Long-Distance and Rural Travel Transferable Parameters for Statewide Travel Forecasting Models (2012)

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.

68 The discussion on Trip Distribution will focus largely on the process of developing friction fac- tor gamma functions and trip length frequency distribution curves from travel survey data, how a transferable set of friction factors was developed for this study, and provision of transferable friction factor gamma functions along with guidance on which to use and when. Benchmark sta- tistics will also be provided on average trip lengths by purpose for rural and long-distance travel. Additional guidance will be provided on the variability of mean trip length and how to normalize distribution patterns. Chapter sections will follow the same order and content as Chapter 4 on trip generation; however, background text on analytical procedures will not be repeated here. 5.1 Long-Distance and Rural Trip Distribution Benchmark Statistics from Statewide Models and Other Sources Section 4.1 provided background information on the summary of available statewide model statistics on rural and long-distance travel, along with statistics borrowed from other sources. This section provides trip distribution statistics from statewide models and other sources. Statewide Model Statistics Table 5.1 is a summary of average trip lengths for long-distance trips by statewide model (in time and distance). Average trip lengths vary by long-distance purpose and model/state; some states report this statistic by miles, others by minutes of travel, and some by both; long-distance trip lengths vary from a low of 122 minutes to a high of 304 minutes. Some states included multistate trips in reporting these statistics while other states did not, resulting in yet another issue affecting transferability. Bureau of Transportation Statistics As noted in the previous chapter, the Bureau of Transportation Statistics (BTS) published findings in May 2006 from the 2001 NHTS on long-distance trip-making. Table 5.2 is derived from the same 2006 BTS report and depicts long-distance trips by distance category. According to BTS’ analysis, 90 percent of long-distance trips are less than 500 miles. Recent and Ongoing GPS Surveys in the United States Table 5.3 depicts several different measures of trip length, including distance, travel time, stopped time, and time spent idling, from recent GPS surveys described earlier in Section 2.3 C h a p t e r 5 Trip Distribution Parameters and Benchmark Statistics

trip Distribution parameters and Benchmark Statistics 69 Average Trip Length By Purpose (Minutes or Milesa) Business Tourist Other Total Minutes Total Miles Arizona (Passenger) – – – 213 206 Arizona (Truck) – – – 228 257 Florida – – – 127 – Georgia – – – 131 – Indiana – – – 121 – Louisiana – – – 168 – Ohio 146 Texas (Miles) 200 – 199 – 200 Utah 89 – 81 85 – Virginia (Interstate) 284 308 318 303 – Virginia (Intrastate) 127 124 126 126 136 Source: BTS. a Listed in minutes unless otherwise indicated. Table 5.1. Average trip length of long-distance trips in statewide models. Distance Trips 50-499 Miles 90.0% 500-999 Miles 5.0% More Than 1,000 Miles 5.0% Source: BTS. Table 5.2. 2001 long-distance trips by trip distance. Trip Distance (Miles) Trip Duration (Minutes) Stop Time (Hours) Idle Time (Minutes) L on g- D is ta n ce R u ra l U rb an L on g- D is ta n ce R u ra l U rb an L on g- D is ta n ce R u ra l U rb an L on g- D is ta n ce R u ra l U rb an Overall 88.59 8.12 4.69 91.13 14.36 11.11 3.91 3.93 3.32 0.93 1.40 0.86 Atlanta 95.46 8.08 5.36 93.84 13.89 11.50 3.98 4.03 3.54 0.74 0.43 0.64 Denver 83.30 7.99 3.98 85.74 14.05 9.78 4.41 4.25 3.35 0.86 3.04 0.85 Massachusetts 79.12 8.47 3.72 89.57 16.58 10.53 2.55 2.48 2.21 1.74 1.56 1.86 Chicago 86.39 8.64 4.70 99.28 16.68 13.17 Source: Geostats based on recent GPS-based travel surveys. Table 5.3. Trip length statistics from recent GPS-based surveys.

70 Long-Distance and rural travel transferable parameters for Statewide travel Forecasting Models of this Guidebook. Statistics are provided for all four surveys as well as the sum of all by long- distance, rural, and urban trip-making categories. Mean long-distance trip length, for the four surveys summarized, exceeds 85 minutes or 79 miles. In virtually every survey and trip length measure, rural trip lengths are longer than urban averages. 5.2 Analytical Approach to Estimating Long-Distance and Rural Trip Distribution Parameters and Benchmarks The same analytical procedures described previously in Section 4.2 on trip generation were applied to trip distribution. Mean trip lengths and standard deviations for each long-distance trip purpose are found in Table 5.4, based on analysis of ATS data. Although average trip lengths in urban models are usually summarized by travel time, mileage is a more appropriate measure for long-distance trips since travel times were generally reported as hundreds of minutes with a wide variation in travel time. Average trip lengths for business trips were highest while lowest for personal business trips. The average long-distance trip length for all households was 836 miles. 5.3 Long-Distance Trip Distribution Model Parameters ATS trip records were sorted by travel time increment in order to calculate a set of friction factor gamma functions for each trip purpose. ATS trip length frequency distribution curves are depicted in Figures 5.1 through 5.3 for each trip purpose. Table 5.5 depicts the resulting long-distance friction factor gamma function parameters by purpose. It should be noted that these gamma Mean Standard Deviation 01 Business 477 1,208 02 Pleasure 414 1,097 03 Personal Business 352 961 Total 418 1,105 Table 5.4. ATS average trip length in miles by purpose, one-way. Figure 5.1. Trip length frequency for long-distance business trips.a a For readability purposes, the curves depicted in Figures 5.1 through 5.2 were capped at 1,000 miles and 2 million trips. - 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000 100 300 500 700 900 Th ou sa nd s o f P er so n T rip s Person Trips Travel Time in Miles

trip Distribution parameters and Benchmark Statistics 71 Figure 5.2. Trip length frequency for long-distance pleasure trips. - 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000 100 300 500 700 900 Th ou sa nd s of P er so n Tr ip s Person Trips Travel Time in Miles Figure 5.3. Trip length frequency for long-distance personal business trips. - 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000 100 300 500 700 900 Th ou sa nd s of P er so n Tr ip s Person Trips Travel Time in Miles LD Business LD Pleasure LD Personal Business “b” “c” “b” “c” “b” “c” Long-Distance Trips -0.421 -0.0022 -0.578 -0.0023 -0.567 -0.0024 Source: 1995 ATS. Table 5.5. Long-distance trip distribution gamma function parameters.

72 Long-Distance and rural travel transferable parameters for Statewide travel Forecasting Models functions reflect reported trip lengths, without any smoothing or confirmation against network skims. Therefore, these gamma functions should only serve as potential starting points for Gravity Model calibration. It should also be noted that the Gravity Model has significant shortcomings when used to distribute long-distance trips, particularly as it relates to the attractiveness of zones. 5.4 Rural Trip Distribution Model Parameters Similar to the rural trip rate calculations, the NHTS URBAN attribute was used to isolate nonurban households for analysis of trip length frequency distributions and gamma function parameters. The 2009 NHTS trip records were sorted by travel time increment, excluding urban households, in order to calculate a set of friction factor gamma functions for each trip purpose. NHTS 2009 trip length frequency distribution curves for all modes combined are depicted for each trip purpose in Figures 5.4 through 5.6. Although these graphs clearly depict some trips that Figure 5.4. Trip length frequency for rural home-based work trips. - 50 100 150 200 250 300 350 400 450 500 0 20 40 60 80 M ill io ns o f P er so n Tr ip s Person Trips Travel Time in Minutes Figure 5.5. Trip length frequency for rural home-based nonwork trips. - 50 100 150 200 250 300 350 400 450 500 0 20 40 60 80 M ill io ns o f P er so n Tr ip s Person Trips Travel Time in Minutes

trip Distribution parameters and Benchmark Statistics 73 would otherwise be calculated as “long-distance,” the occurrence of these trips is purely coinci- dental because there was no sampling strategy for long-distance trips in NHTS 2009. Graphs are scaled to match by purpose, meaning that curves were depicted with trips limited to 500 million along the y-axis and 90 minutes in length along the x-axis. Table 5.6 depicts preliminary long-distance friction factor gamma function parameters by purpose. These gamma function parameters reflect weekday trips by rural households only from 2009 NHTS data. The same precautions noted earlier for the long-distance gamma functions apply here (i.e., lack of smoothing, etc.). Average trip length, although not a model parameter, is an important benchmark statistic used in assessing the validity and reasonableness of trip distribution model results. Table 5.7 depicts average trip lengths in minutes of travel time for home-based work, home-based nonwork, and nonhome-based trip purposes. These tables, originally prepared for NCHRP Report 716, depict average trip lengths for a variety of urban area sizes and travel modes, in minutes of travel time (although low sample sizes for some markets, such as transit trips in smaller areas, meant that NCHRP Report 716 does not include this detailed breakdown). NCHRP Report 716 analyses also excluded weekend trips. The urban statistics are only provided for comparative purposes against rural trip lengths depicted in bold text. On average, rural HBW trips are longer than those in urbanized areas with populations of less than 1 million. For areas in excess of 1 million population, due to excessive congestion, HBW trip lengths are higher than those in rural areas. HBNW auto trips are longer in rural areas than urbanized areas; however, the reverse is generally the case with other trans- portation modes. Finally, rural NHB trips are greater than or equal to comparable trip lengths in urbanized areas, with the exception of very large urbanized areas with rail and for modes other than auto, consistent with trip-chaining studies (sample sizes for transit trips in areas under 1 million population are too low for the means to be statistically significant). Figure 5.6. Trip length frequency for rural nonhome-based trips. - 50 100 150 200 250 300 350 400 450 500 0 20 40 60 80 M ill io ns o f P er so n Tr ip s Person Trips Travel Time in Minutes Home-Based Work Home-Based Nonwork Nonhome-Based “b” “c” “b” “c” “b” “c” Rural Trips -0.053 -0.0495 -0.079 -0.0635 -0.049 -0.063 Source: 2009 NHTS. Table 5.6. Rural trip distribution gamma function parameters.

74 Long-Distance and rural travel transferable parameters for Statewide travel Forecasting Models Urban Size Auto Mean Transit Mean Nonmotorized Mean All Modes Mean Home-Based Work Trips 1 million or more with subway or rail 29 53 18 32 1 million or more without subway or rail 25 64 19 26 Between 500,000 and 1 million 22 46 17 22 Between 200,000 and 500,000 21 63 18 22 Less than 200,000 20 33 18 20 Not in an urbanized area (Rural/Small Urban) 24 58 11 24 All trips 24 55 16 25 Home-Based Nonwork Trips 1 million or more with subway or rail 18 48 16 20 1 million or more without subway or rail 17 48 15 18 Between 500,000 and 1 million 15 46 16 16 Between 200,000 and 500,000 16 50 15 16 Less than 200,000 17 32 15 17 Not in an urbanized area (Rural/Small Urban) 19 45 14 19 All trips 18 47 15 18 Nonhome-Based Trips 1 million or more with subway or rail 21 39 13 20 1 million or more without subway or rail 19 38 14 19 Between 500,000 and 1 million 18 45 15 18 Between 200,000 and 500,000 18 58 15 18 Less than 200,000 17 50 18 18 Not in an urbanized area (Rural/Small Urban) 19 47 14 19 All trips 19 42 14 19 Source: 2009 NHTS/NCHRP Report 716. Analysis also excluded weekend trips. Table 5.7. Rural versus urban average trip lengths in minutes.