søndards proposed by the FHWA in l97g; $41,000 for new con- struction standards; and $56,000 for rhe l97z AAsHTo RRR design guide(Chapær 5). For all federal-aid, two-lane rural highways combined, the recommended minimum values imply approximately the same overall investment as the standards proposed in 1978-a total of roughly $13 billion if all of the lane and shoulder improvements were made at curent cost levels. Application of these values, however, would eliminate about 10,000 1+o percãn4 aaoi- tional accidents annually, The recommended minimum values are stringent enough to ensure that few cost-effective lane and shoulder improvements that enhance safety are excluded from consideration during design, but are not so stringent as to imply an investment level in geometric improvements substantially beyond curent practice. For high-volume roads, the recommended minimum lane and shoulder widths are generally more stringent than special RRR søndards, but less stringent than new construction stândards. For low-volume roads, the recom- mended sønda¡ds are less stringent than the standa¡ds now used in most states. The recommended minimum lane and shoulder widths explicitly take into account vehicle speed and amount of truck Eaffic, which influènce the benefits of wider lanes and shoulders although no reliable quantitative estimates of

196 DESIGNING SAFER ROADS their incremenøl effects are available. Distinctions based on vehicle speeds and fuck percentages are common in the RRR standards currently in use in many states. These distinctions recognize that on roads with higher vehicle speeds and a larger number of trucks, wider lanes and shoulders have a greater safety payoff. I¡ss is lnown about the safety cost-effectiveness of widening urban and multilane rural highways, and minimum values have not been proposed that highway agencies can adopt as standafds. The minimum widths recommended for rural two-lane highways can be used as a guide to safety cost-effective imnrmramanrs fnr mnltilane nrrnl and urban hishwavs. Howgvgr. routinely upgfading lane and shoulder widttrs in urban areas to the minimum widths recommended for rural two-lane highways is likely to produce some widening projects that a¡e not safefy cost-effective, particularly when physical con- s6aints or high right-of-way cosfs afe involved. In such situations, designers must determine the scope of widening improvements on a case-by-case basis. Horizontal Curvature and Superelevation Recommendation 6: Highway agencies should increase tþc superelevøtion of horizontal curves when the design speed of an exisling curve is below the running speedst of approaching vehicÌes and the existing superelevation is betow the allowable mtximum specifted by AASHTO new construction pol- icies. Highway agencies shoutd evaluate reconstruction of horizontal curves when the design speed of the existing curve is more than 15 mph below the running speeds of approaching vehicles (assuming improved superelevation cannot reduce this difference below 15 mph) and the average daily trffic volume is greater than 750 vehicles per day. Neither minimum RRR søndards proposed for nationwide use in the past (other ttran now consfuction standards) nor special RRR standards currently in use set flrm requirements for feconstruction of horizontal curves. FurtÏer- more, where staæs apply new construction standards to federal-aid RRR projects, design exceptions for substandard curves afe common and curve reconstruction is infrequent. However, a number of snþ highway agencies have recognized the need to evaluate alignment improvements during RRR design and have incorporated requirements similar to the preceding recom- mendation in their special RRR standards. Current RRR standards and practices generally emphasize lane and shoul- der width improvements and do not pay enough attention to alignment lTh" 85r¡ percentile speed (the speed below which 85 percent of the vehicles travel) of approachingvehicles(beforeslowingforthecurve)shouldbeusedforthiscomparison.

FINDINGS AN D RECOMMENDATTONS r91 improvements. Review of current state RRR practices revealed that lane and shoulder widening is relatively routine but alignment improvements are uncommon. Moreover, once applicable RRR søndards have triggered a lane or shoulder improvement, designers often go beyond the RRR minimum stândard and specify enough widening to meet new construction standards (Chapter 2). Shifting RRR investments toward more horizontal curve improvements is waranted. At trafflc volumes greater than 750 vehicles per da¡ reconstruction of horizontal curves can be more cost-effective than lane and shoulder widen- ing and can reduce vehicle operating costs and travel time. Curve reconstruc- tion is rarely cost-effective at Eaffic volumes lower than 750 vehicles per day even if construction costs are low and potential user benefrts are fully consid- ered (Chapter 5). Because of the variation in costs (and safety cost-effectiveness), however, minimum geometric standards are inappropriate. The cost of reconstructing similar curves varies widely from site to site because of differences in terrain and right-of-way requirements. Thus requiring highway agencies üo evaluate curve reconstruction-estimating the added costs, safety benefits, and other user beneûts that would result-is the best means available to emphasize the need for selected curve improvements. Good design minimizes inconsistencies in highway geometry that require motorists to make abrupt or frequent changes in speed. To identify such inconsistencies on existing horizontal curves, designers must compare the design speed of the curve as presently constructed with some measure of the running speeds of trafñc on the highway outside the influence of the curve. This snrdy recommends use of the S5th percentile running speed, measured where approaching rafûc has not yet reduced speed. Designers should also consider whether successive curves should be analyzed singly or as a group. Recommendation 7: At lnrizontal curves where reconstruction is unwar- ranted, highway agencies should evaluate less costly sSety measures. Such measures include widening lanes, widening and paving shoulders, flattening steep sideslopes, removing or relocating roadside obstacles, and installing traftc control devices, raised pavement markings, and reflective guideposts. At present highway agencies apply these measures inconsistently. In many cases, safety can be improved at horizontal curves \ilithout costly reconstruction. Depending on site conditions, improvements to curves, short of reconstruction, can be an inexpensive and effective means of reducing the severity and frequency of accidents.

198 DESIGNING SAFER ROADS Vertical Curvature and Stopping Sight Distance Recommendation 8: Highway agencies should evaluste the reconstruction of hill crests when (a) the hi| crest hides from view rnajor hazards strch as intersections, sharp horízontal curves, or norrow bridges; (b) the average daily trffic is greater than 1500 vehicles per day; and (c) the design speed of tlß hill crest (bdsed on the minimwn stopping sight distance provided) is more tlan 20 mph below the running speeds2 of vehicles on the crest. Vertical curves a¡e seldom reconstructed to increase stopping sight disønce on hill crests. Neitle.r minimum standards proposed earlier for nationwide use by FIIWA nor special RRR søndards adopted for use in particular states set ûrm requirements for stopping sight distance. This recommendation is similar to Recommendation 7 (horizontal curve improvements) in that it requires designers to evaluate an improvement when there is a reasonable chance that it will be safety cost€ffective. The study revealed that reconstruction of vertical curves may be cost- effective at average daily trafftc volumes gfeater than 1,500 vehicles per day depending on site conditions. Generally, to be safety cost-effective, vertical curve reconstruction must correct a subsøntial sight disønce resEiction that affects drivers' ability to anticipate a hazardous situation-turning vehicles, sharp curves, or other conditions that demand specific driver responses. Unlike horizontal curve reconstruction, hill crest improvements do little to reduce user costs; therefore, reconstruction must be justified primarily on the basis of safety. Whether or not an evaluation of reconstruction is required, designers should routinely examine the following: the nature of potential hazards hidden by a hill crest, the location of the hazard in relation to the portion of the highway where sight distance falls below the AASHTO new construction standard, and other options such as relocating or correcting the hazard or providing warning signs. Bridge Width Recommendqtion 9: Highway agencies should evaluate bridge replacement or widening if the bridge is /ess than 100 ft long and the usable width of the bridge is less than the following values: 2Ttt" 85tÌ, percentile speed of vehicles passing over the crest should be used for this comparison.

FIN DING S AN D RECO MMENDATION S Design Year Volume (ADT) Usable Bridge Width (Í)a 0-750 751-2,000 2,001-4,000 Over 4,000 199 Width of approach lanes Width of approach lanes plus 2 ft Width of approach lanes plus 4 ft Width of approach lanes plus 6 ft olflane widening,is planned as part of ttre RRR project, the usable bridge widttr should be compared with the ptannãa wiat of the approaches afær they are widened. In most states, highway agencies generally do not widen bridges as pârt of a RRR project. special state RRR standards ofren do not add¡esi bridge width requirements, but RRR srandards proposed by FTIWA in 19zg and the 1977 AASHTO Geometric Design Guide for Resurfacing, Restoration and Rehabilitation of Highways and streets (2) both address minimum bridge widths. The safety cost-effectiveness of bridge width improvements depends on the usable width of tfre bridge, the width of approach lanes, trafûc volumes, and the lengttr of the bridge (costs for replacement or widening will vary in proporfion to length). Designers following this recommendation will analyze bridge replacement or widening in most situations in which bridge width improvements might be justified on the basis of safety cost-effõdu"n"r,(Chapter 5). Recommendation 9 falls between the proposed 197g FHWA standards for bridges on RRR projects and AASHTO new construction policy for bridges to remain in place on highway reconstruction projects. undeì the-proposed l97g FHWA standards, bridges as wide as or wider than the approactres coutd remain in place regardless of average daily traffic. otherwise, an evaluation of widening or replacing the existing bridge was required. under the AAsI{Topolic¡ in order for bridges to remain in place when arterial highways are being reconstructed, they should be at least 4 ft wider than the apprãaches and should be considered for ultimate widening or replacement if their usable width is not at least 6 ft greater than the approachei. Like the proposed 197g FHWA standards, Recommendation 9 requires highway ugun.i.. -to evaluate bridge widening when the existing widrh is less than .p.inø values. At low traffic volumes the recommended values are similar to those proposed by the 1978 FFIWA standards, and at high trafûc volumes they areìirnilar to those specified by the AASHTO policy for bridges ro remain in place on arterial highways. when evaluating bridge replacement or widening, highway agencies should estimate the following:

200 DEsIcNINcSAFERRoADS . Cost of replacing the existing bridge with a wider bridge designed to AASHTO standa¡ds for new bridges, o Cost of widening the existing bridge (if widening is practical), and o Number of accidenß that would be eliminated by replacement or widen- ing (if widening is practical) (see Chapter 3). whether or not evaluation of bridge widening is wananted, designers should routinely consider insølling Eansition guardrails at bridge approaches, rehabiliøted or new bridge rails, and warning signs. Sideslopes and Clear Zones Recommendation 10: State highway agencies slnuld develop consistent pro- cedures for evaluating and improving roadside features with the following objectives: . Flatten sideslopes of 3:1 or steeper at locations where run'off-road accidents are likzty to occur (e.g., on the outside of sharp horizontal curves); . Retain current slope widths (without steepening sideslopes) when wid¿n- ing lønes and. slnulders unless warranted by special circumstances; and . Remave, relocate, or shield isolated roadside obstacles' Neither the RRR standards proposed by AASI{TO nor the stândalds pro- posed by the FFIWA address sideslopes or set clear zone width requirements. However, new construction standafds adopted by state highway agencies commonly address both of ttrese roadside characteristics, and a number of states have incorporated numerical sideslope or clear zone width requirements into their special RRR standards approved by the FHWA. Accident data firmly establish that roadside characteristics are important in determining the overall level of safety provided by a highway. Accident rates are lower and accidents are less severe on highways with few obstacles near the roadway (Chapter 3). Studies of vehicle encroachments onto roadsides and estimates of roadside accident severities, coupled with typical unit costs, indicaæ that removing isolated Eees and relocating utility poles can be more safety cost-effective than widening lanes or flattening horizontal curves (Chapær 5). Despite these frndings, the study revealed no basis for nationwide standards addressing either sideslopes or clear zone widtlt. The safety cost-effectiveness of particular roadside improvements appeårs highly dependent on site-specific conditions, including not only conditions affecting cost but also interactions

FINDINãSAI,IDRECOMMENDATIONS 201 between different roadside features that influence the safety benefits of a particular improvement. Instead of proposing nationwide standards, Recommendation l0 requires highway agencies fo deveþ and apply fheir own procedures for identifying and selecting sideslope and clear zone width improvements on RRR projects. In some cases, highway agencies might conclude that numerical sfandards are the best approach for improving safety and promoting efficient design. They could either adopt numerical standards for maximum sideslopes and minimum clea¡ zone width, as some state highway agencies have done already, or ttrey could adopt more flexible design policies that place additional responsibility on the designer. \Vhichever approach is laken it should reflect the following: r Wherever possible, sideslopes should not be steepened when widening lanes and shoulders. When the initial slopes are relatively flat, however, the slope can be steepened to 6:1 with little effect, and steepening to 4:1 may be reasonable. ¡ Permissible sideslopes can be linked to fill heights, as they are by AASHTO new construction standards, recognizing that as fill height increases, sideslope improvements become less costæffective. simila¡ly, side- slope requirements could be more stringent as traffic volumes increase because the cost-effectiveness of flatter sideslopes generally increases as trafflc volumes increase. o The cost-effectiveness of removing or relocating a roadside obstacle depends on several factors, including (a) the distance of ttre obstacle from flte roadway edge (the shorter the distance, the greater the safety cost-effective- ness); lå) the presence of other obstacles nearby (removing an isolated obstacle is more cost-effective than removing an obstacle locaæd between or in front of other obst¿cles); (c) sideslopes on which ttre obstacle are located(removing an obstacle on a gentle, traversable sideslope is generally more safety cost-effective than removing an obstacle on a steep slope); (d) alisn- ment (clear roadsides are more imporønt at curves); (e) ftaffrc volume (the geafer the traffic volume, the greater the safety cost-effectiveness); and (f) speed (the higher the running speeds, the greater the safety cost-effective- ness). . clear zone width policies can be tailored no particular types of obstacles commonly encountered in a stâte to reflect differences in tfre costs of removal, relocation, or shielding. Pavement Edge Drop and Shoulder Type Recomtnendation ll: To reduce pavement edge-drop hazards on highways with narrow unpaved shoulders, highway agencies should either

202 DESIGNINGSAFERROADS c Selectively pave shoulders at points where out-of-lane vehicle encroach- ments and pavement edge-drop problems are lilæly to develop (e.g., at hori' zontal curves); or . Construct a beveled or tapered pavement edge shape at tlrcse points. The FFIWA cunently sets no nationwide standard or specifrcation for type of ShOulder construction, for either ner¡/ constfuctiOn or RRR, nor does it prescribe requirements for edge shape on resurfacing projecs. Pavement edge drops (i.e., vertical drops or ruts) often develop between the pavement surtace and adjacent unpavui shouiders or roacisides. füese drops can prevent drivers whose vehicles cross over the lane edge from successfully retuming to their original lane without encroaching on an opposing lane or losing confol. In addition, pavemenf edge drops are a common source of tort claims against highway agencies (Chapter 6). Research sponsored as part of this study indicaæs that pavement edge drop hazards are gteater than previously believed (Chapær 3). However, no basis exists for estimating how often pavement edge drops contribute to accidents or the cost and safety trade-offs involved in preventing or correcting them. The FTIWA requires that edge-drop problems be conected on completed federal-aid RRR projects at the time of frnal inspection, but depending on the type of shoulder construction used, resurfacing can increase the likelihood that edge drops will develop later and require repeated maint€nance to correct. Shoulders constructed of gravel, turf, or ea¡th afe the most susceptible to edge-drop problems; paved shoulders mitigate the problem by moving the edge drop from the lane edge O the outside shoulder edge. On roads with wide unpaved shoulders, paving 2 to 3 ft of the shoulder adjacent to the through traffic lane and sfiping the edge of the through naffic lane works well at a cost substantially below that of full shoulder paving. In addition, at a given edge drop height, test results show that drivers can recover much more easily when the edge shape is tapered instead of vertical (Chapter 3). RRR projects can reduce the potential for edge drop-related accidents. Recommendation ll gives highway agencies considerable latitude in deciding how this can best be accomplished. Intersections Recommendation 12: State highway agencies slnuld develop consistent pro- cedures and checklists for evaluating intersection improvernents on RRR projects. Neither minimum RRR standards proposed by F[{WA and AASHTO for nationwide use nor most special RRR søndards cunently in use add¡ess

FINDINGS AND RECOMMENDATIONS 203 intersections. Nevertheless, accidents tend to be concentrated at intersec- tions-more than one-half of all accidents in urban a¡eas and about one-third in ru¡al areas occur at intersections 13) (Chapter 3). Reliable information about the cost and safety trade-offs of individual intersection improvemenß is generally unavailable because of the large num- ber of physical and operational features affecting intersection safety and because intersection projects typically address multþle infersection safety problems simultaneously. Nevertheless, many intersection improvements can be made at relatively low cost and are safety cost-effective, particularly as traffic volumes increase. Designers must tâilor intersection improvements to site-specific conditions and rely heavily on professional judgment and experience. Useful procedures for selecting safety improvements at intersections include the following: o Collision diagrams showing vehicle paths, time of occurrence, and weather conditions for individual accidents; r Condition diagrams showing important physical features that affect úaf- fic movement at the intersection; and o Field review of the intersection to detect hazards not apparent from the collision and condition diagrams. Although numerical standa¡ds for RRR proje.cts are inappropriate, state high- way agencies should develop criæria for identifying intersections that warrant ca¡eful evaluation and checklists of improvements to be considered. The criteria might encompass accident frequencies and rates, traffic values, design characæristics, and type of existing trafflc control. Improvements could be organized on the basis of three primary design objectives: (a) reduction of poæntial vehicle conflicts (e.g., raffic signals and turning lanes); (å) improve- ment of driver decision-making (e.g., longer lines of sight and lane markings); and (c) improvement of the braking capability of vehicles in the intersection (e.g., warning signs to reduce approach speed and increased pavement skid resistance). Normal Pavement Crown Recommendation 13: On resurfacing projects, highway agencies should con- struct pavement overlays with normal pavement crowns that match new c anstr uc tio n I t dndards. Both the earlier AASHTO and FFIWA proposals for narionwide RRR standards required that the normal pavement crown---{ross slopes from the centerline on straight sections of two-lane roads that allow rainfall to drain to

2M DESIGNINGSAFERRoADS the roadside--be restored to generally match new construction requirements. Resurfacing projec6 provide highway agencies the oppornrnity Ûo correct deûcient cross slopes at little or no additional cosL Although the safety effects have not been measured, restoring crOSs slOpes to matgh neril construction standards is a good practice that highway agencies should routinely follow when resurfacing. OTHER DESIGN PROCEDURES AND ASSTJMPTIONS Different highway agencies sometimes design RRR projects differently even when thei¡ minimum RRR sønda¡ds are the same and project conditions are practically identical. Such differences are justified in some cases; for example, one state might have highways with better geometric cha¡acteristics and more funds available for RRR work than another. However, differences mây occur simply because highway agencies use different procedues and assumptions to apply RRR stândards; for example, different assumptions about design traffic volume can change the minimum lane width stândard applied. Procedures ate recommended that will encourage a more uniform applica- tion of RRR standa¡ds and a more consistent approach to safety. Greater uniformity will not prevent highway agencies from tailoring RRR designs to meet the unique conditions of ttreir overall systems or individual projects. state highway agencies have considerable latitude; they can develop their own RRR søndards, incorporating the recommendations in the previous section, or seek design exceptions for specific projects. Design Tlaffic Volume Recommendstion 14: The design trffic volume for ø given highway feature should ruttch tlæ average trffic anticipated over the expected performance period of tlwt feature. Although projected traffic volumes for some future year a¡e used to select Ståndards for new construction projects, current RRR praCtice varies among ståtes. Some s[ates use current-year taffic, others use projecæd future-year traffic. The majority of states, however, use current-yea¡ traffic even though the expected performance period of the pavement rehabilitation work is 5 to 15 years and the performance period for geometric improvemenß may exceed 25 years. The study committee concluded ttrat design decisions for particular highway features should be based on conditions that reflect the anticipated service life of the feanre.

FINDINGS AI,I D RECO MMENDATI ONS 205 Speed Recommenduion 15: When evalwting geometric improvements where vehicle speed is a key factor, highway agencies should estimøte running speeds in a manner appropriate for the feature under consideration. Review of highway agencies' practices revealed that most agencies select a single "design" speed for a RRR project based on highway type, terrain, or the posted speed limirs. The FI{WA cunently requires that the design speed on federal-aid RRR projects equal or exceed the posæd or regulaory speed limits. However, the appropriaæ speed measure for design varies depending on the feature under consideration. When selecting minimum lane widths, speed should be handled differently than when evaluating a horizontal curve. Fur- tlermore, unlike new construction, RRR project design can use speed param- eten based on actual running speeds. Accordingl¡ the design practices for key features include recommendations on how speed should be øken into account. When selecting minimum lane and shoulder widths, designers should use a measure of average running speed throughout the projecf For horizontal curves, designers should use the 85th percentile speed ofvehicles approaching the curve, estimated at a point where drivers have not yet reduced speed. For vertical curves at hill crests where stopping sight distance is limited, designers should use the 85th percentile speed measured on the hill crest Pavement resurfacing reduces surface roughness and improves ride quality, which in turn may lead to increases in average speed (Chapter 3). However, these speed increases are usually slight unless the pavement was seriously deteriorated before resurfacing. Therefore, adjusting measured speeds to account for the effect of resurfacing is seldom necessary in RRR project design. Design Values Recommendation 16: Highway agencies should estimate the incremental safety cost-ffictiveness of improvements that exceed the minimum standard. Designers should consider overall highway geometry, design of adjacent segments, and expected trends in traffic growth and truck use when selecting design vølues. Highway designers use minimum RRR søndards, such as minimum lane and shoulder widths, to screen existing highway cha¡acteristics to determine if lane or shoulder widening, or both, is required. When an improvement must

206 DESIGNING SAFER ROADS be made, designers must choose the specific design values to be used, which can range from minimum RRR standards to new construction søndards. Many highway agencies choose design values based on new constnrction standards, reasoning that once an improvement has to be made it is sensible to use new construction søndards in order to reduce the need for future improvements. However, improvements beyond the RRR minimum stândards may not be safety cost-effective and may create inconsistencies between the level of safety provided by the features improved on the RRR project and the featu¡es not improved. For example, upgfading cross-section geometry to new con- -.,-^t---t- ----- l--r ji----- ¿:^- ^^-¡:Â:^-^SlruOlfon SIZln(IafU$ IIIAy rçaU UIIYçl¡i tU t ^lrç{rt t¡9w Uullùuu'vuvrl vulrl¡tllulrù even though commensurate alignment improvements were not made. Designers are encouraged to make a more deliberate selection of design values and explicitly add¡ess issues of increment¿l safety cost-effectiveness and overall highway consistency in geometric design. In evaluating increment¿l cost and safety rade-offs, highway agencies sometimes must make difficult judgments about how much additional cost should be incuned to improve safety. The amount of money one state agency is willing to spend to eliminaæ an accident may be inappropriate in another state because of differences in existing highway conditions and the financial resources available for systemwide improvements. For rural highways, which typically have grcater proportions of more severe accidents, the study com- mittee concluded that when an improvement can eliminate an accident at a cost less than $10,000, it is usually safety cost-effective, but when the cost exceeds $50,000 the improvement is seldom cost-effective. An improvement mây or may not be judged safety cost-effective depending on factors such as systemwide safety needs, available financial resources, and the assumed monetary values assigned to highway fatalities, injuries, and property damage. Design Exceptions Recomrnendation 17: When a highwoy agency requests an exception to a standard, the request should explicitly address the expected safery con- sequ¿nces, along with cost and other impacts. The review of RRR practices conducted at the outset of this study revealed that the cited justiûcations for exceptions to design standards were often imprecise and varied from state to stiate, indicating some confusion over legitimate grounds for design exceptions. To conect this situation, the FTIWA has det¿iled specifrc requirements for design exceptions consistent with this recommendation. The design practices recommended in this study should reduce the fre- quency of design exception requests, but site-specific ci¡cumstances will arise for which design exceptions are justified.

FI N DINGS AI{ D REC O M MENDATI ON S 207 PLANMNG AND PROGRAMMING RRR PROJECTS Highway agencies select RRR projecs primæily on the basis of pavement repair needs and seldom consider safety needs until preliminary design begins. Given curent budget levels and existing highway conditions, pavement repair needs will continue to be the dominant factor in the selection and scheduling of RRR projects. Nevertheless, highway agencies can take safety inø account earlier in the overall RRR process in practical ways that will lead to safer highways. Recomtnendation 18: Highway øgencies shotid screen the existing charac- teristics of highways prograntnædfor RRR projects to identify locations where desirable geometric improvemcnts would require additional rìght-of-way. For such cases, highway agencies should expedite design to d¿termine actual right-of-way requirements and schedule acquisirton of tlu necessary real estate so tlnt it will be available when needed. After cost, the time required for right-of-way acquisition is the major obstacle to RRR geometric improvements such as reconstruction of horizontal curves, which frequently require additional right-of-way. In a design process gearcd to completing resurfacing projects from design through construction in one yeår, additional time is usually not available. Consequentl¡ when right- of-way is needed for a geometric improvement under consideration for a RRR project, highway agencies often must either delay the project or neglect the geometric improvement. The problem of acquiring right-of-way could be avoided if highway agen- cies expedited right-of-way acquisition for RRR projects. Although right-oÊ way acquisition problems are the principal concern, the recommended project screening could be extended to address adverse aesthetic or community impacts that occasionally delay or prevent geometric improvements. Highway agencies could also screen projects for these impacts and work in advance with affected parties to develop an acceptâble balance between environmental and safety concems. Recommendøtion 19: Highway agencies should periodically assess the systemwide potential for improving safety through upgraded design. State highway agencies do relatively little safety planning on a sysremwide basis, particularly with respect to geometric design. Safety-oriented planning is usually confined to the federal hazard elimination progrrim for which state highway agencies identify high-accident locations and use federal aid to undertake mainly spot improvements (Chapær 2). Søæwide highway needs studies, which many states conduct periodically in one form or another, usually mention safety and may include safety in some manner in developing "sufficiency ratings" of existing highways, but these studies tend to focus on

208 DESIGNING SAFER ROADS capacity and pavement preservation issues. Thus, most highway agencies have not deærmined where geometric improvements to existing highways would have the greatest safery payoffs and where such improvements would be the most cost-effective. Although the absence of safery-geometric design relationships and other analysis procedures may have handicapped such efforts in the past the safety cost-effectiveness analyses conducted in this study illusEate that statewide analysis is practical for some key geometric features such as lane and shoulder widths. Assessment of the safety cost-effectiveness of making geometric improve- ñêñto ^n ø ctofarrri¡la l^asic ¡n.'l¡l in¡raaca tha mcitir¡c imnqef nf ÞPÞ wnd¿ nn¡¡rv¡¡Þ v¡¡ @ ù@w w ^uv leù¡ù wvs¡s srv l,vùrB r v safety in several ways. o The results would deæct any unusual opportunities for safety cost- effective geometic improvements (e.g., highways with narow lanes and shoulders and high traffic volumes) that might warrant earlier RRR project programming. o The results could be used o help øilor design practices and standards to the circumstances of a particular søte. o Along with statewide analyses of capacity and preservation needs, the results could be a basis for establishing future state highway programs and funding requirements. . The assessment could be linked ûo reviews of other state safety programs (hazud elimination, special state safety ptograms, rail¡oad grade-crossings, seat belt laws, etc.) to gauge overall progess toward improving highway safety. Systemwide assessments could also serve as the basis for agreements between state highway agencies and FFIWA division offices, such as those discussed in Chapær 2 for California and Ohio, where state spending on special safety improvement programs is linked to requirements for upgraded safety in RRR projects. In California, the state highway agency has earma¡ked funding for a staæwide priority bridge rail upgrading progam in lieu of any requirement in the state's federal-aid RRR standards that bridge rail automat- ically be upgraded in the course of RRR work. Ohio has a similar arrangement with its FFIWA division office for guardrail improvements. SAFETY RESEARCH AND TRAINING The preceding study recommendations are aimed principally at establishing a more safety-conscious design process. They call for better safety engineering in the design of RRR projecs. Better safety engineering requires knowledge

FINDINGS AI,IDRECOMMENDATIONS 209 about the safety effects ofdesign opportunities and choices, designers with the training and methods fo apply this knowledge, and, finall¡ enough resources devoted to design to permit a thorough design process on each project. Despiæ more than one-half century of modern road building, knowledge of the safety consequences of highway design decisions is limiæd. Furthermore, designers often lack the capability or time to apply the existing knowledge. A clear need exists to expand ttre knowledge about the relationships between safety and highway design so that designers will be better able üo identify safety problems and select cost-effective solutions. Equally important, the highway community must ensue that such knowledge is translated into appropriaæ methods, manuals, and design aids and that designers receive the safety engineering education and training necessary to apply such design tools. Taking these steps will mean that additional resou¡ces will be needed for safety research, training, and design. The study committee believes the payoff in long-term highway safety gains will be worth the added cost. The recommendations that follow offer the first step toward meeting long- term research and training needs (Recommendation 20) and also suggest steps that can be øken to improve research and training (Recommendations 2l-23). Recommendation 20: Congress should direct the Secretary of Transportø- tion to establish a special task force to assess highwøy safety engineering needs and to establish research, education, and funding priorities. The Intersfate highway system is nearly complete, and the Uniæd Søtes has shifted emphasis from building new highways to repairing and rebuilding existing highways. As the country proceeds wittr this enornous tåsk and invests billions of dollars, improvements can be made to the highway system that will reduce t¡affrc fatalities and injuries for decades to come. The recommended task force should outline research and education agendas to increase safety engineering knowledge. In addition, the task force should consider the question of how well design resources, in terms of staff and funding, are matched to the øsk of incorporating more extensive safety engineering in highway design and how to promote the use of rigorous statistical controls in safety research. Recorunendation 2l: Tlrc Federal Highway Adminisrrqtion shauld deverop, distribute, and periodically update a compendium that reports tlæ most probable safety fficts of improvemcnts to k¿y highway design features. Critical reviews of existing safety research are generally unavailable, and organizations such as FHWA and AASHTO, which have a major interest in highway design, have not reported most probable relationships that could be used by designers and others interested in the trade-offs between safety and cost in highway design.

2r0 DESIGNING SAFER ROADS The primary objective of the recommended compendium is to provide designers with the best available safety data and simple application meth- odologies. Its contents might include the following: o Background information on lhe use of accident data and accident models to estimate the safety effects of highway design improvements; ¡ Easy-to-apply procedures for estimating the safety effects of improve- ments to specific design features, including a description of the information needed to apply the procedures; and o How to use thc e"stimates of safery effects. together with information on costs and other impacts, to assess the cost-effectiveness of design improve- ments. The recommended compendium would build on eadier research reviews and the findings of this study and would provide designers a common starting point for their work, which could then be adjusted to take into account local accident histories or other highway circumstances. In addition, a safety com- pendium could help redi¡ect cwrent research efforts and focus attention on the value of research in at least three ways: o Designers will be able o apply research results sooner because (c) results would be keyed to most probable relationships contained in the compendium and (b) the compendium would be periodically updaæd to incorporate new research findings. ¡ Major sponsom of safety research (FHWA, AASI{TO, and individual state highway agencies) will identify the principal gaps in current knowledge and determine where research might be able to fill these gaps. . Researchers will have a frame of reference for conducting their work; study objectives could be aimed, for example, at validating a safety relation- ship, testing an assumption in a recommended procedure, or ûlling a key gap in current knowledge. Recommendation 22: The Federal Highway Administration and the National Cooperative Highwøy Research Program (NCHRP) should increøse resesrch on the relalionships between sSery and highway design. Except for a modest FHWA research program and occasional NCHRP- sponsored resea¡ch studies, few opportunities exist for coordinated, purpose- ful safety research aimed at highway design. To a large extent, the highway community has relied on uncoordinated research without rigorous statistical controls to expand knowledge about the safety effects of road design. Cer- tainly some state highway agencies have evaluated the safety impacts of various types of highway improvements and reached conclusions that have

FINDINGSAI'IDRECOMMENDATIONS 2TT been usefully applied in subsequent design work. However, as a rule, such efforts lack the statistical controls necessary to develop relationships that can be reliably transferred to other locations or generalized for nationwide application. signifrcant progress will come only ttrrough research programs that care- fully define objectives, select an appropriate experimental setting, collect the necessary data, and apply appropriate analysis techniques. safety research is especially complex because of the number of factors other than highway design responsible for accidents. Research projects do not always fully achieve their objectives, but when they do the benefits can be subsøntial. Although a new study of research priorities is recommended, existing FtrwA and NCHRP programs can begin to increase knowledge of the rela- tionships be¡reen safety and highway design. Especially important topics that merit further research include o Safety effect of physical and operational featu¡es of intersections, o Safety effecs of lane and shoulder conditions on urban highways and streets, o Safety effects of different sideslope and other roadside conditions, o safety effects of low-cost safety treatments such as warning devices at hazardous locations or shoulder widening af horizontal curves, and¡ combined safety effects of changes in horizontal and vertical alignment. Recommendation 23: The Federal Highway Administration, the American Association of state Highway and Transportation officials, state and local highway agencies, snd other organizations of public works professionars should support continuing training activities to keep design engineers abreast of sSety-conscious design. Regardless of how RRR projects are selected, which sønda¡ds are applied, or how ttre design process is organized, individual design engineers ultimately play the major role in determining ttre degree ro which RRR projects will enhance safety. Although engineers can rely on standards and guidelines to make many design decisions, some decisions must be based on site-specifrc ci¡cumstances and judgment. The quality of these decisions depends on how well the engineer is prepared for designing RRR projecc. The special task force recommended @ecommendation 20) should address the long-term needs for greater safety skills and awareness among designers through university curricula and in-service raining. In the meantime, FFrvy'A, AASHTO, and individual highway agencies can use a combination of formal and informal training techniques to increase the skills of design engineers. Formal naining includes short courses and seminars similar to those con- ducted through FFrwA s National Highway Institute. Informal training could

2I2 DESIGNINGSAFERROADS include reviews of completed RRR projects or conferences that provide forums for designers to sha¡e experiences. Both types of activities can be effective, and can offer highway agencies a range of options to frt time and budget constraints. REFERENCES L Highway Statßtics 1985. FHWA, U.S. Department of Transportation' 1986. 1 lloan¿¡ri¡ I1a<íoa Cuì¡lo fnr Pcsurfa¡Íno Restorntion anÅ Relnbilítatíon of Hiph- ways and Streets. American Association of Staæ Highway and Transpo.rtation Officials, Washington, D.C., 1977. 3. Accident Facts-1985 Edition. National Safety Cowcil, Washington' D.C,

Appendix A Summary Comparison of Nonfreeway Geomeffic Design Standards and Guidelines Table A-1 contains a description of three sets of geometric design søndards and guidelines for nonfreeway highways. The AASHTO RRR guidelines are from the American Association of State Highway and Transportation Offr- cials' Geometric Design Guide for Resurfacing, Restoration, and Rehabilita- tion (RRR) of Híghways and Streets (1). This guide, commonly referred to as the "purple book," conûains minimum design values for lane and shoulder widths, cross slopes, superelevation, and bridge widths, as well as advisory information on grades, curvature, sight distance, and clear zones. Overall, the AASHTO RRR guidelines are considerably less stringent than new con- struction standards. In August 1978, after opposition to the AASHTO RRR guidelines surfaced, the Federal Highway Administrarion (FIIWA) proposed RRR standards (2). These proposed sønda¡ds are generally more stringent ttran ttre AASHTO RRR guidelines, but still less stringent fhan new construction standards. In June 1982, the FHWA issued regulations permitting stâtes to develop thei¡ own RRR standards, subject to FF{WA approval. However, some states continue to use new construction standards for RRR projects, with design exceptions on a case-by-case basis. AASHTO's Policy on Geometric Design of Highways and Streets, 1984 (3) has been approved by the FI-IWA for the design of new and major reconstruction federat-aid projects. The policy also is applicable to RRR work in states fhat use new constructon standards for RRR projects. 2r3

TABLE A-1 Summary Comparison of Nonfreeway Geometric Design Standards and Guidelines 1978 FHWA RRR AASHTO RRR Guidelines Proposed Standards AASHTO, Policy for New Construction Tiaffic data ADT DHV percent trucks, and ADT DHY percent trucks, accident ADI DHY directional distribution,(current) turning movements at locations, and descriptions, traffic composition (percent trucks) signalized intersections must be including collision diagrams, collected and analyzed. should be collected and analyzed. At signalized intersections, turning movements and pedestrian volumes should be collected and analyzed. Future traffic 5 to lO-year traffic forecasts lor Not specified. Design year 20 years ahead "widelyprojections major rehabilitation. used"; <lesign year 5 to 10 years may be l3f#iïå,:"i,ïf ffi lo""'io" "" o Design speed No minimum, but design should Average free-flow running speed plus Rural accommodate cur¡ent running 10 percent. Arterials: 50-70 mph depending on speed. terrain. Collectors: 20-60 mph depending on terrain, ADI DHV Local: 20-50 mph depending on terrain, ADTD'HV Urban Arterials: "generally" 40-60 mph; "occasionally" 30 mph under restrictr:d conditions. Collectoni: 30-60 mph depending on terrain, trafrc, intersection spacing. Local: 20-30 mph depending on terain, traftc, development.

TABLE A-1 continued 1978 FHWA RRR AASHTO RRR Guidelines Proposed Standards AASHTO Policy forNew Construction Superelevations Use new construction rates unless Rates for new construction apply. Rarøl function ofDS, terrain, climate; constraints do not permit. Use Design speed may be reduced if maximum 0.10;0.08 where snow and l0 degrees BBI reading to necessary, but speciai signing ice are factors. determine maximum safe required. Such reductions Urban: 0.04 to 0.06 maximum on higher speed. considered infrequent for rural speed streets with few restrictions; areas; 10 degrees BBI reading used generally no superelevation on low- to establish safe speed. speed curbed streets. Rural two-lane Minimum lane and shoulder widths ADT : 350 l0 f1(2 fÐ 10 lt (2 fÐ Local: 10 ft (2 ft) DS : 40 mph 9 ft (2 lÐ il Collector: l0 ft (2 ft,4 ft "minor road" if barrier) Arterial: NA (arterial recommended DS > 50 mph) ADT : 500 10 ft (2 ft) 11 ft (2 lÐ Local: 1l ft (4 ft) DS : 40 Collector: 11 ft (4 ft, 6 ft) truck ) l0 Arte¡ial: NA percent ADT : 500 10 ft (2 ft) l0 fr (2 fÐ Local: 1l ft (4 ft) DS : 40 Collector: 11 ft (4 ft, 6 ft) truck ( I0 Arterial: NA percent ADT>400 l0ft(2ft) Dft( ft) Local:12ft(8ft) DHV > 400 Collector: 12 ft (8 ft) truck ) l0 Arteriai: 12 ft (10 fÐpercent (for all design speeds)

TABLE A-1 continued 1978 FHWA RRR AASHTO RRR Guidelines Proposed Standards AASHTO Policy forNew Constn¡ction Rural multilane Minimum lane and shoulder widths DS < 50 and trueks l0 ft (2 ft) < l0percent DS > 50 or üucks 10 ft (2 ft) > 10 percent Urban arterials Minimum lane widths Throughlanes 10 ft Parkinglanes 7 fl Turriing lanes DS <40 mph 9 ft DS) 40 mph 10 ft r0 ft (2 ft) 11 ft (4 ft) 10 ft 8ft 10 fr l0 lr 12 ft (10 ft-shoulder "preferable j' 8-ft rninimum) 12 ft (10 ft.shoulder '¿preferatrle," 8-ft minimum) l0 ft, hiehly restricted conditions, low truck traffic 1 1 ft "adequate" 12 ft "desirable," ¿'generally used" on highersp,eed (> 40 mph), free-flowing principal arterials. 10- 12 ft (8 ft acceptable ifnever used as a traffi.c laue). 10-ft left-tu¡n laneq 1 I -ft coritinuous two-way, left-turn lanç. l0-ft left-tu,rn lanes, I l -ft continuous two-way, le ft-tuq lane,

TABLE A-1 continued AASHTO RRR Guidelines 1978 FHWA RRR Proposed Standards AASHTO Policy for New Construction Horizontal curvature, maximum grade, and minimum stopping sight distance Bridges Minimum width (existing bridges) Improvement should be considered at high-accident iocations. Considerable engineering judgment must be exercised. Use signing when sight distance less than AASHTO standard for new construction. No absolute minimum speciñed. One-way operation permitted. Minimum guideline of 18 lt provided lor low-volume (ADT < 250) minor roads with few trucks. rüy'hen curve DS is no more than 15 mph less than roadway DS, signing required. When curve DS exceeds roadway DS by more than 15 mph, corrective work should be undertaken unless "impractical." 18 ft for bridge requiring minor rehabilitation. 20 ft lor iow-voiume (ADT < 250) - bridges requiring major rehabilitation. Minimum radius a function of design speed, maximum superelevation, and side friction. Maximum grade a function of terrain, design speed, percent trucks, tramc volume, lunctional class: Rural (o/o) Urban (o/o) Local 5-16 4-15 Collecror 4-12 5-14 Arterial 3-1 5- 1l Minimum stopping sight distance a function ofdesign speed and grade. On ievel wet pavement: Minimum D,t Assumed stopping mph speed, mph distance, ft30 28-30 200-200 50 44-50 400-475 70 58-70 625-850 Collectors DHV WidIh <200 22fr 200-400 24ft > 400 28ft Rural arterials Tiavel lanes plus 2 ft each side Urban arterials Curb-to-curb width l..){

TABLE A-1 continued AASHTO RRR Guidelines 1978 FHWA RRR Proposed Standards AASHTO Policy for New Co:nstruction Clear zone Rural Urban Salety appurtenances From edge ofpavement, 30 ft desirable, but there must be many exceptions. Emphasis on removing frxed objects identiñed as hazardous by accident analyses. Minimum setback should be behind the pared shoulder or curbing. Provide traffic barriers for clear zone hazards that cannot be eliminated. Review accident data to define dangerous obstructions. Considerable judgment must be used because of existing topographic and right-oÊway limitations. Minimum setback should be behind the pared shoulder or 2 ft behind the curb. Desirable to upgrade to cuffent criteria as part ofRRR project. Arteriat; and high-speed (> 50 mph DS) collectors: follow 1977 AASHTO Guideþr Selecting, Locating, and Desigining Tiffic Barriers (4). Allfved objects should be outside clear zone.a Low-speed collectors, local: minimum l0 ft; exception may be made where guardrail provided. Consider safety, environmental, aesthetic concerns in deciding whether to rem'ove all trees, poles, and so forth. With curbs: 1.5 ft beyond face of curb; without curbs: same as rural. Essential to renrove only "very vulnerable" frxed objecls on urban collectors. Follow 1977 AASHTO Guideþr Select,íng, Locat ing, and Des igning Tiafrc Barriers (4) Norr: ADT - average daily tramc; DHV - design hour volume; DS - design speed; BBI '¡Clear zone as a function of DS, slope, and curvature. - ball bank indicator; and NA - not applicable.

2r9 REF'ERENCES t. Geatnetríc Desígn Guídcs fot Resurfacing, Restoratíon, a¡ú Rehabitítation (RRR) of Highways and Streets. American Assoçiation of State Highway and Transporø- tion Ofñcials, Washington, D.C,, L977, 2. "Design Standards for Highways, Notice of Proposed Rulemaking." Federal Reg- is¡ø; tfol. 43, No. 164, Aug. 23, 1978. 3. A Polìey on Geometric Design of Highways ønd strc¿ts. American A,ssociation of Søte Highway and Transpøtation Of[cials,'Washington, D.C., 1984. 4, Guide for selecting, Læøting and Designìng TrSfic Barriers. American Associa- tion, of Søte Highway and Transportation Ofûcials, WashÍngtor¡ D,C., 1971..

Appendix B Case Study State and Local RRR Programs The øbles in this appendix contain a description of resurfacing, restoration, and rehabilitation (RRR) ûnance and expenditures; programming methods; and design standards and practices in the ståte highway agencies and local governments chosen as case studies. The case studies are discussed in det¿il in Chapter 2. Most of the information presented in the tables was assembled in 1984 and was the most recent available at the time. However, Tables B-9 and B-10 give the special RRR design standards in effect in October 1986 in the case study states. Follow-up interviews were conducted witi highway agency officials in each of the 15 case study stâtes in the fall of 1986. Officials were asked to list the changes that had occurred in their RRR programs since the original interviews. The follow-up interviews indicated no major changes had occuned in the states' RRR activities in the 2-year period and that trends observed in 1984 have continued. Several stâtes reported that emphasis on safety in RRR projects has continued to grow. Some changes have been made in RRR standards, and although most have been simply fine-tuned, a few søtes (Ohio and Texas) have made substantial revisions, and one state (Ari- zona) is in the process of developing special RRR standards for the û¡st time. A number of state officials reported that they have received increased guid- ance from the FFIWA on RRR design matters and that the FFIWA has given more attenton to reviewing exception requests during the past 2 years. In finance as in design, the stâtes experienced few major changes, although some reported an increase in the amount of money available for fully state-funded RRR projects over the period. 220

TABLE B-1 Characteristjcs of the Case Study States Federal-Aid System Program Administration Special CertificationRRR AcceptanceStandards forNon-Interstate Approved Primary DesignState 1984 Mileage (thousands) Percent State Àdmin- istered Fiscal'{ear 1984 Federal-Aid Apportionment (rnillions) Percent Urban Arizoda California Florida Illinois Michigan Mississippi Missourì NewHampshire NewJersey NçwYork Ohio South Dakota lexas Virginia Washington lbtal U.S. total or average 9 4t 20 31 31 20 29 J 9 25 27 18 60 20 l7 360 838 22 46 40 28 2T 11 11 30 69 45 36 3 16 r8 27 22 62 )t 55 49 30 5t 94 88 26 60 67 50 93 93 42 63 239 9Q7 419 s44 336 r52 277 55 288 617 396 78 796 304 283 5,69 1 n,740 No No No Yes No No Yes No Yes Yes No No No ks lês No Yes lôs Yes, Yss Yes No No Yes No No Yes Yes No Yes 26, yes 15, yes Sounc¡s: Highwaysrarisüesl9å4rFHWA,'IablesHM-14HMls;andFB22l;interviewswithhighwayagenc.yoñcialsincâsestìrdystate$FH\lå\recordssfRRRsta¡dards aíd ceitiûcatiqu a.cceplar-úes.

TABLE B-2 Percent of Federal-Aid Highway Miles Under State Administration, by Federal-Aid Systern, Case Study States Federal-Aid Highway System Interstate and Primary Secondary ïbtal Federal- Urban Aid System Arizona California Florida Illinois t f:^L:-^-Mtu¡uË,alt Mississippi Missouri New Hampshire New Jersey New York Ohio South Dakota Tèxas Virginia Washington r00 98 94 99 100 r00 98 97 95 r00 99 100 100 r00 38 8 9 t4 1aLL 30 100 100 5 66 73 t2 99 99 l5 6 5 26 26 aJ L4 29 38 10 l6 2Q 4 4L 53 J 62 37 55 49 JU 5t 94 88 26 60 67 50 93 93 42 United States 63234998 Souncn: Highway Statistics /984 FHWA, Tiable HM-14.

TABLE B-3 Distribution of Non-Interstate Rderal-Aid Project Expenditures by Project Category in Case Study States Percent olTotal Non-Interstate Federal-Aid Expenditures State Reconstruction Resurfacing and Bridge Safety Im- Intersection and New construction Minor widening work provements Improvements other Commentso Arizona California Flo¡ida 58.0 37.'7 71.8 38.0 32.2 13.2 29.0 33.9 I 1.3 13.0 8.9 4.0 7.0l 1.8 14.9 4.2 0.6 FY 1984 programmed; federal-aid primary and federal-aid secondary only. FY 1985 programmed; excludes local federal-aid secondary and federal- aid urban; based on projects more than $250000. FY 1985 programmed; safety and "other" in construction or resurfacing categories; includes local federal-aid secondary and federal- aid urban. FY 1984 programmed; includes local fedêral-aid secondary and federal- aid secondary. FY 1984 programmed; does not include local federal-aid secondary. FY 1984 programmed; federal-aid primary and federal-aid secondary only; includes local federal-aid secondary. FY 1984 programmed; derived from data on state-defined systems (state and federal dollars may be mixed). FY i984 programmed; includes local federal-aid urban and federal-aid secondary. lllinois Michigan Mississippi Missouri New Hampshire 46.2 40.0 73.2 t3.2 28.1 1 1.6 8.0 t6.2 2.5 14.9 66.9 6.2 5.2 3.2 0.1 4.2 0.6 51.1 2.5 6.6 2.2

1..) N) .À TABLE B-3 continued Pe¡cent of Total Non-Interstate Federal-Aid Expenditures State Reconstruction and New Construction Resurlacing and Minor Widening Bridge Safety Im- Work provements Intersection Improvements Other Commentso New Jersey 43.9 New York 22.9 Ohio 20.5 South Dakota 32.9 Texas 6'7.4 Virginia 59.0 Washington 29.0 12.2 15.5 42.4 50.5 28.8 1.0 42.6 18.0 2.3 _ 1.3 0.1 13.5 0.7 t.7 0.8 2.2 3.7 27 .2 8.1 13.5 0.9 14.4 9.8 50.8 2.6 26.0 20.6 6.1 5.9 0.5 0.4 FY 1984 programmed; includes local federall-aid secondary and federal- aid urban. FY 1985 programmed; includes local federal-aid secondary and federal- aid urban. FY 1984 contract awards. FY 1983 contract lettings; excludes local lederal-aid urban. FY 1983 contract lettings; bridge and safety projects in other categories; includes local projects. FY 1984 programmed, federal-aid primary and federal-aid urban only; safety projects in other categories. FY 1985 programmed; excludes local federal-aid secondary and federal- aid urban. NorEs: Amounts include state match and construction phases only. Dashes in cell indicat€ the project category could not be separated fi:om other types in the data supplied oy LIIË 5LA!E. I ll9 lrçlrçulaée urùL its enti¡e cosr is in tn" .".ur¡uJingãnã -iñå, *i¿"niriá "ut.-go.v, ""ån thougtr the'ancillary improvementJmay have_accounted lor a substantial portion ofthat cost' ,,Expendilures are aenne¿ as tne iuîl coits ofall proje"t, ã"truùy "ãrn-enced ii the year, or costs ofprojects programmed (i.e., planned) for the year. Programmed âmounts are àstimated expenditures and may not have been incurred in the given ñscal year.

TABLE B-4 Distribution of State DOT Expenditures for Fully State-Funded Projects, Selected Case Study States State Reconstruction and Resurfacing and Seal Coats and New Const¡uction Minor Widening Thin Overlays(7r) (o/o) (o/o) Bridge Safety Work Improvements e/o) (o/o) Intersection Improvements Other(o/o) (Vo) Notes California 7.2 39.2 0 27.6 2.9 12.8 1.2 Norrs: Expressed as percent oftotal expenditures for state-funded projects. Amounts in Total Expenditures column include state-funded construction and (unless otherwise noted) costs ofseal coats and thin overlays âs maintenance activities. Dashes in cells mean that no information is available for that category; expenditures may be included in Other câtegory. aln lllinois, seal coats and thin overlays are a maintenance activity; dollar amount is unavailable and not included in total expenditures. òSeal coats and thin overlays included in Other category. 'Seal coats and thin overlays included in Resurfacing and Minor Widening category. 0.4 53.2 FY 1985 pro- grammed; includes only projects more than $250000 ex- cept all seal coats included l0.l FY 1984 programmed 39.9 FY 1984 programmed 7.3 FY 1984 programmed 15.9 FY 1984 programmed 9.9 FY 1984 programmed; safety included in Other category39.3 FY 1983 actual contract awards 0 FY 1983 actual contract lettings 6.0 FY 1983 actual contract awards 4.6 FY 1985 programmed Iilinois Michigan Mississippi New Jersey New York Ohio South Dakota Tèxas Washington 0 0.1 29.3 I J.ô 0 0 51.8 65.s 75.8 60. l 12.9 7.'7 45.9 56.8 21 .4 41 .0 12.4 t9.3 30.8 _b 57 .5 _c 8.3 2.0 7.40- 0 0.4 16.4 3.1 2.4 1.0 '7.0

TABLE B-5 RRR Project Programming Procedures in the Case Study States State Programming Procedures Arizona Central office selects projects using its pavement management system, an automated p¡ocedure for determining the least-cost schedule ol pavement repairs that will maintain a specifred systemwide minimum performance standard. Inputs to the process include annual measurements ofpavement deflection, cracking, and roughness. RRR project selection is guided by a pavement management system (PMS), a biennial pavement survey, and priority ranking. Districts recommend major projects (more than $250,000) to centrai office lollowing PMS priorities; central office assembles a 3-year RRR program based on projects'statewide priority rankings. Ride quality is an important consideration in assigning priorities. Almost all major RRR is federal-aid. Districts receive funding allocations for minor const¡uction RRR and maintenance overlays and select their own projects in these programs, subject to central office review. Central ofrce allocates resurlacing funds by formula to each district. Districts receive separate allocations for federal-aid and state- funded resurfacing. Districts select projects guided by PMS ratings, but are allowed considerable leeway to exercisejudgment. Districts choose between federal aid and state funding for each project, considering standards requirements and urgency. District offices annually propose RRR projects for inclusion in the 5- year improvement program based primarily on prolessional judgment, local knowledge, and a visual pavement condition survey. The central office screens these proposals with respect to pavement condition estimates, traffic levels, and geometric condition, and designates projects that balance statewide needs. District ofûces may make project substitutions provided overall costs are about the same. In addition, central oflñce undertakes resurlacing projects as part of a mostly state-funded winter damage program. Annually, the central office allocates funds lor this program to the districts, where project selections are made. Central office selects RRR projects mainly judgmentally, based on district recommendations and a biennial subjective sufficiency rating ofstate roads. Formal geographic or program allocations of funds are not used. Nearly all resurfacing is federal-aid. PMS is under development. Districts determine resurfacing project needs judgmentally, then choose which projects to recommend to the central office as federal-aid RRR construction and which lo conduct as stâte- funded resurlacing using funding allocations for maintenance overlays. Resurlacing-only projects are usually state-funded and federal-aid RRR is usually on roads that need service upgrading in addition to resurfacing. Central office chooses federal-aid RRR projects from among district recommendations on the basis of a state\ryide priority ranking. California Florida Illinois Michigan Mississippi

TABLE B-5 continued State Programming Procedures Missouri Central office assigns a mileage allocation to each district. Districts select their "worst miles" of pavement up to their allocations. A central office team inspects and visually rates these worst miles. Districts then program RRR projects within thei¡ total New Hampshire New Jersey New Yo¡k Ohio construction program funding allocations, following centrai office rules as to the rating scores that warrant a RRR project. Nearly all construction resurfacing is federal-aid. Maintenance resurfacing(mainly on low-volume roads) is selected by central office maintenance division (lor larger projects) or by districts (for small projects) within their maintenance resurfacing funding ailocations. Federal-aid RRR projects are selected by the central officè from a list ofcandidates compiled at the beginning ofthe federal-aid RRR program by the state and the FHWA. Most resurfacing currently is state lunded and selectedjudgmentally by the central òñce maintenance division on the basis of dist¡ict ¡ecommendations. Central office design division selects and designs more extensive RRR projects, which are usually funded using federal-aid. projects typically are identified because ofneeded pavement repairs. The central office maintenance division handles state-fundèd resurfacing work selected on the basis of systemwide roughness measurements. Central office undertakes RRR-related work in threeprogramming categories: regular federal-aid capital, state reconditioning and - preservation, and transportation improvement materials(maintenance resurfacing). For each, the central office allocates available funds to regions and regional offices then selects projects that match available lunds. Project selections are motivated mostly by pavement repair needs that are assessed by professionaljudgment, local knowledge, and a statewide visual pavement survey. Performs RRR-related work in three programs administered by different central office bureaus: location and design, traffic, ánd maintenance. In each case, at least some federal aid is used. The central omce bureaus establish program guidelines and fund allocations, district ofûces select individual projects, and flnally, the central office bureaus review and schedule all projects. Except ior special.traffic or salety projects, distdcts select projects primarily on the basis of pavement repair needs assessed by profeisionaljudgment and local knowledge.

228 TABLE B-5 continued State Programming Procedures South Dakota Tèxas Virginia Washington Central office selects all RRR projects through an annual process that updates the state's 5-year construction program' Project priorities are initialty established on the basis ofpavement design and condition (from visual surveys, roughness measurements, and deflection measurements), drainage adequacy, and traffic characteristics and adjusted after freld reviews.Projects are scheduled in order ofpriority to match allocations ofavailable funds among different functional classes and between resurlacing and other types ofconst¡uction work. Central office undertakes RRR projects principally in its rehabilitation program category. Normally, the highway commission sets the amount of lunding and allocates lunds to districts. District offices then select projects that in total match their fund allocations. These project selections are based mostly on professional judgment and knowledge of local conditions and typically address pavement repair needs, as well as geometric deñciencies. Central office performs RRR improvements through its construction and maintenance programs. On the state primary system, the central office selects construction projects, with input from district ofices, that in total match funds allocated to each district. On the state secondary system, county governments select projects, in consultation with district offices, that match construction funds allocated to each district. Under the maintenance program, the central office allocates resurlacing funds to each district and district offices then selects projects that match their funding. Central office allocates resurfacing [unds by formula to districts and also assigns resurlacing priorities based on PMS ratings. Districts assemble resurfacing programs and must address all high priority roads, but can choose between federal-aid-eligible major RRR or state-funded light overlays. About one-hallolmajor RRR projects are also state-funded, but in general federal-aid and state RRR projects are not treated separately during programming.

TABLE 8-6 Comparison of Federal and State Funding for Resurfacing, Case Study States Federal-Aid Projects' Share of Total State Highway Agency Non-Interstate Resurfacing Federal-Aid Projects as Pe¡cent of Expenditureso Federal-Aid Projects as Percent of Miles Resurfaced Percent ofState- Maintained Non- Interstate Federal- Aid Mileage Resurfaced in Fiscal Yea¡ Federal-Aid Projects Percent ofTotal State- Maintained Non- Interstate Mileage Resurfaced in Fiscal Year, State and Federal FundingState Fiscal Year Arizona California¿' Florida Illinois Michigan Mississippi Missourió.' New Hampshireô. New Jerseyå New York Ohio South Dakotaá lbxasá,. Virgini¿2" Washington l 984 l 985 l 985 l 984 1 984 I 984 r 985 1 985 1 984 1984 1 983 l 983 l 983 1 985 I 984 68 68 26 t9 98 19 35 Jð 78 43 7l 82 40 8 60 31 20 23 t2 99 6 7 2 NA NA o¿ NA 8 2 3'7 1.9 2.5 0.9d 0.7 J.J 1.0 0.9 0.2 0.6d NAd 3.gd NAd 1.2d 0.5 3.6 6 12 6 NA J 16 t2 7 NA NA NA NA NA 8 r0 Nor¡: NA indicates percentage is not available. "State match included in federal-aid expenditures.óMore than 50 percent ofstate resurlacing expenditures are for thin overlays and seal coats costing less than $40000/mi. cstate system includes substantial nonfederal-aid mileage. dlncludes mileage and resurfacing olsone locally administered federal-aid secondary or urban roads.

TABLE Bt Funding for Special Safety Improvement Projects, Case Study States Ariz. Califl Fla. Ill. Mich. Miss. Mo, N.H. N.J. N.Y Ohio S.Dak. Tex. Vy'ash. Federal-aid pn mary, secondary, or urban funds used for safety projçcts? No 100 Bercent state construction funds used for safetyprojects? No Earmarked mâintenance budget item lor safety 'improvements?å No Percent of federal hazard elimination funds routinely m¿de available to local governmen!s Unobligated hazard elimination balance as of June 3Q 1987 as percent of annual apportionment Nore: Special safety improvement projecl i5 priûarily motivated by a safety concern.ai'id inænded to coÍreet a speciåc hazardous condition. ¿With rare exceptions. 'D¿ta not avåilable. Yes No 60 l15 No )lesNo No No No No YesNoYes ks No No¡ No NoNoYes No 50 Yes No Yes No 5025 Yes Yes No No 180 63 1,46 53 0 163 33 t29 't| r04 168 206 4t153

TABLE B-8 Desþn Practices for Federal-Aid RRR projecrs, case study srates Anz. calif. Fla. IIl. Mich. Miss. Mo- N,H. N.J. N.]I ohio s.Dak. Tex. va. Wash. Use Special RRR standards? No Yes Yes Yes Yês Yès ì'lo No yes No No \ès y"s l-f. Mostdesignatdistrictoffices? l¡{o Yes Yes Yes No No Yes No No yes ye$ No yes yesCertifrcation acceptance for design and construotton on primary highways? Safe-ty and geometric needs identified and reviewed Predesign report submitted to ÇenÎral oftce or the FHWA Predesign teld review with theFHWA Review ofaceident data Review of state traftc or safety staff No Yes )Õs Yes YesYesNo o o x o o No + o + o x NoYes + x + Õ No +x* +xx +++ ooo ooo NoYesNoNo o + N9No o+ x*++ + oooo o + + + + + + + o ¡( + + o o o+ ox o+ xo oo Selected use of formal cost- effectivenessanalysis o o o o x Nore: * : always or usually, x = occasionally, and o : rarcly or nevel

TABLE B-9 Minimum Geometric Requirements for Rural Highways in Case Study States V/ith Special RRR Standards Condition California Ftorida Illinois Michigan Highway Design iSpèed(mph) i.qO Notspecifled ARS:4Omph whereposted 35mphposted ì sPeed : 40 mPh i I : 45 Not speciñed ARS : 45 mph Where posted 40 mph posted sPeed : 45 mPh 50 Not specified ARS : 50 mph Whe¡e posted 45 mph posted speed : 50 mPh Lane width (It)l0 -c ARS : 40 mph DS : 40 and TwoJane highways ARS - a0, ADT = 1,000 where4DT(750 ADT : 1,000 ADT : 400 r: t0% (ailDs) I I -c ARS > 40 mph DS : 50 and All highwaysADT : 3,000 ADT: 1,000 ADT : 3,000 and e]iïinfi road'¡/ay Shoulder lltídth (f0 , -c Notpermitted DS : 50and NotpermittedADT : 250 4 -c Not permitted ADT : 3,000 All highways except on two-lane highways where ADT> IO,QOO

Mississippi New Jersey¿ South Dakotaá 'Ièxas Washington ARS: 36 mph ARS: 4I mph Resurfacing and Restoration Posted speed = 40 mph Rehahilitation Posted speed : 35 mph Resurfacing and Restoration Posted speed : 45 mph Rehabilitqtion Posted speed : 40 mph Resurfacing and Restoraliotl Posted speed : 50 mph Rehabilitation Posted speed : 45 mph Not speciñed Not specified Not specified Undivided multilane highways in rolling terrain Twolane highways except in flat terrain with ADT = 1,500 Undivided multilane highways in rolling terrain Two-lane highways except in flat terrain with ADT = 1,500 All highways Not speciñed (should be logical with respect to terrain and type of highway) Not speciñed (should be logical with respect to terrain and type of highway) Not specifred (should be logical with respect to terrain and type of highway) ARS : 45 mph Not permitted Not permitted Not permitted Two-lane CoLlectors Minor reconstruction ADT < 250 Resurlacing: ADT> 5OO Not permitted All multilane highways except where ADT> 4,000 and T : l0o/o All twoJane highways except where ADT > 1,000 and T : l09o Two-lane highways where ADT < 750 All highwaysAll multilane highways Two-lane highways where ADT : 750 All highways for resurfacing and restoration projects Multilane Highways on left side DS:50 ADT: 4000 DS:5Q ADT> 4000, and T < l0% All highways On left side of multilane, divided highways On left side of multilane divided highways Principal ArteriaLs Resurfacing: ADT : 1,500 Minor Arterials Minor reconstruction: ADT: 500 All resurlacing projects Collectors All resurfacing and minor reconstruc- tion projects Principal Arterials Minor reconstruction: ADT : 1,000 All resurfacing; Minor ArteriaLs and Collectors All resurfacing and minor reconstruclion Two-lane highways with ADT < 750 All highways On left side of multilane divided highways On twolane highways where ADT < 1,000 or DS( 50 mph and ADT < 2,000 All two-lane highways Multilane highways where ADT < 4000

TABLEB-9 continued Condition California Michigan Shoulder Iridth (ft)6- All highways Althighways. Allhighwaysexcept DS > 50 on two-lane ADT : 5¡000 highwap where ADT> t0.000 SpoÍ Horizontal CgtveDesign òpeeu tmpn.l35 Notspecifred ARS : 50 mphd DS : 45 mph DS : 35-50 mph ând ûot a high- accident locãtion d 40 45 Notspecifred ARS : 55mph¿ DS : 50rnpb Not speciûod ARS : 60 mphd DS : 55 mph 50 Notspecifred ARS = 65mphd DS:60mph DS : 40-55 mph and not a hþh-accident location DS : 45-60 mph and not a high-accident location DS : 50-65 mph and nota high-acciderft location 40 Notspecified ARS : 55 mphd Spot Vertical Curve Design Speed (mph) 35 Notspecified 45 Notspecifred 50 Not specified ARS: 5Omphd ARS: 6ûmphd ARS : 65 mphd Crest DS : 45 ¡nph ,l4g DS: 50mph Crest DS : 50 mph sag DS : 55 mph Crest DS : 55 mph sog DS : 60 mph Crest DS : 60 mph sag DS : 65 mph DS : 35-50 mph, not â high-acc! dèût locatioo, and no geometric ieatu¡es warranting special considera- tion (e.9., inter- section) DS : 40-55 mph, not a high- accident location, and no geomet¡ic features warranting special considera- tion (e.g., inter- section) DS : 45-60 mph DS : 50-65 mph

Mississippi New Jerseya Souttt All highways All highways for resurfacing and restoration projects All highways All highways All highways DS : 50 mph DS : 35 mph and not a high- accident locationd DS : 55 mph DS = 40 mph and not a high- accident locationd DS : 60 mph DS : 45 mph and not a high- accident locationd DS = 65 mph DS = 50 mph and not a high- accident locationd Not specified (adequate for DS or signs should be provided) Not speciñed (adequate for DS or signs should be provided) Not specified (adequate for DS or signs should be provided) Not speciñed (adequate for DS or signs should be provided) Not specified (consider reconstruc- tion ifhigh- accident location) Not specifred (consider reconstruc- tion ifhigh- accident location) Not speciñed (consider reconstruc- tion ifhigh- accident location) Not specified (consider reconstruc- tion ifhigh- accident location) DS = 50 mph and not a high-accident locationd DS : 55 mph and not a high-accident locationd DS : 60 mph and not a high-accident Iocationd DS = 65 mph and not a high-accident location¿ DS : 50 mph DS : 35 mph and not a high- accident locationd DS : 55 mph DS : 40 mph and not a high- accident locationd DS : 60 mph DS = 45 mph and not a hieh- accident locationd DS : 65 mph DS : 50 mph and not a high- accident Iocationd Not specifred (if l0 mph posted speed, signing required) Not specified (if l0 mph posted speed, signing required) Not specified (if l0 mph posted speed, signing required) Not specified (if 10 mph posted speed, signing required) Not specifred (consider reconstruc- tion ifhigh- accident location) Not specilìed (consider reconstruc- tion if high- accident location) Not specifred (consider reconslruc- tion ifhigh- accident location) Not specified (consider reconstruc- tion ifhigh- accident location) DS : 50 mph and not a high-accident locationd DS : 55 mph and not a high-accident locationd DS : 60 mph and not a high-accidenl locationd DS : 65 mph and not a high-accident locationd Õ

TABLE B-9 continued Condition California Florida Illinois Michigan Bridge Clear width(fi) 20 Not permitted Not Permitted Not permitted Not Permitted Two-Lane Highwavs ADT = 250 Two-Lane Highways ADT : 400 DS : 50, ADT : 1,000 Two-Lane Highways ADT: 1000 DS=a0 ADT : 3,000 Minor Rehabílitation Where approach : t6 ft Major ñ ^I^ ^L: l:. ^.: ^- ADT ) 750 and where approach : l4ft Minor Rehabilitation Where approach : 18 lt Major RehabiLitation ADT ) 750 and where approach : l6lt Minor Rehabíli¡ation Where approach : 20 fr Major RehabìLitation ADT ) 750 and where approach : l8ft 22 24 ADT : 250 Where rehabili. tated approach roadway : 20 ft Stopping Sight DisÍance Less than AASHTO New Construc- tion Standards (Horizontal Curves) e When mitigated When properly by regulations, signed signing, and so forth Not specifred Not specifred Nores: ARS : average running speed, DS : design speed, ADT : average daily traffic, DHV : design hourly volume,ãnd T : percent trucks in traffrc stream. Table entries are based on minimum valuãs and Specify the situations under which that stated condition may remain. In many cases, "not specifred" mèans that although an exact number is not given in the state's standards, subjective treatment must be considered. Case study states fatl into three general categories with regard to distinguishing between urban and rural standards: lal states in which urban standards have not been developed (South Dakota); fål srateswith separate urban provisions(Mississippi, Florida, Washington, Illinois, andTèxas); and (c/ iiates with essentiãlly similar urban and rural slandards (California, New Jersey, and Michigan). Most ofthe AASHTO new construction standards differentiate between urban and rural.

Mississippi New Jerseya South Dakotaå Tèxas Washington Mi:nar Rehabilitation Permitted only ifno accident problern exists and roadway is taÞered before approach Major Rehabílitation Not permitted Minor Rehqbilitation Where approach : 22 rt Mqio, .Rehabilitation ADT: 750or DHV : 200 Minor R¿hsþilÌtøíon Where app¡oach = z4 ft Major Rehabllítation DHV: 400 Not permitted Noi permitted All Colleetors DHv = 200 Not permitted AII Collectors DHV:400 All highways TWo-Lane Highways ADT : 400 Iwo-La:ne Highways ADT : 400 Two-Lane Highways ADT: 750 ADT: 250 (exceÞt not permitted when rehabil- itation is done on bridge) ADT - 1000 (except not permitted when ¡ehabil- itation is done on bridge) ADT:4000 (except not permitted if rehabilitation is done on bridge) When properly Not permitted signed When properly signed Not specified When properly unless high- signia ' accident location a New_Jersey distinguishes between resurfacing and restoratioû projects that are minor in nature and rehabilitation projects that involve moreextensive safety-rel4ted imprçvements such as median ba¡rie¡s and minor widening. b In South Dakota, "minor reconsúuction" can involve minor widening, pavement rehabilitation, or shoulder improvements and thus are consistent with RRR defrnition. - ¿ California¡s standards specify minimum ¡'roadway" widths (lane plus sh.oulder) as follows:ADT Minimum Roadway,ft3,000 24 3,000-5000 28 . 5000 32 a In these states, il the ARS or DS is greater lha¡ that shown, consideration must be given to reconsfructing the eurve. If cost is prohibitive, additional signing is required.¿ Stopping sight distance is also implicit in veiticat curve deiþ.

TABLE B-10 Comparison of Roadside Tieatment Requirements in Special RRR Standards for Rural Highways, Case Study States IllinoisCondition California Florida Michigan Clear zone Not sPecifred (from edge of (removal of lane) obstacles should be considered) ARS> 40 mph 18 ft (if moved, 30 fÐ ARS< -- 40 mPh 14 lt (if moved, 20 f0 Federal-AÌd Secondary Routes DS(50orADT> 750: l0 ft or to ditch line Otherwise: l0 ft plus shoulder FederalAid Primary/Federal- Aid Urban (uncurbed) Posted speed):45:l8ft Otherwise: l0 ft Review accident history and potential for spot corrections Maximum side Not speciûed Not specified slopec (flattening of sloPes should be considered) Not specified 3:1 Twolane, ADT < 750 Multilane undivided, ADT < 10,000 4:1 All other highways Culvert tredt ments Not specifred Not speciñed Extend to meet Not specìfied RRR roadway width lf within clear zone, blend into slope Ifopening 30 in., use grates Ifopening 54 lt, protect with guardrail Review headwall positions UtiLities Not speciñed May be retained if not a safely hazard; otherwise treat as under "clear zone" Governed by state Governed by state clear zone policy clear zone policy

Mississippi New Jersey South Dakota¿ Tèxas Washington ARS> 40 Same as AASHTO Principal and Rural Multilane Not specified, Headwalls: 4 ft new construction Minot Atterials Highways but must be from standardsb Minor reconstruction: 16 ft considered in shoulder 25 ft ifADT RuralTwo-Lane a roadside Other: 14 ft < 1,000, 30 ft Highways hazard review If moved: 30 ft otherwise 7 ft, ADT < 750 report ARS < = 40 Resurfacing: 20 ft 16 ft, ADT > ADT> 750 Collectors : 750 Headwalls: 2 ft Minor reconstruction: from 20 ft if ADT < 500 shoulder 25 ft otherwise Other: l0 ft Resurfacing: 10 ft If moved: 20 ft (ADT < 250) 15 ft ARS<= 40, (250<ADT<500) ADT < 750 20 ft (ADT > 500) Headwalls: I ft from shoulder Other: 6 ft If moved: 12 ft Not specified Same as AASHTO If 3:1, slope must be Retain existing Do not steepen new construction justified, corrected, sideslope existing standardså or protected except where slopes; grade or consider crown modifying if changes 3:1 dictate otherwise Governed by Same as AASHTO Not specified ) 36 in. Mitre end state clear new construction zone policy standardsc Governed by IfROW available, Governed by state Governed by Governed by state clear locate outside clear clear zone policy state clear state clear zone policy zone unless cost is zone policy zone policy excessive or no accident problem IfROW not available, conduct analysis for relocation If inside clear sections zone or creates a safety problem, treat by: l. Grates, 2. Extending, or 3. Guard fence 1:36in. If inside clear zone, replace with 3: I or flatter culvert ends that blend into sideslope

240 TABLE B-10 continued Condilion California Florida Illinois Michigan Atherroadside None Spgcimentreesand Renoveorupgrade Tlee¡emovalifprovisians unique histori- guardrail frequent accident cal/environmen- Signorlight ortargetposition talfeatules,ifa supportswithin ofhorizont4l hazard, do not clear zoRe should curve harotobe bebreakaway Obstructingsight removed if - Remow 4 in. o¡ distance at inteÊprotected greaterdiameter section troes or protect \úslurlteer trees in Remove4 in. or elearzone higherconcrete Breakexisting signalboxes freeline within clear.zone Retain ifunique, Whe¡e practical sceniq or historic use impact value attenuâtors instead of guafdfail

Mississippi NewJersey South Dakotaø Texas Washington Specimen trees and unique hisforicaV environl1ten- tal features, ifa hazard, do not have to be removed if protected Upgrade safety appurte- nanceS (desirable) All safety appurte- nances should conform to state gEidelines Roadside obstacles eliminated or shielded by longitu" dinal bariers Upgrade or rcmovÊ guardrail Evaluate existing barriers and end trôat- me!ts Smooth traüsi- tion from guardrail to bridge rail Relocate, protect, o¡ provide breakaway sign and lishtiûs supports Protect bridge piers and ¿butments Modify raised dropiqlets in clear zone None Nore:-ARS: averagerunningspeed,ADT: averagedailytrafûc,DS : designspeed,andROW: rightoÊway. a. In South Dakota, "minpr reqonstruçtiod' is equivalent to minor widening and pavement rehabilita- tion. å New Jerseyb RRR staÍdards state that AAsHTo new c.onstruotion standards are to be applied iladesign feature is not explicitly mentioned in its standards- ¿ Refers to the rcgion adjacent 1o the shoulder more specifically known as the foreslope, I I

TABLE B-11 Federal-Aid Highway Projects, Case Study Cities City Population Federal-Aid TypesofProjects Category Perlormed With Received FederalAid Columbus, Ohio San Antoniq TÞxas Taiiahassee, Fioricia O¡lando, Florida Phoenix, Arizona Cherry Hill, New Jersey Tioy, New York Kansas City, Missouri Columbia, Missouri Moline, Illinois 62,000 46,000 EAU All types, including resur- lacing, reconstruction, intersections, and bridges Resurfacing, reconstruc- tion, and intersections r r^: - l,, -^-., -f^^: -^ - - Ilv!4lr1rJ¡ r!¡ur l4uruË ar¡u bridge rehabilitation because annual allot- ments are small Exclusively reconstruc- tion due to capacity needs Exclusively reconstruc- tion due to capacity needs Exclusively reconstruc- tion and intersections because ofcapacity and safety needs Ali types, including resurfacing, recon- struction, bridges, and intersections Exclusively reconstruc- tion (FAU) and bridge replacements and reha- bilitations (FAU and FA Bridge) Exclusively reconstruc- tion and intersections (favors large projects to reduce number of fed- eral-aid projects and associated federal granl procedures) Exclusively reconstruc- tion, intersections, or safety hazards due to metropolitan planning organization technical ranking procedure (favors projects with big benefits) 565,000 785,000 ö/!uuu r 30,000 765,000 64,000 56,000 448,000 EAU FAU FAU FAU FAU EAU FAU, FA Bridge FAU

TABLE B-12 continued County Federal-Aid Category Population Received Types ol Proj ects Perlormed With Federal Aid Schenectady County, New York St. Louis County, Missouri Union County, South Dakota Ingham County, Michigan Los Angeles County, Calilornia Shasta Count¡ California FAU, FAS EAU EAS, FAU, HES, RR Crossing FAU FAS, EAU 60,000 (unincor- porated) 200,000 (unincor- porated) l1,000 90,000 (unincor- porated) 7,4'17,000 (total) r r 5,000 (total) EAS, EA Bridge Exclusively reconstruction and bridges because ofcapacity and safety needs plus impression that state favors larger projects Tiaditionally reconst¡uction because ofcapacity needs but resurfacing and rehabilitation projects are now programmed Exclusively resurfacing and bridges because of small annual allotment offederal funds Mostly resurfacing and rehabilitation because ofstate's 90 percent preservation requirement Tiaditionally reconstruction because ofcapacity needs but now all types including reconstruction, RRR, intersections and bridges Mostly resurfacing because of need; some reconstruction Nor¡: MPO - metropolitan planning organization, FAU-federal-aid urban fund, FA bridge-federal- aid bridge replacement and rehabilitation funds, HES-federal-aid hazard elimination fund, FAS- federal-aid secondary fund, and RR crossing-federal-aid rail-highway grade crossing lund. TABLE B-13 Federal-Aid Highway Projects, Case-Study Metropolitan Planning Organizations MPOs Population Federal-Aid Category TypesofProjectsPerformed Received With Federal Aid Capital Districr Tiansportation Authority, Albany, New York Erie County, New York Planning Commission Chicago Area Tiansportation Study 430,000 (urbanized area) 850,000 (urbanized area) 6,000,000 (approxi- mate) EAU Mostly intersections, traffi c operations, and bridges but also reconst¡uction and RRR-type minor widenings (no resurfacing); project selection primarily because oIMPO technical ranking procedure Exclusively reconstruction primarily because ofcapacity needs but also past problems with standards capacity Tiaditionally reconstruction and new construction but shilting now to a balance among all types FAU Nora: MPO - metropolitan planning organization, FAU-federal-aid urban fund, FA bridge-federal- aid bridge replacement and rehabilitation funds, HES-federal-aid hazard elimination fund, FAS- federal-aid secondary fund, and RR crossing-federal-aid rail-highway grade crossing fund.

TABLE B-14 Allocation of Rderal Aid to Local Governments, Case Study States State Highway Agency Practices Ariz. Calit Fla. n- Mich. Miss. Mo. N.}I. N.J. N.Y. Ohio S.D.ak. Ibx. \¡a. Wasb, Allocates funds to local governments by formula Attributable urbân system Nonattributâþ[e urban systern Secondary Provides matching funds lo¡ local federal-aid projects Urban systeui Secondary Providesdesign and construction assistance for local fcderal-aid projects Performs most design Lets most construclioll contracts Yes Yes Yes Yès Yes No Yes Yes Yes Yes No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes No Yes Yes Yes 1ês' Yes Yes Yesa Yes Yes NA Yes" Yes Yes YesNAá NAll No.' No NAå NA¿ NA¿, NAá Yes I.iAá Yes N0NAå NoYes Yesd Yes YesYes Yes 'ltes NoNo No Yes No No NoNo No Yes No No No No No Yes No No No No Yes Yes Yes No No Yes Yes No No No )ê$ Yæ Yes No No Yes Yes Yes Yes Yes Yes Yes No ,rSome Nonattributable funds are reøined by state and dispensed !o smaller cities on a discretionary basis. r,Ñot applicable because neady all secondary highways are stats maintained; stâte agency may still aliocate fu¡ds by formula among its administratire substate aieas. , lf the federal-aid urban highway is also a state route, the state provides the full màtch, 'lEighry pçrcent of matching share is provided by state.

244 TABLE B-ll continued City Population Federal-Aid TypesofProjects Category Perlormed With Received FederalAid Various cities in Mississippi 10,000- 120,000 EAU Tiaditionally intersec- tions, but shifting to- ward RRR; lavored because olsmall an- nual allotment of FAU funds. (In Mississippi it :- ^^--^- --^^+i^^ Ê^-ls uuu¡r¡ru¡l pl 4vlruç lvr one consulting engi- neer to manage under contract several citieJ or counties' road pro- grams. In the cases described here, one consulting engineer was responsible for sev- eral cities and another lor several counties.) Nors: MPO- metropolitan planning organization, FAU-federal-aid urban fund, FA Bridge- federal- aid bridge replacement and rehabilitation funds, HES - federal-aid hazard elimination fund, FAS - federal-iid seìondary fund, and RR crossing - federal-aid rail-highway grade crossing lund. TABLE B-i2 Federal-Aid Highway Projects, Case Study Counties County Federal-Aid Category TypesolProjectsPerformed Population Received With Federal Aid Franklin County, Ohio 100,000 (uni ncor- Porated) Dallas County, Texas 150,000 (uni ncor- porated) Four small counties in 30,000- Mississippi 50,000 Coahoma County, 37,000 Mississippi Maricopa County, 300,000 Arizona (unincor- porated) Cape May County, 82,000 New Jersey Middlesex County, 50,000 New Jersey (unincor- porated) Tiaditionally reconstruction but shilting toward RRR and intersections because ol reduction in capacity needs Exclusively reconstruction (at state's recommendation) Exclusively reconstruction because of restrictiveness of standards Tiaditionally on bridges because of small allotment olannual FAS funds; shifting toward RRR FAU: reconstruction because ol capacity needs; some resurfacing FAS: bridges because ofsafety needs; some resurfacing Exclusively reconstruction and bridges because of restricliveness ol standards Exclusively reconstruction and bridges because of needs FAU, trAS EAU FAS FAS FAU, FAS FAS FAU, EAS

TABLEB-ll continued City Population Federal-Aid TypesofProjectsCategory Performed WithReceived FederalAid Decatur, Illinois Rapid City, South Dakota Sioux Falls, South Dakota Lansing, Michigan Detroit, Michigan Madison, Wisconsin Knoxville, Tennessee 47,000 EAU EAU, FA Bridge FAU EAU, HES, EA Bridge FAU, HES, FA Bridge FAU, HES FAU, FA Bridge 94,000 Exclusively reconsûuc- tion (favors large proj- ects to reduce number of federal-aid projects and associated federal grant procedures) Almost exclusively recon- struction (one overlay project in last 5 years); favors reconstruction because ofcapacity needs Exclusively reconstruc- tion because ofcapac- ity needs Tiaditionally reconstruc- tion and intersections but shifting toward re- . surfacing and pave- ment rehabilitation. Michigan state law re- quires that 90 percent of all highway funds be spent on "preserva- tion"; this also applies to federal aid. All types, including resur- fticing, reconstruction, intersections, and bridges Exclusively reconstruc- tion and intersections because ofcapacity needs Almcict exclusively inter- sections because of local budget process, small annual amount of EAU funds, and.re- strictiveness of design standards 8 1,000 130,000 t,203,000 r7l,000 175,000

TABLE B-15 Design Standards Used for Study States Local Federal-Aid RRR Projects, Case Has State Developed Special Federal-Aid RRR Standards? Urban System Secondary System Standards Applied to Local Federal RRR Projects State Arizona California Florida Illinois Michigan Mississippi Missouri New Hampshire No NewJersey Yes New York No Ohio No South Dakota Yes Tèxas Virginia Washington AASHTO new construction State's RRR standardso State's RRR standards Either state's RRR standards or AASHTO new construction Special local RRR standardsó State's RRR standards. AASHTO new construction AASHTO new construction State's RRR standards AASHTO new construction AASHTO new construction AASHTO new construction State's RRR standards AASHTO new construction Special local RRR standards¿ AASHTO new construction State's RRR standards State's RRR standards Either state's RRR standards or AASHTO new construction Special local RRR standardsó State's RRR standards NAd NA,/ State's RRR standa¡ds AASHTO new construction AASHTO new construction Special RRR standards specified in secondary road plan NAl AASHTO new construction Special local RRR standardsó No Yes Yes Yes Yes Yes No Yes No Yes ¿Los Angeles county.has deveroped its own RRR stanclards for rederar-aid projects. "standards that are distinct from the ones applied to state-administe.ed projecis. 'Arterials.onlyl AASHTO new construction standards apply to urban "oil"cio., and local streets."r\ol apprrcabre because 90 percenr or more ofsecondary highways are state mâintained.

Appendix C Summary of Detailed Safety Relationships As a par[ of this study, detailed relationships were developed that describe the tikely effects of the following design features on highway accidents: (a) lane and shoulder conditions, (Ã¥) bridge width, (c) horizontal and vertical curva- ture, and ld) roadside obstacles. The study focused on two-lane rural high- ways of ttre type eligible for resurfacing, restoration, and rehabilitation (RRR) funding. Considerable judgment was required in developing these relationships. Although they have been useful in several phases of the study, none has been validated in a critical way. Until such testing has been completed and appro- priate modifications made, the relationships must be considered as only iepresenting reasonable, most likely safety effects of these roadway and roadside elements. Their most promising interim use is to estimate ttre acci- dent reductions likely to result from incremental roadway and roadside improvements. The purpose of ttris appendix is to summarize ttrese safety relationships and to pressnt a method for estimating the combined effects of simultaneous improvements. LANE AND SHOULDER WIDTH AND SHOULDER TYPE Lane and shoulder conditions have been found to influence the frequency of accidents but not necessarily accident severity. Safety is enhanced by increases in lane and shoulder widths and improvements in shoulder surface type. '248

249 The safety effect of lane and shoulder width and shoulder type can be estimated as follows (1): ,4 = 0.0019 (AD710'tzz (0.879)w (0.919)PA Q.%z)uP (1.236)H (0.882)rEru 0322)rER2 (1) where A = number of run-off-road, head on, opposite-di¡ection sideswipe, and same-di¡ection sideswþ accidents per mile per ye¿r; ADT = two-directional average daily traftc volume; W = lane widttr in feet; PA = width of paved shoulder in feet; UP = width of unpaved (gravel, turf, earth) shoulder in feet;H = median roadside hazard rating for the highway segment, measured subjectively on a scale from 1 (east hazardous) to Z (most hazardous); TERI = 1 for flat tenain,0 otherwise; and TER2 = 1 for mountainous terrain, 0 otherwise. This accident model is limited in rhat it applies to ¡ Lane widths of 8 to L2 ft and shoulder widths of 0 to l0 ft. Combinations of lane and shoulder widths that can be reasonably modeled a¡e Iimited to those shown in Figure 3-2, Chapter 3; o Two-lane, two-way paved rural roads on state primary and secondary systems; and¡ Homogeneous roadway sections, and does not include the additional accidents expected at intersections. BRIDGE WIDTH Although the hazard associated with narrow bridges has been recognized for many years, efforts to quantitatively esøblish the influence of bridge cross section and geometry on accident frequency and severity have realized limited success. The more acceptable of these efforts use relative bridge width as the appropriate physical measure. Relative bridge widrh is defined as the dif- ference between the clear bridge width, including both trafûc lanes and usable shoulders, and the total width of the traffic lanes, excluding the shoulders, on the approach to the bridge.

250 The rate of bridge-related accidenß on two-lane highways can be estimated as follows (2): AR = 0.50 - 0.061 (AtÐ + 0.OOL2(RW2 for 0 < Rw < 14 A) where ,4lt is the number of accidents per million vehicles and RW is the relative bridge width in feet. Equation 2 does not apply in situations where the width of the approach traffic lanes exceeds the clea¡ bridge width: in this region, the accident rate is geatly increased by further constriction in the rnfÊ¡ lnncc nn the. hridoe. Nor does it annlv for relative bridse widths in excess of about 14 ft, aregion where the upturn in computed accident rates is more likely an artifact of the model-buitding process than a valid indication of impaired safety. Note that the measure of exposure to nanow-bridge hazard, as expressed in Equation 2, is the total number of vehicle traversals and not the more commonly used vehicle miles of travel. Factors other than relative bridge widtlt, such as bridge length and type (e.g., deck versus truss), the presence or absence of curbs, approach align- ment, pavement surface condition, and so forth, may also affect the accident rate at bridges. These factors are not included in Equation 2 because definitive information on thet safety effects is unavailable in the published literature. No evidence exists to suggest a relationship between the severify of constric- tion at bridges and the severiry of bridge-related accidents. HORIZONTAL CURVATURE Accidents are more likely to occur on horizont¿l curves than on sEaight, or tangent segments of roadway because of increased demands placed on the driver, the vehicle, and friction at the tire-pavement interface. Accident frequency on a segment of roadway con[aining a single horizontal curve and its tangent approaches can be estimated as follows (Appendix D): A = AÃ" (L) (V) + 0.0336 (D) (V) for L 2 L" (3) where A= .4R" = L= l= D= L.= total number of accidents on the segment, accident rate on comparable straight segmenß in accidents per million vehicle miles, length of highway segment in miles, traffic volume in millions of vehicles, curvature in degrees, and length of curved component in miles.

251 The accuracy of Equation 3 may be diminished for curves sharper than about 15 degrees, the approximate limit recorded in the data base from which the model was calibrated. simila¡ to the most likely relationship for accidents at narrow bridges @quation 2), the total number of vehicle traversals rather than vehicle miles of travel is used [o represent exposure to the extra hazards of travel at curve locations. Equation 3 does not capture ttre effects ofother physical featu¡es of curves-superelevation and superelevation runoff, cross-slope break, spiral transitioning, curve length, and so forth-that likely have significant, but unquantified, effects on accidents. Neither does it reflect the effects of road- way uniformity on driver expectations: a sharp curve immediately following an extended sEetch of straight highway will experience more accidents than a simila¡ curve situated within a generally winding section. Possible effects of horizontal curvature on accident severity have not been identified in the literature. The propensity for off-roadway encroachments onto the outside of horizonøl curves has been well esøblished, however, and it is clear that roadside conditions in this region can substantially affect accident severity and, quite possibly, accident frequency as well. VERTICAL CURVATURE The primary effect of vertical curvature on highway safety is related fo possible restrictions on sight distance that adversely affect emergency avoid- ance maneuvers. Both fhe severity and length of the sight-distance restriction are significant determinants of accident frequency. Accident frequency on a segrnent of roadway containing a single crest vertical curve and its tangent approaches can be estimated as N = AR¡U U) + AR¡ (L,) (V) (F",) where (4) N= ARh = number of accidents on a segment of highway containing a crest curve, average accident rate for the specific highway-or alternatively for the related general highway class-in accidents per million vehicle miles, length of highway segmenr in miles, traffrc volume in millions of vehicles, length of restricted sight disønce in miles, and a hypothetical accident rate facûor that varies according ¡o bottr L= V= Lr= F=

252 ttre severity of the sight restriction and the nature of the hidden hazard. Procedures for estimating tr" and for selecting Fo, are detailed in Appendix E. Because Equation 4 has not been validated by comparison with actual accident experience, it clearly must be used with considerable caution' Log- ically, it does appear to incorporate the primary effects of restricted lines of sight, although only superficial Eeatment is given to characterizations of the severity of the hidden hazard. With the possible exception of drainage on -^^:..,^-,^ ..,irt. ^,,-L^ ti¡rla ^- -^ Ã"i,{ôñ^â avicto ln crrftftacf fhof fo¡ln¡c nfherruauw4Jù wlul vuruùr uluv u¡ ¡¡v vYruw¡rvw w^rùe rv ùuó6vÙ! than restricted lines of sight affect the safety of operations on vertical curves. The geometry of vertical curves is not lnown to have a signiûcant effect on accident severity. ROADSIDE FEATURES The frequency of roadside accidents can be reduced by geometric improve- ments such as lane widening and curve lengthening and by EafÍic conrol measures such as edge lines and post-mounted delineators, all of which are designed in part to reduce the frequency of inadvertent roadside encroach- ments. Perhaps the most beneficial roadside improvement is the provision of a clear zone adjacent to the travel lanes, an obstacle-free area of traversable slope. The likelihood of a run-off-road accident diminishes rapidly as poten- tial hazards are further displaced from the roadway: even relatively narrow recovery areas are remarkably effective. Recent developments in the design of safer drainage systems provide important opportunities for enhancing safety on RRR projects l3). Additional roadside improvements include flattening of sideslopes to reduce the likelihood of overturn, and redesign of roadside hardwa¡e to reduce the severity of impact. When necessary, the danger of a hazardous roadside condition is lessened by the provision of guardrail to cont¿in and redi¡ect the errant vehicle. The impact of the roadside environment on highway safety has proven to be diffrcult to establish. Useful composite me¿ìsures of the degree of roadside hazard have been elusive, and ttre diversity of the many potential hazards has been almost overwhelming. However, a review commissioned by this study in conjunction with FÉIWA (/) revealed a signiûcant relationship between road- side recovery distance (measured from the outside edge of the shoulder to the nearest roadside obstacles or hazards) and the number of single vehicle, head- - on, and sideswipe accidents on two-lane rural highways:

253 Roadside Recovery Distance (Jt) Percert Reduction In Acciderts 13 25 35 44 5 10 15 20 Roadside encroachment models arc often used to examine the safety effects of speciflc roadside features. The following model is recommended for such use: Ex(A ¡) = o'07":#)r )o'slls + Ðxre4'08224t¡'' J where P r (A ¡lC ) (Ð x' r'0'o822 4Y, (s) Ex(A¡) = exPêcted annual number of accidents involving the hazard; ADT = nvo-di¡ectional average daily naffic volume; Pr(A¡lC¡) = probability that an encounter or collision with the hazard will result in an accident (Table F-1, Appendix Ð; and .x's and y's = projected length along the Eavel lanes and offset to elements of the hazard, respectively. The ¡'s and y's reflect anticipated encroachment angles of errant vehicles, as well as the offset to and dimensions of the roadside hazañ and are chosen in part-following the example of Figure F-l and Table F-2, Appendix F-for computational convenience. The subscript, i, in Equation 5 pertains to near- side encroachments whereas the subscript, j, refers to far-side encroachments. Accident reduction factors, computed using Equation 5, are applicable only to the run-off-road portion of the total accident population. They are likely to undersiate the reduction achievable at particularly critical locations such as the outside of sharp, horizontal curves. In addition Ûo the number of accidents, the severity of run-off foad accidents can also be estimated (Täble F-1, Appendix F). COMBINED EFFECTS The øsk of sorting out possible safety effects of incremental roadway and roadside improvements is best performed through ttre application of accident

254 reduction factors to historical accident data collecæd at the location being improved. Should actual accident data be unavailable, however, the analyst must use models, calibrated from other accident-data sources, that directly estimate either accidents or acciden[ ntes. These models are not expected to be as reliable as the use of accident reduction factors because of their additional abstraction from the location in question. For either type of analysis, the major question involves the future; that is, how the accident pattern is likely to differ if the improvement is or is nor made. When the likely change in accidents is estimated by applying accident tq¡trr¡tìn- f¿¡tnrc +¡ +h^ | i¡+ai¡^l ^^^:.1^-G -^^^-.1 ^^-^ ....^¿ L^ .^t-^- ¿^ ------^¡wuvuv.¡ rswtvrù Lv u¡v ¡¡rùu¡rv4r qw¡uçtlt lw\rtu, t,drft ltluùL uç ku\¡tlt LU ¿1,5sulç that anticipated changes in raffrc volume are properly accounted for. More difficult to treat than the volume influence is the estimation of accident effects for situations in which two or more highway features are improved simultaneously. Interactive effecs of such combined üeaûnents can result in any of three possible consequences: improvement to one highway feature c Diminishes the relative safety effect of an improvement to another feature; o Has no infiuence on the relative safety effect of improvement to another feature; or o Enhances the relative safety effect of improvement to another feature. Because of difficulty both in experimental design and model calibration(4),few interactive accident models are available. Of the models developed for use here, none incorporates interactive effects. In tl¡e absence of interac- tive models, ad-hoc procedures are required to estimate the accident reduction likely to result from two or more simultaneous improvements. A fust-order approximation of combined effects, in use for many years (5), is illustrated for tt¡ree simultaneous improvements as follows: ARF^ ARF, ARF. ARF" roo-= I - (1 -i¡o.l tr -lõõ-) (r - rõd) (6) where ARF" is the accident reduction factor, expressed as a percentage, expected from the combination of improvements, and.A.,i?F 1, ARF2, and ARF, are similar factors representing the effects of the ûrst, second, and third improvements, respectively. Although use of Equation 6 is recommended for estimating combined effects, it is necessary to emphasize that it does not account for interactions

255 among the combined improvements (ó). Consequences of this deficiency will be ameliorated by r Categorizing the accident pattern into subsets, consonant with the types of accidents likely to be affected by the improvements being made; . Applying Equation 6 only for that subset(s) iointly influenced by the combined improvements; and¡ Applying a single accidentreduction factor-usually the more signiflcant one-in situations where one improvement is likely to substantially negate influences of another. lSuch a situation is most likely when joint improve- ments are intended to affect the same subset of the accident population by simila¡ means. A good example is ttre addition of both centerline and edge striping together with post-mounted delineators at a haza¡dous horizonøl curve. In this example, each mitigation measure is designed to improve delineation.l Also, a minimum estimated post-improvement accident rate-based on accident experience for the best designed and operated facilities-is useful to guard against overestimating the safety beneflts of improvements (7). REFERENCES l. C. V. Zegeer, J. Hummer, D. Reinfurt, L. Hurf, and W. Hunter. Safety Effects of Cross-Section Design for Two-Lane Roads, Vols. I and II. Report FHWA- RD-87/008 and 009. FHWA, U.S. Department of Transporøtion, 1986. 2. D. S. Tumer. "Prediction of Bridge Accident Raæs." Journal of Transportation Engìræering, Vol. 110, No. 1, American Society of Civil Engineers, New York, Jan. 1984. 3. H. D. Robertson, S. Basu, K. Colpitts, S. Steiru F. Johnsoru and G. K. Young. Saler Drainage Systems, FHWA Olfice of Research and Development, U.S. Department of Transporøtion, July 1986. 4. F. M. Council et al. Accident Research Manaal. Report FFIWA-RD-80/016. FHWA, U.S. Department of Transportation, Feb. 1980. 5. Roy Jorgersen Associates. Evaluation of Critería for Safety Improvements on the Highway. Gaithersburg, Md., 1966. 6. H. D. Robertson and K. S. Opiela. Feasibility of Determining the Incremental Effectiveness of Accídent Countermeasures. Report FHWA-RD-85/043. FHWA, U.S. Departnrent of Transportation, Feb. 1985. 7. T. N. Tamburri and R. N. Smith. "The Safety Index: A Method of Evaluating and Rating Safety Benefits." In Highway Research Record 332, TRB, National Resea¡ch Council, Washington, D.C., 1970, pp. 28-39.

Appendix D Relationship Between Accidents and Horizontal Curvature Although a number of resea¡chers have suggested a relationship between accident rates at horizontal curves and the radius or degree of cuwature, the validiry of these relationships is often unknown because of questionable experimental design and the imprecise definition of both curve-related acci- dents and vehicle exposure (/). Among the resea¡ch studies currently avail- able, a recent study by Glennon et al. (2), sponsored by the Federal Highway Adminisration, is one of the best. In this study proper experimental design was applied in assembling a data base that includes accident, geometric, and traff,c data for two-lane rural highway segments, each of which included a horizontal curve. The relationship between accidents and horizontal cuwature reported in Chapter 3 and used in Chapter 5 for cost-effectiveness analyses of curve flattening projects is based on these data and the statistical analyses reported by Glennon et al. For site-specific design, the relationship provides a helpful estimate of the accident reduction that may be possible by flattening a curve, but it must be used judiciously atong with other pertinent information such as traffic characteristics, adjacent highway alignment, shoulder and roadside characteristics, and prior accident experience. A summary of the datå set, and the most likely quantitative relationship is presented next. GLENNON, NEUMAN, AND LEISCH DATA SET Highway sites (or segments) in this data base were typically 0.61 mi long or longer as necessary to assure sraight (øngent) components at least 650 ft long 256

257 at each end of a horizontal curve. Care was taken to select sites with uniform lane and shoulder conditions and O avoid influences of major bridges and intersectionS, curbs, and Other nearby hOrizontal curves. FurthermOre, selec- tion ',vas limiæd to sites without reconstruction or signifrcant trafflc gowth during ttre 3-year study period. Sites were located in four states (Florida, Illinois, Ohio, and Te¡as) and limited to two-lane rural roads with daily Eaffic averaging 1,500 or more vehicles. The limited number of straight, control sites was compalable to the curved sites both in traffic volume and in lane and shoulder conditions (Table D-1). TABLE D-l Summary Statistics of Glennon et al' Data Sase (Z) Type olSite Straight Curved 351 3,4t7 0.632 2.22 0.902 t 1.50 7.20 Number of sites Average ADT (VPD) Average site length (mi) Average No. olaccidents (3 Years) Average accident rate (PMVM) Average lane width (ft) Average shoulder width (ft) Average degree of curvature Average curve length (mi) 1 to? 3,184 0.631 3.93 1.823 I 1.45 6.87 3.4'l 0. l7l Norrs: ADT : average daily traffic; vPD : vehicle per day; PMVM = per million vehicle miles' Dashes (-) indicate not applicable. In terms of safety, the average curved site was distinctly inferior to the average súaight sio (Table D-1). An approximate proration of the accidents on thJaverag:e curved site to its straight and curved components indicates that the accidentlaþ on the curve Section was three to four úmes the rate on the straight sections. tricept for what may be an anomaly at the stratification for the smallest degree õf ru*utur" (which persists in later tabulations as well), the average acãidentrate signifrcantly increased with increased degree of curvature (fable D-2). Even larger increases would be expected had it been possible to separate the effects of the curved and sfraight elements within each of the nominally 0.61-mi sites. It is likely that accident rates, such as those reported in Table D-2, ue biased by unknown influences of other safety-related variables. It is reason- able to ã"pa"t, for example, that sites containing the sharper curves had geometric and roadside condiúons that were also likely to have been more ñazardous. Although this possibility is generally confumed by the ftends in average lane and shoulder widths as reported in Table D-2, the differences do

258 TABLE D-2 Average Accident Rate (PMVM) on Nominally 0.61-mi Site as a Function of Degree of Curvature Degree of Curve Average Average Accident Rate ADT(PMVM) (VPD) Average Average Lane Width- Shoulder Width(fÐ (fÐ 0 0.01-0.74 0.7 5-1.49 t.50-2.49 2.50-3.49 3.50-4.49 4.50-6.49 6.50-8.49 8.50- 10.49 10.50- 12.49 12.50 or more 0.90 l.38 1.0ó t.24 1.6 I 2.4t 2.79 2.89 3.59 4.03 4.19 3,400 3,100 3,300 3,200 3,400 3,000 3,200 3,300 3,000 3,200 2,900 I 1.5 tt.'7 I 1.9 l 1.8 tt.7 I 1.3 r0.9 r0.4 r0.2 r 0.3 r0.0 7.2 7.5 7.4 7.3 6.3 5.9 4.8 4.8 4.8 4.8 Nor¡: Täbulations are based on data from Glennon et aI. database (2). not appear as large as might have otherwise been anticipated. Based on the analysis by Zegeer and Deacon (3), the reduced lane and shoulder widths for the sharper curves could be expected to account for only about 10 percent of the difference in the accident rates between the sha¡pest and flattest sites. Additional cross tabulations indicate ttrat lane width may have had a minor effect on reported accident rates (Table D-3) and that volume effects are likely to have been small (Table D-4). In addition, once degree of curvature is øken into account, the data show no consistent and pronounced relationship TABLE D-3 Average Accident Rate (PMVM) on Nominally 0.61-mi Site as a Function of Degree of Curvature and Lane Width Lane Width (lt) Degree ol Curve 0.96 r.69 1.54 1.76 1.95 2.64 3.08 3.65 3.90 4.34 4.15 Nor¡s: Italicized numbers indicate 30 or more sites. Tabulations are based on data lrom Glennon et al d^ta base (2). 9 or Less l0 I I t2 or More 0 0.01 -0.74 0.7 5-1.49 |.50-2.49 2.50-3.49 3.s0-4.49 4.50-6.49 6.50-8.49 8.50- 10.49 t0.50-12.49 12.50 or more 1.17 1 21 1 .01 r.30 1.67 2.45 2.9 2 2.69 3.44 4.t I 4.76 0.90 I .J 1 1.35 1.30 1.93 2.49 2.82 5.29 2.24 4.57 2.08 0.78 r.40 0.98 I.t8 L52 2.24 2.6 I 2.tt 3.82 3.66 2.69

259 TABLE D-4 Average Accident Rate (PMVM) on Nominally 0.61-mi Site as a Function of Degree of Curvature and Volume Volume (VPD) Degree of Curve 2,099 or less 2,100-3,099 3,100-4899 4900 or more 0 0.0 I -0.74 0.'15-t.49 t.50-2.49 2.50-3.49 3.50-4.49 4.50-6.49 6.50-8.49 8.50- r0.49 10.50-12.49 12.50 or more 0.8s 1.93 t.06 1.3 I L85 2.36 2.90 2.68 3.7 3 3.64 4.r 5 0.74 1.10 1.08 L21 1.49 2.12 2.91 2.61 3.56 4.7 t 3.94 Nor¡s: Italicized numbers indicate 30 or more sites. Tabulations are based on data from Glennon et al. database (2). between accident rate and either curve length or curve cenüal angle, a measure of the total change in direction of the highway (Tables D-5 and D-6). DEVELOPMENT OF ACCIDENT MODEL Concemed with the possible confounding effects of other variables (state, lengttr of curve, lane width, and shoulder width), Glennon et al. (2) used analysis-of-covariance techniques to isolate the incremental effects ofchanges in degree (or sharpness) of curvature on accidents. The estimated reduction in the number of accidents per million vehicles for a one-degree change in crlrvature was found to be 0.0336, a number that served as the basis for development of a complete accident model. To develop úe model, the number of accidents at the combined straight- curved sites, .4, can be represented as the sum of two components: A=A"+4" where A is the number of accidents on the straight segments and,A" is the number of accidents on the curved segment. If ,4-R" is the accident rate (PMVlvÐ on the straight roadwa¡ then A =,4R. (L")(V) + A" or A =,4R" (L)(V) + [4, - Æ, (L")(V)] 0.72 r.00 1.05 1.22 1.71 2.73 2.49 3.46 3.51 3.74 3.91 1.55 1.32 1.01 LI4 1.26 2.8t 2.77 2.9'7 3.4'7 4.45 6.1'1

260 (1) l and A = AlR" (L)(V) + M" where length of site in miles, length of sraight segments in miles, length of curved segment, in miles, and traffie volume in millions. Â4", the increase in accidents as a result of curvature beyond that antici- pated for a straight roadway of the same length, was calibrated as follows: M. = 0.0336(DXÐ where D is the degree of cuvature. The complete accident model can then be expressed as A = AR"(L)(V) + 0.0336(D)(Iz) for L > L" and, on the curved segment alone, A" = ^AJ?" (L"Xv) + 0.0336(D)(Iz) From the data base, Æ, = 0.902 (Table D-1). Equation 3 has two components: the ûrst repressnts a steady-ståte turning effect and the second represents transitional (entry and exit) effects. The steady-state-turning component is directly proportional to the vehicle miles of travel on the curve but is insensitive to degree of curvature. The transitional component is directly proportional to both degree of curvature and traffic volume. Ottrer model formulations are possible, and a number were calibrated as paft of ttris study. It was concluded, however, that significant improve- ments to the Glennon et al. model (using the existing daø base) are unlikely and that this model @quations 2 and 3) represents the most acceptåble technique currently available for estimating the safety effects of curve flatten- ing. MODEL APPLICATION The model described was used to analyze ttre safety-cost effectiveness of different RRR policies for improvement of horizontal curves (Chapter 5). As L= L"= L"= v- (2) (3)

TABLED-6 Average Accident Rate (PMVM) *}{"*'"* ----------==- Degree of Curve 2.49 ot less 1.08 1.26 I .13 l.I4 2.42 2.87 ?q? 3.35 ).t+ 1,40 0.96 1.39 1.20 3.38 3.03 2.98 3.24 5.04 |.4'7 0.98 1.25 1.54 2.41j.00 2.87 3.98 3.83) 5R 0.01-0.74 0.?5- 1.49 1.50-2.49 2.50-3-49 3.s0-4.49 4.50-6.49 6.50-8.49 8.50-10.49 10.50- 12.49 ',"1)2 ',¡iot, ti;'ri, iä ".i,": ),,12 '¡11 tåi \t:, ,;n i.ii \\iiffi 2iå riz T1:, ',^J, \T:^ 'i'J TilI i:;å i{r 'ir: i:'iõ'.: \,'^1 ,rl ,urú. '," 'i:f, iq îm ffii "rTA ,,::J Ti¿- "-: .24 128" r')r ;.iõ 3.83 1:9 iiz 1.eø 1-6.2-- z rt Z;gt I :i t.lo It: u,to, ^r'f, "t.ii 1 oi z'st 1.87 2.03 2.37 12.50 or moreäió ì,not. ,.lo , ,o 1'.¿1 183 't' _:',,,,,_, rionetar. database(2)' no sites' Tabulations ¿ Nores: ltalicized numbers indrr All

TABLE D-5 Average Accident Rate (PMVM) on Nominally 061-mi Site as a Function of Degree and Length of Curvature Length (mi) Degree of Curve 0.000- 0.049 0.050- 0.069 0.070- 0.089 0.090- 0.r09 0. r r0- 0.139 0.140- 0.1 69 0, l 70- a.209 0.210- 0.259 0.260- 0.320 0.3 19 or more 0.0 r -0.74 0.7 5-t.49 r.s0-2.49 2.50-3.49 3.50-4.49 4.50-6.49 6.50-8.49 8.50- r 0.49 t0.50-12.49 12.50 or more AIl 1.78 1.24 1.43 2.90 3.54 2.99 2.85 3.46 4.8A 4.63 ¿,J/ 1.19 1.03 1 .13 0.92 2.61 3.24 3,24 4.06 3.86 4.43 2.74 1.60 0.91 L6s 1.57 2.i6 2.53 3.02 ?50 3.58 3.s6 2.10 1.0I 1.41 0.95 1.44 2.t'l 2.40 2.01 2.43 2.19 3.24 1.49 1.57 1.04 L8s r.34 L77 3.19 3.56 4.57 2.25 2.55 /.79 1.65 ô.sa 0.98 L73 2.59) <¿, 2.02. 4.14 3.41 3.65 t.59 L4l r.14 r.32 TÁ\ 2.05 2.33 1.7 1 2.70 1.09 r.4/ Ì.15 0.94 I.t6 1.36 2:..74 t.62 3.49 1,26 /,]4 0.87 0.83 t.05 2.48 2.74 2.2A 0.00 1.19 I.16 T,14 1.03 2.44 2.89 ,u: t,túI/9Norrs: Italicized numbers jndicatej0 o¡ mo¡e sites; Or*

263 noted earlier, this relationship should not be applied to speciûc project design situations without careful consideration of a variety of site-specific factors. When it is used, the coeffrcient, 0.902, should be replaced where possible by an accident rate on straight segments representative of local conditions for the highway under consideration. Equation 2 cantr- used to estimate the reduction in accidents from flatten- ing a horizontrl curve while maintaining is lines of tangency (or central angle). Because curve flatæning reduces the overall lengttr of highway-as well as increases the lengttr of its curved component-before-and-after com- parisons must be made between two points on the original alignment whose locations remain unchanged by the improvement. Any two points, each separated from Íhe point of intersection by at least the tangent distance of the new alignment, will be sufficient. Using Equation 2, the net reduction in the number of accidents, Â4, is as follows: M= Ao- Ao M = ÁlRn Ø" - L)V + 0.0336 (D" - D)V M = AR"(M)V + 0.0336 (^D)Y (4) where the subscript, o, refers to the original alignment and the subscript, n, the new or improved alignment. The amount by which the alignment length is shortened, A.L, can be determined as LL = ÍQ.170 tan Il2) - (t/s2.8)l l(r/D") - (rlD")) (s) where Âl is expressed in miles and,I is the central angle in degrees. At central angles of 90 degrees and less, ÁL is suff,ciently small that it can be ignored. Origlnal Curve Ao = accldents on orlglnal allgnment botwoen PCn and PTn An = accidents on new allgnment between PCn and PTn a _An lroo)percent reduction = _ï" FIGURE D-l Basis for calculating accident reduction percentiles reported in Table D-7. v/here

2& TABLE D-7 Percentage Reduction in Accidents Due to Horizontal Curve Flattening Original Degree of Curve New Degree ofCurve Central Angle (deg.) 1020304050 30 25 20 15 l0 5 ZU l5 l0 5 15 r0 5 l0 5 5 16 16 16 16 1632 33 33 33 3349 49 49 49 5065 65 66 66 6682 82 82 82 83 tn rn a^ 1^ 1^IA L1 LV LV LV39 39 39 39 4058 58 s9 59 6078 78 78 't9 7924 24 24 24 2547 48 48 49 49 '12 72 '72 73 '7430 31 31 32 336t 62 63 64 6s4t 42 44 45 48 ¿) 20 t5 l0 Nor¡: Based on lR, : 0.902 accidents PMVM, the average accident rate at straight sites from Tâble D-1. Percentages based on the number ofaccidents with original curve between the points oftangency of the new curve (see Figure D-l). For a given change in the degree of curvature, ttre greatest accident reduc- tions-when expressed as a percentâge of the number of accidents on the original alignment, which occu¡ between the tângent points of the new curve (Figure D-l)---¿re expected for the smallest central angles and the sharpest original curvature (Table D-7). For given initial conditions (Do and I), accidents decrease with reductions in the degree of curvature (D). REFERENCES L J. C. Glennon. "Effect of Alinement on Highway Safety: A Synthesis of Prior Research." In TRB State-of-the-Art Report. TRB, National Research Council, Washington, D.C. (forthcoming). 2. J. C. Glennon, T. R. Neuma¡r and J. E. Leisch. Safety and Operational Considera- tionr for Design of Rural Highway Curves. Report F[{WA-RD-86/035. FHWA, U.S. Department of Transportation, Aug. 1983. 3. C. V. Zegeer and J. A. Deacon. "Effect of Lane Width, Shoulder Width, and Shoulder Type on Highway Safety: A Synthesis of Prior Research." In TRB State- of-the-A¡t Report. TRB, National Resea¡ch Council, Washington, D.C. (forthcom- ing).

Appendix E Relationship Between Accidents and Sight Distance at Crest Vertical Curves Neuman and Glennon have developed a hypothetical model for estimating the effects of restricted sight distances at crest vertical curves on accident rates û,2). Although the model has not been validated by comparison with accident data, it provides a useful first approximation of beneûts that can be expected from improvements to substandard vertical curves. The model is intuitively appealing because it accounts not only for the severity and length of ttre sight restriction, but also ttre relative h.azard of that portion of the hidden roadway. According to Neuman and Glennon, the model is likely to overståte the detrimental effects of restricted sight disønce. Pending more substantial validation, the model should be used with caution and the understânding that it probably yields an upper bound for accident reductions resulting from increasing sight distance where existing conditions do not meet American Association of State Highway and Transporøtion Officials (AASHTO) søn- dards. As postulated by the model, the number of accidents attributable to a sight resriction is directly related to both its severity and its length. The severity of the restriction is measured by the difference between the speed at which vehicles operate on the curve and the speed at which, according [o AASHTO procedures, they could safely stop in emergencies. The length of the restric- tion is the distance over which the available sight distance is less than that considered adequate by AASIITO procedures for the actual highway operat- ing speed. 265

266 For a highway segment containing an isolated vertical curve, the accident model can be expressed as N = AR¡ (L) (V) + AR^(L,) (v) (F*) where N = number of accidents on a segment of highway contâining a crest curve, ARa = âr/erâgê accident rate for the specifrc highwa-v---or alternatively for the related general highway class-in accidents per million vehicle miles, L = length of highway segment in miles,V = Eaffic volume in millions of vehicles,L, = length of restricted sight disønce in miles (the distance over which sight distance is below or equal to the value specifled by AASHTO), and Fo, = a hypothetical accident rate factor that varies according to the nature of ttre sight restriction and the nature of the hidden hazud. In applying Equation 1, the average accident rafe, AR¡, is preferably based on historical data collected for a substantial length of the highway under consideration. In ttre absence of such data, the use of statewide averages for the particular highway type is recommended as an alternative. For their cost- effectiveness evaluation, Neuman and Glennon used a rate of abow 2.4 accidents per million vehicle miles to represent, average conditions on two- lane rural highways (/). The length of sight restricúon, L,, is a complex function of the highway operating speed and curve geometrics. As an approximation, it can be esti- mated as L, = (ao + ar{) (V5,280) A) where the a's are the constants idenúûed in Table E-l and .4 is the absolute value of the grade difference in percent. To use Table E-1, the highway design speed must first be determined. This requires computing the stopping sight distance (SSD) for existing curvature conditions as follows: SSD = [7.017 x 106 il,"¡7410's for SSD < L,, (3) or SSD = 2,&0(L") + 664.51A for SSD > Lu" (4) (1)

267 TABLE El Constants Used for Determining Length of Restricted Sight Distance (L,) by Equation 2 Highway Operating Speed on Vertical Curve (mph) Highway Design Speed (mph) 60 55 50 45 35 2530 Values ofao 60 55 50 45 40 35 30 25 -524 -138 -452 1 13 20211 lll -65 45 -332 -'76 _)7) 256 305t72 221 1 15 1692t 82 -55 15 -231 -74 - 193 -25 - 163 -405 382 301 248 t6't 110 5l l9 - 130 No sight restriction Values ofa, 207.3 152.6 t20.9 80.2 56.6 38.6 29.4 where SSD is the stopping sight distance in feet and Lu" is the length of the vertical curve in miles. The design speed is then found by inærpolation from Table E-2. The appropriate accident rate factor, Fo,, is selected from Table E-3. Using Equation 1., the accidents attibutåble to a speciflc curve (excluding its straight approaches) can be estimated as N"= AR¡(L,")V + ARh&) U) F",) (5) where N, is the number of accidents attributable to the curve. Again using Equation 1 and making comparisons as necessary over a common tength of highwa¡ the change in accidents expected from improving the stopping sight distance is as follows: ÂN = ái?¡ (n W (L,F",)) TABLE E2 AASHTO Stopping Sight Distance as a Function of Design Speed (! 15.3 (6) Design Speed (mph) Stopping Sight Distance (ft) Design Speed (mph) Stopping Sight Distance (lt) 45 50 55 60 25 30 35 40 r50 200 225 275 32s 400 450 525

Severity ofSight Restriction (mph) 268 TABLE E-3 Accident Rate Factors (F,l Degree of Hazard in Sight-Restricted Area' Minor Significant 0 (0.3) 0.5 t.2 2.0 Nor¡: Numbers in parentheses were interpolated from Glennon (2,/' aSee Table E-4. TABLE B4 Classifrcation of Degree of Hazard in SightRestricted Area Relative Hazañ Geometric Condition Major 0 5 l0 t5 20 0.4 (0.8) 1.1 2.0 3.0 i.0 ( r.4) 1.8 2.8 4.0 Minor Significant Major Tängent horizontal alignment Mild curvature (less than 3 degrees) Mild downgrade (less than 3 percent) Low-volume intersection Intermediate curvature (3 to 6 degrees) Moderate downgrade (3 to 5 percent) Structure High-volume intersection Y-diverge on road Sharp curvature (greater than 6 degrees) Steep downgrade (greater than 5 percent) Narrow bridge Narrowed pavement The fractional change in accidents is expressed by the quotient of Equations 5 and 6 as follows: N,l =A(L,F,) e)N, Lu" + LrF o, where the denominaûor on the right represents conditions on the original curve and the numerator represents the change from original conditions. The effect of curve improvements on accidents can best be estimated by applying the ratio of Equation 7 to the known number of accidenfs on the original curve. Should the historical accident record be unavailable, Equation 6 provides the next best estimate, the most reasonable estimate being made for ARh. Equations 6 andT are useful primarily for evaluating the safety benefits of incremental improvements in sight distance over practical ranges. They are

269 not applicable when the actual operating speed is less than the design speed for either the unimproved or improved condition. In such a circumstance, further improvements to the design speed are not expected to result in substantial added benefits to safety. REFERENCES 1, T. R. Neuman and J. C. Glennon. "Cost-Effectiveness of Improvements to Stop- ping Sight Distance Safety Problems." InTrarsportation Resea¡ch Record 923. TRB, National Research Council, Washingøn, D.C., 1983' pp.26-34' 2. J.C. Glennon. "Effect of Alinement on Highway Safety: A Synthesis of Prior Research." h TRB State-of-the-Art-Report. TRB, National Research Council, Washington, D.C. (fo¡thcoming). 3. A Policy on Geom¿tric Design of Highways and Streets. American Association of State Highway and Transportation Ofûcials, Washington, D'C.' 1984.

Appendix F Relationship Between Accidents and Specific Roadside Features Described in this appendix are ttre development and calibration of a roadside encroachment model, one of two types of models used to quantify the safety effects of the highway roadside envi¡onment. The second type, regression modeling, was used in work by Zegeer et al. (I) and is briefly summarized in Chapter 3 and Appendix C. Using gross measures of roadside condition- such as "percentage exposure length to objects within 6 m" (2) or "number of discrete objects, 0 to 5 ft from pavement edge" (3)--regression models are most useful for explaining the overall contribution of the roadside environ- ment to highway safety. At the same time, they have been effectively used in analyzing the effects ofspecific types ofroadside objects such as utility poles (4). In contrast, roadside encroachment models are an attractive altemative for capturing the effects of a variety of roadside objects while permitting detailed investigation of either a specific highway site characterizedby a homogenous roadside condition or an extended segment with a mixture of roadside ele- ments. Because of these advantâges, an in-depth investigation of encroach- ment models was undertaken to supplement the work by Zegeer et al. (I). Conceptually, roadside encroachment models capture the sequence of events culminating in a roadside accident and described as follows: o An out-of-cont¡ol vehicle leaves the Eavel lanes and encroaches on the roadside, . Location of encroachment is such that the path of travel is directed toward a potentially hazardous object or slope, 270

27r . Hazardous object is sufficiently close to the travel lanes that conEol is not regained before encounter or collision between vehicle and object, and . Collision is suftciently severe enough ûo result in an accident. 'When expressed in the nomenclature of mathematical probability, the encroachment model assumes the following form: Ex(A) = Ex(E) Pr(E¡lE) Pr(C^lE) Pr(A¡lC¡) (Ð where Ex(Ar) Ex(E) Pr(E¡lE) Pr(C¡lE) Pr(A^lC) = expected annual number ofroadside accidents involving a speciflc hazud (h); = expected annual number of encroachments on the highway segment encompassing the hazæd (typically I mi long); = conditional probability that, given an encroachment, iß location is such that an impact with the hazard is possible; = conditional probability that, given an encroachment in the potential impact area, a collision between vehicle and object will occur; and = conditional probability that, given a collision, its severity will be so greåt as to result, in an accident. The fotlowing extension is useful for examining more severe types of acci- dents: Ex(CAì = Ex(A¡) Pr(CA¡lAn) Q) where Ex(CA¡r) is the expected annual number of casualty (injury or fatality) accidents involving the hazard and Pr(CA^þ¿) is the conditional probability that, given an accident" an injury or fatality will occur. Development and calibration of roadside encroachment models were effec- tively delayed until the mid-1960s when field daø describing the natffe and rate of roadside encroachments became widely available (5, ó). Quickly these data were incorporated into an encroachment model used in this special case for evaluating the safety effects of luminaire supports (7).l-atrr models 18-10) have continued to rely in large part on this original data base. Unfortunately, most of the original daø were collected at a single, low-volume (from 2,000 to 6,000 vehicles per day) fteeway site. The freeway had two 12-ft lanes in each di¡ection and alignment was predominantly strâight. Most important, only median encroachments beyond a 3-ft stabilized shoulder were investigated. In this study no evidence was found that any of the available models had been

272 validated and it was concluded that, given the inappropriateness of the encroachment data, no available model could be recommended for use in analyzing the safety effecß of roadside haza¡ds on two-lane highways. As a result, an independent calibration was undertaken based primarily on accident rather than encroachment data. Fortunately, an appropriate data base developed in the earlier investigation of utility pole accidents was available(4).The data bæe was extensive, involving about 9,500 accidents over 2,500 mi of roadway in four sates; the majority of tl¡e 1,500 highway segments was twolane roads located in rural areas with a wide range in taffic volumes and, to a some.what !esæ.r e.xtent, pole offseæ r','ere included from the traveL lanes. The calibration and t€sting of this roadside encroachment model for two-lane, rural conditions is described in the remainder of this appendix. Referring to Equation I, the ûrst required element of the model is Ex(E),rhe expected annual number of encroachments per mile of highway. For encroach- ments on both sides of a two-lane roadway, the assumed model is Ex(E) = Ex(EXC) tt + pr(r > rfl 2 where Ex(EXC) is the expected annuâl number of lane encroachments per mile and Pr(Y ) r) is the probability that an errant vehicle, veering !o tl¡e left, will cross the adjacent lane of width, L, and encroach on the roadside. As used here, lane encroachment describes an out-of-control vehicle that travels onto an adjacent lane or shoulder, an occrrrrence assumed to be equally likely for left and right directions. The û¡st of the two components of the bracketed term in Equation 3 is for movement to the righÍ here a lane encroachment always results in a roadside encroachment. The second is for movement ûo the left in which some probability exists for safe recovery before reaching the roadside. The expected number of encroachmenæ, Ex(EXC), is assumed to be related only to trafûc volume as follows: Ex(EXC) = a(ADT)b () where ADr is the twodirectional average daily traffrc volume in vehicles per day and a and b are calibration constants. Reûnement of Equation 4 to reflect influences of roadway elements such as curvature and lane width was imprac- tical for this preliminary efforr The next element of Equation 1 to be modeled is pr(E^p), the conditional probability that, given an encroachment, its location is such that an impact with the roadside hazard is possible. Letting x represent the distance in feet along the roadway within which an encroachment, if continued sufficiently far, will result in a collision with the haza¡d. then (3)

Pr(E¡p) = l0F6-0 The denominafor of Eçation 5 is changed fo 5,280 when Ex(E) rcpresents encroachments on only a single roadside. The impact envelope used for computing X is shown schematically in Figure F-1. The path of the e¡rant vehicle is assumed to be straight and X is a Edge of Travel l¡n€s Zore l - Cojllslôn ón pilallel side of objæt Zonè 2 - Colìisl'on on neâr corn6r ol cbiæl Zone 3 - Collision on perpendicular s¡de otobiect FIGURE F-l Envelope of potential hazard based on trace of right, front corner of enøoaching vehicle.

274 function only of the angle of departure, ttre dimensions of the object, and the width of the colliding vehicle. For this investigation, each utility pole was assumed to have a square cross section with 8-in. sides; the departure angle, Q, was taken as 6.1 degrees for right-side departures and 11.5 degrees for tefçside departures (8). The width of the colliding vehicle, d, was raken as 6 ft. Using these parameters, the projected length along the roadway of potential hazard from a single utility pole is 63.4 ft for a right-side deparrure and 34.0 ft for a left-side departurè. Returning to Equation 1, the next element to be modeled, is Pr(C rpr), the conditional probability that, given an encroaehment in the potential impaet area, a collision between vehicle and object will occur. Again assuming that the path of Eavel is sraight, this probability reflects the likelihood that conrol of the vehicle will be regained before the vehicle reaches the object. Mathe- maticall¡ the expression is Pr(CrlE) = Pr(Y >-t) (6) where Pr(Y ) y) is the probability that the vehicle-more precisely its ourer front fender-will continue beyond a lateral distance of y from the lane boundary if its ravel is not impeded by a prior collision or overturn and if control is not regained. In this investigation, the following three, one-param- eter distributions of lateral travel were considered: Linear Pr(Y>y¡=(l-ylc) Pr(Y >il = 0 Exponential Pr(Y2y)=ec(Y) fory<c fory>c Sinusoidal ;;¿i;;' =r.ë&a] ror y s rsoic (7) (8) (e) Pr(Y2.))=0 fory>180/c where c is a calibrarion constant. The final two necessary components of the model are Pr(A¡lCr), Ihe conditional probability of an accident given a collision, and Pr(CAolA), the conditional probability of a casualty given an accident. Although the frrst of these has not. þen extensively add¡essed in the literature, Zegeer and Parker (4) have presented usable estimates until more substantive values become

275 available (fable F-1, accidents per collision). However, in view of ttre Mak and Mason frnding of approximately one unreported collision with utility poles for every two reported accidents (11), the Zegenr and Pa¡ker estimates (4) may eventually prove to be excessive. Also presented in Table F-l are recommendations by Glennon and Wilton for the fraction of accidents that result in a casualt¡ either a fatality or a nonfatal injury (8). These recommendations served as a guide for extrapolat- ing the Zegeer and Parker estmates of accidents per collision to other types of objects and a¡e suiøble for direct use with the roadside encroachment model. Roadside encroachmens in the vicinity of a utility pole are generated by near-side vehicles departing to the right and far-side vehicles departing to the left. For raffic distributed equally in both directions, ttre expected annual number of encroachments, Ex(Ep), in the impact zone of a utility pole is Ex(E) = a(ADT)b (0.25X63i5,280) + (0.25)(34/5,280) Pr(Y > L)l (10) where the bracketed term includes both near- and far-side vehicles, 0.25 represents the fraction of total encroachments directed toward a single side of the roadway and associated wittr t¡affrc moving in one direction, and I is the lane width in feet. The expected annual number of collisions with the utility pole, Ex(Cr), reflects not only ttre number of encroachments in the impact zone but also the lateral offset of the pole from the Eavel lanes. As eadier explained, increasing the offset reduces the number of collisions because of the greater chance that control can be regained before vehicles reach the pole. Exarnination of Figure F-l reveals that the offset to the near-most front fender of colliding vehicles is constånt for an impact at any location within Zonel but that it varies with the speciflc location of impact in Zones 2 and 3, Because "w" is small for utility poles, the offset for Zone 3 impacts can be assumed to occur at the midpoint location. Similar treatment can be given to Zone2 by dividing it into six, l-ft strips and using the midpoint offset of each strip. The eight segments thus considered for each direction of travel are described in Table F-2. The expected annual number of collisions with a pole can then be approximated as Ex(cr) = ffit,0.2Ð t;, x¡Pr(Y>y) + !.x¡Pr¡Y>y,¡1 (11) where the subscripts i and j refer to near-side and far-side encroachments, respectively. Assuming on âverage 0.9 accidents occur for each collision (4), the expected annual number of collisions with the utility pole, Ex(A), is

TABLE F-l Most Likely Consequences of Encounters with Roadside Hazards Type of Hazard Accidents per Collision CasualtyAccidents per Accident /8/Zegeer and Parker Extrapolated /4,/ 0.90 0.45 0.50 0.50 0.50 0.30 0.20 0.40 0.20 0.35 0.35 0.45 0.30 0.35 0.rs 0.15 4,25 0.35 0.45 0.60 0. l5 0. 15 0,25 0.35 0.45 0.60 0,45 0.45 0.45 0.60 0.50 0.35 0.40 0.50 0.50 0.35 0.30 Utility pole Trees Þ 6 irt.) R.igid signposts Steel (> 6 in.) Timber(>10 in.) Small Breakaway Light or signal pole Risid Breakaway Fixed object Nonclear zone Curb Gu¿rdrail Short(< 100 ft) Safety end Nonsafety erid Longþ I00 ft) Safety end Nonsalety end Fill slope 10: I ll: I 5:l 4:l 3:I 2:l otsleeper Cut slope 6:l 4:I 3:l 2:l 1,5: I 1:1 orsteeper Washout ditch Culvert (lateral or longitudinal) Raised drop inlet Bridge abutment or pier Roadway over bridge Structure Open gap between parallel Lrridges Bridge rail Smooth Parapet-type End Gore .abutüent Retainingwall or fence Fireplug 0.e0 0.50 0. r0 0.05 0.20 0.30 0.60 0.05 0.20 0.30 0.60 0.95 0.95 0.95 0.55 0.20 0"75 0,20 0.35 0.45 0.30 0.35 0.25 0.70 0.90 0.85 0.85 1.00 0.95 0.35 0.40 0.95 0.95 0.65 0.55 I L

277 TABLE F-2 Length Of and Offset To Utility Pole Near-Side Encroachments Far-Side Encroachments Zone (Figure F-1) Segment No. Hazard Length Offset(fÐ (f0 Hazard Length Offset(fÐ (f0 I 2 2 2 2 2 2 3 I 2 3 4 5 6 7 8 0.67 9.4t 9.41 9.4r 9.4t 9.41 9.41 6.27 0.67 s.02 5.02 5.02 5.02 5.02 5.02 3.29 yo y + 0.50 y + 1.49 y *2.48 y + 3.48 y * 4.41 y + 5.47 y * 6.30 y+ 12.00 y * 12.49 y + 13.47 y * 14.45 y+ 15.43 y + 16.41 y + r7.39 y + 18.21 ay is the perpendicular distance from the lace ofthe pole to the near side ofthe travel lanes. o$ Ex(A | = {Ðo li x¡ Pr¡Y > yí) + \ x¡ Pr(Y > v,)l (r2)uLt^Pt - n,467 -È-l J=r The data in Table F-3 were used to evaluate the three lateral tavel distrìbu- tion models, Equations 7-9, and to calibrate the unknown constants, a, b, and c, for each. Thì actual number of accidents per pole (AA), computed by TABLE F-3 Mean utility Pole Accident Frequency (Accidents per Mile per V.liut u Function of Traific Volume, Pole Offset, and Pole Detsity (4) Pole Density, Pole Offset (lt) r5.07.5 22.5 26.0ADT Low (16 Poles per Mile) 2,800 7,200 1 2,500 19,400 38,000 0.38 1.93 0.57 2.28 0.30 0.75 0.19 0.99 2.20 0.12 0.27 0.3 l 0.52 0.77 0.05 0. 16 0.32 0.r3 0.t2 0.03 0.08 0.10 0.t2 t).Õ / Medium (40 Poles Per Mile) 2,800 7,200 r 2,500 l 9,400 38,000 t.l6 t.46 l.98 2.34 2.57 0.42 0.71 0.96 1.20 2.02 0.22 0.35 0.59 0.73 1.46 0.1 1 0.15 0.2 r 0.1 I 0.67 0.09 0.r8 0.t2 0.21 High (63 Poles per Mile) 2,800 7,200 l 2,500 19,400 18,000 0.66 2.02 t.'7'l 3.00 2.99 0.60 r,00 0.94 1.70 2.13 0.38 0.62 0.63 0.81 1.05 0.24 0.3 3 0.38 1.15 0.09 0. 14 0.33 0.29

278 dividing each mean accident frequency in Table F-3 by the corresponding pole density, replaced the expected number, Ex(A), in Equation 12. One of the th¡ee lateral travel distributions was then selected and values of its constant, c, Feated paramerically. For each value of c, a new variable Z was computed for each of the 71 data points as follows: Z = 23467(AAp)/l .t x, pr(y > y,, + L x¡ pr(y > yr)l (13)Èl j=l ' such that z = a(ADr)b (14) Following logarithmic transformations, the consiants, a and b, in Equation 14 were determined using least-squares procedures. Results of the calibration process are given in Table F-4. For each type of lateral travel distribution, the best calibration is one that (a) yields a predicted mean accident frequency that equals the actual frequency, (å) produces the largest correlation coefficient in the calibration of the transformed Equation 14, that is, between ln(Z) and ln(ADT), and (c) most accurately predics accident frequency as measured by the correlation coefûcient between the actual and predicted accident frequencies. For each of the three dist¡ibutions, the first two criteria are satisfied by approximately the same calibration (Thble F-4). This is considered to be the calibration of choice: at this point, the ttri¡d çriterion-maximum correlation between actual and predicted accident fre- quencies-is not unduly compromised. The three models, so calibrated, are given in Täble F-5 and shown in Figures F-2 and F-3. As measured by the correlation coefficients, the three lateral travel distribu- tions offer approximately the same accuracy. The calculated percentages of right-side departures ue73,61, and 57 for the best exponential, linear, and sinusoidal models, respectively. Each of these compares favorably with prior research showing that 60 to 70 percent, of roadside encroachments result from right-side departures (11-12). A further accuracy comparison, using nine test sites specifically reserved for that purpose by Zegenr and parker (4), revealed little significant difference among the three disributions but suggested that any of the three models could be used, at least. as reliably as the best Zegeer-Parker model, to estimate utility pole accidents (Table F-6). The exponential model is recommended bott¡ for iß ease of use and for the intuitive appeal of its greater sensitivity to lateral offset in regions near the travel lanes (Figures F-2 and F-3). A primary advantage of the roadside encroachment. model over the regres- sion type module as calibrated by Tngeer and Parker (4) is its potential

TABLE F-4 Calibration of Roadside Encroachment Model Predicted Mean Lateral Accident Frequencyó Distanceo (Utility Pole Accidents(fÐ per Mile per Year) Correlation Coefrcient ln(Z) versus Actual versus Predicted ln(ADT) Accident Frequency Linear Distribution 20 22 24 26 28 30, 35 40 50 70 0.830 1.381 2.131 t.t67 0.93'7 0.832 0.734 0.688 0.654 0.636 0.7 t6 0.319 0.206 0.502 0.667 0.727 0.7 t2 0.672 0.62t 0.583 0.878 0.871 0.869 0.872 0.866 0.858 0.831 0.801 0.748 0.685 Exponential Distribution 20 22 1À 26 28" 30 35 40 50 70 0.989 0.923 0.874 0.836 0.807 0.7 83 0.142 0.716 0.685 0.6 57 0.709 0.733 0.'t44 0.7 47 0;745 0.739 0.720 0.701 0.668 0.628 0.882 0.88 r 0.879 0.876 0.8'72 0.868 0.855 0.842 0.815 0.768 Sinusoidal Distribution 20 22 24 26 28 30, 35 40 50 10 1.949 2.7 63 l.5ló 1.133 0.955 0.860 0.'745 0.693 0.653 0.633 0.2'70 0. l8l 0.314 0.549 0.668 0.718 0.715 0.67 6 0.6 l9 0.572 0.875 0.868 0.873 0.871 0.865 0.857 0.8 30 0.799 0.73't 0.65 7 "Lateraldistancebeyondwhich l0percentoferrantvehiclescontinueasdeterminedforeachdistribution by its calibration constant, c. åActual mean accident frequency is 0.8 utility pole accidents per mile peryear. cRecommended calibration.

280 TABLE F-5 Summary of Calibrated Models Laæral Travel Distribution Model Excursion Rate (Equation 4) Lateral Travel Distribution (Equations 7-9) Linear Exponential Sinusoidalo 0.05409 (ADT¡o'stoz 0.07285 (AD1¡o'sr:s 0.05117 (ADT)o's7s6 Pr(Y > y) = Pr(Y > y) = P(Y > y) (1t133.33) e4.08224(y) 1 + cos [4.7710 O)] -2 alhe a¡gom"trt of the cosine function in rhe sinusoidal distribution is expressed in degrees. TABLE F-6 Comparison of Predicted and Actual Utility Pole Accidents Accident Frequency (Accidents per Mile per Year) Pole Density (Poles per ADT Mile) Zegeer- Actual Pa¡ke¡ Predicted Pole Ibst Offset Site (ft) Exponential Linear Sinusoidal J 7 2 8 4 6 I 9 5 20.7 3,700 25.014.0 2,150 2t.724.6 14,500 41.5 r2.2 1,875 48.622.7 4,610 36.318.1 4,900 37 .94.8 41,000 19.0t4.9 37, 100 2t.61.9 11,000 76.7 0.04 0.160.07 0.16 0. 15 0.380.23 0.380.25 0.230.57 0.28L00 1.80 1.48 0.833.06 2.s9 r.47 0.12 0.120.15 0.160.26 0.230.36 0.380.16 0.150.29 0.29 | .28 t.340.72 0.'742.74 2.82 0.89 0.85 0.1 I 0.r2 0.30 0.29 0. 15 0.24 t.28 0.60 3.00 r .01Sum of squared residuals NorE: Actual data and predictions using Zegeer-Parker model l4l. applicability to hazards other than utility poles. For general application, the expected annual number of accidents, Ex(A), involving any roadside hazard, is given by Ex( A, | - 0'07285(ADDo't"t o r¡o^1c n) l1,xpa'orzxtiu^tr.hr - 2L,IZO + \r\ea'o822ati1 (15) where all va¡iables are as previously deûned- An analysis, similar to that used for utility poles and summarized in Figure F-l and Table F-2, is necessary to determine the .r's and y's for the particular hazard under review. The proba- bility of an accident given a collision, Pr(A¡lC ¡), and, if desi¡ed, the proba- bility of a casualty accident given an accident, Pr(CA^lAr), can be t¿ken from Table F-1.

28t É UJ e.a fL l¡lJ5 u¡ fL ant- u¿ =To ocÉroz t¡J 0 LATERAL DISTANCE (fI) FIGURE F-2 Comparison of lateral travel disribution models on the basis of the frequency of roadside encroachments. As a preliminary test of the general applicability of Equation 15 and Table F-1, predictions of roadside accidents for extended highway segments typical of those evaluated by Graham and Harwood (13) were compared wittr actual average conditions (Iable F-7). For accidents at all levels of severit¡ predic- tions exceed actual rates by up to 160 percent, a margin believed to be only slightly magnified by the inclusion of multiple-vehicle, run-off-road accidents in the predicted quantities. For casualty accidents, the level of overprediction

282 LATERAL DISTANCE (fI) FIGURE F-3 Comparison of lateral travel dist¡ibution models on the basis of the rate of roadside encroachments. shrank to a maximum of about 85 percent. Given the diversity of these two data sources and the numerous assumptions required in applying the encroachment model to the Graham-l{arwood conditions û3), agreement between predicted and actual accident rates is considered reasonably good. One frnal comparison is in order, namely, befween the encroachment model developed here and the models advanced earlier by other researchers. Prior models are similar only in the effect of the extent of lateral movement on froJ E l!JIfi6 z o JJ =4É IJJ fL atF w2 =!o o tc2o t¡J Ø!c =lr¡J9 ! lJ.¡ t z o =J =¡2 l¡¡À UIt-z IJJ =1Io o tÉofro

TABLE Fl Comparison of Predicted and Actual Roadside Accidents All Accident Severities Casualty Accidents Clear Zone Policyo Predicted ROR Actual SVRORAccident AccidentperMVM perMVM Predicted ROR Actual SVRORAccident AccidentperMVM perMVM Nonclear zone l.l7 4:1 clearzone 0.84 6:l clearzone 0.65 0.68 0.40 0.25 0.54 0.33 0.18 0.32 0. l8 0.10 Norrs: Actual data derived irom Graham and Harwood ûl). ROR : run-oflroad, MVM : million vehicles miles, and SVROR : single vehicle run-oÊroad. aAs defined by Graham and Harwood (13). ADT=2,000VPD - Exponenllal ---- AASHTO Barrler Guldc i/9,/ -- Glennon ând wltton (wlde) la., -\ ,. . ¡ .. .. Glennon and Wllton (Narrow) /8/\- -. -.- McFarland and Rolllns Ito,, \\\ E uJ tr uJÀ l¡lJ =tr lJl o- oF t! Et o otro uJ tc l! cÉ ul o- l¡JJ =É lrJ ô.(t F !J EIo o CC oz uJ 510152025 LATERAL DISTANCE (ft) ADT=6,000VPD - Expôñentlål ---- AASHTO Bar¡ler Gulde 19./ -- Glennon and Wllton (Wdo) l8l . ... . ... Glennon and W¡lton(Narow) l8l McFarland and Rolllns //r/ 051015202530 LATERAL DISTANCE (ft) FIGURE F4 Comparison of roadside encroachment models on the basis of the frequency of roadside encroachments.

284 attt! 10 J =l!J98I ¡¡l zo =oJ =É,Ia an ,t! =t =o o5o lrr ttqJ =l¡JJo =3ur o =J =2É, UJ fL Ø 2i t¡t =To oEOoz l¡J 10 15 20 LATERAL OTSTANCE (tr) ADT = 6,399 YPp a----- "'..\- \- '.. -l -'. \- -r "..\\-\ - Exponentlal -- -- AASHTO Barrler Gulde /9i - - ctennon and Wlhon (Wde) l8l .. .. .... Glennon and Wllton (Narrow) l8.i -.-.- McFarland and Rolllns lt0,/ s\ 15 20 25 LATERAL DTSTANCE (tr) FIGURE F-5 Comparison of roadside encroachment models on the basis of the rate of roadside encroachments. encroachment rate, a reflection of thei¡ sensitivity to the early study and analysis of freeway-median encroachments (Figures F-4 and F-5) (5-7). The exponential model developed here-as well the linear and sinusoidal mod- els--demonstrates much greater sensitivity of encroachment rate úo the extent of lateral travel. All models differ radically in their estimates of the general level of encroachments and of the influence of traffic volume on encroach- ment level. At a volume of 2,000 vehicles per day, the model recommended in the AASHTO Guide for Selecting, Locating, and Designing Trffic Barriers

285 (9) yields particularly extreme estimations. At all úaffic volumes, the McFa¡land-Rollins model yields the lowest estimates of both the frequency and rate of roadside encroachments (10). Although it appears clear that much remains to be leamed about roadside encroachment models, the study concluded that the model described herein was likely to be superior to other alternatives and that it could be used to provide approximate estimates of the reduction in accidents expected from roadside improvements. REFERENCES L. C. Zegeer, J. Hummer, D. Reinfurt, L. Herl and V/' Hunter. Safety Effects of Cross-Sectíon Design for Two-Lanc Roadç-Volunæs I and 11' Report FHWA- RD-S7/00S and 009. FHWA, U.S. Department of Transporøtion, Dec. 1986. 2. D. E. Cleveland and R. Kitamura. "Macroscopic Modeling of Two-Lane Rural Roadside Accidents." In Transportatíon Research Record ó8l. TRB' National Resea¡ch Council, Washington, D.C., 1978, pp.53-62. 3. P. H. Wright and K. K. Mak. "single Vehicle Accident Relationships"' Traffic Engíneering, Vol. 46, No. t, Jan. 197 6, pp. l6-2L. 4. C. V. Zegeær and M. R. Parke¡ !r. cost-Effectiveness of counfermeasures for Utitity Pole Accidets. Goodell-Grivas, Inc., Southfiel{ Mich.' Jan. 1983' 5. J. W. Hutchinson and T. W. Kennedy. Medians of Divided Highways-4requettcy and Nature of Vehicle Encroachments. Bulletin 487. Engineering Experiment Staúon, University of Illinois, Urbana, 1966. 6. J. V/. Huæhinson and T. W. Kennedy. "Safety Considerations in Median Design." ln Highway Resea¡ch Record 162. HRB, National Research Council, Washington, D.C.,1967, pp.L-29. 7. T. C.Edwardsetal. NCIIRP ReportTT:Developmertof DesignCriteriaforSøfer Lumínaire Supports. TRB, National Resea¡ch Council, Washingtor¡ D.C., 1969. 8. J. C. Glennon and C, J. Wilton. Effectíveness of Roadside Safety Improvements: Vol. I, A Methodology for Determíning the Safety Effectiveness of Improvemenfs on AII Classes of Highways. Report FHWA-RD-75/23' Midwest Research Institute, Kansas City, Mo., Nov. 1974. 9. Guidefor Selecting, Locating, and DesigningTrafic Barriers. American Associa- tion of State Highway and Transportation Officials, Washington, D.C., 1977. 10. W. F. McFarland and J, B. Rollins. Cost Effectiveness Techniques for Highway Safety: Resource Allocation. Final Report. Report FHWA-RD-84/011. Texas Transportation Institute, College Station, June 1985. il. K. K. Mak and R. L. Mason. Accidet Analysis-4realtaway and Nonbreakaway Poles Including Sign and Light Standards Along Highways: Vol. II, Teclnical Report. Report DOT-HS-805-605. Southwest Research Institute, San Antonio, Tex., Aug. 1980. 12. K. Perchonok et al. Hazardous Effects of Highway Features and Roadside Objects: Vol. II, Findings. Report FHWA-RD-781202. Calspan Field Services, fnc., Buffalo, N.Y., Sept. 1978. 13. J. L. Graham and D. V/. Ha¡wood. NCHRP Report 247: Effectiveness of Clear Recovery A¡eas. TRB, National Research Council, Washington, D.C., 1982.

Appendix G Physical and Operational Features Affecting Safety at Intersections Although simple quantitative relationships to predict the effects of specific intersection improvements are generally not available, nonetheless, a substan- 'ial body of information exis's that designers use in remedying deficiencies ai intersections. This appendix contains a summary of the effects of physical and operational features on safety at intersections. NT.JMBER OF LEGS The hazard of at-grade intersections increases as the number of legs (approaches) increases. Thus, three-legged, T-type intersections are less haz- ardous than four-legged, cross-type intersections which, in tum, are less hazædous than ûve-legged intersections, and so forth. The increasing hazard results from a number of factors, including a large increase in the number of poæntial conflict points, an increase in ttre number of options underlying d¡iver choice, diffrculty in proper signing and pavement marking, including the delineation of appropriate paths of travel, and impaired surface drainage because of increased surface area within the intersecúon. ANGLE OF INTERSECTION The prefened angle between intersecting legs at an intersection is 90 degrees. At angles deviating significantly from this standard, drivers of crossing vehi- cles are unable to detect úre presence or judge the speed of vehicles approach- ing on conflicting paths, and the time (and area) of potential conflict is increased. Also it becomes increasingly difficult to make tuming maneuvers, in part because of larger vehicle off-uacking. NUMBER OF THROUGH LANES The intersection accident rate, expressed as fhe number of accidents per million entering vehicles, is typically greater when tle approaching roadways have a larger number of lanes. Less clea¡ is whether this phenomenon is 286

287 directly the result of some aspect related to number of lanes, such as the number of potential conflict points, or whether it is indicative of indi¡ect effects, such as larger trafûc volumes. In any event, selection of ttre number of through lanes on intersection approaches is determined predominantly by capacity rather than by safety considerations. SIGHT DISTANCE In the context of at-grade intersections, sight distance refers to the view of a driver about to enter the intersection of traffrc approaching on the cross road or street. As the view is enlarged by the removal of obstructions in the line of sight, the driver is better able to judge the hazard of entering the intersection, and, as a result, safety is enhanced. Although greater corner sight distance is viewed as beneficial in all circumsüances, it is considerably less significant under traffic-signal control where decisions to enter the intersection are based primarily on the signal indication, and perception of threat from crossing traffrc is usually a secondary consideration. Improvements to intersection sight distance impinge primarily on angle collisions. ALIGNMENT From a safety søndpoint, intersecting roadways should be flat and straight. Cu¡vature in either the verúcal or horizontal plane that is so great as to impair sight distance will increase intersection hazañ. A small gradient does not appear to be harmful and may even be advanrageous in improving surface drainage. More substantial gradient becomes a liability as stopping distance on downward approaches increases, and conflict is intensified by reduced acceleration following a stop on upgrade approaches. Horizontal curvature on approaches to at-grade intersections is always harmful. Not only is it more difficult for drivers to discern the proper paths of Favel but their visual focus is directed along lines øngential to Íhese paths. Horizontal cu¡ves add further complexity to an already difficult driving environment. AUXILIARY LANES Auxiliary lanes reduce disruption to through naffic by accommodating special needs of turning vehicles--deceleration, acceleration, and waiting. Although auxiliary lanes are benefrcial in virtually every situation, the extent of their impact is dependent on the volume of turning movements, volume of possibly

288 conflicting movements, and approach speed- The primary impact of auxiliary lanes is on collisions between vehicles on the same approach, particularly the rear-end type. They are typica[y more effective in reducing hazard when accommodating left-tuming vehicles than right-turning ones, and a separate left+urn phase at signalized inærsections is often necessary for full benefits to be realized. One disadvantage of auxiliary lanes is increased pedestrian crossing fime as a result of the added roadway widttr. CHANNELIZATION The advantages of channelization with respect to safety have been flrmly established l2). Channelization can be used to delineate proper paths of Eavel, separate points of conflict, conEol angles of conflict, and provide pedesrian refuge. The esøblishment of left-turn lanes, a fundamental element in many channelization improvements, provides significant reductions in accident rates, particularly at unsignalized intersections 1.3). Channelization is required for the provision of protected left-turn lanes, is often necessary to eliminate excessive painted areas otherwise associated wittr turning roadways, and is usually considered when the intersection is large. Although curbed islands provide more positive control than paved ones, hazards associated with sriking raised islands suggest that painted, flush islands are prefened when approach speeds are high and fixed illumination is not, provided. TIRE.PAVEMENT FRICTION Of all road and street locations, intersecúons place the greatest demands on the tire-pavement interface. The most critical conditions exist when a large number of vehicles must stop, ttre approach speed is greât, and stopping must be accomplished quickly. The type of surfacing material, its prior wear or polishing by raffic, and the slope of the surface are important pavement attributes that affect tire-pavement füction. Grooving can be used to partly compensate for deficient pavement surfaces. Tire-pavement friction affects primæily multivehicle accidents that occur when the pavement is wet or icy. TURNING RADU Safety is degraded when vehicles must either encroach on adjacent lanes or slow excessively in order to execute turning maneuvers. From a geometrical

289 standpoint, right turns are more critical ttran left turns, and the degree of hazañ is related o both vehicle size and traffic volumes. FIXED LIGHTING At night, flxed lighting provides advance warning of the presence of at-grade intersections and allows the approaching driver to view objects outside the freld of headlight illumination. Although the eye's delay in adjusting to the changing level of background illumination, diffusion of light on dirty or damp windshield surfaces, and the roadside hazards of lighting stândards are undesirable influences, on balance the overall effect of overhead lighting is benefrcial. In specific situations, such as intense pedestrian activity, installa- tion of ûxed lighting can provide subsøntial gains. LANE AND SHOULDER WIDTHS Incremental changes in lane and shoulder widths have little effect on the accident pattern at intersections. The types of accidents affecæd by lane and shoulder conditions on open roadways, namely, opposite-direction and run- off-road accidents, ale not a significant problem at intersections. At the same time, the presence of a shoulder may well be beneficial not only because of the space it provides for collision-avoidance maneuvers but also because of sight- distance, obstacle-offset, and radius-of-turn implications. DRIVEWAYS Driveways in the vicinity of intersections place additional demands on the driver both in terms of the level of information that must be processed and the complexity of required decision-making. The hazard of nearby driveways, expressed primarily in their effect on multivehicle accidents, is less for entrances designed for easy and rapid use and for those located farther from the intersection. Safety is further enhanced if the site served by the driveway can properly accommodate parking and off-roadway circulation needs. ASSIGNMENT OF RTGHT.OF.WAY Common rules-of-the-road assign the right of intersection use to the vehicle that enters first or, in the event of two vehicles approaching simultaneously,

290 the vehicle on the right. V/hen sight distance is restricted or as traffic volumes increase, safety demands more positive control of traffic. Yield and sop signs and raffrc signals represenf progressively more definitive means for right-of- way assignment. The optimal technique from a safety stândpoint depends heavily on siæ-specific characteristics and can best be determined by using warrants presented in the Mantnl of Uniform Trffic Control Devices (1). Signalization, in particular, is not necessarily a safety panacea. Compared with other forms of conFol, signalization is often accompanied by fewer angle collisions but more rea¡-end collisions. The net effect of signalization depends strictly on site-specific cha¡acteristics: usually intersections having complex geometry and large ftafûc volumes respond best to traffic signal conûol. The first intersections encountered on approaches to urban areas and isolated rural intersections are difûcult to safely signalize because they are not expected by approaching drivers. APPROACH SPEED The hazard of at-grade intersections increases as approach speed increases. Time available for driver decision-making and response is less, braking requires longer distances, and, in the event of collision, the kinetic energy dissipated is much g¡eater. Furthermore, high approach speeds may indicate that approaching drivers do not expect the intersection and, hence, are ill- prepared to deal with the decisions it may demand. To the extent that they are capable of lowering approach speeds, traffic control devices such as advisory speed signs, flashing beacons, and rumble strips improve safety at intersec- tions. Higher speeds intensify the problem of dilemma zones at signalized inter- sections. These a¡e zones in which motorists approaching a trafflc signal that turns yellow tnd they are too close to stop before reaching the intersection and too far to get through the intersection before the light turns red. Special remedies that have been proposed for ttris problem include all-red intervals, longer yellow lights, and systems that detect the presence of a vehicle in is dilemma zone and extend the green phase (2). ON.STREET PARKING Intersections with adjacent on-street parking are more haza¡dous than inter- sections with no parking. Vehicles parked at the roadside restrict the sideward visibility necessary for safe operadon at intersections, and parking and

29r unparking maneuvers disrupt through-raffic movements. The degree of haz- ard is intensifled at locations of concentrated pedestrian activity. ONE.WAY OPERATION At least in the environment in which it is feasible, one-way operation is considerably safer than two-way operation. The advantage stems primarily from a reduction in the number of conflict points but extends to other factors such as improved signal timing, reduced headlight glare, and so forth. MISCELLANEOUS TRAFFIC CONTROL MEASURES Before-and-after studies demonstrate that significant safety benefits can some- times be realized at hazardous intersections by minor changes in trafûc control, including ¡ Improved delineation, o No-passing signs and markings, ¡ Flashing beacons, o Advance warning of intersection hazards, o Advance directional signing, ¡ P¡ohibition of left tums, o Enlarged signs and signal lens, . Additional signal faces, o Removal of roadside distractions, and . Adjustment of signal timing. The degree of improvement achieved by such measures is obviously dependent on the extent to which a specific deficiency can be ameliorated. REFERENCES 1. National Advisory Committee on Uniform Trafûc Connol Devices. Manual on UnifarmTrafftc Control Devices For Streets and Highways (as revised). FHWA, U.S. Department of Transportation, 1986. 2. Synthesis of Safety Research Related to Trffic ConÍrol and Roadwoy Elernents, Volume 1. Report FHWA-TS-82-232. Ofûce of Research and Development, FHWA, U.S. DeparÍnent of Transportation, Dec. 1984. 3. T. R. Neuman. NCHRP Report 279: Intersection Chanrwlization Design Guide. TRB, National Resea¡ch Council, Washington, D.C., 1985.

Appendix H Highway Accidents on the Federal-Aid System The characteristics and frequency of highway accidents point not only to the role that highway design plays in safety generally, but also to the potential for accident reduction through incremental improvement to design elements. Discussed in this appendix are accident classifications, the characteristics of accidents on federal-aid systems, and accident ¡ates. ACCIDENT CLASSIFICATION Highway safety is usually measured in terms of accidents or accident rates, and available data often include accident severity and accident type. Accidens generally are classifìed into three categories of severity: fatal, nonfatal injury and property-damage-only accidents. Based on the number of vehicles involved, accidents are first classified into two categories by type of acci- dent-single-vehicle and multivehicle-and then often further subdivided based on the nature of the accident. For example, single-vehicle accidents can be classiûed on the basis of the "ûrst harmful event," such as hitting a frxed object or a vehicle overturn, whereas multivehicle accidents can be classified on the basis of type of vehicle interaction such as head-on collision, side- swipe, rear end, angle, and so forth. In addition to accident severity and type, accident data a¡e sometimes stratiûed in a variety of other ways-by time of day, weather and surface conditions, and driver and vehicle characteristics--depending on t}re purpose. 292

293 For the analysis of a speciûc highway feature, researchers often focus on accident types believed to be most influenced by the feature in question. For example, skid-resistånce studies focus on wet-weather accidents, and roadside studies concentrâte on single-vehicle, run-off-road accidents. ACCIDENT TYPES ON FEDERAL.AID SYSTEMS Each year the Federal Highway Administration assembles and reports nation- wide estimates for fatal and nonfatal injury accidents occurring on federal-aid highway systems (Table H-1). On rural highways, nonfaral injury accidents oumumber fat¿l accidents by about 25 to l. On urban sEeets this ratio is much higher, roughly 70 to 1, demonsEating that accidents arc more severe on rural than on urban roads. On a nationwide basis, accident data are most complete and accurate for fatal accidents; thus different federal-aid highway systems can be best com- pared with respect to fatal accident cha¡acteristics. Summa¡ized in Table H-2 a¡e the types of fatal accidents that occur on non-Intersøte federai-aid sys- tems. Differences among systems reflect variations in location (urban and rural), function (arterial versus collectors), and design. Single-vehicle acci- dents comprise approximately one-half to two-thirds of all fat¿l accidents. Of these, ûxed-object accidents, particularly those involving trees and utility poles, account for about 50 to 60 percent of accidents on rural roads and about 40 percent on urban roads. These data suggest that design improvements that remove, relocate, or shield fixed obstacles would have signiflcant potential for improved safety by reducing the severity of accidents. Among u¡ban multivehicle accidents, about two-thirds a¡e rear-end or angle collisions. These accident types are particularly related to intersections and TABLE H-l Fatal and Nonfatal Injury Accidents on Federal-Aid Systems, 1984 (t) System Fatal Accidents Nonfatal Injury Accidents Ratio olNonlatal Injury Accidents to Fatal Accirlents Rural Interstate Primary Secondary Urban Interstate Primary Urban System 1,813 8,557 5,573 r,897 3,902 8,463 37,089 tg't,097 t5t,041 Q'l )¿.1 287,640 656,213 20.2 23.0 21 .t 51.3 7 3.7 11.5

294 TABLE H-2 Characteristics of Fatal Accidents on Non-Interstate Federal-Aid Systems Single-Vehicle Accidents Multivehicle Accidents System Percent Percent of Fixed Total Object Most Common Objects Percent Pedes- trians 4844 46 58 3t Percent Percent Percent Head-on Rear-End of Total Collision and Angle Rural Primary 45 Secondary 60 Urban Primary 50 Arterials 55 Collectors 66 Tiees Guard¡ail Utility Poles Tiees Ditches utility Poles Utility Poles Tiees Guardrail Utility Poles Curbs Tiees Tiees Utility Poles Curbs 552T 56363747 Source: 1985 Fatal Accident Reporting System Data (as summarized by L. Griffin). turning movements, indicating that intersection improvements and access controls might also have signiûcant potential for improved safety. ACCIDENT RATES Accident rates are most often calculated with respect to vehicle miles of travel ryNfD, recognizing that VMT is the most readily available measure of exposure to accident occl¡lrences. Such rates are also a convenient means of comparing the hazard associated with different highway systems (table H-3). From the stândpoint of allocadng resources for improving safety on a limited number of highway miles, however, highway agencies seek reductions in accident rates only as a step toward achieving the greatest reduction in accidents per dollar invested. Thus, the potential for enhancing safety on a given highway segment is measured in terms of accidents per year (or other unit of time), rather than accidents per VMT. Although urban Interst¿tes have the lowest fatal accident rates in terms of VMT, they have by far the highest rate in terms of fatal accidents per year per mile because of their high traffic volumes. Urban primaries have tle next highest rate in terms of fatâl âccidents per year per mile (Tables H-3 and H-4).

295 TABLE H-3 Fatal Accidents and Fatal Accident Rates on Federal-Aid Highways [Fatal Accidents per 100 Million Vehicle Miles (MVM)] (1984) (1) System Fatal FatalAccidents 100 MVM Accidents per Year per Year per 100 MVM Rural Interstate Primary Secondary Urban Interstate Primary 3,902 Urban System 8,463 I,837 8,557 5,573 t,Òv I 1,485 2,767 1,5 l6 2,036 2,300 3,744 r.24 3.09 3.68 0.93 t.70 2.26 TABLE H-4 Fatal Accidents and Fatal Accident Rates on Federal-Aid Highways (Fatal Accidents per Year per 100 System MileÐ (198a) û) Fatal Fatal Accidents Accidents Highway per Year System per Year Miles per 100 mi Rural Intefstate Primary Secondary Urban Interstate Primary Urban System 8,463 1,837 8,5s7 5,573 1,897 3,902 32,67 6 5.62224,868 3.81397,'t96 1.40 10,615 17.8731,859 12.25140,492 6.02 Rural primaries and secondary roads have only about one-ûfth to one-tenth as many fatal accidents per year per mile as urban Interstates. Consequently, although rural primary and secondary roads are substantially more hazardous than rural Interstates from the perspective of accidents per vehicle mile traveled, their relatively low potential for accident reduction per year limis the level of investment that can be justified for safety reasons alone. REFERENCE 1. Highway Safery Performance-L9\4, Faral an¿ Iniury Accilent Rates on Public Roads in the United Srø¿s. FHWA, U.S. Departrnent of Transportation, Jan. 1986.

Appendix I Initial Cost to Flatten Highway Curves The development of models for estimating the cost of flattening highway curves is described in this appendix. using the engineering approach, the cost models are based on estimates of hypothetical quantities of typical pay items and their related unit construction costs. unit construction costs were based on dat¿ from the rüashington state Departrnent of Transportation for RRR projects and are cônsidered represenüative of 1983 conditions. In developing "most reasonable" estimates, an attempt was made to incorporate possible economies of scale that can result in lower unit costs as construction quantities increase. However, costs are estimated as though curve reconstructiãn is not part of a larger RRR project. Pay items and unit construction costs are given in Table I-1, and the typical cross section deflning assumed conditions ii shown in Figure I-1. construction quanúties for flattening horizontal curves and a brief explana- tion of the assumptions necessary for making the estimates are given inTable I-2. These quantities were used with rhe unit cost daø in Tlrble I-1 to estimate FIGURE I-1 Typical cross section used for developing cost models. 296 Cenlerline

297 TABLE I-1 Pay ltems and Unit Construction Costs Item Unit Cost ($) Simpliñcation lor Vertical Curve Cost Model Other Costs Earthwork Cost/CY : 10-0.00025 CY for CY < 14000 Cost/CY : a9,000/CY+ 3 for CY > 14000Surfacing Cost/T: 30.5-0.00027T S27lT Base (CSTC) Cost/T : 10.7-0.00014T $9/T ¡¡¡einqqe 15 percent ofåarthwork' surfacing, and base costs Clearingand Cost/A : 2,500 grubbing Seed,fertilize, Cost/A:975-11.154 $700/A and mulchGuardrail Cost/LF : 22-0.001 LF Pavement striping Cost/LF : 0.06 $400 ñxed cost Wire lence Cost/LF : 2 Tiafficcontrol, Cost/PL:500 $10,000signing lJif"".il","' curves; $20,000 fixed cost for vertical curves Remove fence Cost/LF : I Removepavement Cost/SY : 5-0.00014 SY $3/SY Remove guardrail Cost/LF : 2 Remove*miscel- $1,000/A laneous Miscellaneous 32 Percent oltotal construction and removal costs Nors: A : acres, CY : cubic yards, LF : lineal feet, PL : project length in 100 ft.' SY : square yards, T : tons. totâl construction costs for a variety of typical cttrve projects having cenftal angles of 10 to 50 degrees, original curvature of up to 48 degrees, and improved or new curvature down to 2 degrees. The following model approxi mâtes results of the detailed computâtions: c = 18,395(Oo'eo2 (Dùa @zlDt)b (1)

298 where C = construction cost in 1983 dollars,Dr = original curvature in degrees, D2 = new curvature in degrees, .I = central angle in degrees, a = 4.0944 - 0.405 (00'1ola, andå = -0.0758 (O¡o'e<8. For the wide range of realistic conditions investigated, the maximum error in the approximation of Equation I was found to be + 10 percent. Construction quantities for lengthening crest vertical curves and a brief explanation of the assumptions necessary for making the estimates are given in Table I-3. Excavation quantities were estimated for a variety of typical TABLE I-2 Construction Quantities for Flattening Horizontal Curves Item Quantity Remarks Earthwork CY : Lz (4 + 0.5Dr) Surfacing Base Clearing and grubbing Pavement striping Wire lence Remove lence Remove surfacing Remove guardrail T:0;7Lz T: 1.8L2 A : (0.011 Lr, right-oÊ way acres) LF : 3Lz LF : 2Lz LF :ZLt SY:4L, LF : (0.2 + 0.0lDr) Lr Cross section in Figure I-1, balanced earthwork; typical centerline excavation ol 5 ft at Dr : 9 degrees; adjustment to reflect greater grading with poorer initial condition 4-in. pavement, 2-in. paved shoulde¡ 8-in. pavement, lO-in. shoulder plus wedge Area added to new inside right- oÊway line but limited to 25 ft each side of new centerline 15 ft each side of new alignment Minimum ol l0 percent of new length with increases lor larger D, Edge lines, centerlines, and non- passing zones Both sides Both sides 36 ft ofsurlacing Slightly greater removal than installation 40 ft, shaping, cleanup, seeding, and so forth Seed, fertilize, and mulch A : min/0.0007 L2Gua¡drail LF : (0.1 + 0.01D?) L? Removal-miscellaneous A : 0.0009Lr Nor¡: A = acres, CY = cubic yards, D, : original degree ofcurvature, D, : new degree of curvature,Lr: feetalongoldalignmentbetweenpointsolbeginninganðendolnewcurve, Lz : feet from beginning to ending olnew curve, LF : linèal feet, SY : square yards,T : tons.

299 TABLE I-3 Construction Quantities for Lengthening Crest Vertical Curves Item Quantity Remarks Earthwork Equation 2 Cross section in Figure I-l' all excavation; grade and ground elevations the same at beginning and end of both cufves Surfacing T : 0.7 Lz 4-in pavement, 2-in' paved shoulder Base T : 1.8 Lz 8-in. pavement, 10-in. shoulder plus wedge Clearing and grubbing A : 0.0014 Lz 30 ft each side Seed, fertilize, and mulch A : 0.0011 Lz 25 lt each sideGuardrail None Pavement striping LF : 3 Lz Edge lines,. centerlines, and nonpassrng zones Wire fence LF : 2 Lz Both sides Remove fence LF : 2 Lt Both sides Remove surfacing None Removal - miscellaneous None NorE: A : acres, Lz : leet from beginning to end ofnew curve, LF : lineal feet, T : tons. improvement projects, including those having grade differences of up to 14 percent, initial curve lengths as smâll as 200 ft, and new curve lengths as large ãs 1,800 ft. The estimates ranged from a low of 452 yd3 to a high of 98,923 yd3. The following model was found to reproduce the computations with an error no greater than + 10 Percent: CY=aØt-Lb Q) where CY = cubic yards of excavation,Lr = original vertical curve length in feet, L2 = new or improved vertical curve length in feet,G = absolute value of the percent difference in grades, a = 0.10392 (G)-l'ffils, andb = 1.4385 (G¡o'tot++. The simplifìed construction cost model for crest curve projects is summa¡ized as C = i.52 (CÐ (UC) + 73.62 L, + 26,928 (3)

300 where C is the construction cost in 1983 dolla¡s and UC is the unit cost of excavation as given in Table I-1. Acquisition of additional right-of-way is an anticipated consequence of horizontal curveprojeets and, deponding on initial right-of-way width, may be required on vertical curve projects. as well. However, the extreme variation in right-of-way costs makes ttrem impossible to model with any reliability; thus, thoy havebeen ontitted from Equations 1 and 3.

Appendix J Relationship Between Cost per Accident Eliminated and Benefit-Cost Ratio Approaches IllusUated in this appendix is the relationship between cost-effectiveness analyses for a hypothetical horizontal curve improvement using cost per accident eliminated and rhe benefit-cost ratio approach. Assumed conditions for the example are r 1,000 ADT, ¡ 2Q-deg¡ee central angle, o 35 mph design speed before improvement, . 55 mph design speed after improvement, . 0.20 combined side friction and superelevation factor, . 7 percent discount rate, . 3O-year project life, and . $30,000 value imputed to each accident eliminated. The imputed value for each accident eliminated is required only for the benefit-cost approach. As noæd in Chapter 5, cost per accident eliminated was selected as the principal measure of cost-effectiveness (instead of the benefit- cost ratio) to avoid the arbitrary imputation of dollar values to accidents eliminated. To determine cost per accident eliminated

302 . Estimate øccidents eliminated-4.10/year, based on assumed conditions before and after improvement and the accident relationship for horizontal curves presented in Appendix E; o Estimate added cost--$111,000, based on assumed conditions and the cost relationship presented in Appendix I; o Annualize the added cosl--$8,950lyear, bæed on the assumed discount rate and project life; and c Calculate cost per accident eliminated-$8g,500, calculated as 8,950/0.10. The preceding calculation does not account for the highway user travel time and operating cost savings associated wittr flattening horizontal curves. These savings can be taken into account within the cost per accident eliminated framework by subracting ttrem from the cost required to implement the improvement: o Estimate user savings--$6,150/year, based on the methodology pre- senfed in AASHTO's Manual on (Jser Benefit Analysß of Highway and Bus- Transit Improvements (1 );t calculate net (implementation minus user) cosr-$2,800/year, calculated as 8,950 - 6,150; and o Calculate net cost per accident eliminated-$28,000, calculated as 2,800i0.10. For the same horizontal curve, the beneflrcost ratio is calculated as follows: o Es timat e ac ciden t s eliminate d-O.10 /y ear, as previously; c Express safety benefits in dollar terms--$3,000/year, based on the imputed value of $30,000 per accident eliminated; o Estimate highway user tavel tim¿ and operating cost savings__$6,150, as previously; c Calculate total benefits-$9,150/year, calculated as 3,000 + 6,150; . Calculate the annualized added cos¡-$8,950, as previously; and . Calculate benefit-cost ratio-I.)2, calculated as 9,150/g,950. As shown by this example, when net cost per accident eliminated is close to the value imputed to each accident eliminated--$28,000 versus g30,000 in this case-the benefit-cost ratio is close to 1.00. REFERENCE 1. Manual on user Beræfit Analysis of Highway and Bus-Transit Improvemcnts. American Association of state Highway and rransportation ofñcials, washingto4 D.C., t977.

Appendix K Effects of Lane and Shoulder \Midths on Travel Time Narrow lanes and shoulders on two-lane roads cause motorists to drive closer to vehicles in the opposing lane. They must compensate for driving closer to opposing rafûc by slowing down and allowing larger headways between nèttictes in the same lane. Thus, drivers experience more delay and lower speeds on roads with narrow lanes and shoulders than on roads with wider lanes and shoulders. A procedure for estimating the effects of lane and shoulder width improve- rent" on travel úme is presented in this appendix. This procedure is based on methodology presented inthe Highway Capaciry Manus.l (1) and accounts for the combined effects of lane and shoulder width and traffic volumes' These combined effects can be important because the effects of narrow lanes and shoulders on speed are exacerbated when ftafflc volumes ale gleater. PROCEDURE For two-lane rural highways, the effects of lane and shoulder width and average daily naffic (ADT) on travel time can be estimated as VHT|VMT = (1/58) exp [0.01 ADT|(SFD f*)] (1) In ttris equation 303

3M . VHTIVMT is hours per vehicle mile-the inverse of speed; . exp is the exponential operator; o ADT is annual average daily traffic; o ,SFD is the maximum I¡vel of Service D flow rate two-way (in vehicles per hour) for the facility under consideration, assumingl2-ft lanes and 6-ft shoulders. The value of SFD depends on terrain, percent no-passing, vehicle mix, and directional split of trafûc. If information on these various factors is not available, a default value of 819 can be used for SFD; and o fi is a width adjustment factor used to adjust capacity downward if lane widths are less than 12 ft or shoulder widttrs a¡e less than 6 ft. values of/as a function of lane and shoulder width a¡e given in Table K-1. These values can be approximated wittr the equation f. = min [1, 0.085 (LW + SWf2) - 0.275] (Z) where rl{¡ is one-way lane width (in feet) and sw is one-way shoulder width -1iin_feet). The assumptions used in developing ttre default value of 819 for,sFD in the preceding equation are as follows: e Rolling tenain; . No-passing zones, 40 percent; o Directional distibution for raffic is 6G-40; r Vehicle mix is 14 percent trucks, 4 percent recreational vehicles (RVs), and 1 percent buses; o Design speed is 60 mph or greater; and . Speed limit is 55 mph. Table K-2 provides altemative values of SFD, which can be used if specific information on terrain, percent no-passing, di¡ectional distribution, or vehicle mix is available for the highway segment under consideration. TABLE K-l Lane and Shoulder Width Adjustment Facrors Usable Shoulder Width ( ft) 12-1T. Lanes a I ll-fr Lanes l0-fr Lanes 9-lr Lanes 6 4 1 0 l,00 0.92 0.8 r 0.70 0.93 0.85 0.7 5 0.65 0.84 0.77 0.68 0.58 0-70 0.6s 0.57 0.49 Souncr: Highway Capacity Manual, Table 8-5 (f

305 The "free flow" speed (i.e., the speed at low trafûc volumes) of 58 mph is based on an assumed design speed of 60 mph or gleater and a speed limit of 55 mph. For lower design speeds, the Highway Cøpacity Msnunl (1) recom- mends that speeds be reduced by 4 mph for each iO-mph reduction in design speed below 60 mph. Application of this procedure to hypottretical RRR projects indicates that travel time savings can be an important consideration in making decisions about lane and shoulder widening, particularly on highways with more than 2000 ADT. However, as noted in Chapter 5, these savings will be parrþ offset by higher motor vehicie operaiing cosis at higher speeds. A.iso, higher specds can lead to increases in the severity of accidents. DERIVATION Equation 1 is based on methodology for estimating level of service and speed on two-lane rural highways in Chapter 8 of the Higlrway Capacity Manual (l). The 0.01 and 1/58 constants in the equations were derived using the Worksheet for General Terrain Segments, Highway Capacity Manual (1, pp'8-31)' fogether with the default assumptions presented here. The worksheet provides a basis for determining average speed as a function of hourly flow rate for a given highway segment. Specifically, it provides frve points on a speed-volume cgrve. Speeds at other traffic volumes can then be determined by interpolation. Because the manual provides speeds as a function of hourly flow rate, it is necessaly to assume a temporal distribution of traffic, in order to calculate travel time as a function of ADT, For this purpose, a representâtive rural weekday traffrc distribution was used (2, p.167). The relationship between travel time and ADT was investigated for three cases: 10-ft lanes with 2-ft shoulders, ll-ft lanes with 4-ft shoulders, and 12-ft lanes with 6-ft shoulders. The values of fi and the lævel of Service D maximum llow rates for the three cases are Level of Service D Maximum Flow Lara Widrh Shoulder Width f- 5s7 696 410 10 2 0,6811 4 0.85t2 6 1.00 For each of the three cases, travel time þer mile) was calculated for ADT ranging from 500 to 7900 (in steps of 200), and regression analysis was applied !o estimate coefficients for the equation

306 Ln(VHTIVMT)=a+b(ADI) The three values of "å" obtained were o 1.80 x 10-s for 10-ft lanes with 2-ft shoulders, . 1.50 x lg-s for ll-ft lanes with 4-ft shoulders, and o 1.28 x 10-s for 12-ft lanes with 6-ft shoulders. In all three cases, the value of "b" carlbe approximated as 0.01 divided by the Level of Service D maximum flow rate. The value of "c" was -4.06 in all ttnee cases. This is consistent with a free flow speed of 58 mph; that is, when ADT is zero, the inverse of speed is calculated as exp (-4.06) or 1/58. The alærnative values of sFD given in Table K-2 were calculated from the Highway Capacity Manual (Tables 8-1, 8-4, and 8-6). Tirble K-2 provides TABLE K-2 Level of Service D Maximum Flow Rates Percent No-Passing Zone 7 Percent Tiucks, 2 Percent RVs, and 0.5 Percent Buses Level 1,553 1,504 1,455Rolling 1,222 1,123 t,025Mountainous 8l I 699 629 14 Percent Tiucks, 4 Percent RVs, and I Pe¡cent Buses Level Rolling Mountainous 2l Percent Trucks, 6 Percent RVs, and 1.5 Percent Buses Level Rolling Mountainous (3) r0080604020 |,431 946 560 r,3 83 841 462 1,401 907 518 t,440 977 553 1,395 898 476 r,350 8i9 429 1,32'l 756 381 1,305 t,282725 677352 314 |,342 813 4r9 1,300 748 361 1,258 682 325 |,231 630 289 r,216 1 , 195603 564267 238 NorEs: The maximum flow rates are for l2-ft lanes with 6-lt shoulders. The width adjustment lactors given in Tâble K-l are applied to these flow rates if lanes are less than 12 ft or shoulders àre less than 6 ft. The maximum flow rates were calculated for an assumed 60-40 directional distribution during individual hoursoftheday. Fora 50-50 directionaldistribution, the flowratesare muhipliedby 1.057 Foi a 70-30 directional distribution, the fìow rates are multiplied by 0.943. values of SFD for three different vehicle mix distributions, corresponding to high, medium, and low shares for the ttrree types of heavy vehicles. The

307 information given in Table 8-6 of the manual can be used to develop more precise estimates of SFD for a given vehicle mix. REFERENCES t. Specîal Repart 209: Hìglway Capacity Manual. TRB, National Resea¡ch Council, Washington, D.C., 1985. 2. Allocation af Life Cycle Hígltway Pavement Cosfs. Report FHWA-RD-83/080. EUIIT^ lI C n^-^-.-^-. ^C r-^----r-.!-- l^oô¡ lrrrÃ, u.J. vEpau¡rçttt ur rlilttptiltrulL LyóJ.

Appendix L Alternative Lane and Shoulder Width Standards Used in System-Level Analyses Described in this appendix are the four sets of lane and shoulder width standards examined in the national system-level analyses presented in Chapter 5. These standards apply to two-lane rural highways only. The American Association of Søte Highway and Transportation Ofûcials' (AASIITO) new consruction st¿ndards (Table L-1) are the most stringent søndards analyzed in this study. For highways with design speeds of 60 mph and above, minimum lane widths a¡e 12 ft for arterials and 11 or 12 ft for collectors, depending on traffrc volumes. Under these standards shoulder widths for high-volume collectors and arterials are set at 8 and 10 ft, respec- tively. AASHTO resurfacing, restoration, and rehabiliøtion (RRR) søndards (Table L-2) are the most lenient examined; they never require more than ll-ft lane and 2-ft shoulder widths and are the only standards that do not account for traffrc volume. The 1978 FHWA proposed stândards (Table L-3) usually fall between AASI{TO new construction and AASHTO RRR standards in terms of strin- gency. Lane widths vary depending on traffic volumes, percent trucks, and design speed. Lane widths can be as low as 9-ft on low-volume, low speed "minor" roads and increase to 12 ft for highways with greater traffic volumes and higher speeds. Shoulder widths range from 2 to 4 ft. The modified 1978 FI{WA standards (Table L-4) were developed to improve the original 1978 FHWA proposed standards. The committee found 308

309 t2 t2 t2 t2 t2 t2 t2 t2 t2 t2 t2 T2 ll t2 t2 50 60 70 TABLE L-l AASHTO New Construction Standards for Lane and Shoulder widrh ø Width of Road Feature (ft) by Projected Design Traffic Volume Design Speed ADT ADT4OO DHV DHV DHV Under400 andOver 100-200 200-400 Over400 Arterials: Lanes Arterials: Shoulders All speeds Collectors: Lanes l0 l0 1l 1t 1l ll Collectors: Shoulders (Graded) All speeds Minimum widths(f0 Average Running Speeds (mph) Percent Tiucks Lane Shoulder l0 20 30 40 50 60 70 t0 r0 t0 r0 ll l1 T2 T2 t2 t2 t2 t2 1l ll l1 t2 ta t2 r0 l0 NorEs: Usable shoulders should be paved; graded shoulders need not be paved; 4 ft is minimum shoulder width ifa roadside barrier is used; and DHV is design hour volume, usually the traffic volume projected for the 30th highest hour ofthe design year. DHV is generally 10 to 15 percent ofADT TABLE L-2 AASHTO RRR Standards lor Lane and Shoulder Widfhs (2) 40 or less Over 40 Al1 15 or less Over 15 2 z 2 l0 l0 ll NoTE: The standards specily a range of widths lor all cases. Only the minimums are reported here. that the concept of strarification by average daily rafflc (ADT), truck com- position, and design speed was sound but that minor changes could reduce ãost per accident eliminated; this was verified by subsequent analyses. The modified standffds used in the national sysæm-level analyses differ from the original 197& FHWA ståndards: 1. Ali ranges vler€ shifted, 2. Shoulder widths for high-volume highways were increased by 2 ft'

310 3. Shoulder widths for highways in mountainous terrain were reduced by 1 fr, 4. Average running speed was substituted for design speed because the concept of design speed is difflcult to apply to an existing highway, and 5. The concept of minor roads was eliminated because it was not clearly defined in the original 1978 FHWA proposal. The flrst three changes were made to improve cost-effectiveness; the last two were made to clarify the standa¡ds for users. TABLE L-3 1978 FHWA RRR Proposed Standards for Lane and Shoulder widrhs 6) width (ft) 10 Percent or More Tiucks Less Than l0 Percent TiucksCurrenlTiaffic (ADT) Design Speed (mph) Lanes Shoulders Lanes Shoulders l -400 40 I -4,000 Over 4,000 50 or less Over 50 50 or less Over 50 Ali l0 l0 ttIt 12 t2 2 2 2 3 4 2 2 2 3 4 9a l0 l0 ll ll Nor¡s: The 1978 FHWA proposed standards were actually defined in terms of lane width and total roadway (lane plus shoulder) width. The standards are shown here in terms oflane and shoulder width so that they can be more easily compared with AASHTO standards. In the actual standard, I l -ft lanes and a l-ft shoulder are perrnitted where IO-ft lanes and 2-ft shoulders are specified in the table. a"Mino¡ roads" only; otherwise lO-ft [anes. TABLE L-4 Modified 1978 FHWA RRR Standards for Lane and Shoulder Widths width (fr) Design Year AveragcVolume Running(ADT) Speed (mph) l0 Percent or More Tiucks Less Than l0 Percent lucks Lanes Shoulders Lanes Shoulders l -750 751 -2,000 Over 2,000 Under 50 50 or more Under 50 50 or more All r0 l0 ll T2 t2 2 2 2 3 6 2 2 2 3 6 9 r0 10 ll l1 Nores: Shoulders may be I lt less lor highways in mountainous terrain. The standards were actually defined in terms oflane width and total roadway (lane plus shoulder) width. They aregiven here in terms oflane and shoulder width foreasiercomparison with AASHTO standards. Forthe purpose olsimplicity, weighted average design speed was used for average running speed in the analysis.

311 REFERENCES t. A Policy an Geometríc Desígn of Hþhways and Streets. Americ¿n Associ¿tion of Staæ llighway and Tbansportation Ofñcials, lVashingtor¡ D.C., 1984. 2, Geonetric Design Guide þr Reswføcing, Resto'.øtion and RehabítitsÍìon (RRH of Híghwøys and Streeß, American Association of State Highway and Transpo,rtation Ofûcials, Washingto¡r" D.C., 1977. 3. Desigrr Standards for Highways. Federal Regßter, \61. 43, No. 164, Aug. 23,1978.

Study Committee Biographical lnformation Pernn G. Korrxow, Co-Chairman, is a consulting engineer and Counselor to the President, American Trucking Associations. He received his bachelor's degree from Antioch College and his master's degree from the University of California. Mr. Koltnow has worked in transportation and traftc safety for both the public and private sectors during his career. He served as chairman of the TRB Executive Committee in 19'79, and has served on the Executive Council of the Institute for Transportation, American Public Works Associa- tion, and the Executive Committee, National Committee on Uniform Trafflc Laws and Ordinances. He was President of the Highway Users Federation for Safety and Mobility from 1974 to 1984. A registered professional engineer in California and Ohio, Mr. Kolt¡ow is a member of the American Society of Civil Engineers and a Fellow of the Institute of Tianspo¡tation Engineers. HsnsErr H. Rrcs¿R¡s oN, C o -C hairmnn, is Deputy Chancellor for En gineer- ing, and Director, Texas Engineering Experiment Station, Texas A&M Uni- versity System, where he also holds the positions of Dean of Engineering and Distinguished P¡ofessor of Engineering. Dr. Richardson eamed bachelor's and master's degrees from the Massachusetts Institute of Technology, where he also completed his doctor of science degree in mechanical engineering, graduating in 1950. Before joining Texas A&M Universit¡ Dr. Richardson was Professor and Associate Dean of Engineering at MIT, becoming Professor Emeritus in Mechanical Engineering in 1984. During this period he also served with the U.S. Army Ordnance Corps at the Ballistics Research Labora- 312

313 tory Aberdeen Proving Grounds in Maryland. Before joining the faculty of MIT, he was Chief Scientist, Office of the Secreøry of ttre U'S. Department of Transportation. To this background, he adds varied experience in consulting, listing the Engineering Societies Commission on Energy, Intemational Har- vester, Inc., Foster-Miller, Inc., and the Caterpillar Tractor Company among his many clients. Cunently, Dr. Richardson also serves on numerous commit- tees and panels, such as the Council, National Academy of'Engineering; the Committee on Engineering Education, National Academy of Engineering; the Governing Board of the National Research Council; and the Executive Com- mittee of the Transpcrtatjon Research Boa¡d. He is also a mennber of the National Science Foundation's Engineering Advisory Committee. Dr. Richardson is active in the work of the American Society of Mechanical Engineers and in that of ottrer scientific societies. He was a cowinner of the Moody Award in 1970, and a recipient of the Pi Tau Sigma Gold Medal for "Outstanding Mechanical Engineer Ten Years After Graduation" in 1963. He was also awarded ttre ASME Centennial Medallion in 1980 and the Rufus Oldenberg Medal, ASME, in Dynamical Systems and Cont¡ol in 1984. Dr. Richa¡dson is the auttror of many publications, contributing a number of books and innumerable major reporß and magazine articles to his fleld. Rov W. ANoERsoN, a registered professional engineer, is currently President of TranSafety, Inc. He received his bachelor's degree from Texas Tech University and his master's degfee from the University of California, Berkeley. Mr. Anderson has served as Director of the Offlce of Safety Studies, National Transportation Safety Board, and has worked as a civil engineer for the California Department of Transportation. He was a member of the National Cooperative Highway Research Program Panel on Evaluation of Traffic Controls for Súeet and Highway Work Zones from 1978 to 1980. A former member of the FIIWA Safety Review Task Force, Mr. Anderson is the recipient of the California Society of Professional Engineers' award for outsønding performance and a special service award from the National Transportâtion Safety Board. He is a former ståte director of the National Society of Professional Engineers, a member of the American Association for Automotive Medicine, and a member of the American Society of Civil Engineers. He also serves on the Safety Coordinating Committee. of the Institute of Transportation Engineers. LsoNano Ev¡Ns is a physicist and Principal Research Scientist in the Operat- ing Sciences Department of the General Motors Research Laboratories' A gtaduate of Queens University, Belfast" and Oxford Universit¡ Dr. Evans'

3t4 more that 70 technical publications cover such diverse subjecs as physics, matl¡ematics, traffic engineering, Eansportation energy, human factors, trauma analysis, and trafûc safety. His main professional interests focus on faffic safety research. He is a member of the Society of Automotive Engineers (SAE) and is past chairman of the SAE Human Factors Committee. Dr. Evans is also a member of the Human Factors Societ¡ having served as chairman of its Southeastern Michigan Chapter. He is a member of the American Associa- tion for Automotive Medicine, Sigma Xi, the American Association for the Advancement of Science, the Society for Risk Analysis, and the Editorial Advisory Boards of Accident Analysis and Prevention and Human Factors. JouN C. Gr-sNNoN is a transportation consulting engineer. He is a graduate of the University of Illinois with a bachelor's and a master's degree, and the University of Kansas with a DE degree. Dr. Glennon joined the California Division of Highways where he worked as an Assistant Highway Engineer and later as Assistant State Transit Planning Coordinator before joining the Texas Transportation Instituæ as a Research Engineer. Following his tenure at the Texas Transportation Institute, Dr. Glennon joined the Midwest Research Institute QvIRI) where he served first as Manager of the Traffrc Safety Center and later as Manager of the Design and Operation Program. He is a member of the TRB Task Force on Tort Liability and a former member of the TRB Committee on Vehicle Characteristics. He was a member and subcommittee chairman of the 'IRB Committee on Operational Effects of Geometrics from 1972to 1980, and is currently Chairman of the TRB Committee on Geometric Design. In addition, he has chaired the Special Advisory Panel úo Review FHWA Research Programs on Geometric Design, ttre ITE Committee on Criteria for Installation of Concrete Barrier Wall Versus Guardrail, and the ITE Committee to Review Standards for Urban A¡terial Streets. Dr. Glennon has also served on several FHWA advisory panels dealing with such topics as size and weight of heavy vehicles and highway barrier need indices. He received the TRB D. Grant Mickle award (wirh D. Harwood) for an outstand- ing paper on highway traffic and maintenance and the ITE Missouri Valley Section President's Award for his outstanding contribution to naffic engineer- ing. Dr. Glennon is a member of the National Society of Professional Engineers, National Association of Forensic Engineers, American Academy of Forensic Sciences, and the Insútute of Transportation Engineers.

315 Ezn¡ Heurn is an engineer and Professor, Depætment of Civil Engineering, University of Toronto. He received his bachelor's and master's degrees from Technion University in Israel and his Ph.D. from the University of California' He has served as a special consultant to the Midwest Reseafch Institute; the World Bank; the Department of Transport, Road, and Motor Vehicle Trafflc Safety; and Delæuw, Cather and Company. Dr. Hauer is a member of the Assoðiaúon of Professional Engineers of Ontario, Institute of Transportation Engineers, American Association for Automotive Medicine, Operations Reiearch Society of America, Canadian Operations Research Society, and the israeii Ässociation oi Transporiaiiorr Research. 'W. RoNnlo HuosoN is a Professor of Civil Engineering and Dewitt C. Greer Centennial hofessor of Transportation at the University of Texas, Austin. He received his bachelor's degree from Texas A&M University and his master's degee and Ph.D. ftom the University of Texas' Dr' Hudson served as Re'search Engineer for the Highway Research Boa¡d AASHO Road Test from 1958 to 1961. He also served as Research Engineer for the Center for Highway Research and as Design Research Engineer for the Texas Highway Depart- ment. He was an Assistant Project Engineer for the National Cooperative Highway Research Program from 1963 to 1964. Since 1979 he has been Chãrman of TRB's Pavement Management Section. He is a member of the TRB Committee on Monitoring, Evaluation, and Daø Storage and the Execu- tive committee, Highway Division, American Society of civil Engineers (ASCE). Dr. Hudson received the Highway Research Board's outstanding Èaper Á*arO in 1964 and the ASCE Texas Section's Outstanding Paper Award in iqoS. He is a member of the American Association for the Advancement of Science, National Society of Professional Engineers, American Concrete Institute, and the New York Academy of Sciences. Dr. Hudson has conducted research on pavement management and performance, improved and stabilized materials, and application of e"onotnics and ståtistical methods to engineering problems. fte ii itre author of Pøy ement Management Systems (McGraw-Hill' 1978). Jacr T. KassEL, an engineer, recentþ retired from the position of Project Development Division Chief for the California Department of Transportation' Mr. Kassel received his bachelor's degree from the university of california, Los Angeles, and has had a long, distinguished career at the california Deparunent of Transportation, having served as Project Design Engineer, Los

316 Angeles District; Disrict Traffic Engineer; Assistant Tlaffic Engineer, Divi- sion of Highways; Computer Systems Engineer; Chief, Office of Local Assis- trnce, Caltrans; Chief, Office of State Planning; Chief, Division of Equip- menq and Chief, Division of Value Engineering. Mr. Kassel is a member of the Institute of Transportation Engineers. J¡vrs L. Manrw, an engineer and administrator, recently retired as public works Director for the city of Fresno, california. He holds a bachelor's degree in civil engineering from George washington university. He began his career as a civil engineer for the u.s. Bureau of Reclamation, and has served as Bridge Engineer for the state of California; Civil Engineer, Richmond, California; Assistant City Engineer, San Leand¡o, California; and public works Director, Berkeley california. He cunently serves on the Transporta- tion Advisory committee, california Department of rransportation, and has served on the boards of directors of the League of california cities; the Municipal and Aþorts Division, American Road and rransport Builders Association; the Fresno Metropoliøn Flood conrol Dist¡ict; and the Fresno county water Advisory committee. Mr. Martin was a member of the boa¡d of di¡ectors of the American Public works Association from 1916 ro 19g5 and president from 1983 to 1985. He is an American society of civil Engineers Fellow. Bnoors o. Nrcnom is chief Engineer of Design for fhe Arkansas state Highway and Transportation Department where he served as Engineer of Highway Design, Roadway Design Section Head, Assisønt Roadway Engi- neer, and Roadway Design Engineer before assuming his current position. Mr. Nichols is a graduate of the university of Arkansas with a bachelor's degree in civil engineering. He is a member and former Chairman of the AASHTO Task Force on Geometric Design and a member of the AASIilO subcommittee on Design and the AASIilo Joint Task Force on pavements. Mr. Nichols is also chairman of the Arkansas Highway and rransportation Department Resea¡ch council and a member of the university of Arkansas Academy of Civil Engineering. In 1981 he received the AASHTO Region II Design Award.

3t't Bnra¡¡ o'Nern is President of the Insurance Institute for Highway Safety (IIHS) and its associated organization, the Híghway Loss Data Institute iUt-Ot¡, two independenr organizations dedicated to reducing the lisses-deaths, injuries, and property damago-resulting from motor vehicle crashes. He received his bachelor's degfee in mathematics from the Bath university of Technology, England. Before becoming president of IIHS and HLDI, Mi. O'Neitt served as vice president for research, senior vice president, and executive vice president of IIHS and senior vice president of HLDI. He was responsible for the research programs of both organizations, and over the years irc has 'oeerr per-sorraiiy invoived in resea¡ch covering virtually all ãrpoß of highway loss reduction, including vehicle and highway design, emergency medical care, the effectiveness of traffic laws, and driver behavior. He is the author of numerous scientific papers and coauthor of the Injury Fact Book. Mr. O'Neill is Chairman of the National Safety Council's Committee on Alcohol and Other Drugs. He has served on the Advisory Committee for the U.S. Department of Transportation's National Accident Sampling System and the Naiional Academy of Sciences' Committees on Trauma Research. RosERr H. RnyvoNp, JR., is presently Assistant Chief Counsel in charge of the General Law Division, Commonwealttr of Pennsylvania Department of Transportation. He received his AB degree from Bucknell University and his JD from Dickinson School of Law. Mr. Raymond has held the following positions in his 23 years with the Pennsylvania Department of Transportation: -Chief, Environmental Section, General Law Division; Assist¿nt Counsel, Regional Attorney, Land Acquisition Division; and Trial Attorney, Eminent Domain and Right-of-way Division. He was a partner in the law firm of Ziegler and Raymond from 1966 to 1967. He has been admitted to practice befõre the U.S. Supreme Court; U.S. Court of Appeals for the Third Circuit; u.s. court of claims; u.S. District court, Middle Disrict of Pennsylvania; U.S. District Court, Eastern District of Pennsylvania; Supreme Court of Pennsylvania; Superior Court of Pennsylvania; and the Commonwealth Court of Pennsylvania. Jornv H. Ss¡,ren is currently an executive with the New York Staæ Thruway Auttrority. He was formerly Assistant commissioner and chief Engineer for the New York State Depaftment of Transportåtion. He received his bachelor's degfee in civil engineering from the University of Denoit. He has held field aná office positions in the New York Department of Public Works (Rochester)

318 and has served as Director of the Project Deveþment Bureau and ttre safety and rraffic Division for the New York søte Deparrnent of rransportation. Mr. shafer represented the New York State Department of rransporøtion Commissioner on ttre Governor's Traffic safety committee andis a cofounder of the Upstate Section of the Institute of Transportation Engineers. He has served on National Cooperative Highway Research program panels on high- way capacity and is a member of the AASHTO Standing Committee on Highways and the TRB Committee on RRR Standards. He is a registered professional engineer in the søte of New York. Rrc¡¡eno R. Sr¡¡lom, JR., is a businessman and hesident of Mansfield Asphalt Paving Company. At Mansûeld Asphalt Paving Company, he served as construction crew foreman, civil engineer, and vice president. He ¡eceived his bachelor's degree and MBA from ohio state university and his master's degree from MIT. Mr. stander served as a part-time lecturer in construction management at ohio state university from 1969 ûo t971. He is a member of the TRB construction Management committee, the ohio conEactors Asso- ciation, and the Association of Asphalt paving Technologists. He is a registered professional engineer in Ohio. J¡'rurs I. T.qyl-on is Associate Dean of Engineering, university of Notre Dame. He received his bachelor's and master's degrees from fhe Case Institute of rechnology and his Ph.D. from ohio state university. He was formerly chairman of ttre Department of civil Engineering, university of Notre Dame, and Director of the Bu¡eau of Highway Traffic at pennsylvania state university. Dr. Taylor has worked as a consultant to a number of organizations, including the TRB, Goodell-Grivas Incorporated, HRB-Singer, Incorporated, and the Institute For Research. He was formerly president of the North cenEal Indiana Branch of the American Society of civil Engineers and the Educational Division of the American Road and rransportation Builders Association. Dr. Taylor is a member of the American soiiety for Engineering Education, Inslitute of rransportation Engineers, Transportation Research Board, Sigma Xi, and Tau Beta pi. E. DSAN Trsoale is Director of the Idaho Department of rransportation. He is a graduate of the university of Idaho with a bachelor of science degree in forestry (1950) and in civil engineering (1955). since 1953 Mr. Tisdate has

319 served in the following positions at the Idaho Department of Transportation: Planning Engineer, Deputy State Highway Engineer, State Highway Admin- isrator, and Chief of Engineering Services. He was Vice Chairman of the AASIilO Standing Committee on Highway Traffic Safety, and formerly Chairman of the AASHTO Subcommittee for Traffrc Engineering and the AASFIIO delegation to the National Committee on Uniform Trafûc Control Devices. He is also a member of ttre AASI{TO Executive and Policy Committees, AASIITO Reorganization Tiask Force, Chairman of the Federal Mandate Review II Task Force, and a member of the University of ldaho Engineering advisory Boarti. He is currenfiy Chaimian of irie idaho Traffic Safety Commission and a member of the National Society of hofessional Engineers and ttre Idaho Society of hofessional Engineers.

Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation -- Special Report 214 Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB Special Report 214: Designing Safer Roads: Practices for Resurfacing, Restoration, and Rehabilitation examines the cost-effectiveness of safety-related geometric design elements such as lane and shoulder widths, crest vertical curves, stopping sight distances, and intersections.

Safety is a central design consideration for modern highways. For roads receiving federal aid, safety is incorporated through the design policies established by the American Association of State Highway and Transportation Officials (AASHTO), adherence to which is required by the Federal Highway Administration for roads funded with federal aid. Many older and rural roads, however, were built before AASHTO’s modern guidelines had been established. When federal aid is used to improve these roads—when they are resurfaced, for example—safety advocates have urged that they be upgraded to incorporate modern design standards, which might include wider lanes, improved provision for driver sight distance, and other enhancements. State and local officials, by contrast, often contend that raising these roads to the current standards would greatly reduce the number of miles of road that could be resurfaced or upgraded, which would itself be detrimental to safety.

Unfortunately, the safety benefits of such design features have not been well established. Moreover, the variability in local conditions, the amount of daily traffic, and other considerations undermine the usefulness of specific standards. Nonetheless, the committee that produced this report recommended a number of safety-conscious design practices and improvements, including minimum lane and shoulder widths for two-lane roads and bridge widths. The committee also recommended analytical approaches that could be used by state and local officials to determine when safety improvements should be considered, and outlined an approach for assessing the safety cost-effectiveness of doing so. The report produced by the committee has become a standard reference for engineers designing improvements to rural roads.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!