Appendixes to TCRP RRD 84: Audible Signals for Pedestrian Safety in LRT Environments (2007)

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.

Appendix C C-1 Analysis of Pedestrian/LRV Collision Data (Prepared February 2005) C-2 Review of Accident Data (Prepared September 2004)

APPENDIX C-1 ANALYSIS OF PEDESTRIAN/LRV COLLISION DATA (Prepared February 2005) INTRODUCTION AND SUMMARY This report summarizes the results of our analysis of pedestrian/LRT accident data for Task 3: Evaluation of the Effectiveness of Audible Devices Based on Historical Data for TCRP Project D-10: Audible Signals for Pedestrian Safety in LRT Environments. Specifically, the KORVE Team gathered accident data from both transit agencies and the National Transit Database, analyzed this data, and identified trends regarding the number, location, severity, and potential cause of pedestrian/LRT accidents. The preliminary conclusions are: ƒ The annual number of pedestrian/LRT accidents is very low relative to the number of vehicle/LRT accidents. ƒ There is substantial variability in accident rates among transit agencies. ƒ Pedestrian/LRT accidents are much more likely to result in fatalities than vehicle/LRT accidents. ƒ About half of pedestrian/LRT accidents occur at grade crossings. ƒ Most of the at grade crossings where collisions occur have active crossing control devices. ƒ Of the accidents at grade crossings with active crossing control devices, traffic signals and a combination of gates, flashing lights, and bells are the most common active devices. ƒ Risky or inattentive behavior appears to be a frequent factor in pedestrian/LRT accidents. ƒ Annual revenue service miles and directional route miles are good predictors of the number of pedestrian/LRT accidents. ƒ Site- or alignment-specific factors that are unique to a particular transit agency may be significant contributors to pedestrian/LRT accidents. ƒ It is not possible to directly evaluate the effectiveness of audible warnings in preventing pedestrian/LRT accidents based on the available data. This report is organized into three sections. The first section summarizes our analysis of the National Transit Database for the years 2002 and 2003. The second section includes detailed analyses of more historical data obtained from three transit agencies: (1) Los Angeles County Metropolitan Transportation Authority; (2) the Santa Clara Valley Transportation Authority; and (3) San Diego Trolley, Inc. The final section includes general conclusions regarding the effectiveness of audible devices based on all the data sources. NATIONAL TRANSIT DATABASE Transit agencies receiving Federal Transit Administration (FTA) Urbanized Area Formula Program grants are required to submit data to the FTA regarding transit service and safety. The FTA maintains this data in the National Transit Database (NTD). The Volpe National Transportations System Center (Volpe Center) is responsible for maintaining this database. The KORVE Team contacted the Volpe Center and requested pedestrian/LRT accident data for the past several years. Due to a significant change in reporting requirements, year 2002 and 2003 data is far more detailed than in previous years. Although this is helpful in evaluating the locations and causes of pedestrian/LRT accidents, it does prevent us from drawing conclusions

based on accident trends over longer periods of time. As a result, most of the analysis reported herein is based on data over the past two years. In order to qualify for reporting, the current standard is that a pedestrian/LRT collision must result in “injuries requiring medical attention away from the scene for one or more persons.” Prior to 2002, the threshold for reporting was lower, resulting in a higher number of accidents. In terms of the pedestrian/LRT accident data collected, some of the key parameters related to the study are1: ƒ Agency. In 2003, there were 22 transit agencies operating LRT systems that reported to the NTD. ƒ Alignment Type. Possible alignment types include: o Exclusive right-of-way (tunnel, elevated structure, at-grade). o Semi-exclusive right-of-way. o Non-exclusive right-of-way (mixed traffic/LRT, transit mall, and LRT/pedestrian mall). o Shared track/corridor (temporal separation, non-temporal separation). o Other non-exclusive. ƒ Collision Location. Possible locations include: o Revenue facility (transit center, platform, other revenue facility). o Non-revenue facility. o Right-of-way/roadway (grade crossing, intersection, other). o Private property (shopping center, residential, commercial, nonprofit facility). o Other. ƒ Grade-Crossing Controls. Possible control options include: o Active devices (two quadrant gates [median barrier, no median barrier], four quadrant gates, flashing lights, traffic signal, and train approaching sign). o Passive devices (stop sign, cross bucks). o No control device. ƒ Event Description. This includes narrative summaries of the event. ƒ Injury Severity. Generally classified as fatal or non-fatal. ƒ Action Description. Although not provided for all events, this identifies possible causes for the accident. ƒ Party Involved. Collisions occurred with trespassers, employees, revenue facility occupants, and others. Following are findings from the NTD data with respect to the number of accidents, fatality rates, accident locations, types of crossing control, possible accident causes, and potential predictor variables for pedestrian/LRT accidents. 1 Data categories are from the NTD Safety and Security Manual, Major Incident Reporting form (S&S-40), 2004.

Number of Accidents ƒ There are relatively few annual pedestrian/LRT accidents The total number of pedestrian/LRT incidents by transit agency, including both fatal and non- fatal incidents, in the years 2002 and 2003 are presented in Table 1 below (detailed incident statistics from 2002 and 2003 can be found in Appendix A). There were relatively few pedestrian/LRT incidents over this two year period, particularly with respect to fatalities. Of the 56 total incidents, 55 resulted in 1 injury and 1 incident had 2 reported injuries, for a total of 57 injuries. Table 1. Summary of Pedestrian/LRT Incidents by Transit Agency (2002 and 2003) 2002 2003 Transit Agency Fatal Non-Fatal Total Fatal Non-Fatal Total City of Detroit DOT 0 0 0 0 0 0 Cleveland-RTA 0 0 0 0 0 0 Dallas Area Rapid Transit 0 1 1 0 0 0 Denver-RTD 0 0 0 0 0 0 Kenosha Transit 0 0 0 0 0 0 King County DOT 0 1 1 0 0 0 Los Angeles County MTA 1 8 9 1 4 5 Maryland MTA 2 1 3 0 0 0 Mass. Bay Transp. Auth. 0 1 1 0 0 0 Memphis Area TA 0 0 0 0 0 0 New Jersey Transit 0 0 0 0 0 0 New Orleans-RTA 0 0 0 0 0 0 Niagara Frontier Transp. Auth. 0 0 0 0 1 1 Pittsburgh Transit 0 0 0 0 0 0 Portland Tri-Met 0 2 2 2 5 7 Sacramento RT 0 0 0 1 0 1 San Diego 0 2 2 2 0 2 San Francisco MUNI 1 4 5 0 8 8 Santa Clara VTA 1 0 1 0 0 0 SE Pennsylvania Trans. Auth. 0 3 3 0 2 2 St. Louis Bi-State Dev. 0 0 0 0 0 0 Utah Transit Authority 2 0 2 0 0 0 Total 7 23 30 6 20 26 Source: National Transit Database The 22 transit agencies listed in Table 1 with qualifying LRT systems included in the NTD operated a total of 38.7 million annual revenue service miles in 2002. Table 2 lists accident rates by transit agency in both accidents per thousand revenue service miles and number of revenue service miles per accident. The accident rates are based on the average yearly accidents in 2002 and 2003.

Table 2. Accident Rates by Transit Agency 2002 & 2003 Accident Rates (million miles) Transit Agency Annual Revenue Service Miles (million) 1 Average Accidents (per year) 1 Per Mile Miles between Accidents City of Detroit DOT 0.01 0 0.00 0.00 Cleveland-RTA 0.94 0 0.00 0.00 Dallas Area Rapid Transit 2.64 0.5 0.19 5.28 Denver-RTD 1.56 0 0.00 0.00 Kenosha Transit 0.02 0 0.00 0.00 King County DOT 0.04 0.5 12.09 0.08 Los Angeles County MTA 3.07 7 2.28 0.44 Maryland MTA 1.66 1.5 0.90 1.11 Mass. Bay Transp. Auth. 4.26 0.5 0.12 8.52 Memphis Area TA 0.40 0 0.00 0.00 New Jersey Transit 1.24 0 0.00 0.00 New Orleans-RTA 0.69 0 0.00 0.00 Niagara Frontier Transp. Auth. 0.38 0.5 1.32 0.76 Pittsburgh Transit 1.33 0 0.00 0.00 Portland Tri-Met 3.27 4.5 1.38 0.73 Sacramento RT 0.95 0.5 0.52 1.91 San Diego 2.64 2 0.76 1.32 San Francisco MUNI 5.49 6.5 1.18 0.85 Santa Clara VTA 1.73 0.5 0.29 3.46 SE Pennsylvania Trans. Auth. 3.08 2.5 0.81 1.23 St. Louis-Bi-State Dev. 2.65 0 0.00 0.00 Utah Transit Authority 1.10 1 0.91 1.10 Total 39.16 28 -- -- Average 1.78 1.27 1.03 1.22 Notes: 1 Based on an average of 2002 and 2003 data. ƒ There is substantial variability in accident rates among transit agencies. As can be seen, there is significant variability in the accident rates. The agencies with the highest rates include the Los Angeles County MTA (highest), Portland Tri-Met, and San Francisco MUNI.2 There are nine agencies that did not report any pedestrian/LRT accidents in either 2002 or 2003. Five of the nine agencies operate less than 1 million annual revenue service miles (Detroit, Cleveland, Kenosha, Memphis, and New Orleans). The four agencies with more than 1 million annual revenue service miles but no reported accidents are Denver, New Jersey, Pittsburgh, and St. Louis. One possible reason for the low accident rate with these four agencies is that they operate primarily on semi-exclusive rights-of-way with relatively few grade crossings. However, as explained later in the analysis of predictor variables, neither of these two factors are strongly correlated with accident rates. 2 This excludes King County DOT and Niagara Frontier Transportation Authority due to the low number of annual revenue service miles.

Fatalities ƒ Pedestrian/LRT incidents are much more likely to result in fatalities than vehicle/LRT incidents. An average of 23% of all pedestrian/LRT incidents in 2002 and 2003 resulted in fatalities. Figure 1 compares the total number of pedestrian/LRT fatalities to total number of pedestrian/LRT injuries over the past seven years. Note that the reporting requirements for accidents changed in 2002, likely resulting in a lower number of reported accidents in 2002 and 2003. Figure 1. Fatal vs. Total Pedestrian/LRT Accidents (1997 to 2003) 1997 1998 1999 2000 2001 2002 2003 46 56 56 26 31 53 63 3 11 7 67 11 13 0 10 20 30 40 50 60 70 # of In ci de nt s Year Fatal Total Incidents As discussed in TCRP Report 69, relative to vehicle/LRT accidents, pedestrian/LRT accidents are much more likely to result in fatalities in both high-speed and low-speed settings.3 For example, along alignments with speeds greater than 55 km/hour (35 mph), 29% of pedestrian/LRT accidents result in fatalities whereas only 19% of vehicle/LRT accidents result in fatalities. Where LRT speeds are less than 55 km/hour, the disparity is even more dramatic, with fatality rates of 18% and 1%, respectively. Table 3 lists the total incidents, fatalities, and injuries for vehicle/LRT, object/LRT, and pedestrian/LRT collisions for the period between 1997 and 2001. As shown in the table and in Figure 2, most incidents were between vehicles and light rail vehicles. Although pedestrian/LRT accidents represent less than 20% of the total LRT incidents, they average 67% of the total LRT fatalities. 3 Transportation Research Board, Transit Cooperative Research Board, TCRP Report #69 – Light Rail Service: Pedestrian and Vehicular Safety, 2001.

Table 3. Summary of LRT Incidents (1997 to 2001) Year Collisions 1997 1998 1999 2000 2001 Total % of Total Total Incidents 352 297 276 333 301 1559 100% w/ vehicles 281 223 206 260 234 1204 77% w/ objects 25 18 14 10 14 81 5% w/ people 46 56 56 63 53 274 18% Fatalities 3 14 13 22 15 67 100% w/ vehicles 0 2 6 9 4 21 31% w/ objects 0 1 0 0 0 1 2% w/ people 3 11 7 13 11 45 67% Injuries 316 332 404 359 299 1710 100% w/ vehicles 263 254 339 261 244 1361 79% w/ objects 8 14 11 37 11 81 5% w/ people 45 64 54 61 44 268 16% Figure 2. LRT Incidents Involving Vehicles, Objects, and People (1997 to 2001) 0 50 100 150 200 250 300 In ci de nt s 1997 1998 1999 2000 2001 Year Vehicles Objects People

Figure 3. Percent Pedestrian/LRT Incidents vs. Total Incidents and Percent Pedestrian/LRT Fatalities vs. Total Fatalities 0% 20% 40% 60% 80% 100% Pe rc en t o f T ot al 1997 1998 1999 2000 2001 Year Ped/LRT Incidents Ped/LRT Fatalities Location of Accidents ƒ Approximately half of pedestrian/LRT accidents occur at grade crossings. Table 4 below separates the total number of pedestrian/LRT injuries by crossings, stations, and other locations. As can be seen, 27 of the 57 total injuries resulting from pedestrian/LRT collisions in 2002 and 2003 occurred at crossings, including either vehicle grade crossings or pedestrian-only crossings. The 22 accidents occurring at “other” locations are typically associated with trespassers and occurred at mid-block locations or within exclusive rights-of- way. Table 4. Fatal and Non-Fatal Pedestrian/LRT Injuries by Location (2002 & 2003) Location Fatal Non-Fatal Total Crossing 5 22 27 Station 1 7 8 Other 7 15 22 Total Injuries 13 44 57 Note: Crossings include grade crossings & intersections. The incidents at stations include all the accidents in the NTD that occurred at revenue facilities. Figure 4 and Figure 5 provide a percentage breakdown of the fatal and non-fatal pedestrian/LRT injuries by crossing, station, and other locations. As shown by the graphs, fatal injuries are most likely (54%) to occur at other locations whereas non-fatal injuries occur more commonly at crossings (49%) than at stations or other locations.

Figure 4. Percent of Fatal Pedestrian/LRT Injuries by Location Figure 5. Percent of Non-Fatal Pedestrian/LRT Injuries by Location Accidents occurring at “other” locations are generally beyond the scope of the D-10 study. When they are removed from the analysis, the difference between the percentage of pedestrian/LRT accidents occurring at grade crossings compared to stations is more pronounced, with 83% of fatalities occurring at grade crossings and 75% of non-fatal accidents occurring at grade crossings. Crossing Controls ƒ Most pedestrian/LRT accidents at grade crossings occur at locations with active crossing control devices. As illustrated in Table 5, 42% of the pedestrian/LRT injuries in 2002 and 2003 (24 in total) occurred at locations with active crossing control devices. In addition, nine injuries occurred at grade crossings where the type of crossing control device was not listed. From the data collected as part of Task 2, we know that most grade crossings have some type of active device. Therefore, the number of injuries at locations with active devices is likely to be higher than listed in Table 5. 38% 8% 54% Xings Stations Other 49% 16% 36% Xings Stations Other

Table 5. Fatal and Non-Fatal Pedestrian/LRT Injuries by Crossing Control Type (2002 & 2003) Protection Type Fatal Non-Fatal Total Active 5 19 24 Passive 1 1 2 Other 0 0 0 Not Listed 4 21 25 None 3 3 6 Total 13 44 57 Table 6 provides a breakdown of the aggregate injury data (fatal and non-fatal) by protection type and location. Once again, when accidents at “other” locations are removed, the percentage of pedestrian/LRT accidents at active crossing controls increases to 55%. Table 6. Total Pedestrian/LRT Injuries by Protection Type and Crossing (2002 & 2003) Protection Type Crossings Stations Other Total Active 17 3 4 24 Passive 2 0 0 2 Other 0 0 0 0 Not Listed 8 5 12 25 None 0 0 6 6 Total 27 8 22 57 Active Crossing Control Devices ƒ At locations with active crossing control devices, most accidents occur where there are traffic signals or crossing gates. Table 7 is a breakdown of all the injuries (fatal and non-fatal) that occurred at locations with active crossing control devices. The major categories include crossing gates, traffic signals, flashers/lights/bells, and other. Gates and traffic signals comprise 83% of all the accidents at locations with active crossing control devices, with slightly more at the traffic-signal controlled crossings. It is assumed that there are no flashing lights or crossing bells at the traffic-signal controlled locations. TCRP Report 174 found that LRV accidents in shared rights-of-way account for the largest proportion of accidents even though this alignment type constitutes the smallest proportion of route miles. Although the analysis in TRCP Report 17 included vehicle as well as pedestrian accidents, this supports the finding of this Task 3 report in that a greater number of 4 Transportation Research Board, Transit Cooperative Research Program, TCRP Report 17: Integration of Light Rail Transit into City Streets, 1996.

pedestrian/LRT accidents occur along shared rights-of-way where traffic signals are used to control pedestrian movement across the tracks. Table 7. Total LRT/Pedestrian Injuries at Different Locations by Type of Active Crossing Control Device (2002 & 2003) Type of Active Device Crossings Stations Other Total Gates 7 2 0 9 Traffic Signals 9 0 2 11 Flashers/Lights/Bells 1 1 0 2 Other 0 0 2 2 Total 17 3 4 24 As can be seen in Figure 6 and Figure 7, there is a substantially larger percentage of fatal accidents (60%) occurring at traffic signal controlled crossings compared to non-fatal accidents (42%). Figure 6. Percent Fatal Pedestrian/LRT Injuries by Active Device 40% 60% Gates Traffic Signals Flashers/Lights/Bells Other

Figure 7. Percent Non-Fatal Pedestrian/LRT Injuries by Active Device 37% 42% 11% 11% Gates Traffic Signals Flashers/Lights/Bells Other Cause of Accidents ƒ Risky or inattentive pedestrian behavior is a frequent factor in pedestrian/LRT accidents. The NTD does not provide a root-cause analysis for each accident. However, based on the available data, in many instances risky or inattentive pedestrian behavior appears to have contributed to the accident. The operator was at least partially at fault in only a few incidents. Observations from the NTD data are supplemented and supported by conversations with transit agency staff. In no particular order, many of the reasons for the 57 pedestrian/LRT injuries in 2002 and 2003 are: 1. Rushing to catch trains or get across intersections. This primarily includes accidents near stations or on station platforms. Coupler jumping also contributed to a few accidents. 2. Ignoring audible and/or visual warnings at grade crossings. In many instances, pedestrians purposefully walked around crossing gates or disregarded other active warnings. The reasons for this behavior are not known. 3. Distractions. The use of cells phones and headsets were contributing factors in four of the accidents. 4. Not paying attention in transit malls. Although most of these incidents do not result in serious injury and therefore were not reported in the NTD, several agencies indicated that this is their most common type of accident. For instance, people walk in front of trains as they leave the station – even after an audible warning is sounded. 5. Intoxication. At least five serious accidents were attributed to intoxicated pedestrians. 6. Trespassing. There were several accidents near tunnel portals or within exclusive rights-of- way. These locations are clearly off limits to pedestrians and are relatively inaccessible. Accidents due to trespassing are beyond the scope of the D-10 study.

Pedestrians with Disabilities Although there was no specific category in the NTD identifying LRT collisions with persons with disabilities, there was no indication in either the Event Description or Action Description categories that any of the pedestrian/LRT accidents involved hearing impaired, visually impaired, or other disabled persons. Therefore, based on the available data, it can not be concluded that disability is a significant factor in pedestrian/LRT accidents. Accident Predictors ƒ Annual revenue service miles and directional route miles are good predictors of the number of pedestrian/LRT accidents. In order to better predict where pedestrian/LRT accidents are likely to occur, a regression analysis was performed looking at a number of possible predictor variables. Specifically, single- variable regressions on the average total accidents in 2002 and 2003 (excluding the 22 total accidents occurring at “other” locations, which are not covered in this study) for each transit agency were run against: ƒ Annual Revenue Service Miles ƒ Directional Route Miles ƒ Number of Stations ƒ Number of Grade Crossings ƒ At-Grade Track Miles Table 8 gives the relevant regression statistics for each variable. The variables are organized by the degree of statistical significance with the number of accidents. Generally speaking, T- Statistics greater than 2 are considered statistically significant. The strongest relationships are annual revenue service miles and directional route miles. These two variables have the highest R squared values, F-statistics, and T-statistics. The usefulness of the statistical analysis is limited because of limited available data. However, including data prior to 2002 is problematic because of different reporting requirements. Figure 8 shows the linear regression of annual revenue service miles, which has the strongest relationship with pedestrian/LRT accidents. There also appears to be somewhat of a weak relationship between accidents and at-grade track miles and accidents and the number of grade crossings. However, it is clear that the number of stations is not a good predictor of pedestrian/LRT accidents. Table 8. Summary of Regression Results F-Statistic R Squared Value Significance T-Statistic Annual Revenue Service Miles 0.37 11.74 0.003 3.4 Directional Route Miles 0.32 9.51 0.006 3.1 At-Grade Track Miles 0.17 4.11 0.056 2.0 Number of Grade Crossings 0.14 3.24 0.087 1.8 Number of Stations 0.07 1.55 0.228 1.2

Figure 8. Linear Regression of Pedestrian LRT Accident Rate and Annual Revenue Service Miles 0 1 2 3 4 5 6 0 1000 2000 3000 4000 5000 6000 Annual Revenue Service Miles A cc id en ts Accidents1 Predicted Accidents1 There also appear to be relationships between many of the variables. For example, systems with a greater number of annual revenue service miles are likely to have more grade crossings and systems with more grade crossings are likely to have a greater number of at-grade track miles. Table 9 is a correlation matrix of these variables. Generally, correlations above 0.6 can be considered “significant.” Of the 10 co-variants, the correlation coefficient was above 0.6 for: ƒ Number of Grade Crossings and At-Grade Track Miles ƒ Annual Revenue Service Miles and Directional Route Miles Table 9. Correlation Matrix for Regression Variables Annual Revenue Service Miles Directional Route Miles No. of Stations No. of Grade Crossings At-Grade Track Miles Annual Revenue Service Miles 1.00 Directional Route Miles 0.82 1.00 Number of Stations 0.55 0.59 1.00 Number of Grade Crossings 0.35 0.31 0.42 1.00 At-Grade Track Miles 0.42 0.40 0.49 0.99 1.00 At-grade track miles and the number of grade crossings have the highest correlation among the predictor variables. Both of these variables were also found to have somewhat weak relationships with the number of pedestrian/LRT accidents. The correlation between annual revenue service miles and directional route miles was also relatively high (0.82). Therefore, either variable would be equally effective at predicting the number of pedestrian/LRT accidents.

TRANSIT AGENCIES Accident data was also requested from the transit agencies surveyed as part of Task 2 of this project. Of the 11 agencies who responded to the survey, detailed accident data was obtained from: ƒ Los Angeles County Metropolitan Transportation Authority (LA Metro) ƒ Santa Clara Valley Transportation Authority (VTA) ƒ San Diego Trolley, Inc. Other agencies did not have the information available or had too little data to provide a meaningful analysis. Follow-up phone calls with staff were also conducted to gather anecdotal information and ask questions regarding specific pedestrian safety issues facing each agency. Table 10 presents the aggregated historical pedestrian/LRT data for the three agencies based on accident location. Similar to the NTD data, accidents are classified as occurring at crossings, stations, or other. Consistent with the national data, a majority (64%) of all accidents occurred at crossings. However, more accidents occurred at/near stations for VTA and San Diego than at grade crossings.5 These and other trends are further analyzed in the following section. The information that is reported by the transit agencies is limited. Table 10. Summary of Historical Pedestrian/LRT Accident Data for San Diego, VTA, and LA Metro Location San Diego VTA LA Metro Total Crossings 9 18 89 116 Station 12 22 17 51 Other 4 0 6 10 Total 25 40 112 177 Notes: San Diego (August 1999 – May 2004) VTA (September 1988 – March 2004) LA Metro (July 1990 – March 2004) Los Angeles County Metropolitan Transportation Authority (LA Metro)6 LA Metro operates approximately 55 route miles of light rail service on three lines, the Blue, Gold, and Green lines. Figure in Appendix B is a map of the Metro Rail system. The Green line is fully within an exclusive right-of-way and therefore is not included in this analysis. The 14- mile Gold Line opened for revenue service in July 2003. Historical accident data from this new system is not yet available. The Long Beach Blue Line, which runs approximately 22 miles from downtown Los Angeles south to downtown Long Beach, has experienced a high level of vehicle/LRT and 5 It is important to note that the classifications for accidents at stations, grade crossings, etc. may not be the same between transit agencies or with the NTD. 6 Accident data in this section comes from the Los Angeles County Metropolitan Transportation Authority, Summary of Metro Blue Line Train/Vehicle and Train/Pedestrian Accidents (July 1990 – March 2004), April, 2004, and communications with Vijay Khawani, Metro Rail Operations Safety Department.

pedestrian/LRT accidents. The Blue Line opened for revenue service in July 1990. As of the end of the first quarter of 2004, there have been a total of 679 accidents on the Blue Line (excluding 12 suicides). Of these accidents, 112 have involved pedestrians. While pedestrian/LRT accidents represent only 16% of the total accidents, they account for slightly over half (51%) of the 55 total fatalities. The Blue Line has three distinct route segments defined as the Los Angeles Street Running (LASR), Cab Signal (CS), and Long Beach Street Running (LBSR). A vast majority of the pedestrian/LRT accidents (91 of the 112, or 81%) have occurred along the Cab Signal segment, which extends from approximately the Washington Station in Los Angeles to past the Willow Station in Long Beach. The Blue Line generally operates on a semi-exclusive/fenced right-of- way along this 15-mile segment which has approximately 31 grade crossings (including pedestrian only crossings). A portion of the Cab Signal alignment is adjacent to a freight rail corridor, which means that there are up to four parallel tracks (1-2 freight, 2 LRT). In terms of safety devices, all of the motor vehicle crossings along the Cab Signal segment have warning devices installed such as gates, flashing lights, and crossing bells. The vehicle crossing at 124th Street has four quadrant gates whereas all other crossings have two quadrant gates. LA Metro is currently expanding the four quadrant gate system to 11 additional crossings on the Blue Line, the design for which is currently in progress. There are also at least three pedestrian only crossings, which do not have flashing lights, or bells. In addition to these active warnings, LA Metro sounds a train horn in advance of each crossing (vehicle and pedestrian). The horns are set to meet the CPUC minimum sound level of 75 dBA at 100 feet. Although the exact sound level of the horn is not known, it is believed to be about 95 dBA at 100 feet. In addition to the warning devices listed above, LA Metro has also installed supplementary safety devices at several of the crossings. Specifically, there are pedestrian swing gates at the Imperial Station, Artesia Station, and Willow Street Station. Also, a fiber-optic “2nd train coming” sign has been installed at the Vernon Avenue pedestrian grade crossing as part of a TCRP demonstration project.7 Table 11 summarizes the fatal and non-fatal accidents by location. The vast majority of accidents occurred at grade crossings (89 out of 112, or 79%). All of the accidents were at crossings with active visual and audible warnings, with 58 of the 89 accidents at crossings with gates.8 7 TCRP Project A-5A, “Active Train Coming/Second Train Coming Sign Demonstration Project.” 8 Crossing accidents do not include those accidents identified as at stations or median barriers.

Table 11. Summary of LA Metro Blue Line Pedestrian/LRT Accident Data (July 1990 – March 2004) Location Fatal Non-Fatal Total Stations 16 1 17 Gated Crossings -- -- -- Gates, Flashing Lights, Bells/Traffic Signal 23 9 32 Gates, Flashing Lights, Bells 12 7 19 Double Gates, Flashing Lights, Bells/Traffic Signal 2 5 7 Signalized Crossings -- -- -- Signal, No Left Turn 3 -- 3 Walk/Don’t Walk 2 -- 2 Traffic Signal 10 2 12 Ped Crossings (Flashing Lights, Bells) 12 2 14 Other (Median Barrier) 4 2 6 Total 84 28 112 Note: Double gates = four quadrant gates Figure 9 shows the frequency of accidents per crossings, including crossings at stations. It excludes the other (median barrier) pedestrian/LRT incidents. As can be seen, a disproportionately large number of accidents (37 accidents) have occurred at only 4 locations, all of which are along the Cab Signal segment. All four of these crossings are also at stations: ƒ Vernon Avenue (14 accidents). This grade crossing is protected by gates, flashing lights, bells, pedestrian gates, 2nd Train Coming sign, and traffic signals. ƒ Florence Avenue (8 accidents). This grade crossing is protected by gates, flashing lights, bells, and pedestrian gates. ƒ Artesia Pedestrian (8 accidents). This pedestrian crossing is protected by flashing lights, bells, and swing gates. ƒ 103rd Street (7 accidents). This grade crossing is protected by gates, flashing lights, bells, pedestrian gates, and traffic signals.

Figure 9. Histogram of LA Metro Blue Line Pedestrian/LRT Accidents per Crossing (July 1990 through March 2004) 0 2 4 6 8 10 12 14 16 # C ro ss in gs 1 2 3 4 5 6 7 8 9 10 11 12 13 14 # of Accidents Figure 10 shows the number of accidents for both vehicle/LRT and pedestrian/LRT accidents by route segment. Vehicle and pedestrian accidents are occurring along different route segments. This difference becomes even more pronounced when comparing accidents by segment on a percentage basis, which is shown in Figure 11. We can not determine why there are so many accidents on the CS segment of the Blue Line, and without doing extensive research we can not derive any meaningful conclusions. Figure 10. No. of Vehicle/LRT and Pedestrian/LRT Accidents by Route Segment (July 1990 through March 2004) 0 100 200 300 400 500 600 LASR CS LBSR Total Vehicle Pedestrian

Figure 11. Percent Vehicle/LRT and Pedestrian/LRT Accidents by Route Segment (July 1990 through March 2004) 0% 20% 40% 60% 80% 100% LASR CS LBSR Vehicle Pedestrian Figure 12 shows the accident rate by segment and in total on the Blue Line since just after it opened in July 1990. Although there is a spike in the Cab Signal segment and total accidents in 1998 and a dip in 1996, pedestrian/LRT accidents rates have been fairly steady over the past 13 years, averaging slightly over 7 per year on the Cab Signal segment. Although there may have been minor increases in total annual revenue service miles over the years (i.e. change from 1 to 2 and 2 to 3 car trains), the general operation of the Blue Line has remained consistent. To our knowledge LA Metro has not installed any innovative audible devices or wayside horns. Figure 12. Pedestrian/LRT Accidents over Time

There are several factors that may be contributing to the high accident rate on the Blue Line. The first is speed. The Blue Line generally operates at speeds greater than 55 km/hour along the Cab Signal route segment. The second is the adjacent freight railroad (UPRR). These factors are discussed in more detail in section 2.3.6.3 of TRCP Report 69 and in the TCRP Research Results Digest (November 2002, Number 51). As previously discussed, LA Metro has installed swing gates and fiber optic active warning signs to improve pedestrian safety at certain crossings. Other pedestrian safety measures include pedestrian gates, flashing train headlights, fencing, and delineated crosswalks. From an education standpoint, LA Metro has provided safety education seminars to all schools, and at events such as fairs, community meetings, church meetings and other social gatherings. LA Metro has also produced 9 PSAs and computer generated animation videos that are currently being shown on local cable and TV stations. LA Metro has also invested in a mobile safety education theatre called the Metro Experience that is used to spread the rail safety message. Despite all of LA Metro’s efforts, accidents continue to occur. The data indicates that pedestrian/LRT accidents in the CS segment continue to be a challenge to mitigate. Santa Clara Valley Transportation Authority (VTA) The VTA operates three LRT lines: Guadalupe, Tasman, and Almaden. Figure in Appendix B is a map of the VTA rail system. A summary of the mileage by alignment type for all three lines is given in Table 12. A majority of the alignment is in a street median with no fence where the maximum speed is 55 km/hour (35 mph). The crossings in these locations are typically traffic signal controlled and do not have crossing bells. Also, consistent with California Public Utilities Commission requirements, on-vehicle audible warnings are not routinely sounded in advance of the crossings. The maximum speed is 15 km/hour (10 mph) in the pedestrian mall. Table 12. VTA Mileage by Alignment and Line Alignment Guadalupe Tasman Almaden Total Exclusive 9.5 1.58 0 11.1 Semi-Exclusive 0 4.47 1.18 5.7 Street Median/Curb (no fence) 8.35 9.89 0 18.2 Non-Exclusive/Mixed Traffic 0.1 0 0 0.1 Non-Exclusive/Pedestrian Mall 1.33 0 0 1.3 Total 19.3 15.9 1.2 36.4 Note: The Guadalupe line opened in 1997, the Almaden extension in 1991, and the Tasman line in 1999. Source: VTA, 2004. Table 13 is a breakdown of the fatal and non-fatal accidents by alignment type. All 40 of the pedestrian/LRT accidents that have occurred since revenue service began in 1987 have been on the Guadalupe Line. This is somewhat surprising since 12 of the 40 accidents have occurred since 1999 (when the Tasman line opened). Also, the Tasman line has nearly 10 miles of street median/curb (no fence) alignment. Crossing treatments are similar on all three lines and no unique audible devices have been installed. Figure 13 shows the yearly number of accidents since 1988. It should be noted that there is a discrepancy between the VTA data in Figure 13 and the NTD data in Table 1 for VTA. We believe the discrepancy exists because of differences in the reporting requirements.

Table 13. VTA Pedestrian/LRT Accidents by Alignment Type (September 1990 through March 2004) Non-Fatal Alignment Fatal Major Minor Total Exclusive 0 0 0 0 Street Median/Curb (no fence) 3 9 7 19 Non-Exclusive/Mixed Traffic 0 0 0 0 Non-Exclusive/Pedestrian Mall 0 5 16 21 Total 3 14 23 40 Note: No information provided regarding difference between major and minor non-fatal injuries. This table includes one accident that occurred in 2004. Figure 13. Number of VTA Pedestrian/LRT Accidents over Time (1988 to 2003) 0 1 2 3 4 5 6 7 # of A cc id en ts 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 Year In addition to the fact that all of the reported accidents have been on the Guadalupe line, all of the accidents have occurred along segments where the operational speed is less than 55 km/hour (35 mph), which represents only 50% of the total line miles. Table 14 includes the historical accident rate per mile for each alignment segment. In particular, the accident rate per mile for the downtown pedestrian mall is very high, but the severity of the accidents is low. One possible reason is the lower operating speed in this area. The pedestrian mall is effectively a continuous crossing in a busy/pedestrian heavy urban environment.

Table 14. VTA Pedestrian/LRT Accident Rate by Alignment Type (September 1990 through March 2004) Alignment Miles Accidents Accident Rate (#/ Route Mile) Exclusive 9.5 0 0 Street Median/Curb (no fence) 8.35 19 2.3 Non-Exclusive/Mixed Traffic 0.1 0 0 Non-Exclusive/Pedestrian Mall 1.33 21 15.8 Total 19.3 40 2.1 Table 15 shows the number of pedestrian/LRT accidents at either traffic signal controlled grade crossings or at/near stations, with the majority (63%) of the accidents at/near stations. Table 15. VTA Pedestrian/LRT Accidents by Location (September 1990 through March 2004) Alignment Traffic Signal Station Total Street Median/Curb (no fence) 9 10 19 Non-Exclusive/Pedestrian Mall 9 12 21 Total 18 20 40 San Diego Trolley The San Diego Trolley is comprised of two lines totaling approximately 60 route miles. Figure in Appendix B is a map of the San Diego Trolley System. The Blue Line runs north-south from Mission Valley to the U.S.-Mexico Border at San Ysidro. The northern portion of this line operates at speeds up to 90 km/hr (55 mph) along a semi-exclusive right-of-way. The downtown segment is approximately 2 miles long and is in a street median/curb (no fence) alignment. South of downtown, the line operates at speeds up to 55 km/hr (35 mph) on a semi- exclusive right-of-way shared with the San Diego and Imperial Valley Railway during non- revenue service hours. There are approximately 35 grade crossings along the southern portion of this route. The Orange Line runs east-west from downtown San Diego to Santee. From downtown east to approximately 32nd Street, the Orange Line operates at speeds up to 55 km/hr (35 mph) on Street Median/Curb (no fence) alignment. To the east, the alignment is semi-exclusive right-of- way at speeds up to 90 km/hr (55 mph). Between August 1999 and December 2003, there have been a total of 25 pedestrian/LRT accidents on the San Diego Trolley System. Only 2 of these 25 accidents have been fatalities. Table 16 lists the accidents by line and location. As can be seen, nearly 80% of all accidents have occurred on the Blue Line even though the Blue Line and Orange Line are similar in route length and have approximately the same number of grade crossings. In addition, nearly half of all the accidents have occurred at or near stations.

Table 16. Summary of San Diego Pedestrian/LRT Accidents by Location and Line (September 1990 through March 2004) Location Blue Orange Total Gates, Flashing lights, Bells 4 1 5 Traffic Signal 2 2 4 Station 9 3 12 Exclusive Right-of-Way 4 0 4 Total 19 6 25 The highest number of reported accidents at any one location was at or near the Palm Avenue Station (4 accidents). According to TCRP Report 69, there is a tall wall separating the freight railroad tracks from the back side of the LRT platform, which limits the ability of pedestrians walking on the sidewalk just outside the LRT station to see an approaching LRV on the tracks on the other side of the wall. We are not aware of the installation of any innovative audible devices at this location. According to the NTD, the San Diego Trolley system includes approximately 97 track miles, 88 of which are either on semi-exclusive or non-exclusive rights-of-way. By grouping incidents together either by using direct information regarding their location or by the types of crossing protections indicated, we were able to compare the incident rates by alignment type to the track miles by alignment type. As can be seen in Table 17, the percentage of track miles and % incidents by alignment are relatively uniform, with most occurring along the semi-exclusive track segments. Table 17. Summary of San Diego Pedestrian/LRT Incidents by Alignment Type & Track Miles Alignment Track Miles Incidents % Track Miles % Incidents At-Grade Semi-Exclusive 80 20 82% 80% Non-Exclusive, Mixed 8 5 8% 20% Subtotal 88 25 90% 100% Exclusive 9 NA 10% NA Total 97 25 100% 100% Notes: August 1999 to May 2004 CONCLUSIONS Based on the data collected, it is difficult to determine the effectiveness of audible warnings in preventing pedestrian/LRT accidents. First, it is not possible to know why the accident occurred. Neither the NTD nor the agency data provide definitive causes for each accident or include a first-hand account from the pedestrian’s perspective. Second, the reporting does not

always include the procedures for using audible warning devices prior to each accident. Some generalized information is available based on the information obtained from Task 2 of this project. Nonetheless, we cannot accurately determine what role, if any, the audible warnings played in the accident. Third, we do not have any data to determine whether, if sounded, an audible warning helped prevent an accident. Near misses are not recorded in the NTD or independently by most transit agencies. In addition, anecdotal information suggests that on a few occasions emergency horns actually contributed to an accident by startling and confusing pedestrians. In one instance, an elderly pedestrian outside the right-of-way was confused by the emergency horn and walked into the path of the train. Following is a review of what we have found in analyzing the accident data. ƒ There are relatively few annual pedestrian/LRT accidents. We can generally conclude that existing grade crossing measures and LRT operating procedures are relatively effective at preventing pedestrian/LRT accidents. ƒ Pedestrian/LRT accidents are more likely to result in fatalities than vehicle/LRT collisions. This is not an unusual result considering the lack of physical crossing control for pedestrians. ƒ The majority of pedestrian/LRT accidents occur at grade crossings. Since most LRT systems operate along alignments with some sort of physical barrier between pedestrians and the train (curbs, fences, etc.), most pedestrians are channeled to designated crossings. As a result, this is where accidents are most likely to occur. The NTD appears to group all accidents together that occur at revenue facilities. This appears to include accidents at grade crossings near stations and on the platforms, therefore it is difficult to determine exactly where some accidents at revenue facilities occurred. Adding accidents at grade crossings near stations would increase the number and percentage of accidents that occur at grade crossings. ƒ Most accidents occur at locations with active protective devices. Based on information obtained in Task 2 of this project, most LRT crossings, including pedestrian-only crossings, have some type of active crossing control device. ƒ The higher number of accidents at traffic signal controlled crossings versus gated crossings appears to suggest that lack of visual, physical, and/or audible measures decreases pedestrian safety. This preliminary conclusion is supported by common sense and casual observation, given that beyond the sounding of an on-vehicle audible warning (which is not always sounded at traffic signal controlled crossings)9, there are no standard visual or audible cues that a train is approaching. ƒ It may not be possible to use audible devices to protect against many causes of pedestrian/LRT accidents. In particular, accidents attributed to intoxication and trespassing are not a function of audible warnings or a lack thereof. Furthermore, distraction from cell phones and headsets may be difficult to overcome using audible devices. Many accidents occurred at locations with physical (gates), audible (bells and horns), and visual warnings (flashers and lights). These accidents are likely due to risky pedestrian behavior that is independent of the degree of crossing protection. However, there seem to be some exceptions to this trend that warrant further consideration. For example, there are 9 The California Public Utilities Commission General Orders, which regulate LRT systems in California and are often used as a model for transit agencies in other states, only require the sounding of an on- vehicle audible warning at gate-protected crossings.

situations where audible warnings are ignored not necessarily because of risky behavior but instead due to other factors such as: o Second train coming. This type of accident occurs when a pedestrian enters a crossing against the active crossing control devices after a train clears the crossing and the pedestrian is unaware of a train approaching from the opposite direction. o Active joint use corridors. In situations where both slower moving and louder freight trains share crossing control devices with faster and quieter LRT systems, some pedestrians enter a crossing against the active protection devices thinking that they are warning the approach of a freight train rather than the LRV. ƒ There is substantial variability in accident rates among transit agencies. Some of this variability is explained by the size of the LRT system (annual revenue service or directional route miles); however, much of it is not. Therefore, site- or alignment-specific factors that are unique to transit agencies may be significant contributors to pedestrian/LRT accidents. The variation in accident trends seems to indicate that nation wide statistics have limited usefulness when looking at individual transit agencies.

ANNEX A: TRANSIT AGENCY DATA

Table A-1. Summary of Transit Agency Alignment and Track Mileage Track Mileage At Grade Agency Exclusive Right-of- Way with Cross Traffic Mixed and Cross Traffic Total Elevated on Structure Elevated on Fill Open- cut Subway Total Miles City of Detroit 0.0 0.0 1.3 1.3 0.0 0.0 0.0 0.0 1.3 Cleveland-RTA 12.5 14.5 0.0 27.0 0.9 2.0 3.1 0.0 33.0 Dallas Area Rapid Transit 9.0 54.0 3.0 66.0 9.0 0.0 2.0 6.0 83.0 Denver RTD 15.2 4.3 0.0 19.5 3.4 7.0 2.2 0.0 32.1 Kenosha Transit 0.0 0.5 1.4 1.9 0.0 0.0 0.0 0.0 1.9 King County DOT 0.0 2.1 0.0 2.1 0.0 0.0 0.0 0.0 2.1 Los Angeles County MTA 4.0 31.0 0.0 35.0 12.0 24.0 15.0 1.0 87.0 Maryland MTA 37.0 10.0 3.0 50.0 2.0 0.0 0.0 0.0 52.0 Mass Bay Transp. Auth. 29.0 29.0 2.0 60.0 4.0 0.0 0.0 14.0 78.0 Memphis 0.0 2.4 3.7 6.1 0.0 0.0 0.0 0.0 6.1 New Jersey Transit 17 7 2 26.0 0 0 3 0 29 New Orleans 16.0 0.0 0.0 16.0 0.0 0.0 0.0 0.0 16.0 Niagara Frontier 1.6 2.8 0.0 4.4 0.0 0.0 0.0 9.7 14.1 Pittsburg Transit 34.1 0.0 4.4 38.5 1.9 0.0 0.0 4.4 44.8 Portland Tri-Met 12.0 45.0 5.0 62.0 2.0 1.0 10.0 6.0 81.0 Sacramento RT 9.9 19.4 6.8 36.1 1.8 1.5 0.0 0.0 39.4 San Diego 0.0 80.2 8.0 88.2 6.0 1.8 0.6 0.0 96.6 San Francisco MUNI 5.0 6.0 48.0 59.0 0.0 0.0 0.0 15.0 74.0 Santa Clara VTA 12.0 37.0 0.0 49.0 1.0 7.0 1.0 0.0 58.0 SE Pennsylvania 1.0 21.0 145.0 167.0 0.0 0.0 0.0 4.0 171.0 St. Louis 44.0 1.0 0.0 45.0 5.0 10.0 11.0 3.0 74.0 Utah Transit Authority 0.0 30.2 4.0 34.2 0.0 0.0 0.0 0.0 34.2

Table A-2. Summary of NTD Transit Agency Alignment and Operating Data Number of Crossings Transit Service Supplied Agency with Cross Traffic Mixed and Cross Traffic Total Directional Route Miles Stations Number of Trains in Operation (Average Weekday) Annual Train Miles Annual Train Revenue Miles Annual Train Hours Annual Train Revenue Hours City of Detroit 0 8 8 1.3 8 0 13.0 11.2 3.1 1.9 Cleveland-RTA 22 0 22 30.4 34 15 951.3 938.3 62.6 61.2 Dallas Area Rapid Transit 70 13 83 71.9 22 24 2,279.3 2,260.2 124.8 122.6 Denver RTD 39 0 39 31.6 20 17 1,691.8 1,463.7 107.4 89.7 Kenosha Transit 7 12 19 1.9 1 1 16.8 16.5 2.6 2.4 King County DOT 14 0 14 3.7 9 3 39.9 39.8 11.6 11.5 Los Angeles County MTA 77 0 77 82.4 36 31 3,143.6 3,114.6 135.5 130.7 Maryland MTA 35 17 52 57.6 32 17 1,668.6 1,649.0 103.6 102.1 Mass Bay Transp. Auth. 56 0 56 51.0 78 80 4,246.7 4,238.2 283.1 282.5 Memphis 0 0 0 5.8 28 10 309.7 308.1 38.7 38.4 New Jersey Transit 16 11 27 25 26 24.0 1,183.8 1,183.8 118.5 118.516 New Orleans 124 0 124 16.0 9 20 681.2 648.2 93.7 77.2 Niagara Frontier 8 0 8 12.4 14 10 428.2 421.1 36.9 35.5 Pittsburg Transit 34 5 39 34.8 13 33 1,454.1 1,394.6 112.8 107.9 Portland Tri-Met 115 81 196 81.3 47 29 3,218.2 3,203.2 202.2 201.0 Sacramento RT 34 56 90 40.7 29 8 971.6 953.5 47.5 46.4 San Diego 70 26 96 96.6 49 27 2,687.7 2,661.3 141.6 138.8 San Francisco MUNI 27 324 351 72.9 11 126 5,458.9 5,458.9 571.3 571.3 Santa Clara VTA 97 0 97 58.4 49 23 2,031.4 1,960.8 137.0 131.9 SE Pennsylvania 43 1,659 1,702 69.3 64 110 3,135.9 3,027.9 320.1 310.3 St. Louis 23 0 23 68.8 26 21 2,679.2 2,658.8 106.0 91.1 Utah Transit Authority 33 32 65 34.2 20 12 1,080.5 1,075.1 95.9 95.4

Table A-3. Summary NTD Pedestrian/LRT Data for 2003 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party Portland Tri-Met Right of way/roadway: Other Pedestrian came down the steps and walked into side of train, making contact just behind operator's right side window. Man limped away. A MOW unit made him wait for a supervisor. Pedestrian was transported. 2-25-03 0 Trespasser Portland Tri-Met Exclusive right of way: At grade Right of way/roadway: Other Operator stated he made service stop 18 ties south of berthing marker which allows sound of bell to radiate from glass on shelter. Passenger detrained and was running on the platform same time train started to depart station. Operator made two bell warn 4-01-03 0 Other Portland Tri-Met Exclusive right of way: At grade Passive Devices : Stop Sign Other (specify in box below) Bicyclist was traveling along side the train and cut in front of it. Bike was knocked clear and pedestrian went under the train. 6-23-03 1 Trespasser Portland Tri-Met Right of way/roadway: Other A senior citizen walked out in front of the moving train. 6-03-03 0 Trespasser Portland Tri-Met Exclusive right of way: At grade Revenue facility: Platform Train was headed eastbound into the Rockwood platform. While dropping speed to enter the intersection, operator was scanning left, right, and forward for pedestrians, etc. Op was paying extra close attention to the west end of the platform because it was 10-10-03 1 Other Portland Tri-Met Right of way/roadway: Other After servicing the Quatama (Hillsboro) platform westbound, operator proceeded on a green indication with two bell warnings. The pedestrian crosswalk was empty. After leaving, the operator heard a noise on the left hand side of the train and pulled the 11-12-03 0 Other

Table A-3. Summary NTD Pedestrian/LRT Data for 2003 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party Portland Tri-Met Right of way/roadway: Other Two girls were "horsing" around on the platform and one of them pushed the other between the coupled cars. Victim dragged perhaps 12 feet. Not life-threatening injuries - minor. Operator saw a hand between platform and side of rear car, towards the fro 11-17-03 0 Other Niagara Frontier Right of way/roadway: Other Intoxicated pedestrian apparently walked 996 feet into train tunnel from portal gate, ignoring signs and flashing lights. Front of train struck pedestrian. 4-27-03 0 Passenger SEPTA Exclusive right of way: At grade Active Devices: Traffic Signal Right of way/roadway: Grade crossing Trolley driver pulled away from stop, lady ran in front of trolley past #4 door, she yelled out a man just ran into trolley, man was laying on sidewalk.. 5-29-03 0 SEPTA Exclusive right of way: At grade Active Devices: Traffic Signal Right of way/roadway: Grade crossing Heading east on Island Road, driver pulled to end of stand, going by crowd, then slowed down and heard a thump. Driver went to see what happened, girl was sitting on the ground near pole. Kids said she was hit by trolley. 5-19-03 0 SF MUNI Non- Exclusive right of way: Mixed traffic/LRT Other (specify in box below) Pedestrian stepped in front of the streetcar. 1-22-03 0 Other SF MUNI Exclusive right of way: Tunnel Right of way/roadway: Other In the subway between Vanness and Civic Center. The pedestrian was walking in the subway and was unable to get fully out of the way of a train. The train sideswiped him twice causing minor injuries. 7-3-03 0 Trespasser

Table A-3. Summary NTD Pedestrian/LRT Data for 2003 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party SF MUNI Non-revenue facility The employee was standing near the track and was hit by the rear of the train as it turned. 0 Employees SF MUNI Non- Exclusive right of way: Transit mall Active Devices: Train Approaching Sign Revenue facility : Other revenue facility (specify below) PCC streetcar traveling south struck a pedestrian that was inattentive while using a cell phone and stepped off the curb and into the approaching streetcar. 9-25-03 0 Revenue Facility Occupants SF MUNI Non- Exclusive right of way: Transit mall Active Devices: Train Approaching Sign Revenue facility: Other revenue facility (specify below) PCC streetcar traveling south struck a pedestrian that was inattentive while using a cell phone and stepped off the curb and into the approaching streetcar. 9-25-03 0 Other SF MUNI Semi- exclusive right of way Active Devices: Traffic Signal Right of way/roadway: Grade crossing Historic streetcar attempting to come to a stop at traffic stop light, made contact with a pedestrian that was crossing in the crosswalk 12-12-03 0 Other SF MUNI Non- Exclusive right of way: Mixed traffic/LRT Passive Devices: Stop Sign Right of way/roadway: Grade crossing Male pedestrian ran into the left front side of the LRV that was moving through an intersection. 12-17-03 0 Other SF MUNI Right of way/roadway: Grade crossing F Market going straight; pedestrian stepped from the curb in front of the streetcar. The traffic light was green for the streetcar, but the pedestrian stepped into the roadway outside the crosswalk. 10-22-03 0 Other

Table A-3. Summary NTD Pedestrian/LRT Data for 2003 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party Sac RT Semi- exclusive right of way Right of way/roadway: Other Trespasser was between the rails as train approached. Trespasser made no attempt to clear the tracks and the train was unable to stop short of the impact. 3-21-03 1 Trespasser San Diego Exclusive right of way: At grade No control device Right of way/roadway: Other 3-13-03 1 Trespasser San Diego Semi- exclusive right of way No control device Right of way/roadway: Other Eastbound Train #5 departed Harborside Station at 11:35PM on the Blue Line and reported observing an object between running rails. The T/O discovered blood and pieces of flesh. Trains emergency brakes were applied but the train was not able to stop short 8-02-03 1 Trespasser LACMTA Semi- exclusive right of way Active Devices: Gates (no median barrier) Right of way/roadway: Grade crossing Northbound Train 12 was approaching 119th St grade crossing in coast mode, when an individual on a bicycle bolted past the flashing lights and bells on the grade crossing warning devices into the grade crossing. 01-08-03 0 Other LACMTA Semi- exclusive right of way Active Devices: Gates (no median barrier) Right of way/roadway: Grade crossing Northbound Train 9 approaching Century Blvd. Crossing gates were down but a bicyclist went around them eastbound. Train Operator activated emergency braking and horn but unable to avoid contact. 2-12-03 1 Other LACMTA Semi- exclusive right of way Active Devices: Traffic Signal Right of way/roadway: Grade crossing Northbound Train 8 gave audible warning prior to entering Long Beach Blvd intersection into Willow Station. Operator proceeded upon receipt of wayside "proceed" indication. Operator observed 3-14-03 0 Other

Table A-3. Summary NTD Pedestrian/LRT Data for 2003 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party bicyclist moving across parking lot slowly, gave audible warning LACMTA Semi- exclusive right of way Revenue facility: Platform Northbound train was approaching Firestone Station in full service braking when he noticed an individual lying between the running rails in a fetal position. Operator went into emergency braking but was unable to avoid contact. 4-15-03 0 Revenue Facility Occupants LACMTA Semi- exclusive right of way Active Devices: Gates (median barrier) Right of way/roadway: Grade crossing Southbound Train 15 was approaching Imperial Pedestrian Crossing. Train Operator observed male pedestrian wearing headsets enter crossing. The pedestrian was not heeding or attentive to on-coming train. Train Operator applied emergency brakes and used both. 7-2-03 0 Revenue Facility Occupants 1 The Event Descriptions were truncated in records obtained from the NTD.

Table A-4. Summary NTD Pedestrian/LRT Data for 2002 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party King County DOT Non- Exclusive right of way: Mixed traffic/LRT Active Devices: Traffic Signal Right of way/roadway: Grade crossing A pedestrian walked into the right front side of the passing streetcar proceeding through the intersection. The pedestrian sustained facial/head injures and was transported to the hospital. 5-9-02 0 Other Tri-Met Right of way/roadway: Grade crossing As I approached SW 9th Avenue, I noticed a bicyclist on the sidewalk on the SE corner headed north. I started showing down as I was not sure if the bicyclist would stop. The bicyclist did stop. A second bicyclist coming from the same direction did not 6-16-02 0 Other Tri-Met Revenue facility : Platform A person in a wheelchair made contact with the side of the train coming into platform WB. Witnesses said person was intoxicated. Was treated and released same day. 8-25-02 0 Other MBTA Right of way/roadway: Grade crossing Bus diversion to allow maintenance crews to make track repairs. MOW crew on ROW as the KGT Boom Truck operator was attempting to place it on the track on a 5.7% grade level. As operator lowered the rear wheels the KGT started to free wheel down the gra 8-11-02 0 Employees SEPTA Right of way/roadway: Grade crossing Trolley was traveling westbound when a pedestrian suddenly walked into path of trolley. 7-20-02 0 Trespassers

Table A-4. Summary NTD Pedestrian/LRT Data for 2002 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party SEPTA Exclusive right of way: At grade Active Devices: Traffic Signal Right of way/roadway: Grade crossing Discharged passenger claimed the operator brushed her arms when taking off. Mentor saw two ladies on the platform when operator pulled off, both ladies were standing.. 9-23-02 0 SEPTA Exclusive right of way: At grade Active Devices: Traffic Signal Right of way/roadway: Grade crossing Going through intersection, 3 people walking across not watching traffic. As I got close, I blew the horn, 3rd person never looked up and walked into the path of the trolley, trolley was put into emergency, stopped as contact was made. 10-17-02 0 MD MTA Non- Exclusive right of way: Mixed traffic/LRT Active Devices: Traffic Signal Right of way/roadway: Intersections Pedestrian stepped of roadway median strip on MLK Blvd@Read street into a southbound train on ML1. 4-3-02 1 Other MD MTA Non- Exclusive right of way: Mixed traffic/LRT Right of way/roadway: Intersections Train 24 had lady walk into side of train 5-17-02 0 Other MD MTA Exclusive right of way: At grade Right of way/roadway: Other Dark area, trespasser 11-10-02 1 Trespassers DART Exclusive right of way: At grade Active Devices: Gates (median Right of way/roadway: Grade crossing A pedestrian disregarded flashing light and bell at grade crossing. 1-23-02 0 Other

Table A-4. Summary NTD Pedestrian/LRT Data for 2002 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party barrier) UTA Shared track/corridor (LRT/FRA): Temporal separation Active Devices: Gates (median barrier) Right of way/roadway: Grade crossing Northbound train (1005,1010,1007) struck a pedestrian walking Westbound across the main line tracks at 9400 South intersection (Grade Crossing) Gates, Lights, Bells were all operational at the time of incident. 1-4-02 1 Other UTA Non- Exclusive right of way: Mixed traffic/LRT No control device Right of way/roadway: Other Train #1016B was arriving at the 900 East Station platform traveling at approximately 20 mph. An individual male (48 yr.) was walking across the street to get to the Light Rail station ("jay-walking). Operator observed the pedestrian walking in the far 11-3-02 1 Trespassers VTA Non- Exclusive right of way: Mixed traffic/LRT Active Devices: Traffic Signal Right of way/roadway: Intersections Pedestrian was crossing against the red light and ran right in front of the train. He was struck by train at the right front corner at the right front headlight turn light/brake light assembly area of LRV. He received medical attention at the scene then w 8-20-02 1 Other SF MUNI Right of way/roadway: Grade crossing Male walked in front of car and was struck as car started out. 2-9-02 0 SF MUNI Right of way/roadway: Intersections 3-9-02 0 SF MUNI Right of 4-1-02 0

Table A-4. Summary NTD Pedestrian/LRT Data for 2002 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party way/roadway: Grade crossing SF MUNI Right of way/roadway: Other alleged intoxicated pedestrian grabbed unto trunk of auto. As auto left scene, pedestrian fell from auto into path of the streetcar. Left front wheel of streetcar made contact with pedestrian resulting in severe injuries. 10-30-02 0 Other SF MUNI Non- Exclusive right of way: Mixed traffic/LRT Other (specify in box below) Pedestrian fell in the coupled area between the two cars of a two-car trains 11-16-02 1 Other San Diego Exclusive right of way: At grade No control device Right of way/roadway: Other 17 year old female jumped into the path of westbound Train #52 east of Amaya Station and the grade crossing. Trespasser sustained major injuries. 5-27-02 0 Trespassers San Diego Exclusive right of way: At grade No control device Right of way/roadway: Other Approximately 182 feet west of the 47th Street Station, a white male adult walked out of the early morning darkness and into the path of the approaching eastbound train #51. The train operator applied the emergency brakes but was unable to stop before co 11-14-02 0 Trespassers LACMTA Semi- exclusive right of way No control device Other (specify in box below) A Latin female patron walked into the side of moving northbound train. The operator was approaching Pico Station and passed the pedestrian crosswalk leading into the station. Witness statement indicates patron walked 2-11-02 0 Other

Table A-4. Summary NTD Pedestrian/LRT Data for 2002 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party into the side of the train. LACMTA Semi- exclusive right of way Active Devices: Gates (no median barrier) Right of way/roadway: Grade crossing Train 9 was southbound and made contact with the head of a man who was bending over. The operator and the pedestrian were transported. 4-8-02 0 Employees LACMTA Semi- exclusive right of way Active Devices: Gates (no median barrier) Right of way/roadway: Grade crossing Train 9 was southbound and made contact with the head of a man who was bending over. The operator and the pedestrian were transported. 4-8-02 0 Other LACMTA Semi- exclusive right of way Active Devices: Gates (no median barrier) Revenue facility: Platform Northbound train departing Compton station sounding audible warning (gong) continuously. The train operator observed a previous passenger run in front of the train eastbound on the pedestrian crossing and then a female pedestrian looking northbound walk w 7-18-02 0 Other LACMTA Semi- exclusive right of way Active Devices: Gates (no median barrier) Revenue facility: Platform Northbound train departing Compton station sounding audible warning (gong) continuously. The train operator observed a previous passenger run in front of the train eastbound on the pedestrian crossing and then a female pedestrian looking northbound walk w 7-18-02 0 Revenue Facility Occupants LACMTA Semi- exclusive Active Devices: Revenue facility: Other Southbound Train 9 was approaching Artesia Station when two pedestrians enter 11-9-02 0 Revenue Facility

Table A-4. Summary NTD Pedestrian/LRT Data for 2002 Agency Name: Alignment Type Grade Crossing Control Collision Location Event Description 1 Date Fatality Involved Party right of way Flashing Lights revenue facility (specify below) the ped gates and went into braking. The male pedestrian ran across and the female pedestrian continued walking and her right calf was scraped by the train. She refused medical Occupants LACMTA Semi- exclusive right of way Active Devices: Traffic Signal Right of way/roadway: Other Southbound Train 17 on approach to berth at San Pedro Station hit a young female who ran in front of the train. 11-22-02 0 Trespassers LACMTA Semi- exclusive right of way Active Devices: Traffic Signal Right of way/roadway: Other Southbound Train 17 on approach to berth at San Pedro Station hit a young female who ran in front of the train. 11-22-02 1 Other LACMTA Semi- exclusive right of way Active Devices: Flashing Lights Right of way/roadway: Grade crossing Southbound Train 5 was braking as it approached Artesia Station and the Ped Crossing. A female pedestrian wearing headsets and reading a newspaper entered the pedestrian crossing. The train went into emergency braking, sounding the horn, pedestrian looked 12-23-02 0 Revenue Facility Occupants 1 The Event Descriptions were truncated in records obtained from the NTD.

Table A-5. Summary of Pedestrian/LRT Accidents between 1997 and 2001 1997 1998 1999 2000 2001 Agency Accidents Fatal Accidents Fatal Accidents Fatal Accidents Fatal Accidents Fatal City of Detroit DOT -- -- -- -- -- -- -- -- 0 0 Cleveland-RTA 0 0 1 0 1 0 1 0 0 0 Dallas Area Rapid Transit 2 0 2 0 1 0 0 0 1 0 Denver-RTD 5 0 2 0 0 0 5 3 0 0 Kenosha Transit -- -- -- -- -- -- 0 0 0 0 King County DOT 0 0 0 0 0 0 0 0 0 0 Los Angeles County MTA 8 2 3 4 4 2 9 1 7 1 Maryland MTA 5 0 10 0 8 0 5 2 4 4 Mass. Bay Transp. Auth. 7 0 7 0 8 0 10 0 10 0 Memphis Area TA 0 0 0 0 0 0 0 0 0 0 New Jersey Transit 1 0 0 0 0 0 1 0 0 0 New Orleans-RTA 0 0 0 0 0 0 0 0 NR NR Niagara Frontier Trans. 0 0 0 0 1 0 0 0 0 0 Pittsburg Transit 0 0 1 1 2 0 2 0 2 0 Portland Tri-Met 3 0 3 0 7 3 4 2 2 0 Sacramento RT 0 0 2 0 4 0 2 0 4 1 San Diego NR NR 4 4 4 1 1 0 3 1 San Francisco MUNI 6 0 9 1 7 0 12 1 6 2 Santa Clara VTA 1 0 4 0 3 0 2 1 3 1 SE Pennsylvania Trans. 8 1 7 1 4 0 7 2 10 1 St. Louis Bi-State Dev. 0 0 1 0 0 0 1 0 1 0 Utah Transit Authority -- -- -- -- -- -- 1 1 0 0 Sum 46 3 56 11 54 6 63 13 53 11 -- indicates the LRT system was not operating during this year and NR indicates the agency did Not Report any accident data

ANNEX B: MAPS OF THE LA METRO, VTA, AND SAN DIEGO SYSTEMS Figure A-1. Map of the LA Metro Rail System

Figure A-2. Map of the VTA Light Rail System

Figure A-3. Map of the San Diego Trolley System

APPENDIX C-2 REVIEW OF ACCIDENT DATA (Prepared September 2004) INTRODUCTION AND SUMMARY This report is an addendum to Task 3: Evaluation of the Effectiveness of Audible Devices Based on Historical Data for TCRP Project D-10: Audible Signals for Pedestrian Safety in LRT Environments. For this addendum, the KORVE Team gathered and analyzed additional accident data from the National Transit Database (NTD) relative to pedestrian accidents at grade crossings and performed supplementary statistical analyses of this data. Because of the differences in the reporting thresholds and parameters, the previous statistical analysis was based on total pedestrian/LRT incidents over a two-year period (2002 and 2003).10,11 The supplementary analysis reported herein incorporates accident data from the periods 1997 through 2001 and limits the data to pedestrian/LRT incidents at grade crossings. The goal of the supplemental analysis is to determine whether the additional data and narrowed focus improves the predictability of pedestrian/LRT incidents based on key operational variables that relate to grade crossings. Three operating variables tested in this analysis are at- grade track miles, number of grade crossings, and total track miles. The primary conclusions are that: y The additional data greatly increased the number of observations (146). However, the total number of pedestrian/LRT incidents at grade crossings over a 7 year period for 23 LRT systems is very low (52). y The relationships between pedestrian/LRT accidents at grade crossing and operating variables are clearer when using the average annual values rather than the data from each LRT system for each year (i.e. all 146 incidents as a separate observation). y When all the LRT systems are included, there is a statistically significant relationship (T-statistic greater than 2) between pedestrian/LRT incidents at grade crossings and total track miles. However, the overall predictive power of this variable (as measured by the R-Squared value) is very weak. The relationships between pedestrian/LRT incidents and at-grade track miles and pedestrian/LRT incidents and the number of grade crossings are not statistically significant. y When one outlier is excluded for the data, there is a statistically significant relationship between pedestrian/LRT incidents and at-grade track miles, number of grade crossings, and total track miles. The predictive power also increases substantially. y Overall, the relatively low predictive power is due to the low number of accidents and the wide variation in operating variables. METHODOLOGY Accident Data For the five year period between 1997 and 2001, the NTD data was already separated between incidents with people, vehicles, and objects. From this data we isolated the “Total Collisions with People” and the “Total Collisions with People at Grade Crossings.” Next, the incidents with “Patrons” and “Employees” were removed, leaving only those pedestrian incidents at grade crossings. For the sequent two years (2002 and 2003), the data was in a different format. The first step was to identify and remove those accidents that didn’t involve “Trespassers”, “Patrons”, “Employees”, or “Occupants” (refer to the NTD for formal definitions for each of these categories). The residual incidents were assumed to involve pedestrians. From this subset of data, we extracted those pedestrian/LRT incidents that occurred at grade crossings. 10 Excluding suicides. 11 Please refer to the Task 3 report for a detailed discussion regarding the differences in reporting thresholds.

Figure 14 is a histogram showing the distribution of the average number of annual pedestrian/LRT incidents at grade crossings. As can be seen, 18 of the 23 LRT systems reported, on average, less than one incident per year between 1997 and 2003. Figure 15 is a graph showing the total number of pedestrian/LRT incidents at grade crossings per year and the average number per year for all LRT systems. More detailed information is included in Table 27 at the end of this report. Figure 14. Histogram of Average Annual Pedestrian/LRT Incidents at Grade Crossings 0 2 4 6 8 10 12 14 16 0 0.5 1 1.5 2 2.5 3 More No. of Incidents N o. o f L R T Sy st em s Figure 15. Pedestrian/LRT Incidents at Grade Crossings 0 2 4 6 8 10 12 1997 1998 1999 2000 2001 2002 2003 Year N um be r of A cc id en ts (T ot al ) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 N um be r of A cc id en ts (A ve ra ge p er S ys te m ) Total Average Operational Data In order to perform a statistical analysis on the pedestrian/LRT incidents at grade crossings for the seven year period from 1997 and 2003, operational data common to both reporting periods (1997-2001 and 2002-2003) was required. In addition, to focus on pedestrian/LRT incidents at grade crossings, the data needed to be relevant to at-grade track and grade crossings. The following data were tabulated for each of the seven years: y Total Track Miles y Total At-Grade Track Miles

y Total Number of Grade Crossings Figure 16 shows the distribution of the operational parameters listed above. These are average values for the period of time each agency reported to the NTD between 1997 and 2003. As can be seen, there is a wide variation in the number of at-grade track miles (AGTM), grade crossings (Xings), and of total track miles (TTM). Figure 16. Histogram Showing the Breakdown of Operational Parameters 0 1 2 3 4 5 6 7 0-5 5-10 10-20 20-40 40-60 60-80 80-100 >100 No. of AGTM, Xings, & TTM N o. o f L R T Sy st em s AGTM Xings TTM Statistical Analysis Single-variable and multi-variable linear regressions were run on the total number of pedestrian/LRT incidents and grade crossings (dependent variable) and the at-grade track miles, the number of grade crossings, and the total track miles (dependent variables). Because of the low number of incidents, separate analyses were not run on fatalities or injuries. Table 17 through Table 20 show the results of the various regressions. Table 17 and Table 18 are based on the average annual pedestrian/LRT incidents at grade crossings and the average annual operational parameters. Table 17 tests each variable independently whereas Table 18 is a multi-variable regression testing all three variables together. Table 19 and Table 20 include the accident data for each system for each of the seven years reported between 1997 and 2003. As can be seen, the multi-variable regressions add little value to the analysis. Table 17. Regression Results – Individual Variables Using Average Annual Values F-Statistic R-Squared Value Significance T-Stat AGTM 0.07 1.57 0.22 1.25 Xings 0.00 0.00 1.00 0.00 TTM 0.17 4.40 0.05 2.10 Table 18. Regression Results – All Variables Using Average Annual Values R-Squared F-Statistic T-Stat

Value Significance Total 0.45 5.27 0.01 AGTM -1.10 Xings -1.55 TTM 2.63 Table 19. Regression Results – Individual Variables Using All Data F-Statistic R-Squared Value Significance T-Stat AGTM 0.02 2.4 0.1 1.5 Xings 0.00 0.0 1.0 0.1 TTM 0.05 7.2 0.0 2.7 Table 20. Regression Results – Individual Variables Using All Data F-Statistic R-Squared Value Significance T-Stat Total 0.13 7.21 0.00 AGTM -1.95 Xings -1.27 TTM 3.58 In terms of outliers, one agency reports a substantially higher number of at-grade track miles, grade crossings, and total track miles. Eliminating that agency from the analysis improves the R-Squared and the T-Statistic for all three single-variable regressions. As a result, the T-Statistic is now statistically significant for at-grade track miles and number of crossings. In terms of the multi-variable regression, although R-Squared increases from 0.45 to 0.50, only the total track miles is statistically significant. Table 21 is a summary of the modified regression results using the annual average values. Figure 17, Figure 18, and Figure 19 are graphs showing the best-fit line for the annual average pedestrian/LRT incidents at grade crossings based on the regression for at-grade track miles, number of grade crossings, and total track miles, respectively (excluding the outlier). Table 22 lists the equations for the best-fit lines shown in these graphs. Table 21. Regression Results – Individual Variables Using Average Annual Values (excluding outlier) F-Statistic R-Squared Value Significance T-Stat AGTM 0.30 8.67 0.01 2.95 Xings 0.18 4.44 0.05 2.11 TTM 0.45 16.51 0.00 4.06

Figure 17. Best-Fit Line for Average Annual Pedestrian/LRT Incidents at Grade Crossings and At- Grade Track Miles 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 10 20 30 40 50 60 70 80 AGTM N o. o f I nc id en ts Total Predicted Total Figure 18. Best Fit Line for Average Annual Pedestrian/LRT Incidents at Grade Crossings and the Number of Grade Crossings 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 50 100 150 200 250 300 Xings N o. o f I nc id en ts Total Predicted Total

Figure 19. Best Fit Line for Average Annual Pedestrian/LRT Incidents at Grade Crossings and the Number of Total Track Miles 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 20 40 60 80 100 TTM N o. o f I nc id en ts Total Predicted Total Table 22. Regression Equations for Best Fit Lines Slope Intercept AGTM 0.016 -0.115 Xings 0.005 0.062 TTM 0.015 -0.214 Table 23 is a summary of the regression results excluding the outlier and using all the remaining data. Again, the R-squared, F-Statistic, and T-Statistic improve when the outlier data is removed. Although the T-Statistic is now significant for all three variables (whereas previously it was only statistically significant for total track miles), the predictive power is very low (R-Squared less than 0.15). Table 23. Regression Results – Individual Variables Using All Data (excluding outlier) F-Statistic R-Squared Value Significance T-Stat AGTM 0.08 11.4 0.00 3.4 Xings 0.09 12.8 0.00 3.6 TTM 0.13 20.0 0.00 4.5 CONCLUSIONS When LRT Systems are Included:

y R-Squared values are very low. y T-Statistics for total track miles is significant. y T-Statistics for at-grade track miles and number of grade crossings are not significant. y Annual average values are better predictors than using all the data. When the Outlier is excluded: y R-Squared values, F-Statistics, and T-Statistics increase. y T-Statistics for all three variables are statistically significant. y Annual average values are better predictors than using all the data. Other Notes A panel member requested that a chi-squared analysis be performed on the accident data for the two year period from 2002 and 2003 originally reported in the Task 3 report. However, this is no longer required as the number of years has been expanded to seven, which increases the sample size and therefore eliminates the need for the chi-squared analysis.

Table 24. Number of At-Grade Track Miles per LRT System by Year (1997-2003) Year System 1997 1998 1999 2000 2001 2002 2003 Average Baltimore-Maryland-MTA 49 49 49 49 49 50 50 49.6 Boston-MBTA 67 67 60 60 60 60 60 62.0 Buffalo-NFTA 4 4 4 4 4 4 4 4.4 City of Detroit DOT -- -- -- -- 2 1 1 1.4 Cleveland-RTA 27 27 27 27 27 27 27 27.0 Dallas-DART 34 34 34 34 40 66 79 45.9 Denver-RTD 10 10 10 16 16 20 20 14.3 Galveston-Island Transit 5 5 -- -- 5 5 5 4.9 Kenosha Transit -- -- -- 2 2 2 2 1.9 King County DOT 2 2 2 2 2 2 2 2.1 LACMTA-Metro 35 35 35 35 35 35 35 34.7 Memphis-MATA 7 7 6 6 6 6 6 6.2 New Jersey Transit-NJT 6 6 6 6 6 26 20 10.7 New Orleans-RTA 14 14 14 14 14 16 16 14.4 Philadelphia-SEPTA 167 167 167 167 167 167 167 166.7 Pittsburgh-PATH 40 40 40 39 39 39 39 39.2 Portland-Tri-Met 24 24 55 55 55 62 62 48.0 Sacramento-RTD 31 31 36 36 36 36 37 34.8 San Diego Trolley-SDTI 44 44 44 88 88 88 88 69.3 San Francisco-Muni 41 41 56 55 58 59 58 52.7 Santa Clara Valley-VTA 32 32 32 47 49 49 49 41.6 St. Louis-Bi-State 15 15 15 15 45 45 51 28.9 Utah Transit Authority-UTA -- -- 30 30 34 34 37 33.0 Average per LRT System 33 33 36 37 36 39 40 35 Note: Systems not reporting are listed as “--"

Table 25. Number of Grade Crossings per LRT System by Year (1997-2003) Year System 1997 1998 1999 2000 2001 2002 2003 Average Baltimore-Maryland-MTA 50 50 52 52 52 52 52 51.4 Boston-MBTA 56 56 56 56 56 56 56 56.0 Buffalo-NFTA 8 8 8 8 8 8 8 8.0 City of Detroit DOT -- -- -- -- 8 8 8 8.0 Cleveland-RTA 22 22 22 22 22 22 22 22.0 Dallas-DART 66 66 66 66 68 83 98 73.3 Denver-RTD 34 34 34 34 34 39 39 35.4 Galveston-Island Transit 57 57 -- -- 57 57 57 57.0 Kenosha Transit -- -- -- 14 26 19 19 19.5 King County DOT 14 14 14 14 14 14 14 14.0 LACMTA-Metro 77 77 77 77 77 77 77 77.0 Memphis-MATA 40 40 40 40 40 0 0 28.6 New Jersey Transit-NJT 1 1 1 1 1 27 26 8.3 New Orleans-RTA 124 124 124 124 124 124 124 124.0 Philadelphia-SEPTA 1702 1702 1702 1702 1702 1702 1702 1702.0 Pittsburgh-PATH 42 42 42 39 39 39 39 40.3 Portland-Tri-Met 74 74 111 111 111 196 196 124.7 Sacramento-RTD 86 86 90 90 90 90 93 89.3 San Diego Trolley-SDTI 70 70 96 96 96 96 96 88.6 San Francisco-Muni 191 191 191 191 351 351 351 259.6 Santa Clara Valley-VTA 64 64 64 93 97 97 97 82.3 St. Louis-Bi-State 12 12 12 12 23 23 24 16.9 Utah Transit Authority-UTA -- -- 46 46 59 65 58 54.8 Average per LRT System 140 140 142 138 137 141 142 132 Note: Systems not reporting are listed as “--"

Table 26. Number of Total Track Miles by LRT System per Year Year System 1997 1998 1999 2000 2001 2002 2003 Average Baltimore-Maryland-MTA 50.9 50.9 50.9 50.9 50.9 52 52 51.2 Boston-MBTA 77.5 77.5 77.5 77.5 77.5 78 78 77.6 Buffalo-NFTA 14.1 14.1 14.1 14.1 14.1 14.1 14.1 14.1 City of Detroit DOT -- -- -- -- 1.6 1.3 1.3 1.4 Cleveland-RTA 33 33 33 33 33 33 33 33.0 Dallas-DART 46.7 46.7 46.7 46.7 53 83 98.4 60.2 Denver-RTD 10.3 10.3 10.3 28.5 28.5 32.1 32.1 21.7 Galveston-Island Transit 4.9 4.9 -- -- 4.9 5 5 4.9 Kenosha Transit -- -- -- 1.9 1.9 1.9 1.9 1.9 King County DOT 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 LACMTA-Metro 85.8 85.8 85.8 85.7 85.7 87 85.7 85.9 Memphis-MATA 6.6 6.6 6.1 6.1 6.1 6.1 6.1 6.2 New Jersey Transit-NJT 8.3 8.3 8.3 8.3 8.3 29 21.3 13.1 New Orleans-RTA 13.7 13.7 13.7 13.7 13.7 16 16 14.4 Philadelphia-SEPTA 171 171 171 171 171 171 171 171.0 Pittsburgh-PATH 46.5 46.5 46.5 44.8 44.8 44.8 44.8 45.5 Portland-Tri-Met 34.4 34.4 71.9 71.9 71.9 81 81.3 63.8 Sacramento-RTD 34 34 39.4 39.4 39.4 39.4 40.7 38.0 San Diego Trolley-SDTI 48.3 48.3 48.3 96.6 96.6 96.6 97 76.0 San Francisco-Muni 54.2 54.2 69 70 73.3 74 72.9 66.8 Santa Clara Valley-VTA 41.1 41.1 41.1 56.3 58.9 58 58.9 50.8 St. Louis-Bi-State 36.2 36.2 36.2 36.2 73.5 74 81 53.3 Utah Transit Authority-UTA -- -- 29.6 29.6 34.2 34.2 37.3 33.0 Average per LRT System 41 41 45 47 45 48 49 43 Note: Systems not reporting are listed as “--"

Table 27. Number of Collisions with People at Grade Crossings by LRT System per Year (1997-2003) Year System 1997 1998 1999 2000 2001 2002 2003 Grand Total Average Baltimore-Maryland-MTA 0 0 0 7 0 2 0 9 1.3 Boston-MBTA 0 0 0 0 2 0 0 2 0.3 Buffalo-NFTA 0 0 0 0 0 0 0 0 0.0 City of Detroit DOT -- -- -- -- 0 0 0 0 0.0 Cleveland-RTA 0 0 0 0 0 0 0 0 0.0 Dallas-DART 0 0 0 0 0 0 0 0 0.0 Denver-RTD 0 0 0 0 0 0 0 0 0.0 Galveston-Island Transit 0 -- -- -- 0 0 0 0 0.0 Kenosha Transit -- -- -- 0 0 0 0 0 0.0 King County DOT 0 0 0 0 0 0 0 0 0.0 LACMTA-Metro 2 4 4 1 0 3 0 14 2.0 Memphis-MATA 0 0 0 0 0 0 0 0 0.0 New Jersey Transit-NJT 0 0 0 0 0 0 0 0 0.0 New Orleans-RTA 0 0 1 -- -- 0 0 1 0.2 Philadelphia-SEPTA 1 0 0 0 0 0 0 1 0.1 Pittsburgh-PATH 0 0 0 0 0 0 0 0 0.0 Portland-Tri-Met 2 0 4 0 0 1 4 11 1.6 Sacramento-RTD 0 0 0 0 0 0 0 0 0.0 San Diego Trolley-SDTI -- 4 1 0 3 0 0 8 1.3 San Francisco-Muni 0 0 0 0 0 3 2 5 0.7 Santa Clara Valley-VTA 0 0 0 0 0 1 0 1 0.1 St. Louis-Bi-State 0 0 0 0 0 0 0 0 0.0 Utah Transit Authority-UTA -- -- 0 0 0 0 0 0 0.0 Total per Year 5 8 10 8 5 10 6 52 -- Average per LRT System 0.3 0.4 0.5 0.4 0.2 0.4 0.3 -- 0.3 Note: Systems not reporting are listed as “--"

Table 28. Summary of Average Annual Operating Parameters and Pedestrian/LRT Incidents Collisions with Pedestrians At Grade Crossings Average Yearly Average Yearly System AGTM Xings TTM % AG 1 Average Yearly 2 Total 3 Total 4 Fatalities Injuries City of Detroit DOT 1.4 8 1.4 100% 0.0 0 0.0 0.0 0.0 Kenosha Transit 1.9 20 1.9 100% 0.0 0 0.0 0.0 0.0 King County DOT 2.1 14 2.1 100% 0.1 0 0.0 0.0 0.0 Buffalo-NFTA 4.4 8 14.1 31% 0.3 0 0.0 0.0 0.0 Galveston-Island Transit 4.9 57 4.9 100% 0.0 0 0.0 0.0 0.0 Memphis-MATA 6.2 29 6.2 100% 0.0 0 0.0 0.0 0.0 New Jersey Transit-NJT 10.7 8 13.1 82% 0.0 0 0.0 0.0 0.0 Denver-RTD 14.3 35 21.7 66% 0.7 0 0.0 0.0 0.0 New Orleans-RTA 14.4 124 14.4 100% 0.2 1 0.2 0.0 0.2 Cleveland-RTA 27.0 22 33.0 82% 0.0 0 0.0 0.0 0.0 St. Louis-Bi-State 28.9 17 53.3 54% 0.4 0 0.0 0.0 0.0 Utah Transit Authority-UTA 33.0 55 33.0 100% 0.2 0 0.0 0.0 0.0 LACMTA-Metro 34.7 77 85.9 40% 5.0 14 2.0 1.6 0.4 Sacramento-RTD 34.8 89 38.0 92% 0.7 0 0.0 0.0 0.0 Pittsburgh-PATH 39.2 40 45.5 86% 0.3 0 0.0 0.0 0.0 Santa Clara Valley-VTA 41.6 82 50.8 82% 0.9 1 0.1 0.0 0.1 Dallas-DART 45.9 73 60.2 76% 0.3 0 0.0 0.0 0.0 Portland-Tri-Met 48.0 125 63.8 75% 2.1 11 1.6 0.9 0.7 Baltimore-Maryland-MTA 49.6 51 51.2 97% 2.1 9 1.3 0.4 0.9 San Francisco-Muni 52.7 260 66.8 79% 5.7 5 0.7 0.6 0.1 Boston-MBTA 62.0 56 77.6 80% 5.9 2 0.3 0.0 0.3 San Diego Trolley-SDTI 69.3 89 76.0 91% 1.3 8 1.3 1.0 0.3 Philadelphia-SEPTA 166.7 1702 171.0 97% 4.4 1 0.1 0.1 0.0 1 Percent of System At-Grade (AGTM/TTM). 2 Average Yearly Pedestrian/LRT incidents. 3 Total Pedestrian/LRT incidents at grade crossings between 1997 and 2003. 4 Average Yearly Pedestrian/LRT incidents – at grade.