Further Development of the Safety and Congestion Relationship for Urban Freeways (2014)

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.

8Research Approach and State-by-State Results To determine a relationship between safety and congestion for use in evaluating design treatments, relationships between crash rates and level of service (LOS) were developed based on traffic operational and crash data obtained from instru- mented directional freeway segments in five metropolitan areas: Seattle, Washington; Minneapolis–St. Paul, Minnesota; Sacramento, California; the Kansas portion of the Kansas City metropolitan area; and the Missouri portion of the Kan- sas City metropolitan area. The selection of these five metro- politan areas was based on the availability of relevant data. The Kansas and Missouri portions of the Kansas City metro- politan area were analyzed separately because the crash data were obtained from different sources. Technical Approach For analysis purposes, the freeway system in each metropoli- tan area was divided into directional segments, usually extend- ing from one interchange to the next. The sections were selected so that a given detector station would be representa- tive of the traffic conditions for all crashes within that sec- tion. All of the detector stations used in the study were located on the mainline freeway, rather than on ramps; and each detector station provided coverage for all through lanes on the directional freeway segment, including any high- occupancy vehicle (HOV) lanes that were present adjacent to the mainline freeway lanes. (Separate HOV roadways in the freeway median were excluded from the analysis.) The most appropriate detector station was selected for each direc- tional segment; whenever possible, a detector station near the center of a segment was selected. In some cases, a detector station on the mainline freeway within the limits of either the upstream or downstream interchange was used. In a few cases, a detector station located in the immediately upstream or downstream freeway segment was used; this was done only in limited cases where the intervening interchange was rela- tively minor in nature. The traffic operational data collected at each detector sta- tion on the directional freeway segments consisted of 5-min volume and average speed data for each travel lane. Speed or volume was missing for some 5-min intervals on one or more lanes. Most missing data were attributed to detector malfunc- tions. No set of loop detectors will function across all freeway lanes all of the time; therefore, some missing volume and speed data are inevitable. A detector that malfunctions is usu- ally out of service for a substantial time period; however, there is no reason to believe that missing data due to a mal- functioning detector lead to a bias in the remaining data set. Data for each detector station were obtained for a specified study period—either 3 or 5 years. Some detector stations were either first installed or taken out of service during the study period. When this occurred, data from the detector sta- tion could only be obtained for time periods when the detector station was actually in service. For such detector stations, the term missing data simply represents time periods when the detector station did not exist. Flow rates in vehicles per hour per lane were computed from the data for each station, both for each lane and for all lanes combined based on the available 5-min volume data. These 5-min flow rates showed some large fluctuations. The speed and volume data were aggregated into 15-min intervals, which provided much more stable data. Once processed, the volume and speed data were used to determine the level of service for each 15-min interval. Crash data for each directional freeway segment were com- piled for the same 15-min periods as the traffic volume and speed detector data on the basis of the reported crash date and time. The crash data included all mainline freeway crashes that occurred within the limits of each roadway section of interest during the study period. Crash severity levels considered in the evaluation are • Total crashes (i.e., all crash severity levels combined); • Fatal-and-injury (FI) crashes; and • Property-damage-only (PDO) crashes. C h A p T e r 2

9 Level of service was computed for each 15-min record using the operational analysis procedure presented in the 1994 Highway Capacity Manual (HCM) Chapter 23 (2). Com- ponents in the LOS calculations included directional vol- ume, directional speed, flow rate, traffic mix adjustment factor to determine flow rates in passenger cars per hour per lane (i.e., heavy-vehicle adjustment factor), and traffic density. Truck percentages for each roadway section were obtained from maps and other data published by the state department of transportation (DOT) or the relevant metro- politan planning organization (MPO). Truck percentages were typically available for the day as a whole (i.e., a typical 24-h period), but were not available for specific peak-hour or off-peak periods. The study periods for the five metropolitan areas ranged from 3 to 5 years. For each 15-min period during the study period, the available data included the following: • 15-min traffic volume (number of vehicles counted) summed across all lanes of the directional freeway segment; • Average spot speed of vehicles across all lanes (weighted by lane volumes) (mi/h); and • Number of crashes that occurred on the directional free- way segment during the 15-min period (generally either zero or one) by crash severity level. Data were used for all 15-min periods during the study period, unless some of the needed data values were missing. Data were used for all available periods, including peak and off-peak periods, daytime and nighttime, weekdays, week- ends, and holidays, as the data for each of the periods rep- resent a valid observation of crash rate. Thus, in a 3-year study period, the number of 15-min periods for which data were available at any given site was calculated as follows (Equation 2.1): 4 15-min periods h 24 h day 365 days year 3 years 105,120 15-min periods (2.1) × × × = For a 5-year study period, the number of 15-min periods for which data were available at any given site was calculated as follows (Equation 2.2): 4 15-min periods h 24 h day 365 days year 5 years 175,200 15-min periods (2.2) × × × = Appropriate adjustments were made for leap year, as needed. Site characteristics data available to compute traffic density and vehicle miles of exposure and to determine LOS included the following: • Directional segment length; • Number of lanes; and • Average truck percentage. The operational measure used to define LOS for free- ways is the traffic density in passenger cars per mile per hour. The traffic density for a 15-min period was com- puted from the available speed and volume data as follows (Equation 2.3): 4 (2.3)15 15 HV 15 D V f nS = where D15 = traffic density for a 15-min period (passenger cars per mile per lane); V15 = traffic volume for the 15-min period (vehicle) summed across all lanes of the directional freeway segment; fHV = heavy-vehicle adjustment factor from HCM Equa- tion 23-3 (assuming site-specific truck percentage, but zero recreational vehicles); S15 = average spot speed across all lanes (weighted by lane volumes) (mi/h); and n = number of lanes on directional freeway segment. It should be noted that Equation 2.3 does not include the peak-hour factor, so D15 is based on the actual 15-min volume and not the highest 15-min volume during a particular hour, as is commonly used in HCM procedures. As specified in the HCM, six LOS categories are assigned by density ranges as shown in Table 2.1 (2): Since the LOS categories are quite broad, a more refined LOS categorization was used to better capture the relationship Table 2.1. LOS Categories by Density Range LOS Traffic Density Range (pc/mi/ln) A 0 to 11 B 11 to 18 C 18 to 26 D 26 to 35 E 35 to 45 F 45+

10 between density and crash rates. The 18 LOS categories selected are shown in Table 2.2. Based on the 15-min crash rate and traffic density data, average crash rates (expressed in crashes/MVMT) were calcu- lated within each of the 18 LOS categories, separately for each severity level and each metropolitan area. Specifically, the crash rate for a given LOS category was determined using Equation 2.4 for all 15-min periods in that LOS category combined. Crash rate number of crashes veh mi of travel (2.4) - ∑ ∑= The median traffic density was simply the median traffic density for all 15-min periods in that LOS category com- bined, with equal weight given to each 15-min period. Similarly, median traffic densities were calculated within each of the 18 LOS categories in each metropolitan area. The results of the analysis of these data for individual metropolitan areas are presented in the next section, Data- base and Results by State. The results across all metropoli- tan areas are subsequently reviewed in Summary of Full Data Set. Database and results by State Seattle, Washington For the Seattle metropolitan area, data were obtained in the original Phase 2 research in Project L07 for 139 freeway sites representing 194 mi of directional freeway segments. The study period for Seattle was 3 years from 2005 to 2007, inclusive. Traffic operational data were provided by the Washington State DOT traffic management center to SHRP 2 Project L03. Project L03 organized and formatted the data and provided them to Project L07 for analysis. Crash data for the study period were drawn from Washing- ton State DOT records provided by the Federal Highway Administration (FHWA) Highway Safety Information Sys- tem (HSIS). Table 2.3 presents a summary of the site characteristics in the Seattle metropolitan area and the number of 15-min records available for analysis. Figure 2.1 presents a plot of crash rate versus traffic density by LOS level for the Seattle metropolitan area for each crash severity level. Minneapolis–St. Paul, Minnesota For the Minneapolis–St. Paul metropolitan area, data were obtained in the original Phase 2 research in Project L07 for 423 freeway sites representing 411 mi of directional freeway segments. The study period for Minneapolis–St. Paul was from 2005 to 2007, inclusive. Because of the unusual flow conditions, a decision was reached to exclude from the study all data in the Minneapolis–St. Paul area after the I-35W bridge collapse on August 1, 2007. While this period might have been interesting (because volumes changed dramatically on many freeway segments), the changed driving conditions were new to many drivers and the Minnesota DOT made many modifications to specific roadways to increase base Table 2.2. LOS Categories Used in Study LOS Traffic Density Range (pc/mi/ln) A+ 0 to 3 A 3 to 7 A- 7 to 11 B+ 11 to 13 B 13 to 15 B- 15 to 18 C+ 18 to 20 C 20 to 23 C- 23 to 26 D+ 26 to 29 D 29 to 32 D- 32 to 35 E+ 35 to 38 E 38 to 41 E- 41 to 45 F+ 45 to 50 F 50 to 55 F- 55+ Table 2.3. Site Distribution Characteristics for Directional Freeway Segments in the Seattle Metropolitan Area Number of Directional Lanesa Number of Sites Length (mi) Number of 15-Min Recordsb 2 62 89.1 5,834,492 3 53 71.6 5,781,601 4 24 33.4 2,522,880 All lanes 139 194.1 14,138,973 a Not including HOV lanes. b Includes records with missing volume or speed.

11 capacity. Thus, the study period for Minneapolis–St. Paul was 2.6 years. Traffic operational data were provided by the Minnesota DOT traffic management center to SHRP 2 Project L03. Proj- ect L03 organized and formatted the data and provided them to Project L07 for analysis. Crash data for the study period were drawn from Minnesota DOT records provided by the FHWA HSIS. Table 2.4 presents a summary of the site characteristics in the Minneapolis–St. Paul metropolitan area and the number of 15-min records available for analysis. Figure 2.2 presents a plot of crash rate versus traffic density by LOS level for the Minneapolis–St. Paul metropolitan area for each crash severity level. Sacramento, California For the Sacramento metropolitan area, data were obtained in the new Task IV-5 research in Project L07 for 319 freeway sites representing 437.7 mi of directional freeway segments. The study period for Sacramento was 3 years from 2009 to 2011, inclusive. Traffic operational data were obtained from the California DOT (Caltrans) Performance Measurement System (PeMS). Crash data for the study period were drawn from Caltrans records provided by the FHWA HSIS. Table 2.5 presents a summary of the site characteristics in the Sacramento metropolitan area and the number of 15-min records available for analysis. Figure 2.3 presents a plot of crash rate versus traffic density by LOS level for the Sacramento metropolitan area for each crash severity level. Kansas City, Kansas For the Kansas portion of the Kansas City metropolitan area, data were obtained in the new Task IV-5 research in Project L07 for 144 freeway sites representing 139.7 mi of directional freeway segments. The study period for Kansas City was 5 years from 2008 to 2012, inclusive. Traffic operational data were obtained from the Kansas City Scout traffic management center, which is jointly operated by the Kansas and Missouri DOTs. Crash data for the study period were provided by the Kansas DOT. Table 2.6 presents a summary of the site characteristics in the Kansas portion of the Kansas City metropolitan area and the number of 15-min records available for analysis. Figure 2.1. FI, PDO, and total crash rates versus traffic density for directional freeway segments in the Seattle metropolitan area. FI  fatal and injury, PDO  property damage only, and pc/mi/ln  passenger cars per mile per lane. Severity level FI PDO Total Cr a s he s pe r M VM T 0 1 2 3 4 5 6 7 8 9 10 11 Traffic density (pc/mi/ln) 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Table 2.4. Site Distribution Characteristics for Directional Freeway Segments in the Minneapolis–St. Paul Metropolitan Area Number of Directional Lanesa Number of Sites Length (mi) Number of 15-Min Recordsb 2 153 147.3 13,742,976 3 185 183.3 16,695,168 4 73 65.3 660,536 5 12 15.0 1,085,184 All lanes 423 410.9 38,124,864 a Not including HOV lanes. b Includes records with missing volume or speed.

12 Table 2.5. Site Distribution Characteristics for Directional Freeway Segments in the Sacramento Metropolitan Area Number of Directional Lanesa Number of Sites Length (mi) Number of 15-Min Recordsb 2 96 146.5 8,382,244 3 99 141.0 8,660,719 4 78 92.9 6,343,977 5 43 52.6 3,381,762 6 1 2.0 105,108 7 2 2.7 210,220 All lanes 319 437.7 27,084,030 a Not including HOV lanes. b Includes records with missing volume or speed. Table 2.6. Site Distribution Characteristics for Directional Freeway Segments in the Kansas Portion of Kansas City Metropolitan Area Number of Directional Lanes Number of Sites Length (mi) Number of 15-Min Recordsa 2 22 26.8 864,552 3 79 68.8 7,448,853 4 32 28.4 3,720,877 5 10 14.7 759,396 6 1 1.0 171,219 All lanes 144 139.7 12,964,897 a Includes records with missing volume or speed. Severity level FI PDO Total Cr a s he s pe r M VM T 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 Traffic density (pc/mi/ln) 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Figure 2.2. FI, PDO, and total crash rates versus traffic density for directional freeway segments in the Minneapolis–St. Paul metropolitan area. Severity level FI PDO Total Cr a s he s pe r M VM T 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 Traffic density (pc/mi/ln) 0 5 10 15 20 25 30 35 40 45 50 55 60 65 Figure 2.3. FI, PDO, and total crash rates versus traffic density for directional freeway segments in the Sacramento metropolitan area.

13 Severity level FI PDO Total Cr a s he s pe r M VM T 0 1 2 3 4 5 6 7 8 9 10 Traffic density (pc/mi/ln) 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Figure 2.4. FI, PDO, and total crash rates versus traffic density for directional freeway segments in the Kansas portion of Kansas City metropolitan area. Figure 2.4 presents a plot of crash rate versus traffic den- sity by LOS level for the Kansas portion of the Kansas City metropolitan area for each crash severity level. Kansas City, Missouri For the Missouri portion of the Kansas City metropolitan area, data were obtained in the new Task IV-5 research in Project L07 for 201 freeway sites representing 184.2 mi of directional freeway segments. The study period for Kansas City was 5 years from 2008 to 2012, inclusive. Traffic opera- tional data were obtained from the Kansas City Scout traffic management center, which is jointly operated by the Kansas and Missouri DOTs. Crash data for the study period were provided by the Missouri DOT. Table 2.7 presents a summary of the site characteris- tics in the Missouri portion of the Kansas City metropoli- tan area and the number of 15-min records available for analysis. Figure 2.5 presents a plot of crash rate versus traffic density by LOS level for the Missouri portion of the Kansas City met- ropolitan area for each crash severity level. Table 2.7. Site Distribution Characteristics for Directional Freeway Segments in Missouri Portion of Kansas City Metropolitan Area Number of Directional Lanes Number of Sites Length (mi) Number of 15-Min Recordsa 2 57 48.7 4,374,666 3 115 110.2 11,106,428 4 26 21.7 3,209,361 5 1 0.9 97,716 6 2 2.7 65,896 All lanes 201 184.2 18,854,067 a Includes records with missing volume or speed. Summary of Full Data Set Table 2.8 presents a summary of the sample sizes in the full data set for all five metropolitan areas/states. The table shows that 1,226 sites were studied for a potential total of 4,191 site-years of data. Table 2.9 summarizes the crash and exposure data during the periods for which volume and speed data were available.

14 Severity level FI PDO Total Cr a s he s pe r M VM T 0 1 2 3 4 5 6 7 8 9 10 11 Traffic density (pc/mi/ln) 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 Figure 2.5. FI, PDO, and total crash rates versus traffic density for directional freeway segments in the Missouri portion of Kansas City metropolitan area. Table 2.8. Summary of the Sample Sizes in the Full Data Set by State Metropolitan Area State Number of Sites Total Length (mi) Number of Years Potential Number of Site-Years Maximum Potential Number of 15-Min Records Actual Number of 15-Min Records with Detector Present Actual Number of 15-Min Records with Nonmissing Volume and Speed Data Missing Volume and Speed Data (%) Seattle Washington 139 194.1 3 417 14,611,680 14,138,973 11,526,511 18.5 Minneapolis– St. Paul Minnesota 423 410.9 2.6a 1,092a 38,252,736 38,124,864 31,986,802 16.1 Sacramento California 319 437.7 3 957 33,533,280 27,001,960 26,720,533 1.0 Kansas City Kansas 144 139.7 5 720 25,242,624 12,964,850 11,858,383 8.5 Kansas City Missouri 201 184.2 5 1,005 35,234,496 18,845,109 15,742,204 16.5 Total 1,226 1,366.6 na 4,191 146,874,816 111,075,756 97,834,433 11.9 Note: na = not applicable. a After excluding the period from August 1, 2007, to December 31, 2007, when traffic conditions were changed due to the I-35W bridge collapse. Table 2.9. Crash, Exposure, and Crash Rate Data by State Metropolitan Area State Reported Crashes During Period with Volume and Speed Data Available MVMT During Period with Volume and Speed Data Available Crash Rate per MVMT During Period with Volume and Speed Data Available FI PDO Total FI PDO Total Seattle Washington 3,863 7,131 10,994 4,793.9 0.81 1.49 2.29 Minneapolis– St. Paul Minnesota 1,289 3,360 4,649 6,298.0 0.20 0.53 0.74 Sacramento California 4,598 8,653 13,251 19,452.0 0.24 0.44 0.68 Kansas City Kansas 1,566 4,885 6,451 5,035.6 0.31 0.97 1.28 Kansas City Missouri 1,914 5,481 7,395 5,565.9 0.34 0.98 1.33 Total 13,230 29,510 42,740 41,145.4 0.32 0.72 1.04