I-10, 35TH AVE TO SKY HARBOR BLVD PHOENIX CORRIDOR SAFETY STUDY FINAL REPORT Kittelson & Associates, Inc. Lee Engineering Works Consulting DiExSys June 2015 This report is subject to the provisions of 23 usc § 409. Any intentional or inadvertent release of this material, or any data derived from its use does not constitute a waiver of privilege to 23 usc § 409. 23 usc § 409 – discovery and admission as evidence of certain reports and surveys Notwithstanding any other provision of law, reports, surveys, schedules, lists, or data compiled or collected for the purpose of identifying, evaluating, or planning the safety enhancement of potential accident sites, hazardous roadway conditions, or railway-highway crossings, pursuant to sections 130, 144, and 148 of this title or for the purpose of developing any highway safety construction improvement project which may be implemented utilizing federal-aid highway funds shall not be subject to discovery or admitted into evidence in a federal or state court proceeding or considered for other purposes in any action for damages arising from any occurrence at a location mentioned or addressed in such reports, surveys, schedules, lists, or data. I-10 Phoenix Corridor Safety Study Final Report TABLE OF CONTENTS Executive Summary ............................................................................................................................ ES1 Introduction............................................................................................................................................ 1 Overview of the Freeway Safety Evaluation Methodology .................................................................... 4 Model Calibration ................................................................................................................................... 6 Assignment of Crashes to Segments .................................................................................................... 20 Evaluation of the Safety Performance of the I-10 Study Section ......................................................... 28 Crash Diagnostic Process ...................................................................................................................... 33 Potential Countermeasures ................................................................................................................. 53 Study Results and Recommendations .................................................................................................. 70 Application of Study Results and Tools for Future Safety Evaluations ................................................ 74 Expansion of Arizona Freeway Crash Prediction Models ..................................................................... 79 References .......................................................................................................................................... 80 Appendix A – Discussion of Safety Performance Measures ................................................................. 81 Appendix B – ISATe Output .................................................................................................................. 89 Appendix C – Segment performance measures ................................................................................... 90 Appendix D – Speed and congestion profiles....................................................................................... 91 Appendix E – 2012-2013 crash data summary by segment ................................................................. 93 Appendix F – Fatal and Serious Injury Crash Details ............................................................................ 96 Appendix G – Variable Ppeed Limit System Literature Review .......................................................... 148 Appendix H – cost estimates .............................................................................................................. 154 LIST OF FIGURES Figure 1. I-10 Study Section Map............................................................................................................... 3 Figure 2. Example of Segmentation........................................................................................................... 9 Figure 3. Example of Pinning to Collect Physical Features Data.............................................................. 13 Figure 4. Region Defining Speed-Change Related Crashes...................................................................... 20 Figure 5. Alternative Safety Performance Measures .............................................................................. 28 Figure 6. Level of Service of Safety (z-score) for Study Section Basic Segments..................................... 31 Figure 7. Expected Average Crash Frequency per Mile ........................................................................... 32 Figure 8. Two-year (2012-2013) Crashes within the I-10 Study Section ................................................. 34 Figure 9. Crash locations on one segment (Deck Park Tunnel to 7th St) ................................................ 35 Figure 10. Distribution of Crashes by Severity (Deck Park Tunnel to 7th St) .......................................... 35 Figure 11. Speed profile for eastbound I-10, MP 146.96 ........................................................................ 37 Figure 12. Speed profile for eastbound I-10, MP 146.06 ........................................................................ 37 Kittelson & Associates, Inc i I-10 Phoenix Corridor Safety Study Final Report Figure 13. Correlation between Speed and Crashes due to Slowing/Stopped Maneuvering (Segment I-10 WB, MP 144.42) ......................................................................................... 39 Figure 14. Correlation between Speed and Crashes due to Slowing/Stopped Maneuvering (Segment I-10 EB, MP 144.42) .......................................................................................... 39 Figure 15. Distribution of z-score and Cumulative CMF of the I-10 Corridor Study Segments ............... 41 Figure 16. Sample Collision Diagram for a Segment (7th St to 16th St) on a GIS Platform ..................... 45 Figure 17. VSL System in Seattle, WA ...................................................................................................... 57 Figure 18. Westbound Tunnel in the Afternoon ..................................................................................... 60 Figure 19. Westbound Lane Drop West of 7th Ave Exit; Note queuing in right two lanes ..................... 63 LIST OF TABLES Table 1. Freeway Crash Prediction Models ............................................................................................... 4 Table 2 Candidate Freeway Calibration Sections ...................................................................................... 8 Table 3. Segmented Freeway Sections ...................................................................................................... 9 Table 4. Calibration Segments ................................................................................................................. 10 Table 5. I-10 Study Section Segments ..................................................................................................... 11 Table 6. Distribution of Calibration Database ......................................................................................... 12 Table 7. Calibration Segments Crash Summary ....................................................................................... 15 Table 8. I-10 Study Section Segments Crash Summary ........................................................................... 17 Table 9. Comparative Urban Freeway Crash Frequency and Rate .......................................................... 19 Table 10. Comparative Urban Freeway Crash Severity Distribution ....................................................... 19 Table 11. Screening of Basic Segments (excluding speed-change lanes) ................................................ 21 Table 12. Screening of Speed-Change Lanes ........................................................................................... 23 Table 13. SPF Calibration for Basic Freeway Segments (excl. speed-change lanes). .............................. 24 Table 14. SPF Calibration for Speed-Change Lanes. ................................................................................ 26 Table 15. SDF Calibration for Basic Segments and Speed-Change Lanes Combined. ............................. 27 Table 16. Crash Distribution by Time-of-Day for Study Section .............................................................. 36 Table 17. Crash Distribution by Time-of-Day 3rd St HOV Ramp to 7th St Ramp..................................... 36 Table 18. Collision Type Distribution (Segment – I-10 EB, MP 145.78) ................................................... 38 Table 19. Distribution of Slowing/Stopped in the Traffic Way by Time of Day ....................................... 38 Table 20. Drivers’ Maneuvering/Action (Segment – I-10 EB, MP 145.78) .............................................. 38 Table 21. List of Selected Segments for Phase 1 Detail Crash Report Review ........................................ 40 Table 22. Sample Crash Summary for Segment I-17 (MP142.99) to 19th Ave (MP143.66) .................... 44 Table 23. Segments in the Phase II Crash Analysis .................................................................................. 47 Table 24. Fatal crash information ............................................................................................................ 50 Table 25. Incapacitating injury (Type A) crash information .................................................................... 51 Table 26. Crash Severity Distribution Used for Benefit/Cost Analysis .................................................... 54 Table 27. ADOT Crash Cost Factors ......................................................................................................... 54 Kittelson & Associates, Inc ii I-10 Phoenix Corridor Safety Study Final Report Table 28. Crash Modification Factors for Reduction in Operating Speed 1,2 ........................................... 55 Table 29 . Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from Lowered Speed Limit .............................................................................................................................................. 55 Table 30. Estimated 20-yr Benefit Value of Reduced K, A, & B Crashes Resulting from Lowered Speed Limit ......................................................................................................................................................... 56 Table 31. 20-yr B/C Analysis of Lowered Speed Limit ............................................................................. 56 Table 32. Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from a VSL System ................................................................................................................... 58 Table 33. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from a VSL System.................. 58 Table 34. 20-yr B/C Analysis of VSL System ............................................................................................. 59 Table 35. Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from an Automated Speed Enforcement System ................................................................................................. 60 Table 36. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from an Automated Speed Enforcement System................................................................................................................................ 60 Table 37. Excess Expected Average Crashes in Tunnel and Adjacent Segments..................................... 62 Table 38. Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from Upgraded Tunnel Lighting ........................................................................................................................................ 62 Table 39. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from Upgraded Tunnel Lighting .................................................................................................................................................... 63 Table 40. 20-yr B/C Analysis of Upgraded Tunnel Lighting ..................................................................... 63 Table 41. Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from Eliminating Lane Drops ............................................................................................................................................... 65 Table 42. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from Eliminating Lane Drops.. 65 Table 43. 20-yr B/C Analysis of Eliminating Lane Drops .......................................................................... 66 Table 44. Estimated Reduced Crashes over 20 years Resulting from Metering the EB 7th Ave and WB 7th St On-ramps....................................................................................................................................... 67 Table 45. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from Metering the EB 7th Ave and WB 7th St On-ramps ......................................................................................................................... 67 Table 46. 20-yr B/C Analysis of Metering the EB 7th Ave and WB 7th St On-ramps .............................. 67 Table 47. Freeway Predictive Models Calibrated for Phoenix Urban Area ............................................. 75 Table 48. Predictive Model Calibration Factors for Phoenix Area Urban Freeways ............................... 75 Table 49. Recommended Data Sources for Safety Performance Evaluations ......................................... 77 Table 50. Calculated Performance Measures and Ranking - Basic Segments ......................................... 82 Table 51. Calculated Performance Measures and Ranking - Speed Change Lanes ................................. 83 Table 52. Crash Cost Factors and PDO Weighting Factors ...................................................................... 85 Kittelson & Associates, Inc iii I-10 Phoenix Corridor Safety Study Final Report EXECUTIVE SUMMARY The downtown Phoenix corridor section of I-10 between 35th Ave and Sky Harbor Blvd is one of the most heavily traveled freeways in the region and state, carrying up to 280,000 vehicles-per-day (vpd) on a typical weekday. As might be expected given the high traffic volume, this freeway section experiences a high number of crashes each year. The crash rate for the 3-year period from 2011-2013 was 3.10 crashes per million vehicle miles travelled (crashes/mvm), compared with an average rate of 1.47 crashes/mvm on other freeways throughout the region. Given these conditions, improving the safety of this section of I-10 is a priority for ADOT. The I-10 Phoenix Corridor Safety Study is an initial step in achieving this objective. METHODOLOGY This study conducted a safety performance evaluation of the I-10 Phoenix corridor section following the urban freeway safety evaluation methodologies described in the AASHTO Highway Safety Manual (HSM). A rigorous process was conducted to calibrate the HSM freeway crash prediction models to accurately reflect conditions on Phoenix area freeways. Using these calibrated models, the study quantified the safety performance of the I-10 study section relative to other freeways in the Phoenix area, identified the primary factors contributing to the high crash frequency in the section, identified opportunities for reducing crashes, and evaluated potential countermeasures to improve safety. Figure ES- 1 summarizes the study process. Figure ES- 1. I-10 Safety Study Process The safety evaluation tools developed from the study can be applied and expanded by ADOT to conduct quantitative safety performance evaluations of urban and rural freeway sections across the state. The application of the HSM methods and these evaluation tools provide ADOT with a quantitative and reliable approach to identify and prioritize freeway safety improvement needs and evaluate safety improvements to achieve the greatest benefit to the public. In addition, the evaluation methods and tools can be applied to assess the safety performance of freeway and interchange improvement alternatives while in the project development process. Kittelson & Associates, Inc ES-1 I-10 Phoenix Corridor Safety Study Final Report SAFETY PERFORMANCE RESULTS Several metrics that describe safety performance were calculated for each freeway segment in the study section. Evaluating the performance of the section by segment, including basic segments and speed-change lanes, is necessary to pinpoint physical and operational factors that are contributing to high crash frequency and which can then potentially be addressed with roadway and operational improvements. The safety performance of the I-10 study section can best be described using two metrics: expected crash frequency per mile and level of service of safety (LOSS). The expected crash frequency per mile describes the magnitude of the safety performance of a segment relative to other segments in the study section and to other freeways in the Phoenix area. The expected crash frequencies per mile for the study section are graphically presented in Figure ES- 2. The red line represents the average expected crash frequency per mile on other Phoenix area freeways. Clearly, the results show many high crash locations within the study section. In the Deck Park Tunnel and the segment to the west, expected crashes per year per mile exceed 400. Other segments, essentially between 19th Ave and the L202/SR 51 TI also have high expected crash frequencies. Figure ES- 2. Safety Performance of Study Section Segments Level of service of safety describes the potential to reduce crash frequency with the application of appropriate countermeasures. The LOSS results presented in Figure ES- 3 show that nearly all of the segments have a moderate (III) to high (IV) potential to find safety improvement opportunities. Kittelson & Associates, Inc ES-2 I-10 Phoenix Corridor Safety Study Final Report Figure ES- 3. Potential to Reduce Crashes with Improvement Table ES-1 provides the average number of crashes, by severity, expected to occur each year in this 7.6 mile stretch of I-10 without safety improvement. When considering the overall safety performance, it is clear that the I-10 study section is a strong candidate for implementing appropriate crash reduction countermeasures. Table ES- 1. Annual Crashes Expected on I-10 Study Section Total Fatal Incapacitating Injury 1718.4 11.5 27.6 Nonincapacitating Injury 190.8 Possible Injury 257.3 Property Damage Only 1231.3 KEY FINDINGS The safety performance evaluation not only identified the magnitude of the safety problem within the study section, but also identified high priority areas to focus further attention to determine factors contributing to high crash rates. A detailed diagnosis of two years (2012-2013) of crash data produced the following findings. Kittelson & Associates, Inc ES-3 I-10 Phoenix Corridor Safety Study          Final Report 6 fatal and 40 incapacitating injury crashes occurred. Rear-end crashes are the predominant crash type. Speed and congestion are contributing factors in approximately 57% of the crashes. Of the 40 incapacitating injury crashes, half noted congested conditions and speed as factors. Congestion is a primary contributing factor, with 70% of the crashes occurring during congested periods. Large speed differentials due to congestion are associated with a notably high number of crashes. During peak periods, crashes associated with high speed vehicles encountering slowed or stopped vehicles is common, particularly with crashes in the HOV lanes. The number of reported crashes associated with left-side HOV entry ramps (3 locations) was low. A combination of several factors likely contributes to the exceptionally high crash frequency in the Deck Park Tunnel. These factors include: o the illumination levels at the tunnel threshold zones are lower than the original design, likely due to the effect of age and environment on the luminaires; o the difference between the day-time ambient illumination level outside of the tunnel and the illumination level at the tunnel portals which can affect drivers vision as their eyes adjust when they enter/exit the tunnel, o lane changing within the tunnel associated with the eastbound and westbound weaving sections between 7th St and 7th Ave, o the horizontal curve within the tunnel, and o congestion occurring within the tunnel and just outside the exit portals. Heavy lane changing associated with the major weaving sections at the I-17 and L202/SR51 system interchanges creates congestion at the exit ramps, resulting in rear-end and sideswipe crashes. Mainline right-side lane drops at two locations, WB at 7th Ave and EB at 7th St are a factor in observed crashes. The lane drops contribute to congestion and produce forced lane changes. In contrast, the EB lane drop at Washington St. does not appear to be contributing to an increased number of crashes. STUDY RECOMMENDATIONS Countermeasures to mitigate the contributing factors and reduce crashes were identified and analyzed. Based on the results of benefit/cost analysis, the following countermeasures are recommended for consideration and evaluation. Lower Speed Limit: Lowering the speed limit (65 mph to 55 mph) along the I-10 Phoenix study corridor, in conjunction with heightened and continued enforcement is expected to lower the average operating speed on the freeway, potentially reducing crash frequency substantially at a high benefit/cost. The limits of the reduced speed zone could extend further to the east and west on I-10. ADOT and DPS should evaluate this countermeasure further. Kittelson & Associates, Inc ES-4 I-10 Phoenix Corridor Safety Study Final Report Variable Speed Limit: VSL systems are effective in lowering primary and secondary crashes, and increasing capacity and throughput on high speed access controlled roadways. Installing a variable speed limit system that operates primarily during weekday peak traffic periods to lower operating speeds on I-10 is estimated to result in potentially 100 fewer total crashes and 30 fatal and injury crashes each year. ADOT should consider developing a VSL system design concept to define system limits, operations, system requirements, and construction, operating, and added enforcement costs. Upgrade Deck Park Tunnel Lighting: It is not possible to definitively quantify the effect of existing illumination levels on the high expected crash frequency within the tunnel, particularly since there are other contributing factors (weaving sections, horizontal curve, and congestion). However, the high crash frequency within the tunnel section supports the need for safety improvement. It is understood that ADOT is evaluating a possible upgrade to tunnel lighting to reduce energy costs. If implemented, a study of the effect of tunnel illumination improvements on crash frequency should be conducted. Meter EB 7th Ave and WB 7th St Entry Ramps: Lane changing is contributing to the high crash frequency in the tunnel. It is recommended that ADOT evaluate installing ramp meters on these two ramps. The evaluation should include study of the effect of ramp metering on traffic operations on the freeway mainline, the ramp terminal intersections, and cross streets. High Friction Surface Treatment: Since rear end-crashes are predominant, application of HFST may prove to be an effective countermeasure for this crash type. ADOT should assess the survivability and service life of this treatment for application on the I-10 study section and consider applying HFST on a test segment. If installed, the safety benefit of this treatment should be evaluated. Other Countermeasures Recommended for Consideration:    Install freeway lane markings in advance of the I-17 and SR 51/SR 202 interchanges where two or more lanes exit I-10. The intent of this low cost countermeasure is to reduce last minute lane changing. Turn-on the ramp meter on the westbound Sky Harbor Blvd on-ramp during morning and evening peak hours to reduce the impact of merging traffic on the congested mainline. This countermeasure should be evaluated to confirm potential impacts on traffic operations within Sky Harbor airport. Implement enforcement and education countermeasures intended to reduce speed and increase driver awareness during peak traffic periods. Potential countermeasures include: o Heightened and visible DPS presence within the corridor. o Using existing dynamic message signs, provide messages reminding drivers to be prepared for upcoming congestion and to drive at a safe speed. o Develop and implement a public outreach campaign designating this section of I-10 as a “safety corridor”. The campaign would inform the public of the need to improve safety on this freeway section, steps being taken by ADOT and DPS to do so, and actions that the public can take to reduce crash potential. o Regional and statewide efforts to reduce distracted driving. Kittelson & Associates, Inc ES-5 I-10 Phoenix Corridor Safety Study Final Report Other Countermeasures Evaluated:    Eliminate Lane Drops at 7th Ave and 7th. This countermeasure is not recommended due to the low estimated benefit/cost based on lowered crash frequency. Restrict HOV Lane Access Points. Based on the limited information available, it is unclear if this countermeasure will reduce crashes. Automated Speed Enforcement. Studies of this technology have shown it lowers average operating speeds, resulting in reduced crashes. The I-10 study has provided a solid foundation for conducting safety performance evaluations on Arizona freeways using the HSM predictive methods. In order to expand the predictive models to cover all freeways throughout the state, it is recommended that the following additional model calibration be conducted.     Develop calibration factors for urban freeways without HOV lanes and without ramp metering. Develop calibrated crash predictive models for 4 and 6-lane rural freeways. Evaluate the need for calibration factors for urban freeway with posted speeds of 55 mph. Evaluate the applicability of calibration factors based strictly on MAG region data to other urban areas in the state. Kittelson & Associates, Inc ES-6 I-10 Phoenix Corridor Safety Study Final Report INTRODUCTION This report documents the safety performance study conducted for the section of I-10 from 35th Ave to Sky Harbor Blvd. The safety performance evaluation was conducted following the urban freeway safety evaluation methodologies described in the AASHTO Highway Safety Manual (HSM). The study results quantify the safety performance of the I-10 study section relative to other freeways in the Phoenix area, identify the primary factors contributing to the crash frequency in the section and opportunities for reducing crashes, and present potential countermeasures to improve safety. The safety evaluation tools developed from the study can be applied and expanded by ADOT to conduct quantitative safety performance evaluations of urban and rural freeway sections across the state. The application of the HSM methods and these evaluation tools provide ADOT with a quantitative and statistically sound approach to identify and prioritize freeway safety improvement needs and evaluate safety improvements to achieve the greatest benefit to the public. In addition, the evaluation methods and tools can be applied to assess the safety performance of freeway and interchange improvement alternatives while in the project development process. STUDY OBJECTIVES The following objectives and goals were achieved in this study: Expand ADOT’s capacity to apply state-of-the art safety evaluation methods.    HSM crash prediction models for urban freeways were calibrated to reflect local conditions so that the models will more accurately assess the safety performance of freeways in the Phoenix metropolitan area. The calibration process applied in this study can be used by ADOT for calibration of predictive models for other freeway facility types in the Phoenix area or around the state. The availability and quality of existing ADOT databases relative to the needs of the HSM crash predictive methods for safety performance evaluation was assessed. Efficient methods for collecting accurate data, not available from existing databases, were demonstrated. ADOT staff received training in data collection methods, application of the predictive models, and the HSM safety performance evaluation methods and process. Apply the predictive and diagnostic HSM methodologies to evaluate the safety performance of the I10 study section.   Using the calibrated crash prediction models, the safety performance of the I-10 study section was evaluated based on a range of performance measures. High crash locations were identified for more detailed evaluation. A detailed diagnostic review of high crash locations within the study section was conducted following the methods prescribed in the HSM. Roadway, traffic, and operational factors contributing to high crash frequencies were identified through the diagnostic process. Kittelson & Associates, Inc 1 I-10 Phoenix Corridor Safety Study Final Report Identify and assess countermeasures intended to reduce crash frequency at high crash locations.  Countermeasures intended to improve safety at high crash locations were identified and evaluated to determine the potential benefit in reducing crashes, as well as cost effectiveness. STUDY TEAM The study was conducted by the consultant team led by Kittelson & Associates, including support from Lee Engineering, Works Consulting, and DiExSys,LLC, managed by ADOT MPD under the direction of the ADOT Traffic Engineering Group. Study technical oversight and support was provided by a Core Team comprised of representatives from the following ADOT organizations and other agencies.    ADOT  Traffic Operations  Traffic Safety  Phoenix Regional Traffic Engineering  MPD – HPMS  Budget  Traffic Records  State Traffic Engineering  Urban Project Management  Traffic Design  Roadway Engineering  Information Technology Group Maricopa Association of Governments – ITS and Safety FHWA As part of the study, two workshops were conducted with the intent of disseminating information on study progress and results as well as the value of implementing the HSM safety performance evaluation methods into ADOT’s safety management program and project development process. A workshop to provide ADOT Safety Section staff with training on the application of the freeway crash prediction methodology and tools was also held. STUDY LIMITS The study limits are shown in Figure 1, extending from 35th Ave (MP 141.67) to Sky Harbor Blvd (MP 149.30). Kittelson & Associates, Inc 2 I-10 Phoenix Corridor Safety Study Final Report Figure 1. I-10 Study Section Map Kittelson & Associates, Inc 3 I-10 Phoenix Corridor Safety Study Final Report OVERVIEW OF THE FREEWAY SAFETY EVALUATION METHODOLOGY The HSM provides a methodology, tools, and guidance for evaluating the safety performance of a freeway network, facility, section, and interchange. This methodology is used to identify and prioritize freeway segments where improvements have a high potential to reduce crashes, diagnose safety issues, assess future safety conditions, and evaluate the impact of design alternatives on crash frequency and severity. The methodology includes detailed models for predicting crashes by type and severity. These models provide several important advantages:    they reflect the effects of roadway geometry, physical features, and traffic volume on crash frequency and severity, they allow for a thorough understanding of safety performance and opportunity to improve performance, they can quantify the benefits of crash countermeasures, and the use of the predictive models in the screening process will more reliably identify segments with potential for safety improvement. The freeway predictive models were developed for the HSM through a research project funded by the Transportation Research Board’s National Cooperative Highway Research Program (NCHRP 17-45)(Ref. 4). Crash predictive models are available for rural and urban freeway segments, speed-change lanes, and ramps. The models applied in this safety study are listed in Table 1. Table 1. Freeway Crash Prediction Models Urban Freeway Speed-change Lanes Urban Freeway Segments  Multiple vehicle crashes – FI  Multiple vehicle crashes – FI  Multiple vehicle crashes – PDO  Multiple vehicle crashes – PDO  Single vehicle crashes – FI  Single vehicle crashes – FI  Single vehicle crashes - PDO  Single vehicle crashes - PDO Each model predicts average crash frequency using Safety Performance Functions (SPFs), Crash Modification Factors (CMFs), and a local calibration factor (C). The general form of the predictive model is as follows: Np = Nspf x (CMF1 x CMF2 x CMF3 x …x CMFn) x C Specific SPFs are available for predicting multiple vehicle and single vehicle crashes on basic freeway segments and speed-change lanes. The SPF predicts the average crash frequency for base conditions. For example, the base conditions for the SPF used to predict multiple vehicle crashes on a freeway segment are: Kittelson & Associates, Inc 4 I-10 Phoenix Corridor Safety Study          Final Report Length of horizontal curve - 0.0 mi (curve not present) Lane width – 12 ft. Inside paved shoulder width – 6 ft. Median width – 60 ft. Length of median barrier – 0.0 mi (barrier not present) Number of hours where traffic volume exceeds 1000 veh/ln/hr – none Distance to nearest upstream entry ramp - > 0.5 mi Distance to nearest downstream exit ramp - >0.5 mi Length of Type B weaving section – 0.0 mi (weaving section not present) The CMFs are applied to adjust the average number of crashes predicted by the SPF to account for differences from the base conditions. For example, the CMF for paved shoulder width would increase the average number of predicted crashes if the width was less than 6 ft. If the width was greater than 6 ft, the predicted average number of crashes may decrease. A calibration factor (C) is also applied to adjust the average number of crashes predicted by the SPF to reflect local conditions and factors that are not covered by the CMFs, but which may affect crash frequency. Development of the calibration factors for Phoenix freeways is described in the following section. The average number of predicted crashes, adjusted for variations from base conditions and for the effect of local conditions are then distributed by severity – fatal (K), incapacitating injury (A), nonincapacitating injury (B), possible injury (C), and non-injury (O or PDO). This distribution is performed using Severity Distribution Functions (SDFs). The SDF distributes crashes by severity based on a number of factors, including presence of median barrier and shoulder barrier, level of congestion, presence of rumble strips, lane width, area type, and horizontal curvature of the roadway. Similar to the SPF, the SDF is also calibrated to local conditions. APPLICATION OF THE FREEWAY PREDICTIVE MODELS FOR THIS STUDY The HSM freeway predictive models were developed using field data from freeways in three states – Maine, Washington, and California. These data were used to calibrate and validate the base models. The sites selected for urban freeway model development covered a range of physical conditions that are reflected in the model factors. Conditions which are not explicitly reflected in the base models include:       urban freeway sections with 11 or more through lanes HOV or managed lanes ramp metering use of shoulders as travel lanes toll plazas reversible lanes Three of these conditions are present within the I-10 Study Section and/or on other Phoenix freeways. Most notably, nearly all of the freeway sections within the Phoenix urban core include HOV lanes and Kittelson & Associates, Inc 5 I-10 Phoenix Corridor Safety Study Final Report ramp metering. Freeway sections with 11 or more through lanes are also present. While further research on the effect of each of these conditions is needed to refine the freeway crash prediction models, their presence does not preclude the use of the current models for the purpose of evaluating freeway safety performance. The models can be applied with reliable results as long as the conditions are consistent between the freeways used to develop local calibration factors and the I-10 freeway section being evaluated. The effect of HOV lanes and ramp metering is then reflected in the local calibration factor and not as a separate CMF. The presence of HOV lanes was given careful consideration, since HOV lane access control can have a marked impact on traffic operations. Since all of the HOV lanes on the Phoenix freeways have continuous access (no access control) and use of the HOV lanes by general traffic is allowed for much of the day during the week and on weekends, it is reasonable to treat these lanes as general through lanes within the model. While the calibrated Phoenix freeway models will not be able to assess modifications to changes in HOV lane access control, the models will be valid for evaluating freeway safety performance and screening freeway segments to prioritize those with the highest potential for crash reduction. MODEL CALIBRATION The eight urban freeway crash prediction models listed in Table 1 were used to evaluate the safety performance of the I-10 study section. Since crash frequencies can vary significantly from one area of the country to another, calibration of these models to conditions on Phoenix urban freeways was necessary so that crash predictions reflect actual Phoenix crash experience. Calibration accounts for differences in climate, driver population, driver behavior (aggressiveness), speed limit, enforcement, crash reporting practices, driving laws, and other unknown differences. The model calibration process involved the following steps: 1. Identify Phoenix urban freeway sections that represent a normal range of physical and operational conditions. Develop criteria for selecting appropriate calibration segments. 2. Segment these freeway sections into homogeneous segments. 3. Randomly select 30 to 50 segments for use in calibration. Avoid segments that are atypical of the Phoenix freeway system, such as one that includes a left-hand entry or exit ramp, or tunnel. 4. Gather and check data for each segment, developing a calibration database. The data required includes detailed roadway geometry and physical characteristics, mainline and ramp traffic volumes (AADT), level of mainline congestion, and crash history. 5. Calculate model calibration factors using the calibration database and the Enhanced Interchange Safety Analysis Tool (ISATe), developed in conjunction with the HSM urban freeway crash prediction methodologies and used to evaluate freeway and interchange safety. Kittelson & Associates, Inc 6 I-10 Phoenix Corridor Safety Study Final Report CALIBRATION SEGMENTS Freeway calibration segments were selected from freeways in the Phoenix area with generally similar operational characteristics as exists on the I-10 study section. Two particularly key features include ramp metering and HOV lanes. Since both are common on Phoenix freeways and are present in the study section, they must also be present in the segments used to calibrate the crash prediction model. The following summarizes the criteria used to identify freeway sections that would provide appropriate calibration segments.         Terrain - Level or rolling terrain such that individual grades of significant length are 4 percent or less. Freeway AADT - AADT data should be available for two or more of the years of the crash period being evaluated in the study section (2011-2013). Ramp AADT - AADT data should be available for each ramp for two or more of the years of the crash period being evaluated in the study section (2011-2013). Freeway crash data - An electronic record should be available of every reported crash occurring on the freeway section during each year of the crash period (2011-2013). Each record should specify the crash severity using the KABCO scale (or be convertible to this scale), the crash type category (i.e., single-vehicle or multiple-vehicle), direction of travel of each involved vehicle, and crash location by roadbed and milepost (or geo-coordinates). Crash location should be specified to the nearest 0.1 mile or less (or its equivalent using geo-coordinates).  Speed limit - The speed limit should be 60, 65, or 70 mi/h.  Shoulder use - No section should allow shoulder use as a travel lane at any time.  Reversible lanes - No section should include travel lanes that are reversed by time of day. Work zone presence - Desirably, no section would have any long-term (i.e., 4 or more days) work zones present during the crash period (2011-2013). It is acceptable to retain a section if it had only a few long-term work zones during the crash period provided that their start and end dates can be identified (so the crash data for these periods can be omitted and the calibration factor adjusted for partial year data). HOV Lanes – Freeway sections should have HOV lanes separated by a single white stripe or white stripe buffer, with no control of where vehicles can enter or exit. Through Lanes – If possible, the selected calibration segments should have an equal number of 6/7, 8/9, and 10/11- lane cross sections. HOV lanes were not counted as through lanes, however were included in each calibration segment. Ramp meters – The selected calibration segments should have ramp metering. Based on these criteria, the freeway sections listed in Table 2 were determined to be appropriate from which to select segments to be used to calibrate the crash prediction models. These freeway sections were clear of long-term construction during the calibration timeframe (2011-2013) as determined from the ADOT construction log, included ramp meters, included HOV lanes, included 6 to 11 through lanes, excluded 55 mph posted speed limits, and had AADT and crash data available for the study period (2011-2013). Kittelson & Associates, Inc 7 I-10 Phoenix Corridor Safety Study Final Report Table 2 Candidate Freeway Calibration Sections Facility BMP EMP I-10 (91st Ave to 35th Ave) 134.68 141.66 Length (miles) 6.98 Comments Long-term construction west of section, study area east of section HOV lanes end south of section I-10 (Buckeye Rd to Ray Rd) 148.93 159.78 10.85 I-17 (Cactus Rd to Union Hills Blvd) 209.96 213.98 4.02 SR-101 (Shea Blvd to Warner Rd) 41.06 58.60 17.54 SR-202 (24th St to McClintock Dr) 0.74 8.80 8.06 South of section posted 55 mph, long-term construction north of section West/north valley 101 due to long-term construction Long-term construction east of section US-60 (Mill Ave to Sossaman Rd) 173.68 189.70 16.02 HOV lanes end east of section BMP – Begin Mile Post, EMP – End Mile Post. SEGMENTATION Each of the freeway sections was segmented according to the following guidelines:     Segments must be essentially homogeneous for the entire length. This requirement includes number of lanes, lane width, shoulder width, median width, and clear zone. Segments should be longer than 0.1 mile. This requirement minimizes the potential of crash location error. New segments should begin where:  Number-of-lanes change  Lane width changes by 0.5 ft. or more  Shoulder width changes by 1 ft. or more  A ramp is present (the gore point defines start of the segment)  Clear zone width changes by 5 ft. or more  Median width changes by 10 ft. or more if the width is ≤ 90 ft.; if more than 90 ft., changes in median width are not considered The presence of a horizontal curve or roadside barrier does not define a segment. The research that developed the freeway crash prediction models concluded that these elements are so common that their consideration as segmentation criteria would result in too many small segments. As a result, their presence is incorporated directly into the models. Approximately 2-4 miles of each freeway section, listed in Table 2 was fully segmented. The segmented sections are listed in Table 3. Segmentation of each freeway section was conducted using Google Earth. The aerial imagery for the Phoenix metro area is high quality and current (2014). Figure 2 provides an image of a section of freeway segmented. The yellow push-pins represent the beginning of each segment. The exact milepost location of each pin was determined using ADOT’s linear referencing system. Kittelson & Associates, Inc 8 I-10 Phoenix Corridor Safety Study Final Report Table 3. Segmented Freeway Sections Facility BMP EMP Length (miles) I-10 (67th Ave to 43rd Ave) 137.36 140.93 3.57 I-10 (Sky Harbor Blvd to 40th St) 149.15 152.04 2.89 I-17 (Cactus Rd to Bell Rd) 210.24 212.71 2.47 SR-101 (Shea Blvd to Indian Bend Rd) 41.06 44.77 3.71 SR-202 (Van Buren St to McClintock Dr) 4.84 6.48 1.64 US-60 (McQueen Rd to Val Vista Dr) 180.41 184.40 3.99 Figure 2. Example of Segmentation Some 18 miles of freeway segments are listed in Table 4. The minimum segment length is 0.04 miles. The range of each cross section element included in these segments is as follows:       Number of through lanes, 6 to11 Lane width, 11.4 to 12.4 ft., Outside shoulder width, 7.3 to 14 ft. Inside shoulder width, 6.1 to 11.9 ft., Median width, 19.2 to 79.2 ft., Clear zone width, 23.6 to 30 ft. Kittelson & Associates, Inc 9 I-10 Phoenix Corridor Safety Study Final Report Table 4. Calibration Segments Facility SR 101 US 60 SR 202 BMP EMP Controlling Attribute 41.070 41.370 Ramp 41.370 41.430 Ramp 41.430 41.980 Ramp 41.980 42.055 Ramp 42.055 42.675 42.675 Facility Controlling Attribute BMP EMP 210.240 210.695 Ramp 210.695 210.920 Ramp 210.920 211.180 Outside Shoulder Width Ramp 211.180 211.230 Ramp 43.040 Ramp 211.230 211.655 Ramp 43.040 43.085 Ramp 211.655 212.260 Ramp 43.085 43.695 Ramp 212.260 212.390 Ramp 43.695 44.140 Ramp 212.390 212.450 Ramp 44.140 44.745 Ramp 212.450 212.560 Ramp 212.560 212.710 Clear Zone Width 137.362 138.002 Ramp Ramp I 17 180.405 180.695 Ramp 180.695 181.135 Ramp 181.135 181.675 Ramp 138.002 138.323 181.675 182.130 Ramp 138.323 138.843 Ramp 182.130 182.685 Ramp 138.843 138.954 Inside Shoulder Width 182.685 183.720 Ramp 138.954 139.375 Ramp 183.720 184.095 Lane Add 139.375 139.923 Ramp 184.095 184.400 Ramp 139.923 140.398 Ramp 4.840 4.890 Ramp 140.398 140.926 Ramp 4.890 5.035 149.152 149.287 Ramp 5.035 5.190 149.287 149.380 Ramp 5.190 5.365 Ramp Outside Shoulder Width Ramp 149.380 149.461 Ramp 5.365 5.690 Ramp 149.461 149.512 Median Width 5.690 5.860 Ramp 149.582 Median Width 5.860 6.225 Median Width 149.582 149.646 Median Width 6.225 6.650 Ramp 149.646 149.818 Median Width 6.650 6.735 Ramp 149.818 149.971 Ramp 6.735 7.392 Ramp 149.971 150.178 Ramp 7.392 7.495 150.178 150.246 Ramp 7.495 7.690 Ramp Inside Shoulder Width 150.246 151.223 Ramp 151.223 151.349 Ramp 7.690 8.005 Lane Drop 151.349 151.775 Lane Drop 8.005 8.050 Ramp 151.775 151.920 Ramp 8.050 8.300 Ramp 151.920 151.982 Ramp 8.300 8.430 Ramp 151.982 152.041 Ramp 8.430 8.650 Ramp Kittelson & Associates, Inc I 10 W I 10 E 149.512 10 I-10 Phoenix Corridor Safety Study Final Report In addition to segmenting the freeway sections to be used for calibration, the I-10 study section was also segmented. The resulting 30 study section segments are listed in Table 5. Table 5. I-10 Study Section Segments Facility I 10 BMP EMP Controlling Attribute 141.670 141.920 Project Begin 141.920 141.425 Ramp 141.425 142.705 142.705 142.880 Ramp Ramp 142.880 143.000 143.000 143.670 143.670 143.945 143.945 144.360 Ramp Outside Shoulder Width Ramp Ramp 144.360 144.420 Ramp 144.420 144.800 144.800 144.920 144.920 144.960 144.960 145.420 Ramp Outside Shoulder Width Ramp Ramp 145.420 145.520 145.520 145.780 145.780 145.820 Ramp 145.820 145.940 Ramp 145.940 146.440 Ramp Ramp Outside Shoulder Width Facility BMP EMP Controlling Attribute 146.440 146.850 Ramp 146.850 146.930 Ramp 146.930 147.510 Ramp 147.510 147.670 Ramp 147.670 147.955 Ramp 147.955 148.065 Ramp 148.065 148.310 Ramp 148.310 148.430 Ramp 148.430 148.660 Ramp 148.660 149.135 Ramp 149.135 149.170 Ramp 149.170 149.300 Ramp Of the segments listed in Table 4, 31 (those highlighted) were randomly selected to be used for model calibration. An additional 9 segments were selected from the I-10 study section and added to the calibration database. These nine segments are highlighted in Table 5. The resulting distribution of calibration segments by freeway and by physical features is summarized in Table 6. Kittelson & Associates, Inc 11 I-10 Phoenix Corridor Safety Study Final Report Table 6. Distribution of Calibration Database Calibration Segment Characteristics 6-7 Thru Lanes 20 8-9 Thru Lanes 12 10-11 Thru Lanes 8 Weave 15 HOV Lanes 40 Ramp Meter 40 Curve 14 Freeway Segments 40 7 Entry Speed-Change Segments 17 Exit ADT High 240,000 ADT Low 129,000 Freeway Facility I-10 20 I-17 4 SR 101 9 SR 202 2 US 60 5 Total 40 DATABASE DEVELOPMENT Databases were developed for both the calibration segments and the I-10 study section segments. The data collection process is described below. Roadway Geometry and Characteristics The required data included:           number of through lanes, length of freeway segments and speed-change lanes, presence of horizontal curve in one or both direction of travel, including length of curve, radius of curve, and length of curve within a segment, width of lanes, outside shoulders, inside shoulder, and median, length of rumble strips on the inside or outside shoulders, length and offset median and shoulder barriers, width of continuous median barrier, if present, presence and length of a Type B weaving section, distance to nearest upstream entry and downstream exit ramps in each travel direction, and clear zone width. In order to efficiently and accurately gather roadway geometry and characteristic data for 31 calibration segments, totaling 10 miles, and 30 study section segments, totaling 7.6 miles, the following data collection process was used. A comparison of roadway geometry and characteristics data obtained from the ADOT Roadway Characteristic Inventory Database (RCID) and extracted from Google Earth aerials was conducted to determine the accuracy of each method. The results indicated that lane width, shoulder width, and median width information, as well as the horizontal curve data (curvature and length) from the RCID compared well with that data collected from the aerials. However, information on barrier location and Kittelson & Associates, Inc 12 I-10 Phoenix Corridor Safety Study Final Report offset, clear zone, presence of continuous median barrier, presence of a Type B weaving section was not readily available from the RCID. Therefore, data from both the RCID and collected using Google Earth were used. A process to partially automate the extraction of roadway geometry and characteristic data from Google Earth imagery was developed to support application of the urban freeway crash prediction methodology. The process involves pinning the location of key cross section points using the Google Earth software. The points located in this manner include: edge of clear zone, edge of shoulder, edge of travel lanes, edge of median, location of barrier, begin/end points of weaving lanes, speed-change lane tapers, and roadway curvature. An example of a pinned segment is presented in Figure 3. Each pin is coded by latitude and longitude. These points were digitized and imported into an in-house, Excel-based software tool (Earth Tools) that generates the cross section and roadway characteristic data for each segment. The resulting data was reviewed and potential errors identified and checked. Since data for the tunnel segment of I-10 cannot be collected using aerials, field measurements were collected by ADOT staff during a schedule tunnel maintenance closure in January 2014. Figure 3. Example of Pinning to Collect Physical Features Data Traffic Volume Data Traffic volume data required for calibration and evaluation of safety performance on the I-10 study section included mainline and ramp AADTs for the years 2011-2013 and level of mainline congestion. Mainline AADTs – Two sources of mainline AADTs were available – the ADOT Traffic Data Management System (TDMS) website and the traffic count data collected by the ADOT Freeway Kittelson & Associates, Inc 13 I-10 Phoenix Corridor Safety Study Final Report Management System. The FMS counts are the basis for the TDMS AADTs. Comparison of the AADTs from each source indicated potential inconsistencies in the 2011 and 2012 data. At the locations reviewed, there were in some cases, substantial differences between the FMS counts and the AADTs for 2011 and 2012. However, the FMS counts compared well with the AADTs in 2013. It appears that this difference is likely due to missing data. For example, at one count site, 251 days of FMS data were available in 2011, 305 in 2012, and 342 in 2013. The reported AADTs likely represent an adjusted FMS volume to account for the missing data. Since it was not possible to verify the accuracy of the adjusted AADT’s, it was decided that only the 2013 AADTs should be used for the evaluation. The AADT’s for 2011 and 2012 were assumed to be reasonably close to the 2013 values. AADT’s for each calibration and I-10 study section segment were calculated based on the AADT from the nearest TDMS mainline count station, adjusted to reflect traffic entering or exiting from adjacent ramps. Ramp AADTs – ADOT collects volume data at most freeway entrance and exit ramps on a bi-annual basis in the form of a 48-hour tube count. The most recent count was in 2012 and that data was assumed to be equivalent to the 2013 analysis year for the purposes of crash prediction and calibration. Level of Congestion – Level of congestion is described as the proportion of high volume hours occurring during a typical day. Specifically, this input is the proportion of freeway AADT volume that occurs during hours where the lane volume exceeds 1,000 vehicles per hour per lane (vphpl). The proportion was calculated by first summing the volume during each hour where the average lane volume exceeds 1,000 vphpl and then dividing this sum by the AADT. Using Microsoft Access, 2013 FMS data was queried to determine the average annual proportion of high volume hours for each calibration and I-10 study segment. Volume data that occurred during lane or ramp closures was eliminated from the calculation as it would impact the high volume proportion, particularly on high volume facilities. Traffic volume in HOV lanes was included in the calculation. The hourly lane volume was computed by dividing the hourly volume by the number of through lanes, where the number of through lanes included all general purpose lanes (excluding auxiliary lanes) and HOV lanes. Crash Data Crash data for the freeway sections containing the calibration segments and the I-10 study section were obtained from ADOT for 2011, 2012, and 2013. These data were subsequently assigned to the individual freeway segments. The reported crashes for each segment are listed in Table 7 and Table 8. The crash records were reviewed to confirm that there were no duplicate records. However, it was noted that about 2 percent of the crashes occurred at the same time and location. An additional 1 percent of the crashes occurred within five minutes and within 500 ft. of one another. These crashes were assumed to be secondary crashes that resulted partly from the initial crash. They were retained in the database. In about 1 percent of the crash records, the RoadCondition attribute indicated that a work zone may have been present. These crashes were removed from the database. The crash records were screened to include only crashes that were located on either “two-way divided roadway without median barrier” or “two-way divided with median barrier.” In about 1 percent of the Kittelson & Associates, Inc 14 I-10 Phoenix Corridor Safety Study Final Report crash records, the TrafficWayType variable indicated that the crash occurred on a “one-way roadways (including ramps).” Since these crashes may have occurred on a ramp or frontage road and not the mainline, they were removed from the database. In about 1 percent of the crash records, the JunctionRelation attribute indicated that the crash was located on a ramp or frontage road. These crashes were removed from the database. Finally, it was noted that the crash milepost location was less precise for 2011 than for either 2012 or 2013. Crashes in 2011 were more frequently assigned to an even milepost (e.g., 279.00) than they were to locations identified to the tenth or hundredth of a mile (e.g., 279.9 or 279.95). In general, the count of crashes at “full mile” locations in 2011 is twice as large as that for 2012 or 2013. This trend is true for all freeway sections represented in the calibration database. It suggests that using milepost location to determine the number of 2011 crashes associated with a particular segment may not lead to an accurate count of crashes for segments less than one mile in length. The more frequent even milepost grouping of the 2011 crashes may be a result of the new GPS-based Traffic and Citation (TraCS) crash reporting system not being fully implemented in that year. Even though there is crash location uncertainty in the 2011 data, it was concluded that data for all three years could be used to quantify the calibration factors. This conclusion recognizes that the calibration factors are based on the total count of crashes for all segments combined. That is, the aforementioned crash location bias would tend to “average out” when considering the total number of crashes for the collective set of segments. However, due to location uncertainty, it was concluded that only data for 2012 and 2013 can be used to evaluate safety performance for the individual segments on the I-10 study section. Crash location uncertainty would be problematic for the evaluation of an individual segment for several reasons (e.g., the evaluation would require an accurate count of crashes occurring on that segment). Evidence of this uncertainty can be seen by comparing the crash rates for adjacent segments in Table 7 and Table 8. The rates often vary widely from segment to segment. Table 7. Calibration Segments Crash Summary Location Segment Begin Milepost Segment Length, mi Exposure, 1 mvm/3years Crash Count, crashes/3 years Crash Rate, crashes/mvm I-10 138.002 0.321 81.0 87 1.07 138.954 0.421 80.8 112 1.39 139.375 0.548 90.3 137 1.52 139.923 0.475 109.1 126 1.15 141.67 0.25 61.1 127 2.08 141.92 0.505 132.5 325 2.45 142.425 0.28 67.6 171 2.53 142.705 0.175 29.3 50 1.71 143.67 0.275 62.3 105 1.69 143.945 0.415 97.5 269 2.76 147.51 0.16 25.2 24 0.95 Kittelson & Associates, Inc 15 I-10 Phoenix Corridor Safety Study Location I-17 L 101 L 202 U.S. 60 Final Report Segment Begin Milepost Segment Length, mi Exposure, 1 mvm/3years Crash Count, crashes/3 years Crash Rate, crashes/mvm 147.67 0.285 44.6 46 1.03 148.43 0.23 45.0 48 1.07 149.287 0.093 15.7 23 1.47 149.512 0.07 9.9 10 1.01 149.646 0.172 24.3 32 1.32 149.818 0.153 30.0 49 1.63 151.223 0.126 28.9 31 1.07 151.775 0.145 34.4 22 0.64 151.92 0.062 15.2 17 1.12 210.24 0.455 83.6 106 1.27 211.18 0.05 8.4 18 2.16 211.655 0.605 92.9 101 1.09 212.56 0.15 24.5 18 0.74 41.37 0.06 9.4 12 1.28 41.43 0.55 93.7 109 1.16 41.98 0.075 12.3 30 2.43 42.055 0.62 95.7 97 1.01 42.675 0.365 65.8 62 0.94 43.04 0.045 7.8 7 0.89 43.085 0.61 103.1 139 1.35 43.695 0.445 81.5 101 1.24 44.14 0.605 102.2 106 1.04 5.19 0.175 25.5 33 1.29 6.735 0.657 113.0 242 2.14 180.695 0.44 105.2 196 1.86 181.675 0.455 105.1 194 1.85 182.13 0.555 102.4 117 1.14 183.72 0.375 76.9 82 1.07 184.095 0.565 95.3 77 0.81 Total: 13.018 2489 3658 Note: 1 - mvm/3 years: million-vehicle-miles (= segment length x AADT x 365 x 3 / 1,000,000). Kittelson & Associates, Inc 1.47 16 I-10 Phoenix Corridor Safety Study Final Report Table 8. I-10 Study Section Segments Crash Summary Location Segment Begin 2 Milepost Segment Length, mi Exposure, 1 mvm/3years Crash Count, crashes/3 years Crash Rate, crashes/mvm I-10 141.67 0.25 61.1 127 2.08 141.92 0.505 132.5 325 2.45 142.425 0.28 67.6 171 2.53 142.705 0.175 29.3 50 1.71 142.88 0.12 17.8 32 1.80 143.00 0.67 99.2 201 2.03 143.67 0.275 62.3 105 1.69 143.945 0.415 97.5 269 2.76 144.36 0.06 12.7 71 5.61 144.42 0.38 79.6 422 5.30 144.80 0.12 25.5 198 7.76 144.92 0.04 9.1 44 4.85 144.96 0.46 112.9 672 5.95 145.42 0.10 22.8 96 4.21 145.52 0.26 60.0 276 4.60 145.78 0.04 9.5 51 5.36 145.82 0.12 29.8 80 2.68 145.94 0.50 150.3 575 3.83 146.44 0.41 119.3 370 3.10 146.85 0.08 20.8 78 3.75 146.93 0.58 64.3 243 3.78 147.51 0.16 25.2 24 0.95 147.67 0.285 44.6 46 1.03 147.955 0.11 15.7 48 3.06 148.065 0.245 40.7 70 1.72 148.31 0.12 21.7 39 1.80 148.43 0.23 45.0 48 1.07 148.66 0.475 76.9 114 1.48 149.135 0.035 5.8 6 1.04 149.17 0.13 21.9 49 2.24 7.63 1581 4900 3.10 Total: Notes: 1 - mvm/3 years: million-vehicle-miles (= segment length x AADT x 365 x 3 / 1,000,000). 2 – Highlighted segments are also included in the set of calibration segments. Kittelson & Associates, Inc 17 I-10 Phoenix Corridor Safety Study Final Report GENERAL DISCUSSION OF OBSERVED CRASH RATES The overall average crash rate for the calibration segments and the I-10 study section is provided in the last row of Table 7 and Table 8, respectively. The rate for the I-10 study section is roughly twice that for the calibration segments. This finding confirms the basis for this safety study regarding the significant potential for safety improvement in the I-10 study section. The overall average crash rates identified in the last row of Table 7 and Table 8 were compared to the crash rates for urban freeways in other states and metropolitan areas, including the three states (Washington, Maine, California) used to develop the HCS urban freeway crash prediction methodology. The objective of the comparison is to understand the extent to which crash rate can vary among jurisdictions and to develop some expectation about the likely magnitude of the calibration factors for Phoenix urban freeways. The crash rates for several jurisdictions are listed in Table 9. The source of the data is indicated in the first column. The total crash rate varies from 0.63 to 1.48 crashes per million vehicle miles (cr/mvm). There are many possible reasons for this variation in rates. These reasons relate to differences among jurisdiction in reporting threshold and process, the manner by which a secondary crash is defined, HOV lane presence/mileage, speed limit, level of enforcement, and so on. The crash rate for “Washington, Maine, and California combined” describes the data used to calibrate the base safety prediction models included in the HSM urban freeway crash prediction methodology. The total crash rate for these three states combined is 0.76 cr/mvm. A comparison of this rate with that for the Phoenix calibration segments suggests that the typical calibration required to yield accurate estimates of the predicted crash frequency for Phoenix urban freeway segments will be in the range of 1.47/0.76 = 1.90, although the magnitude of the factor varied among the eight crash prediction models calibrated. The crash severity distribution for the Phoenix segments can also be compared with the corresponding distributions for several other jurisdictions. The objective of the comparison is to understand the extent to which the crash severity distribution can vary among jurisdictions and to develop some expectation about the likely magnitude of the calibration factor of the severity distribution function (SDF) for Phoenix urban freeways. The crash severity distributions for several jurisdictions are listed in Table 10. The proportion of property damage only (PDO) crashes tends to range from 0.67 to 0.73. The proportion of PDO crashes for the Phoenix calibration segments (i.e., 0.68) is near the low end of this range. It is equal to that for the “Washington, Maine, and California combined” data that were used to develop the SDF for the HSM urban freeway crash prediction methodology Kittelson & Associates, Inc 18 I-10 Phoenix Corridor Safety Study Final Report Table 9. Comparative Urban Freeway Crash Frequency and Rate Location (Reference) Exposure, 1 mvm Maine DOT (1) Fatal (K) and Injury (A,B,C) Total (K,A,B,C,PDO) Crash Count, cr Crash Rate, cr/mvm Crash Count, cr Crash Rate, cr/mvm 536 113 0.21 340 0.63 Utah DOT (1) 3005 1015 0.34 3103 1.03 Illinois DOT (1) 10,213 4698 0.46 15,156 1.48 Minnesota DOT (1) 6919 2054 0.30 7732 1.12 Oregon DOT (2) 5778 not available not available 3694 0.64 Dallas-Ft. Worth (3) 27,226 not available not available 20,530 0.75 Washington, Maine, California combined (4) 21,224 5112 0.24 16,205 0.76 Phoenix metro. calibration segments 2489 (13.0 mi, 3 yrs.) 1167 (3 years) 0.47 3658 (3 years) 1.47 4900 (3 years) 3.10 Phoenix I-10 study 1581 1427 0.90 section (7.6 mi, 3 yrs.) (3 years) Note: 1 - mvm: million-vehicle-miles (= segment length x AADT x 365 / 1,000,000). Table 10. Comparative Urban Freeway Crash Severity Distribution Location (Reference) Crash Severity Distribution1 Property Damage Only, PDO Possible Injury, C Nonincapacitating Injury, B Incapacitating Injury, A Fatal, K Maine DOT (1) 0.67 0.16 0.13 0.04 0.012 Utah DOT (1) 0.67 0.16 0.09 0.07 0.005 Illinois DOT (1) 0.69 0.16 0.08 0.07 0.005 Minnesota DOT (1) 0.73 0.17 0.08 0.01 0.002 Washington, Maine, California combined (1) 0.68 0.21 0.09 0.01 0.005 Phoenix metro. calibration segments 0.68 0.17 0.13 0.02 0.002 Phoenix I-10 study section 0.71 0.16 0.11 0.01 0.002 The proportion of possible injury (C) crashes ranges from 0.16 to 0.21. The proportion of C-type crashes for the Phoenix calibration segments (0.17) and the I-10 study section (0.16.) are essentially the same as the other urban areas. In contrast, that for the “Washington, Maine, and California combined” Kittelson & Associates, Inc 19 I-10 Phoenix Corridor Safety Study Final Report data is 0.21. The proportion of non-incapacitating injury (B) and incapacitating injury (A) crashes on Phoenix urban freeways are higher than the combined data for these three states. These differences may be due to differences in severity definition, severity reporting, operating speed, and so on among jurisdictions. The severity distribution function (SDF) calibration factor will be used to account for these differences so that the crash severity distribution predicted for the Phoenix urban freeways accurately represents what is observed in the crash data. The calibration factor is based on calculations using the non-PDO proportions for the Phoenix calibration segments and those for the “Washington, Maine, and California combined” data. The SDF calibration process is described in the final report by Bonneson et al.(Ref 4). ASSIGNMENT OF CRASHES TO SEGMENTS Forty segments located on five representative freeways in the Phoenix metropolitan area were used for model calibration. The safety prediction procedure explicitly evaluates both speed-change lanes and basic segments, therefore requiring calibration factors for each. To support this type of evaluation and calibration of the underlying models, reported crashes must be appropriately assigned to speed-change lanes and the basic segments. The definitions of “speed-change lane” and “basis segment” are provided in this section to describe how crashes are assigned to each entity. Crashes assigned to a speed-change lane are identified as the crashes occurring on the roadbed served by the speed-change lane, between the taper and gore points. The region defined by these points is shown in grey in Figure 4. Entrance Ramp with Parallel Design Ramp Entrance Length Gore point B B A * Taper point Exit Ramp with Taper Design Ramp Exit Length B A Gore point * B Taper point * Point where marked gore is 2 ft wide (gore point). A All crashes that occur within this region are classified as speed-change crashes. B All crashes that occur within this region are classified as freeway segment crashes. Figure 4. Region Defining Speed-Change Related Crashes Crashes that are not assigned to a speed-change lane were assigned to a basic segment. The sum of the speed-change-related crashes and the basic-segment crashes represents the total number of crashes reported between the begin- and end-mileposts associated with a given segment. That is, all Kittelson & Associates, Inc 20 I-10 Phoenix Corridor Safety Study Final Report reported crashes occurring between a segment’s begin- and end-mileposts must be assigned to either a speed-change lane or a basic segment (but not both). CALIBRATION FACTORS Calculation of calibration factors involved assessing the degree to which each calibration segment is reasonably representative of typical Phoenix-area freeways. Atypical segments were considered for removal from the calibration database based on their influence on the calculated calibration factors. Model calibration factors were then calculated from the screened database. Screening Calibration Segments The screening process compared the reported fatal and injury crashes within a segment for the years 2011, 2012 and 2013 with average number of crashes occurring in the same period based on the a typical crash rate observed in the data used to develop the HSM freeway crash prediction models (0.24 fatal-and-injury crashes per million vehicle miles (cr/mvm). The assessment focused on fatal-and-injury crashes because property-damage-only crashes can be less consistently reported depending upon the level of damage. Screening of the 40 basic segments is summarized in Table 11. Column 3 lists an estimate of the average number of fatal-and-injury crashes occurring in a three-year period based on a typical crash rate of 0.24 crashes per million vehicle-miles. Column 4 lists the reported fatal-and-injury crashes for the basic segment. The segment beginning at MP147.51 consists of two speed-change lanes, which means that a basic segment does not exist at this location. A “z-score” shown in the last column of Table 11 is computed to identify basic segments where the reported crash frequency is much higher than typical. The z-score is calculated by: z-score = (computed – reported crashes)/(reported crashes)0.5 A typical segment would have a z-score near 0.0. A segment with a large positive z-score would have many more crashes than a typical segment with similar volume and length. The basic segment on I-10 beginning at MP 143.95 is noted to have the largest z-score. It has left- and right-side ramps adjacent to one another on the same roadbed which makes this segment unusual. For this reason, it is considered an outlier and was removed from the calibration database. Table 11. Screening of Basic Segments (excluding speed-change lanes) Location I-10 Segment Begin Milepost Fatal-and-Injury Crash Frequency, cr/3 years Computed 1 z-score Reported 138.00 19.4 28 1.9 138.95 19.4 26 1.5 139.38 21.7 47 5.4 139.92 26.2 39 2.5 141.67 14.7 29 3.7 Kittelson & Associates, Inc 2, 3 21 I-10 Phoenix Corridor Safety Study Location I-17 L 101 L 202 U.S. 60 Segment Begin Milepost 141.92 Final Report Fatal-and-Injury Crash Frequency, cr/3 years z-score 2, 3 30.4 89 10.6 142.43 14.5 47 8.5 142.71 7.0 18 4.1 143.67 14.9 38 6.0 143.95 16.8 64 11.5 147.51 0.0 0 includes two speed-change lanes 147.67 8.9 15 2.1 148.43 9.3 11 0.6 149.29 2.6 5 1.5 149.51 2.4 3 0.4 149.65 5.8 8 0.9 149.82 3.6 2 -0.8 151.22 6.6 8 0.6 151.78 7.0 5 -0.7 151.92 3.6 4 0.2 210.24 20.1 38 4.0 211.18 2.0 5 2.1 211.66 22.3 32 2.1 212.56 4.2 7 1.3 41.37 2.2 3 0.5 41.43 21.5 39 3.8 41.98 2.2 3 0.5 42.06 23.0 30 1.5 42.68 15.8 14 -0.5 43.04 1.9 1 -0.6 43.09 24.8 46 4.3 43.70 19.6 29 2.1 44.14 24.5 34 1.9 5.19 5.6 10 1.8 6.74 21.6 67 9.8 180.69 25.2 69 8.7 181.67 23.7 60 7.5 182.13 24.6 24 -0.1 183.72 9.5 16 2.1 184.09 22.9 35 2.5 Total: 552.0 1048 not applicable Notes: 1 – Based on typical urban freeway crash rate of 0.24 fatal-and-injury crashes per million vehicle miles (cr/mvm). 0.5 2 - z-score computed as (Computed - Reported)/(Reported) . 3 - Segments determined to be outliers are identified by highlighted Z-score. Kittelson & Associates, Inc 22 I-10 Phoenix Corridor Safety Study Final Report Screening of the twenty four speed-change lanes in the calibration database is listed in Table 12. The z-score for one of the speed-change lanes on the I-10 segment beginning at MP 143.95 has a high score. The corresponding basic segment was identified in Table 11 as an outlier so all three speed-change lanes associated with this segment were considered outliers and were removed. The z-score for the L202 segment beginning MP 6.74 was also found to have a high score. This speedchange lane occurs partly on a horizontal curve. It is also the first speed-change lane encountered by drivers (traveling in the decreasing milepost direction) for several miles. The upstream interchange traffic is served by auxiliary lanes (as opposed to speed-change lanes) such that the subject speedchange lane may be somewhat unexpected by drivers. For these reasons, the speed-change lanes and associated basic segment are considered outlier and were removed from the database. Calculation of Calibration Factors The ISATe spreadsheet was used to compute the predicted crash frequency for each of the basic calibration segments. The predicted values are summarized in. Also shown in the table is the reported crash frequency for each basic segment. These values are summed for all segments and the results are shown in the second-to-last row of the table. These sums are used to compute the calibration factors for the safety performance functions (SPFs) in ISATe. Specifically, each factor is computed as the ratio of the reported crash frequency to the predicted crash frequency. A similar process was used to compute the calibration factors for speed-change lanes. The data used and the resulting factors are shown in. Table 12. Screening of Speed-Change Lanes Location Segment Begin Milepost Speed-Change Lane Type Travel Direction Computed I-17 212.56 Exit Increasing 1.6 0 -1.3 I-10 151.77 Exit Decreasing 1.3 3 1.5 151.22 Exit Increasing 0.4 0 -0.6 149.82 Entrance Increasing 3.6 11 3.9 149.29 Exit Increasing 1.2 2 0.7 148.43 Exit Increasing 0.5 0 -0.7 148.43 Exit Decreasing 0.9 1 0.1 147.67 Entrance Increasing 1.0 3 2.1 147.67 Exit Increasing 0.9 0 -0.9 147.51 Entrance Increasing 3.0 1 -1.2 147.51 Entrance Decreasing 3.0 6 1.7 143.95 Entrance Increasing 4.5 24 9.3 143.95 Exit Increasing 1.2 4 2.6 143.95 Exit Decreasing 0.9 2 1.1 Kittelson & Associates, Inc Fatal-and-Injury Crash Frequency, cr/3 years 1 Reported z-score 2, 3 23 I-10 Phoenix Corridor Safety Study Location L 202 L 101 U.S. 60 Final Report Segment Begin Milepost 142.45 Speed-Change Lane Fatal-and-Injury Crash Frequency, cr/3 years Exit Increasing 1.7 4 1.7 141.92 Exit Increasing 1.4 3 1.3 6.74 Entrance Decreasing 4.5 27 10.6 6.74 Exit Decreasing 1.0 3 2.0 5.19 Exit Increasing 0.5 0 -0.7 41.98 Exit Increasing 0.7 4 3.9 41.43 Exit Decreasing 1.0 2 1.0 183.72 Exit Increasing 1.6 0 -1.3 183.72 Entrance Decreasing 7.4 15 2.8 181.68 Exit Increasing 1.6 4 2.0 45.4 119 Total: z-score 2, 3 Notes: 1 – Based on typical urban freeway crash rate of 0.24 fatal-and-injury crashes per million vehicle miles (cr/mvm) . 0.5 2 - z-score computed as (Computed - Reported)/(Reported) . 3 - Segments determined to be outliers are identified by highlighted Z-score . Table 13. SPF Calibration for Basic Freeway Segments (excl. speed-change lanes). Location I-10 Segment Begin Milepost Predicted Crash Frequency, 1 crashes/3years Single-Vehicle Reported Crash Frequency, crashes/3 years Multiple-Vehicle Single-Vehicle Multiple-Vehicle PDO FI PDO FI PDO FI PDO FI 138.002 7.5 3.6 50.9 19.0 13 3 46 25 138.954 11.3 5.1 33.6 13.8 26 4 60 22 139.375 12.3 5.7 28.4 12.4 18 7 72 40 139.923 12.2 6.5 60.1 23.5 17 8 70 31 141.67 7.8 4.0 32.2 11.7 12 3 86 26 141.92 14.2 8.1 71.0 29.1 24 8 202 81 142.425 7.1 3.0 30.3 12.5 11 6 83 41 142.705 3.7 1.5 12.4 5.4 2 1 30 17 143.67 11.0 4.4 44.2 17.5 8 5 59 33 143.945 outlier 147.51 0 0.0 0.0 0.0 0 0 0 0 147.67 6.4 2.3 16.9 7.3 2 0 21 15 148.43 4.8 2.2 45.9 21.9 2 0 25 11 149.287 1.6 0.6 7.3 3.4 2 1 8 4 149.512 4.2 1.9 7.6 2.3 3 1 4 2 149.646 5.4 2.1 11.7 4.1 2 2 22 6 Kittelson & Associates, Inc 24 I-10 Phoenix Corridor Safety Study Location I-17 Segment Begin Milepost Predicted Crash Frequency, 1 crashes/3years Single-Vehicle Reported Crash Frequency, crashes/3 years Multiple-Vehicle Single-Vehicle Multiple-Vehicle PDO FI PDO FI PDO FI PDO FI 149.818 1.5 0.7 8.4 3.2 0 0 14 2 151.223 2.7 1.4 15.1 5.7 4 0 18 8 151.775 3.6 1.8 28.4 12.6 2 0 11 5 151.92 1.9 0.7 20.0 8.5 4 0 9 4 210.24 11.8 5.2 59.0 20.7 13 9 55 29 211.18 1.1 0.5 5.2 1.9 0 0 13 5 211.655 14.1 6.3 47.2 17.7 14 6 55 26 212.56 2.5 1.1 13.9 5.7 2 3 9 4 41.37 1.2 0.7 6.0 2.4 3 1 6 2 41.43 14.7 7.9 73.0 28.1 16 11 51 28 41.98 2.2 1.1 8.7 3.0 3 0 18 3 42.055 17.4 8.8 61.3 21.1 9 9 58 21 42.675 9.6 4.8 50.2 17.5 14 2 34 12 43.04 1.2 0.6 6.0 2.1 3 1 3 0 43.085 14.8 7.0 62.8 22.3 10 3 83 43 43.695 9.6 4.6 55.9 20.2 16 3 56 26 44.14 13.1 5.7 55.9 20.6 18 4 54 30 5.19 5.7 2.4 13.5 5.2 7 4 15 6 6.735 outlier 180.695 16.9 7.4 57.3 21.6 24 9 103 60 181.675 14.2 6.7 55.6 22.9 10 6 113 54 182.13 17.3 7.3 34.7 14.2 13 6 80 18 183.72 6.0 2.9 13.7 5.8 9 3 12 13 184.095 16.1 6.7 29.0 12.3 10 6 32 29 308.7 143.3 1232.9 479.2 346 135 1690 782 1.12 0.94 1.37 1.63 L 101 L 202 U.S. 60 Final Report Total: Calibration Factor: Note: 1 - Based on safety prediction models developed for NCHRP Project 17-45 (and in ISATe). Kittelson & Associates, Inc 25 I-10 Phoenix Corridor Safety Study Final Report Table 14. SPF Calibration for Speed-Change Lanes. Location Segment Begin Milepost Predicted Crash Frequency, 1 crashes/3years Entrance Ramp Reported Crash Frequency, crashes/3 years Exit Ramp Entrance Ramp Exit Ramp PDO FI PDO FI PDO FI PDO FI I-17 212.56 -- -- 3.8 1.6 -- -- 0 0 I-10 151.77 -- -- 2.9 1.3 -- -- 1 3 151.223 -- -- 0.9 0.4 -- -- 1 0 149.818 8.2 5.3 -- -- 22 11 -- -- 149.287 -- -- 2.1 0.9 -- -- 6 2 148.43 -- -- 1.5 0.8 -- -- 5 0 148.43 -- -- 1.5 0.8 -- -- 4 1 147.67 1.7 1.1 -- -- 1 3 -- -- 147.67 -- -- 1.5 0.7 -- -- 4 0 147.51 4.7 2.5 -- -- 9 1 -- -- 147.51 4.7 2.5 -- -- 8 6 -- -- 143.945 outlier 143.945 outlier 143.945 outlier 142.425 -- -- 3.3 1.5 -- -- 26 4 141.92 -- -- 2.5 1.2 -- -- 7 3 6.735 outlier 6.735 outlier 5.19 -- -- 1.2 0.6 -- -- 1 0 41.98 -- -- 2.2 1.0 -- -- 2 4 41.43 -- -- 2.5 1.1 -- -- 1 2 183.72 -- -- 3.0 1.3 -- -- 3 0 183.72 10.0 4.6 -- -- 27 15 -- -- 181.675 -- -- 3.1 1.4 -- -- 7 4 29.3 15.9 31.8 14.5 67 36 68 23 2.29 2.27 2.14 1.58 L 202 L 101 U.S. 60 Total: Calibration Factor: Note: 1 - Based on safety prediction models developed for NCHRP Project 17-45 (and in ISATe). The severity distribution function (SDF) is calibrated using the crash data for the injury and fatal crashes. The procedure for quantifying this calibration factor is described in the final report for NCHRP Project 17-45. This factor adjusts the SDF in ISATe such that the predicted frequency of K, A-injury, B- Kittelson & Associates, Inc 26 I-10 Phoenix Corridor Safety Study Final Report injury, and C-injury crashes matches more closely to that for the local conditions (e.g., the Phoenix calibration sites). The data used to compute the SDF calibration factor is shown in Table 15. An examination of these data indicates that the proportion of reported K-fatal, A-injury, and B-injury crashes is 0.14. However, this same proportion for the predicted crashes is 0.082. Thus, the uncalibrated SDF is under-predicting the proportion of K, A, and B crashes. The calibration factor shown in the last row of the table will remove this bias such that the SDF model predictions will more reliably reflect the distribution of K, A, and B crashes on the Phoenix metro area freeways. This factor is used for both the basic segment SPFs and the speed-change lane SPFs. Table 15. SDF Calibration for Basic Segments and Speed-Change Lanes Combined. Severity Category Predicted Crash Frequency, 1 crashes/3years Reported Crash Frequency, crashes/3 years K - fatal 8.5 8 A - incapacitating injury 22.4 57 B - non-incapacitating injury 153.8 382 C - possible injury 458.2 529 Total KAB 184.7 447 Total KABC 652.9 976 Ratio KAB/KABC 0.283 0.458 Calibration Factor: 2.14 Note: 1 - Calibration factor = Reported Ratio / (1.0 - Reported Ratio) x (1.0 - Predicted Ratio) / Predicted Ratio. Application of the Calibrated Model The eight calibrated SPFs and calibrated SDF provide a crash prediction model that accurately reflects safety performance for Phoenix freeways. The model can be used to effectively evaluate the safety performance of freeway sections with physical and operational characteristics similar to the calibration freeway sections. The ISATe spreadsheet tool is now available to ADOT for use in applying the model on future safety performance studies. The model can also be used to accurately assess freeway design alternatives relative to crash frequency, and evaluate crash mitigation measures. Since the model calibration segments included continuous access HOV lanes and ramp meters, the model should not be applied to freeway sections without HOV lanes or with access controlled HOV lanes, as well as freeways without ramp metering. Care should be used in applying the model to freeway sections with posted speed limits below 65 mph, in which case the free-flow speed should be approximately 65 mph. Kittelson & Associates, Inc 27 I-10 Phoenix Corridor Safety Study Final Report The geometry, traffic, and crash data for the I-10 study section were entered into the ISATe software. The error-checking routines were used to verify that the data were entered correctly. The calibration factors were also entered into the ISATe software. The expected crash frequency (by type and severity) was then computed for each speed-change lane and basic segment using the Empirical Bayes Method, described in the HSM. The ISATe output is provided in Appendix B. Other performance measures were also computed and are described in the next section. These performance measures were computed for a study period that was coincident with the crash period (2011 - 2013). The safety performance for future or prior years was not estimated, but the software will support the safety evaluation of other years if the traffic volumes for these years can be estimated. The two segments having left-side (HOV) and right-side ramps on the same roadbed could not be evaluated directly due to limitations of the predictive method. The technique used to overcome this limitation was to disaggregate the two “left+right” segments that could be evaluated with the method. Specifically, each of the two “left+right” segments was entered into ISATe as three simplified segments whose results could be combined to represent the safety of a “left+right” segment. One simplified segment (i.e., segment 1) described the segment geometry, ramp traffic, and crash as if there were no ramps. A second simplified segment (i.e., segment 2) described the segment as if there were only leftside ramps. The third simplified segment (i.e., segment 3) described the segment as if there were only right-side ramps. The crash estimate for the “left+right” segment was then computed as equal to the safety estimate for Segment 2 plus that for Segment 3 minus that for Segment 1. EVALUATION OF THE SAFETY PERFORMANCE OF THE I-10 STUDY SECTION Based on the calculated performance measures, the segments were ranked to help identify those segments which have the greatest potential for crash reduction and were the focus of the detailed crash diagnostic process. PERFORMANCE MEASURES Figure 5 lists the 13 HSM performance measures that were considered for evaluation of the I-10 study section. Moving down the list, the potential for bias that occurs when safety evaluations are based on short term crash histories (e.g. 2 to 3 years) decreases. This bias, known as regression-to-the-mean (RTM), reflects the potential that the short term crash frequency observed at a site may represent either an abnormally high or low crash frequency relative to the long term crash frequency at the site. Reducing or eliminating the RTM bias provides greater Figure 5. Alternative Safety Performance Measures Kittelson & Associates, Inc 28 I-10 Phoenix Corridor Safety Study Final Report accuracy in identifying sites that have the highest potential for crash reduction. Given the intent of this study, only performance measures that account for RTM bias were used to evaluate safety performance on the I-10 study section segments. The following six performance measures were calculated for each of the basic segments and speedchange lanes in the I-10 study section. Each of these performance measures either is not affected by RTM bias or accounts for the bias through an adjustment using the Empirical Bayse (EB) methodology. 1. Probability of specific crash types exceeding threshold proportion (Crash Type Performance Threshold) 2. Excess proportion of specific crash type 3. Expected average crash frequency with EB adjustment 4. Equivalent PDO average crash frequency with EB adjustment 5. Excess expected average crash frequency with EB adjustment 6. Modified level of service of safety (LOSS) Discussion of each performance measure, including their strengths and weaknesses in evaluating safety performance, and results relative to the I-10 study section are provided in Appendix A. The six performance measures describe the magnitude of the safety problem within the I-10 study section in essentially two dimensions: (a) relative to other segments in the I-10 study section (the ranks do this); and (b) relative to other freeways relative to other similar freeway segments in the Phoenix region. In addition to describing the safety performance within the study segment, these measures were used to identify those segments with high potential to reduce crashes. The two primary ranking measures used were the z-score calculated to define the level of service of safety (LOSS) and the expected average crash frequency with EB adjustment (EACFEB). The z-score indicates whether a site has potential for improvement beyond that potential already identified by the CMF values calculated for each segment. For example, the segment with the 90-degree horizontal curve through the L202/SR51 interchange has a combined CMF value of 1.67 and a relatively high z-score of 4.18. The combined CMF value indicates that 67% of the expected crashes in this segment result from the physical characteristics of the roadway (horizontal curve, shoulder width, presence of ramps, etc.). In this particular segment, the horizontal curve is the primary contributing physical characteristic affecting safety performance. While flattening out the curve is not a reasonable solution, other treatments for curvature may be appropriate such as enhanced signing or delineation. Therefore, when using CMF corrected crash predictions, it is important to screen for locations with high composite CMFs and locations with high expected excess values. Still, the high z-score for this segment indicates that there are other factors beyond the physical characteristics described by the calculated CMF values that are contributing to the expected crashes in the segment. Determining these other factors and identifying potential countermeasures requires diagnosis involving a detailed review of site conditions, crash trends, and crash reports. The LOSS results presented in Figure 6 indicate that nearly all of the basic segments and a majority of the speed-change lanes in the study section, either considering all crashes or just fatal and injury crashes, have a moderate to high potential to find safety improvement opportunities through diagnosis Kittelson & Associates, Inc 29 I-10 Phoenix Corridor Safety Study Final Report of site conditions and a review of crash history. These results also suggest that there may be common factors within the entire study section contributing to the crashes occurring. The red lines in each graph delineate the potential (low, moderate, or high) of identifying opportunities to reduce crashes. Figure 7 presents the expected average crash frequencies per mile for each study section segment. The red line in each graph is the expected average crash frequency for the 40 Phoenix area freeway segments used to calibrate the crash prediction models. These calibration values included segments on 1-10, I-17, L101, L202, and US60, which are similar to those segments found within the I-10 study section. Therefore, while it may not be fully accurate to call these values “typical” for Phoenix freeways, they are fairly representative of urban freeways throughout the area and therefore offer a solid reference to assess the performance of the I-10 study section and priority for safety improvement. These results show that approximately half of the segments in the I-10 study section have crash frequencies that are substantially higher than other similar Phoenix urban freeways. The results confirm the intent of this study to focus on identifying potential countermeasures to reduce the high crash frequency within this section of I-10. Considering these two performance measures, LOSS and EACFEB, the safety performance evaluation indicates that not only can this freeway section be considered a priority, but there is a high potential to reduce crash frequency with the application of appropriate countermeasures. The numerical LOSS and EACFEB results for the basic freeway segments and speed-change lanes are provided in Appendix C. Kittelson & Associates, Inc 30 I-10 Phoenix Corridor Safety Study Final Report a. All Crashes (Fatal, Injury, and PDO) b. Fatal and Injury Crashes only Figure 6. Level of Service of Safety (z-score) for Study Section Basic Segments. Kittelson & Associates, Inc 31 I-10 Phoenix Corridor Safety Study a. Final Report All Crashes (Fatal, Injury, and PDO) b. Fatal and Injury Crashes Only Figure 7. Expected Average Crash Frequency per Mile Kittelson & Associates, Inc 32 I-10 Phoenix Corridor Safety Study Final Report CRASH DIAGNOSTIC PROCESS DEVELOPMENT OF GIS BASED CRASH DATABASE The ADOT crash database for the years 2012 and 2013 was imported into a GIS environment. Initially, crashes were located along the center line of the freeway mainline for each direction of travel utilizing latitude and longitude and an appropriate coordinate system. The distribution of the 2-years of crashes within the entire study section is provided in Figure 8. Each blue dot represents a crash. Figure 9 shows the crash locations on one segment of the study area. Using the GIS crash database, the distribution of different crash attributes including severity, manner of collision, first harmful event, lighting condition, and weather condition were reviewed. Figure 10 shows the crash distribution for a segment by severity type. FMS and HPMS counting stations were added to the database so that traffic volume and operations information can be included in the review of crash trends. ANALYSIS OF CRASH PATTERNS AND FREEWAY OPERATIONAL CHARACTERISTICS An investigative analysis on the crash data and freeway operational data was conducted in an attempt to identify a direct correlation between operational characteristics and crash patterns. Crash distribution by time of day, collision manner, and drivers’ maneuvering action were analyzed for each study section segment. FMS traffic operational data was used to develop speed and congestion profiles for the study area for a typical weekday. These data indicate that the I-10 eastbound direction is typically congested in the morning peak hours from 35th Avenue to 19th Avenue, and the eastbound section from 19th Avenue to 16th Street is congested in both morning peak and afternoon peak hours. For the I-10 westbound direction, the section from 16th Street to 35th Avenue is congested during afternoon peak hours. Kittelson & Associates, Inc 33 I-10 Phoenix Corridor Safety Study Final Report Figure 8. Two-year (2012-2013) Crashes within the I-10 Study Section Kittelson & Associates, Inc 34 I-10 Phoenix Corridor Safety Study Final Report Figure 9. Crash locations on one segment (Deck Park Tunnel to 7th St) Figure 10. Distribution of Crashes by Severity (Deck Park Tunnel to 7th St) Kittelson & Associates, Inc 35 I-10 Phoenix Corridor Safety Study Final Report The crash distribution for the entire Table 16. Crash Distribution by Table 17. Crash Distribution by study section by time of day provided in Time-of-Day for Study Section Time-of-Day 3rd St HOV Ramp to 7th St Ramp Table 16 shows that approximately 70% of observed crashes in 2012 and 2013 occurred during the peak traffic hours. Analysis of individual study section segments provided insight into crash patterns. For example, Table 17 presents the time-of-day crash distribution for the segment from the 3rd St HOV on-ramp to the 7th St on-ramp, Figure 11 and Figure 12 show the speed and congestion profile of two adjacent FMS count stations for this segment. These data support the hypothesis that most crashes are related to congestion and high traffic volumes. Further analysis shows that most crashes that occurred in this segment were rearend crashes and that prior to the crash both drivers were either slowing or were stopped in the traffic way. Table 18 and Table 20 show crash distribution by type of collision and drivers’ maneuvering action, respectively. Time of day distribution for the crashes associated with stopped and/or slowing traffic, Table 19 shows that the crashes are directly correlated to the congestion and speed reduction on the freeway. Appendix D includes the statistics for the westbound direction for this segment. Figure 13 and Figure 14 show the direct correlation between the speed reduction and the crashes that occurred in both westbound and eastbound directions in this segment due to vehicles stopped and/or slowing in traffic. Kittelson & Associates, Inc 36 I-10 Phoenix Corridor Safety Study Final Report Figure 11. Speed profile for eastbound I-10, MP 146.96 Figure 12. Speed profile for eastbound I-10, MP 146.06 Kittelson & Associates, Inc 37 I-10 Phoenix Corridor Safety Study Final Report Table 18. Collision Type Distribution (Segment – I-10 EB, MP 145.78) Collision Type REAR_END SIDESWIPE_OPPOSITE_DIRECTION SIDESWIPE_SAME_DIRECTION SINGLE VEHICLE Total Frequency 18 1 2 1 22 Percent 81.82% 4.55% 9.09% 4.55% 100.00% Table 20. Drivers’ Maneuvering/Action (Segment – I-10 EB, MP 145.78) Drivers Action Frequency Percent CHANGING_LANES 4 8.70% GOING_STRAIGHT_AHEAD 17 36.96% SLOWING_IN_TRAFFICWAY 14 30.43% STOPPED_IN_TRAFFICWAY 7 15.22% UNKNOWN 4 8.70% Total 46 100.00% Kittelson & Associates, Inc Table 19. Distribution of Slowing/Stopped in the Traffic Way by Time of Day (Segment – I-10 EB, MP 145.78) Time of Day Frequency Percent 0:00 1:00 0 0 0.00% 0.00% 2:00 3:00 0 0 0.00% 0.00% 4:00 5:00 0 0 0.00% 0.00% 6:00 7:00 2 2 9.52% 9.52% 8:00 9:00 0 0 0.00% 0.00% 10:00 11:00 0 0 0.00% 0.00% 12:00 13:00 0 1 0.00% 4.76% 14:00 15:00 1 3 4.76% 14.29% 16:00 17:00 2 7 9.52% 33.33% 18:00 19:00 20:00 21:00 3 0 0 0 14.29% 0.00% 0.00% 0.00% 22:00 23:00 0 0 0.00% 0.00% Total 21 100.00 % 38 I-10 Phoenix Corridor Safety Study Final Report Speed Profile I-10 WB, MP 144.641 80 23% 70 20.0% 20% 60 Speed (mph) 25.0% 50 15.0% 14% 40 13% 12% 10.0% 30 20 5.0% 10 3% 0% 0% 0% 0% 0% 0% 0% 1% 0% 2% 1% 2% 1% 0% 1% 0.0% 0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 6:00 6:30 7:00 7:30 8:00 8:30 9:00 9:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30 16:00 16:30 17:00 17:30 18:00 18:30 19:00 19:30 20:00 20:30 21:00 21:30 22:00 22:30 23:00 23:30 0:00 0 3% 2% 2% Time of Day HOV lane1 lane2 lane3 lane4 Slowing/Stopped Crash Figure 13. Correlation between Speed and Crashes due to Slowing/Stopped Maneuvering (Segment I-10 WB, MP 144.42) Speed Profile I-10 EB, MP144.955 80 22.0% 19% 70 Speed (mph) 20% 20.0% 19% 18.0% 60 16.0% 50 14.0% 12.0% 40 10.0% 9% 30 8.0% 7% 20 7% 5% 4% 3% 10 0% 0% 0% 0% 0% 4.0% 4% 2.0% 0% 0% 0% 0% 0% 0% 0% 0% 0.0% 0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 6:00 6:30 7:00 7:30 8:00 8:30 9:00 9:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30 16:00 16:30 17:00 17:30 18:00 18:30 19:00 19:30 20:00 20:30 21:00 21:30 22:00 22:30 23:00 23:30 0:00 0 1% 6.0% Time of Day 3/18/2013 3/18/2013 3/18/2013 3/18/2013 3/18/2013 Slowing/Stopped Figure 14. Correlation between Speed and Crashes due to Slowing/Stopped Maneuvering (Segment I-10 EB, MP 144.42) Kittelson & Associates, Inc 39 I-10 Phoenix Corridor Safety Study Final Report CRASH DIAGNOSTIC ANALYSIS OF STUDY SECTION SEGMENTS A detailed diagnostic analysis was conducted in an attempt to determine the primary contributing factors to observed crashes. As a part of this analysis 756 crash reports of 3,255 crashes (23%) occurring in the study section in 2011 and 2013 were reviewed. The crash reports were selected based on the review of the crash data to identify trends or areas of interest within each segment. The diagnostic analysis was conducted in three phases. First, a subset of study section segments which exhibited high crash frequencies per mile when compared with the average for the broader set of Phoenix freeway segments used for the predictive model calibration and/or in which existing roadway characteristics are contributing substantially to observed crashes was selected for review. The criteria used to select this subset were the estimated z-score and the cumulative effect of crash modification factors (CMFs) for each segment. The z-score compares the expected crashes of a study segment with the average expected crashes for the 40 Phoenix freeway segments used to calibrate the crash prediction models. High z-scores reflect segments where there is high potential for crash reduction as a result of factors not explained by the models. The cumulative CMF of a segment reflects the level at which roadway and traffic characteristics included in the model contribute to the expected crash frequency. High potential for achieving crash reduction in a segment is indicated by a high z-score, a high combined CMF, or both Based on the z-scores and combined CMFs, eleven segments were selected for the Phase 1 diagnostic analysis. The selected segments are listed in Table 21. Figure 15 displays the z-scores and cumulative CMF values for all of the study segments, with those selected for further analysis highlighted. Table 21. List of Selected Segments for Phase 1 Detail Crash Report Review Segment th I-17 to 19 Ave th th 7 Ave Exit Ramp to 5 Ave HOV Exit Ramp th 5 Ave HOV Exit Ramp to WB Lane Drop th West of 7 Ave th th WB Lane Drop West of 7 Ave to 7 Ave Entry th th 7 Ave Entry to 7 St Exit (Deck Park Tunnel) th rd Lane Drop East of 7 St to 3 St HOV Entry rd th 3 St HOV Entry to 7 St Entry th th 7 St Entry to 16 St Exit th 16 St Exit to SR 51 HOV Lane Exit L202/SR51 Exit to L202/SR51 Entry SR 202 HOV Lanes to Jefferson St Entry Kittelson & Associates, Inc Begin MP 142.30 144.36 End MP 143.67 - 144.42 Speed Change Lane Combined CMF 1.44 3.15 z-score 2.69 1.54 144.80 1.19 9.22 144.80 144.92 1.33 7.61 144.96 145.38 2.26 6.78 145.52 145.78 145.94 146.44 146.93 148.31 145.78 145.82 146.38 146.64 147.45 - 1.40 1.82 1.74 1.41 1.67 5.04 5.97 1.59 3.72 4.59 4.18 -1.21 Exit Ent. 40 I-10 Phoenix Corridor Safety Study Final Report The second phase of the diagnostic review focused on known geometric or operational features within the study section that were expected to be primary contributing factors in observed crashes. These features primarily included the left-hand HOV entry/exit ramps, lane drops, weaving sections, tunnel lighting, and HOV lane access. Based on these features, 19 additional segments were reviewed. Finally, diagnostic review of all fatal, and serious injury crashes (total of 46 crashes) was conducted. Though the proportion of fatal and serious injury crashes (1.4%, 46 fatal and serious injury crashes out of 3,255 total crashes) is small, and recognizing the importance of more severe crashes, the fatal and serious injury crash reports were reviewed to investigate whether or not they show any specific pattern or causal factors. 10.0 8.0 Phase 1 6.0 z-score Phase 2 Not Reviewed 4.0 2.0 0.0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 -2.0 Combined CMF Figure 15. Distribution of z-score and Cumulative CMF of the I-10 Corridor Study Segments Phase I Analysis – High z-Score and Combined CMF Segments In the Phase 1 analysis, 323 crash reports out of 1,687 crashes that occurred in 2012 and 2013 were reviewed. Approximately 15 collision reports for crashes that occurred during the periods of congestion for each direction of travel were randomly selected for each segment. For each detailed crash report, a short summary was generated. The effort also included identifying the exact position of the vehicles on the roadway including the lane number during the crash, manner of collision, lighting conditions, vehicular conditions, driver behavior, vehicular speed, traffic conditions, and any influencing factors. Based on the information extracted from the crash reports, FMS speed profile, and features of the Kittelson & Associates, Inc 41 I-10 Phoenix Corridor Safety Study Final Report roadway segment, a brief summary (per direction of travel) was developed for each segment. Then the information was imported into the GIS database. A collision diagram for each crash was also drawn based on the information contained in the crash report. These collision diagrams were not intended to duplicate or replace existing techniques adopted by ADOT, but were used only for analysis by the study team. A library in GIS environment was developed for the collision diagrams to represent the exact crash location on the roadway which helped the study team to visually identify traffic issues in the roadway, potential contributing factors, and what could be done to resolve the issues. Table 22 provides an example of the crash report review summary for one of the segments. Figure 16 shows a sample collision diagram identifying the exact location of each crash within a segment. Phase I Findings The following section provides the diagnostic summary and discusses the contributing factors for the Phase I analysis and crash report review. Eastbound, I-17 (MP143.00) to 19th Ave (MP143.66): In this segment, a total of 86 crashes occurred during 2012 and 2013, and 15 crash reports were reviewed. Ramp traffic from both northbound and southbound I-17 merge with I-10 eastbound traffic just east of this segment. Eastbound I-10 traffic west of I-17 travels at a high speed and suddenly encounters slowing and/or stopped traffic. Congestion, speed, and merging traffic were contributing factors to the crashes in this segment. Eastbound, 7th Ave Exit Ramp (MP144.36) to 5th Ave HOV Exit Ramp (MP144.42): In this segment, a total of 32 crashes (533 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. There is a left-side HOV exit to 5th Avenue in this segment. Higher speed approaching traffic rear-ending stopped traffic on this downgrade, mainly in lane 1 (median lane),is a common crash type. Speed and congestion are the main contributing factors for these crashes. Eastbound, 5th Ave HOV Exit Ramp (MP144.42) to WB Lane Drop West of 7th Ave (MP144.80): In this segment, a total of 178 crashes (468 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. This segment is a downgrade and is located west of the tunnel. Higher speed approaching traffic encounters congestion, resulting in rear end crashes. Congestion and speed are contributing factors in this segment. Eastbound, WB Lane Drop West of 7th Ave (MP144.80) to 7th Ave Entry (MP144.92): In this segment, a total of 54 crashes (450 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. This segment is the west end of the tunnel. Higher speed approaching traffic encounters congestion, often resulting in rear-end crashes. There is merging of heavy, unmetered traffic from the 7th Ave entry ramp that contributes to congestion. Congestion, speed, and merging, were contributing factors for these crashes. Eastbound, 7th Ave Entry (MP144.96) to 7th St Exit (MP145.42)(Deck Park Tunnel): In this segment, a total of 202 crashes (439 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. This is a majority of the tunnel segment. Traffic is congested, and higher speed traffic encounters slowing and/or stopped traffic and a change in lighting conditions. The 7th Street exit starts Kittelson & Associates, Inc 42 I-10 Phoenix Corridor Safety Study Final Report at the east end of the tunnel segment. Congestion, speed, lighting conditions, tunnel effects, and horizontal curve are the contributing factors in this segment. Westbound, 5th Ave HOV Entry Ramp (MP144.42) to 7th Ave Entry Ramp (MP144.36): In this segment, a total of 17 crashes (283 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. Traffic entering from 7th Avenue merges just west of this segment. There is also a left-side HOV entrance from 5th Avenue. There is a significant amount of lane changing resulting in slowing traffic which leads to sideswipe or rear-end crashes. Congestion from downstream bottlenecks can extend into this segment. Congestion and lane changing are the primary contributing factors in this segment. Westbound, Lane Drop West of 7th Ave (MP144.80) to 5th Ave HOV Entry Ramp (MP144.42) : In this segment, a total of 118 crashes (311 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. This segment contains the merge area from a right-side lane drop. After leaving the tunnel traffic starts accelerating, but then often encounters slowed or stopped traffic. Merging from the lane drop, often late and forced, contributes to congestion. Congestion, speed, and merging associated with the lane drop are contributing factors in this segment. Westbound, 7th Ave Exit Ramp (MP144.92) to Lane Drop West of 7th Ave (MP144.80): In this segment, a total of 70 crashes (583 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. This segment is at the west end of the tunnel (including the 7th Avenue exit ramp). Westbound motorists exiting the tunnel on a downgrade and on a curve, often accelerating, typically encounter slowed or stopped traffic during peak hours. Rear end crashes are predominant. Several crash reports noted glare from the setting sun as a factor. The effect of the transition from the tunnel illumination conditions to much brighter ambient conditions, particularly in the afternoon, is also a factor. Congestion, speed, differential illumination levels, sun glare, and roadway geometry (downgrade and horizontal curve) are contributing factors in this segment. Westbound, 7th St Entry Ramp (MP145.42 to 7th Ave Exit Ramp (MP144.96)(Deck Park Tunnel)): In this segment, the Deck Park Tunnel, a total of 214 crashes (465 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. Traffic in the tunnel is very congested during peak traffic periods. The 7th Street entry ramp (lane add) and 7th Avenue exit ramp create a Type B weaving section. Heavy, unmetered, traffic enters the westbound tunnel from 7th Street, weaving with traffic destined for the 7th Avenue exit. The illumination levels in the tunnel likely compound the effect of the horizontal curve and grade, downgrade in the westbound direction, and make weaving maneuvers more challenging. Several crash reports noted that the vehicle drifting out of its lane, suggesting the motorists are having trouble tracking the lane lines. Congestion, speed, tunnel illumination, weaving traffic, roadway geometry (horizontal curve, grade), and overall tunnel effects, are the contributing factors. Westbound, 1st St HOV Exit (MP145.78) to Lane Drop East of 7th St Exit (MP145.52): In this segment, a total of 85 crashes (304 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. High speed traffic encounters congestion of either slowing or stopped traffic approaching the tunnel. Motorists attempting to change lanes to avoid congested lanes are often involved in crashes. Congestion, speed, and lane changing were contributing factors. Kittelson & Associates, Inc 43 I-10 Phoenix Corridor Safety Study Final Report Table 22. Sample Crash Summary for Segment I-17 (MP142.99) to 19th Ave (MP143.66) Kittelson & Associates, Inc 44 I-10 Phoenix Corridor Safety Study Final Report Figure 16. Sample Collision Diagram for a Segment (7th St to 16th St) on a GIS Platform Kittelson & Associates, Inc 45 I-10 Phoenix Corridor Safety Study Final Report Westbound, 7th St Exit (MP145.84) to 1st St HOV Exit (MP145.78): In this segment, a total of 22 crashes (367 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. Higher speed traffic encounters slowing/stopped traffic approaching the tunnel. The left-side HOV exit does not appear to be a factor. Congestion and speed are the contributing factors. Westbound, 16th St Entry (MP146.44) to 7th St Exit (MP145.9): In this segment, a total of 151 crashes (280 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. This segment contains a one-lane on-ramp from 16th Street and a two-lane off-ramp to 7th Street. The weaving in this segment, particularly in the morning, is heavy between traffic entering from L202/SR51 and 16th Street, and traffic destined for 7th Street. In the evening peak period, higher speed traffic often encounters congestion. Rear-end and sideswipe crashes are predominant. Congestion, weaving, and speed are contributing factors. Westbound, L202/SR51 Entry (MP146.93) to 16th St Entry (MP146.44): In this segment, a total of 75 crashes (188 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. Traffic from southbound SR-51, westbound Loop 202, westbound I-10, and an additional left-side HOV lane from Loop 202 merge together. Higher speed traffic encounters a significant amount of lane changing and slowed traffic and often fails to stop in time. Congestion, speed, and lane changing are contributing factors. Westbound, L202/SR51 Exit (MP147.52) to L202/SR51 Entry (MP146.93): In this segment, a total of 88 crashes (149 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. The segment includes the 90-degree curve which accommodates high-speeds. High speed traffic exiting the curve often encounter slowed or stopped traffic created by downstream bottlenecks. Congestion and speed are contributing factors. Twenty sideswipe collisions caused by lane changing or vehicles drifting outside their lanes occurred within the curve. Westbound, Jefferson St Exit (MP148.44) to L202 HOV Lanes (MP148.33): In this segment, a total of 17 crashes (142 crashes per mile) occurred during 2012 and 2013, and 15 crash reports were reviewed. In this segment there is a left-side HOV diverging ramp to northbound L202/SR-51. The right-side diverging ramp to northbound SR-51 and eastbound Loop 202 is just north of this segment (one lane must exit and a second lane is an optional exit). Relatively high speed traffic often encounters slowed and/or stopped traffic. Congestion, speed, and lane changing are contributing factors. Phase II Analysis – Known Operational Issues/Hot Spots The segments included in the Phase II analysis are listed in Table 23. The z-scores and cumulative CMFs for these segments are included in Figure 15. Segments included in the Phase II analysis included several Phase I segments, as well as 19 additional segments (highlighted in Table 23). Note that several segments are listed multiple times depending upon the feature that was reviewed. In all, 387 crash reports were reviewed in the Phase II analysis. Kittelson & Associates, Inc 46 I-10 Phoenix Corridor Safety Study Final Report Phase II Findings HOV Lanes; Westbound 5th Ave HOV Entrance and Eastbound 3rd St HOV Entrance, and Loop 202 HOV Entrance: In these segments, a total of 314 crashes occurred during 2012 and 2013, and 84 of them were reviewed. Crash types included the following: Rear-end crashes due to motorists making a sudden lane change from lane 1 to the HOV lane while trying to avoid congestion (18 crashes). The crashes attributed to the left side HOV entry ramps at the 5th Avenue and 3rd Street entrance were lower than anticipated and primarily involved rear-end crashes within the merge areas (5 crashes). Higher speed vehicles encountering slowed and/or stopped vehicles were reported in 41 crashes reviewed. Contributing crash factors appear to include: open access to HOV lanes, congestion, and speed differential. Table 23. Segments in the Phase II Crash Analysis Features/Issues 5th Ave HOV Entrance 3rd St HOV Entrance HOV lane access to/from through lanes I-17 to 7th St Tunnel lighting, Light level differential at tunnel entries/exits WB Lane Drop West of 7th Ave Exit EB Lane Drop, East of 7th St Exit EB Lane Drop at Washington St WB Sky Harbor Blvd to Washington St WB, 16th St Entrance to 7th St Exit EB 16th St to L202/SR-51 EB 35th Ave Entrance to 27th Ave Exit WB 27th Ave Entrance to 35th Ave Exit Kittelson & Associates, Inc Begin MP 144.42 145.78 143.67 143.95 144.30 144.31 144.36 144.42 144.80 144.92 144.96 144.92 144.96 145.38 144.42 144.80 145.42 145.52 147.96 148.07 148.43 148.64 145.82 145.94 146.38 146.44 146.44 146.64 146.76 146.85 146.89 141.92 142.42 End MP 144.80 145.82 143.95 143.30 144.31 144.36 144.42 144.80 144.92 144.96 145.38 144.96 145.38 145.42 144.80 144.92 144.52 145.78 148.07 148.31 148.64 148.66 145.94 146.38 146.44 146.64 146.64 147.76 146.85 146.89 146.93 142.42 141.92 Crash Reports of Interest Weekday morning and evening peak period Weekday morning and evening peak period Weekday morning and evening peak period If possible, screen crash reports for only crashes occurring in lane 1 or HOV lane. Weekday off-peak; 10:00 AM to 3:00 PM Weekday morning and evening peak period Weekday morning and evening peak period Weekday morning and evening peak period Weekday morning and evening peak period Weekday morning peak period Weekday evening peak period Weekday morning peak period Weekday evening peak period 47 I-10 Phoenix Corridor Safety Study Final Report Tunnel Lighting (MP144.92 to MP145.52): In this segment, a total of 479 crashes occurred during 2012 and 2013, and 60 crash reports were reviewed. Several of the crash reports reviewed noted that sun glare (3 crashes) and lighting conditions (5 crashes) contributed to the crash. The noted lighting effect, combined with the horizontal roadway curvature, caused drivers to “drift” into an adjacent lane. Crashes caused by lane changing within the tunnel were noted on 38 crashes reports. Contributing crash factors appear to include: congestion, speed differential, lane changing, tunnel lighting, sun glare, and roadway horizontal curvature. Westbound Lane Drop West of 7th Ave Exit, (MP144.42 to MP144.92): In this segment, a total of 188 crashes occurred during 2012 and 2013, and 18 crash reports were reviewed. This segment begins at the lane drop and ends at the left-side 5th Ave HOV entry ramp. The lane changing at the lane drop was noted for 6 crashes. Further investigation found that 22 out of 38 crashes occurred in lanes 4 and 5 (closest to the shoulder) in the lane drop influence area (200 feet upstream to 100 feet downstream of lane drop point), suggesting that the forced merge created by the lane drop contributed to 22 crashes. Congestion, speed differential, and the lane drop were contributing factors for these crashes. Eastbound Lane Drop, East of 7th St Exit, (145.42 to MP145.78): In this segment, a total of 146 crashes occurred during 2012 and 2013, and 15 crash reports were reviewed. This segment begins at the lane drop starting 400 feet east of the tunnel and contains the merge area, and ends at the left-side 3rd St HOV on-ramp. In four crashes, traffic in lane 4 and lane 5 were involved in rear-end crashes attributed to congestion caused by merging traffic. Further analysis revealed that 14 out of 32 crashes occurred in lanes 4 and 5 in the lane drop influence area (200 feet upstream to 100 feet downstream of lane drop). Congestion, speed differential, and lane changing were contributing factors for these crashes. The forced merge at the lane drop is likely a contributing factor, particularly during high traffic periods. Eastbound Lane Drop at Washington St, (MP147.97 to MP148.31): In this segment, a total of 37 crashes occurred during 2012 and 2013, and 21 crash reports were reviewed. The lane drop and congestion downstream were noted in four rear end crashes. Last minute merging vehicles at the lane drop were noted in three sideswipe crashes. Thirteen of the 21 crashes occurred in the right two lanes. Contributing crash factors appear to include: congestion, speed differential, and the lane drop. Westbound, Sky Harbor Blvd to Washington St, (MP148.43 to MP148.66): In this segment, a total of 73 crashes occurred during 2012 and 2013, and 30 crash reports were reviewed. Lane changing was noted as the primary cause in 10 crashes and slowed/stopped traffic was noted as a contributing factor in 20 crashes. Sideswipe and rear end crashes were predominant. Of the 30 crashes, 20 occurred in the right two lanes. Contributing crash factors appear to include: speed differential, lane changing, and congestion. Westbound, 16th St Entrance to 7th St Exit, (MP145.82 to MP146.64): In this weaving segment, a total of 261 crashes occurred during 2012 and 2013, and 40 crash reports were reviewed. There is an auxiliary lane from the 16th Street on-ramp that terminates at the 7th Street exit. Two lanes must exit at 7th Street. Higher speed traffic encountering slowed traffic was noted in 26 crashes. In 12 crashes, lane changing due to congestion resulted in rear-end and sideswipe collisions. Lane changing, congestion, and speed differential were contributing factors for the crashes. Kittelson & Associates, Inc 48 I-10 Phoenix Corridor Safety Study Final Report Eastbound, 16th St to L202/SR-51, (MP146.44to MP146.93): In this weaving section, a total of 172 crashes occurred during 2012 and 2013, and 42 crash reports were reviewed. This segment contains a left-side HOV exit and right-side advance exit lanes for Loop 202 (two lanes) and one optional exit lane for SR-51. In 33 crashes reviewed, higher speed vehicles encounter slowing traffic and stop-and-go traffic which resulted in a rear-end crash. Weaving and lane changing was noted as a factor in 7 crashes. Congestion, speed differential, weaving, and lane changing were contributing factors for these crashes. Eastbound 35th Ave Entrance to 27th Ave Exit, (MP141.92 to MP142.42): In this Type B weaving segment, a total of 189 crashes occurred during 2012 and 2013, and 28 crash reports were reviewed. Ten rear-end collisions noted stop-and-go conditions as a primary factor. Ten sideswipe crashes were attributed to lane changing in the weaving section. Contributing crash factors appear to include: congestion, speed differential, and lane changing. Westbound 27th Ave Entrance to 35th Ave Exit (MP142.42 to MP141.92): In this Type A weaving segment with auxiliary lane, a total of 126 crashes occurred during 2012 and 2013, and 49 crash reports were reviewed. There is an additional lane drop just west of the WB 35th Avenue exit. Traffic entering I10 from I-17 were noted in 8 rear-end and sideswipe crashes due to lane changing and congestion. Higher speed traffic encountering slowed traffic was noted in 39 crashes reviewed. Congestion, speed differential, lane changing, and merging traffic were contributing factors for these crashes. Phase III Analysis – Fatal and Serious Injury Crashes All reported fatal and serious injury crashes from 2012 to 2013 were reviewed as a part of this analysis. A detailed review of each fatal and serious injury crash is shown in Appendix F. Table 24 and Table 25 summarize the fatal and serious injury crashes. The crash reports provided for the analysis contain a very limited amount of information (no supplemental reports were provided for review); therefore, the possible contributing factors presented were based on information available coupled with engineering judgment and knowledge of the crash location. Phase III Findings The following section provides the diagnostic summary and a discussion on the contributing factors for fatal and serious injury crash reports reviewed. Fatal crashes: No specific patterns or roadway geometry features were identified from the analysis of fatal crashes. Two out of six fatal crashes involved pedestrians, however due to a lack of information provided (no supplemental reports were provided) it was not possible to determine why the pedestrians were walking along or across the freeway. One fatal crash was related to a vehicle malfunction (wheel), two other fatal crashes involved single vehicles and one was a rear-end crash. Not including the two pedestrian fatalities, excessive speed, and potentially other driver related factors (e.g. older driver) were the primary contributing factors. Congestion and high speed contributed to one fatality. Serious Injury (Type A) crashes: A total of 40 serious injury crashes occurred within the entire study section. Approximately half were related to congestion in combination with excessive speed. Twenty Kittelson & Associates, Inc 49 I-10 Phoenix Corridor Safety Study Final Report were rear end crashes, 9 were single vehicle crashes, 3 were sideswipe crashes, and 8 were angle/other/or unknown crashes. The remaining serious injury crashes occurred under uncongested conditions. The study segment with the highest number of serious injury crashes was westbound I-10 MP143.94 to MP 144.36 (between the 7th Ave on-ramp and the 19th Ave off-ramp). Some 18% of all serious injury crashes involved motorcyclists, with speed being a contributing factor. Two of the nine reported motorcycle serious injuries crashes occurred along the sweeping 90-degree curve. Approximately 20% of all serious injury crashes were single-vehicle crashes. Table 24. Fatal crash information Direction and Segment Summary Contributing Factors Eastbound MP141.67 to MP141.92 Crash #2660116: Tractor-Trailer wheel failure Eastbound MP142.43 to M 142.71 Crash #2735405: Unknown. Single vehicle Westbound MP144.42 to MP144.80 Crash #2595077: Pedestrian walking against traffic Westbound MP145.94 to MP146.44 Crash # 2676135: Pedestrian crossing road. Alcohol involved Westbound MP146.93 to MP147.52 Crash #2689544: Single vehicle, negotiating curve at 70 mph during rain, wet surface condition Vehicle: Yes Driver : No Roadway: No Environmental: No Vehicle: No Driver: Unknown Roadway: No Environmental: No Vehicle: No Pedestrian: Yes Roadway: No Environmental: No Vehicle: No Pedestrian: Yes Roadway: No Environmental: No Vehicle: No Driver: Yes Roadway: Yes Environmental: Yes Westbound MP148.07 to MP148.31 Crash #2599442: Driver DOB: 1927. Rear-end crash. Speed too fast for conditions. Congestion present Kittelson & Associates, Inc Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes 50 I-10 Phoenix Corridor Safety Study Final Report Table 25. Incapacitating injury (Type A) crash information Direction and Segment Summary Contributing Factors Eastbound MP141.94 to MP142.42 Total crashes occurred: 3 (2585023, 2622201, 2752470), Crash reports reviewed: 3. Rear-end crash, speed too fast for conditions, congestion. Rear-end crash, speed too fast for conditions, congestion. Object avoidance crash, speed too fast for conditions. Vehicle: No Driver: Yes Roadway: No Environmental: NO Congestion: Yes Eastbound MP142.42 to MP142.43 Total crashes occurred: 1 (2744192), Crash reports reviewed: 1. Rear-end, speed too fast for conditions, congestion. Westbound MP142.42 to MP142.43 Total crashes occurred: 1 (2596917), Crash reports reviewed: 1. Single vehicle, motorcycle, speed too fast for conditions. Eastbound MP142.70 to MP142.71 Total crashes occurred: 2 (2595374, 2595773), Crash reports reviewed: 2. Rear-end, lane change, motorcyclist following too close, congestion, speed too fast for conditions. Rear-end, congestion, speed too fast for conditions. Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Eastbound MP142.87 to MP142.99 Total crashes occurred: 1 (2579668), Crash reports reviewed: 1. Crash information unknown Westbound MP142.99 to MP143.67 Total crashes occurred: 1 (2715241), Crash reports reviewed: 1. Sideswipe. V1 entered HOV lane from lane 1. Speed too fast for conditions. Eastbound MP143.67 to MP143.94 Total crash occurred: 1 (2585064), Crash reports reviewed - 1. Rear-end, speed too fast for conditions, congestion. Westbound MP143.67 to MP143.94 Total crashes occurred: 1 (2596216), Crash reports reviewed - 1. Rear-end, speed too fast for conditions, congestion. Westbound MP143.94 to MP144.36 Total crashes occurred: 6 (2583826, 2671728, 2689512, 2719871, 2757122, 2778115), Crash reports reviewed: 6. Rear-end, speed too fast for conditions, congestion. Single vehicle, alcohol. Single vehicle, fatigue/fell asleep. Motorcycle, speed too fast for conditions, congestion. Rear-end, driver distracted by police vehicle, congestion. Single vehicle (speed 85 mph). Kittelson & Associates, Inc Vehicle: No Driver: Unknown Roadway: No Environmental: Unknown Vehicle: No Driver: Yes Roadway: No Environmental: No Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes 51 I-10 Phoenix Corridor Safety Study Final Report Direction and Segment Summary Contributing Factors Eastbound MP144.42 to MP144.80 Total crashes occurred: 3 (2593799, 2648100, 2696715), Crash reports reviewed: 3. Sideswipe, unsafe passing. Angle, lane change from lane 1 into HOV, speed too fast for conditions, congestion. Single vehicle, motorcycle, speed too fast for conditions Total crashes occurred: 1 (2665932), Crash reports reviewed: 1. Single Vehicle Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Westbound MP144.42 to MP144.80 Eastbound MP144.80 to MP144.92 Westbound MP144.80 to MP144.92 Total crashes occurred: 2 (2752349, 2762880), Crash reports reviewed: 2. Rear-end, congestion, speed too fast for conditions. Driver tried to avoid object, speed too fast for conditions. Total crashes occurred: 1 (2651125), Crash reports reviewed: 1. Rear-end, congestion, speed too fast for conditions. Eastbound MP144.96 to MP145.42 Total crashes occurred: 2 (2624458, 2788793), Crash reports reviewed: 2. Rear-end, speed too fast for conditions, congestion. Rear-end, speed too fast for conditions, failed to remain in proper lane. Westbound MP144.96 to MP145.42 Total crashes occurred: 2 (2585385, 2738892), Crash reports reviewed: 2. Rear-end, lane change, speed too fast for conditions. Rear-end, speed too fast for conditions. Eastbound MP145.52 to MP145.78 Total crashes occurred: 1 (2768280), Crash reports reviewed: 1. Angle, avoiding object, speed too fast for conditions. Westbound MP145.84 to MP145.94 Total crashes occurred: 1 (2661954), Crash reports reviewed - 1. Rear-end, motorcycle, speed too fast for condition, congestion. Eastbound MP146.44 to MP146.84 Total crashes occurred: 1 (2662051), Crash reports reviewed: 1. Rear-end, driver applied brake as well as gas, congestion. Kittelson & Associates, Inc Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes 52 I-10 Phoenix Corridor Safety Study Final Report Direction and Segment Summary Contributing Factors Eastbound MP146.93 to MP147.52 Total crashes occurred: 2 (2595789, 2642347), Crash reports reviewed: 2. Rear-end, congestion, speed too fast for conditions. Single vehicle, motorcycle, negotiating curve, speed too fast for conditions. Westbound MP146.93 to MP147.52 Total crashes occurred: 1 (2638496), Crash reports reviewed: 1. Single vehicle, motorcycle, speed too fast for conditions (75 mph), negotiating curve. Vehicle: No Driver: Yes Roadway: Yes Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: Yes Environmental: No Westbound MP148.05 to MP148.31 Total crashes occurred: 1 (2638496), Crash reports reviewed: 1. Rear-end, congestion. Eastbound MP148.42 to MP148.64 Total crashes occurred: 1 (2665404), Crash reports reviewed: 1. Sideswipe, rear-end, motorcycle, speed too fast for conditions, congestion. Eastbound MP148.64 to MP149.14 Total crashes occurred: 1 (2665404), Crash reports reviewed: 1. Rear-end, speed too fast for conditions, congestion. Westbound MP148.64 to MP149.14 Total crashes occurred: 2 (2706271, 2719776), Crash reports reviewed: 2. Rear-end, congestion, speed too fast for conditions, right lane ends, congestion. Single vehicle, motorcycle, speed too fast for conditions. Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: No Environmental: No Congestion: Yes Vehicle: No Driver: Yes Roadway: Yes Environmental: No Congestion: Yes POTENTIAL COUNTERMEASURES Based on the detailed review and analysis of observed crashes within the study section, potential countermeasures were identified. Considering that speed and congestion were contributing factors in approximately 57% of all crashes occurring within the study section in 2012 and 2013, half of the serious injury crashes, and were noted on the majority of crash reports reviewed, countermeasures intended to address these factors were primarily considered. Countermeasures to address crash issues at specific hotspots, including lane drops, weaving areas, and HOV access were also identified and evaluated. Detailed evaluations of potential countermeasures to estimate their potential effect on reducing crashes, and their benefit/cost were conducted. The countermeasure evaluation results are discussed below. Kittelson & Associates, Inc 53 I-10 Phoenix Corridor Safety Study Final Report CRASH REDUCTION BENEFIT The benefit, or reduced crash costs associated with each countermeasure was estimated for the expected crashes over a 20 year period. The expected crashes occurring in each study section segment were calculated using the calibrated freeway crash prediction models, based on 20-year traffic volume projections provided from the MAG regional traffic forecasting model. Crashes by severity type (K, A, B, C, PDO) were calculated using the crash severity distribution factor calibrated for Phoenix freeways. However, in reviewing the expected number of fatal (K) crashes relative to the crashes observed in 2012 and 2013, there was concern that the expected number of fatal crashes was potentially high when considering the effect of a fatal crash in the benefit/cost analysis. In the observed 2012-2013 crashes, six fatal crashes occurred (Table 24), two of which were pedestrians and one was a vehicle mechanical failure in which a wheel broke free from a truck and caused the fatal crash. Since only three of the observed fatal crashes could realistically be affected by the range of countermeasures considered, it was determined that the proportion of fatal crashes in the expected crashes should be reduced. The adjusted crash severity distribution applied for the cost benefit analysis is provided in Table 26. The number of expected fatal crashes per year was reduced from 11.5 to 4, while the number of incapacitating injury crashes (A) and non-incapacitating injury crashes (B) were increased so that the total number of crashes remained unchanged. Crash benefit costs were calculated using the cost factors provided in the ADOT HSIP manual and in Table 27. Note the these cost factors only account for costs associated with the crash itself and do not reflect the additional cost benefit that may be incurred due to reduction in secondary crashes and in congestion caused by crashes. Table 26. Crash Severity Distribution Used for Benefit/Cost Analysis Number of Crashes Proportion Number of Crashes -Adjusted Proportion - Adjusted Expected Crashes Per Year based on Calibrated Crash Severity Distribution K A B C PDO Total 11.5 27.6 140.8 257.3 1231.3 1718.4 0.0067 0.0161 0.0819 0.1497 0.7165 1.0000 4.0 30.1 145.0 257.3 1231.3 1718.4 0.0023 0.0175 0.0844 0.1497 0.7165 1.0000 Table 27. ADOT Crash Cost Factors Crash Severity K, Fatal A, Incapacitating Injury B, Non-incapacitating injury C, Possible injury PDO Kittelson & Associates, Inc Comprehensive Cost (2) $5,800,000 $400,000 $80,000 $42,000 $4,000 54 I-10 Phoenix Corridor Safety Study Final Report LOWER SPEED LIMIT Description Lower the speed limit on I-10 from 65 to 55 mph. The reduced speed limit zone could include just the study section, 35th Ave to Sky Harbor Blvd, or might extend further east and west. Continual speed enforcement emphasis will be an integral part of this countermeasure to produce the desired reduction in operating speeds. This section of I-10 could be designated as a “Safety Corridor”, with enhanced focus by ADOT and DPS. Potential Crash Reduction The Highway Safety Manual (Chapter 3, Table 3E-2) provides CMFs for changes in average operating speed. Crash reduction varies depending upon the initial average operating speed and the speed reduction achieved, ranging from -1 to -5 mph. Separate CMFs are provided for fatal crashes and injury crashes. No CMFs are provided for property damage only crashes. Since the effect of reducing the speed limit from 65 to 55 mph on lowering operating speed can vary, the crash reduction analysis performed for this study considered a range, from -2 to -4 mph. The speed data collected for this study indicates the average operating speed on I-10 under uncongested conditions is essentially at the 65 mph posted speed limit. The 85th percentile speed is closer to 70 mph. Therefore, based on a 65 mph average operating speed, the expected reductions in fatal and injury crashes using the CMFs provided in the HSM are listed in Table 28. Table 28. Crash Modification Factors for Reduction in Operating Speed 1,2 Fatal Crashes Injury Crashes Reduction in Average Operating Speed CMF CRF CMF -4 mph 0.70 -0.30 0.81 CRF -0.19 -3 mph 0.77 -0.23 0.86 -0.14 -2 mph 0.84 -0.16 0.90 -0.10 Source: Highway Safety Manuel, Chapter 3, Table 3E-2. CMFs for 65 mph interpolated from 60 mph and 70 mph factors. Applying the crash reduction factors from Table 28 to the expected fatal and injury crashes over a 20 year period on the entire I-10 study section produces estimated crash reductions listed in Table 29. Table 29 . Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from Lowered Speed Limit Effect of Lowered Serious Injury Crashes Fatal Crashes (K) Speed Limit on (A) Average Total Reduced Total Reduced Operating Speed Non-serious Injury Crashes (B) Possible Injury Crashes (C) PDO Crashes Total Reduced Total Reduced Total Reduced -4 mph 88.4 26.5 667.3 126.8 4340.8 824.7 5702.6 1083.5 28087.4 - -3 mph 88.4 20.3 667.3 93.4 4340.8 607.7 5702.6 798.4 28087.4 - -2 mph 88.4 14.1 667.3 66.7 4340.8 434.1 5702.6 570.3 28087.4 - Kittelson & Associates, Inc 55 I-10 Phoenix Corridor Safety Study Final Report Potential Crash Reduction Benefit/Cost The estimated benefit value of the reduction in K, A,and B crashes is presented in Table 30. Note that higher benefit is expected if possible injury and PDO crashes are also considered. Estimating the cost of achieving a lower average operating speed is difficult as the greatest proportion of the cost will be in added enforcement. The cost of changing the speed limit signing is negligible, however the cost of additional DPS officers and equipment can be substantial. Assuming enforcement costs of $1 million/yr, Table 31 presents the B/C ratios. Two scenarios are included in the analysis, the first considering crashes occurring during both congested and non-congested periods, and the second assuming that lowered operating speeds would only be achieved during non-congested conditions. Table 30. Estimated 20-yr Benefit Value of Reduced K, A, & B Crashes Resulting from Lowered Speed Limit Effect of Lowered Speed Limit on Average Operating Speed Total Fatal Crashes (K) Serious Injury Crashes (A) Non-serious Injury Crashes (B) K+A+B K+A -4 mph $153,786,714 $50,712,994 $65,979,942 $270,479,650 $204,499,708 -3 mph $117,903,148 $37,367,470 $48,616,800 $203,887,418 $155,270,618 -2 mph $82,019,580 $26,691,050 $34,726,286 $143,436,916 $108,710,630 Crash cost factors from ADOT Highway Safety Improvement Program Manual, March 2010 Based on the distribution of observed crashes by time-of-day, it is estimated that 30% of expected crashes within the study section occur during non-congested conditions. The results show that even when considering only non-congested periods, reducing the operating speed will be a cost effective crash countermeasure. Table 31. 20-yr B/C Analysis of Lowered Speed Limit Reduced Operating Speed Annual Const. Change in Total Cost Total Annual Annual Const. Cost (CRF x Total Const Cost O & M Cost Cost) (1) All Expected Crashes Annual Benefit B/C K+A+B K+A K+A+B K+A -4 mph $50,000 $5,095 $1,000,000 $1,005,095 $13,523,983 $10,224,985 13.46 10.17 -3 mph $50,000 $5,095 $1,000,000 $1,005,095 $10,194,371 $7,763,531 10.14 7.72 -2 mph $50,000 $5,095 $1,000,000 $1,005,095 $7,171,846 $5,435,532 7.14 5.41 Expected Crashes Occurring During Non-congested Periods Only -4 mph $50,000 $5,095 $1,000,000 $1,005,095 $4,057,195 $3,067,496 4.04 3.05 -3 mph $50,000 $5,095 $1,000,000 $1,005,095 $3,058,311 $2,329,059 3.04 2.32 -2 mph $50,000 $5,095 $1,000,000 $1,005,095 $2,151,554 $1,630,659 2.14 1.62 CRF =0.1019; 20 yrs at 8% interest; ADOT HSIP Manual, 2010 Kittelson & Associates, Inc 56 I-10 Phoenix Corridor Safety Study Final Report VARIABLE SPEED LIMIT SYSTEM Description Variable speed limit (VSL) systems have been widely used around the world to improve safety and operational efficiency on congested highways. The principle behind VSL systems is to post a speed limit that is appropriate for current conditions considering time dependent freeway traffic demand, speed profile and/or special conditions like adverse weather and incidents. This provides an opportunity to warn drivers of downstream conditions, reduce speed, decrease headways and encourage more uniform flow. VSL has the capability of increasing safety by reducing both primary and secondary crashes. It can also reduce congestion, travel time, and emissions. VSL systems have been one of the most heavily researched Active Traffic Management (ATM) techniques, and deployments have occurred in the U.S. and internationally. Recently several state DOTs in the United States including Florida DOT, Caltrans, Figure 17. VSL System in Seattle, WA Washington DOT, Minnesota DOT, Virginia DOT, Missouri DOT, Utah DOT, and Colorado DOT have implemented VSL systems. A literature review of VSL systems around the world was conducted and a summary is provided in Appendix G. Relative to the I-10 Study Section, the application of VSL would primarily be intended to reduce operating speeds during peak traffic periods when high traffic volumes produce congestion and when the majority (70%) of crashes occur. In addition to reducing primary and secondary crashes, an ancillary effect would be improved traffic flow with stop-and-go conditions occurring less frequently. Similar to reducing the posted speed limit, the success of a VSL system will also require emphasis on speed enforcement. However, heightened enforcement would only be required during peak traffic periods. As such, the potential application of automated enforcement or some hybrid of automated detection with manual enforcement during these periods within a limited section of I-10 might be an effective alternative. Potential Crash Reduction A recent (2008) application and evaluation of a VSL system installed on an interstate freeway in Missouri (Ref. 7) produced a high quality (four star) CMF included in the FHWA CMF Clearinghouse. The evaluation found that while the VSL is not providing the desired improvement in overall mobility along the freeway corridor, noticeable benefits have been seen with respect to reduction in the number of crashes. The evaluation determined that all crashes decreased 8% with the VSL system in operation. Kittelson & Associates, Inc 57 I-10 Phoenix Corridor Safety Study Final Report Evaluations of other systems installed internationally and in the U.S. have provided additional information relative to the potential benefit of a VSL system. The following benefits have been reported in different published research documents:     Increase in average congested period throughput – 3% to 7% Increase in overall capacity – 3% to 22% Decrease in primary incidents – 3% to 30% Decrease in secondary incidents – 40% to 50% Applying the CMF from the CMF Clearinghouse to the expected crashes (all crash types) over a 20 year period on the entire I-10 study section produces estimated crash reductions listed in Table 32. The CMF was applied to all crashes and to those crashes occurring during peak traffic periods (6:00 – 10:00 a.m. and 3:00 to 7:00 pm) when approximately 70% of observed crashes occurred. Potential Crash Reduction Benefit/Cost The estimated benefit value of the reduction in all crashes and in fatal and serious injury crashes only is presented in Table 33. Table 32. Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from a VSL System Serious Injury Crashes (A) Fatal Crashes (K) Non-serious Injury Crashes (B) Possible Injury Crashes (C) PDO Crashes Total Reduced Total Reduced Total Reduced Total Reduced Total All Crashes 88.4 7.1 667.3 53.4 4340.8 347.3 5702.6 456.2 28087.4 Reduced 2247.0 Peak Traffic Period Crashes 61.9 4.9 467.1 37.4 3038.5 243.1 3991.8 319.3 19661.2 1572.9 Table 33. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from a VSL System Fatal Crashes Serious Injury (K) Crashes (A) Non-serious Possible Injury Injury Crashes PDO Crashes Crashes (C) (B) Total All Crash Severities K+A All Crashes $41,009,790 $21,352,840 $27,781,028 $19,160,756 $8,987,966 $118,292,380 $62,362,630 Peak Traffic Period Crashes $28,706,854 $14,946,988 $19,446,720 $13,412,530 $6,291,576 $82,804,668 $43,653,842 Crash cost factors from ADOT Highway Safety Improvement Program Manual, March 2010 A planning level cost estimate for implementing a VSL system within the I-10 study section was prepared. This estimate represents implementation of a VSL system from just east of the I-17 TI to the L202/SR51 TI. However, a broader application of a VSL system, potentially from 35th Ave to Sky Harbor Blvd can be considered. It may also be appropriate to extend a VSL system east on L202 and north on SR51 as well. Five-year (2009-2013) crash data reviewed for the I-10/I-17 Master Plan (Spine Study) Kittelson & Associates, Inc 58 I-10 Phoenix Corridor Safety Study Final Report indicate high crash rates on segments in the vicinity of system interchanges, which is observed in this I10 safety study. The detailed cost estimate, provided in Appendix H reflects a VSL that includes 12 VSL sign structures and necessary control and communications equipment to operate the system as part of the regional FMS. The estimated implementation cost of the system is roughly $8.2 million or $2.7 million/mile. The electrical cost for the VSL was estimated at $12,500/yr in the update to the tunnel lighting study (3). Assuming an annual maintenance costs of $20,000/yr and $500,000 of annual enforcement costs, the B/C analysis is summarized in Table 34. Table 34. 20-yr B/C Analysis of VSL System $8,200,000 Annual Const. Cost (CRF x Total Const Cost) (1) $835,580 $8,200,000 $835,580 Total Const. Cost All Crashes Peak Traffic Period Crashes Annual Benefit Change in Annual O&M Cost Total Annual Cost $532,500 $532,500 B/C All Crash Severities K+A All Crash Severities K+A $1,368,080 $4,507,183 $3,118,132 3.29 2.28 $1,368,080 $3,155,028 $2,182,692 2.31 1.60 CRF =0.1019; 20 yrs at 8% interest; ADOT HSIP Manual, 2010 AUTOMATED SPEED ENFORCEMENT Description Automated speed enforcement has been shown to be an effective safety countermeasure on roadways of all types. While public perception of the use of these systems is often negative for several reasons, including the increased number of citations that result, they are viewed as revenue generators, and since in the United States, they are typically installed and operated by private companies, their application on selected roadway segments, such as the I-10 Study section, where it is highly desirable to lower operating speed to reduce crashes could prove more acceptable. Automated enforcement could be installed in conjunction with a variable speed limit system to produce greater compliance during peak traffic periods when crash potential is highest. Potential Crash Reduction The Highway Safety Manual (Chapter 17, Table 17-5) provides a very high quality CMF for the potential crash effect of automated speed enforcement. The CMF is 0.83 with a standard error of 0.01. The CMF reflects the effect on fatal and injury crashes only. Applying the Highway Safety Manual CMF to the expected fatal and injury crashes over a 10 year period on the entire I-10 study section produces estimated crash reductions listed in Table 35. The CMF used was 0.83, or a 17% reduction in crashes. Kittelson & Associates, Inc 59 I-10 Phoenix Corridor Safety Study Final Report Table 35. Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from an Automated Speed Enforcement System Fatal Crashes (K) All Crashes Serious Injury Crashes (A) Non-serious Injury Crashes (B) Possible Injury Crashes (C) PDO Crashes Total Reduced Total Reduced Total Reduced Total Reduced Total Reduced 88.4 15.0 667.3 113.4 4340.8 737.9 5702.6 969.4 28087.4 - Potential Crash Reduction Benefit/Cost The estimated benefit value of the reduction in all injury crashes is presented in Table 36. As the cost of installing, operating, and maintaining an automated speed enforcement system within the study section is unknown, a B/C analysis was not prepared. Table 36. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from an Automated Speed Enforcement System All Crashes Total Fatal Crashes (K) Serious Injury Crashes (A) Non-serious Injury Crashes (B) Possible Injury Crashes (C) PDO Crashes All Injury Crashes K+A+B K+A $87,145,804 $45,374,784 $59,034,686 $40,716,608 - $191,555,274 $132,520,588 Crash cost factors from ADOT Highway Safety Improvement Program Manual, March 2010 UPGRADE TUNNEL LIGHTING Description Based on an evaluation of crash frequency within the Deck Park Tunnel relative to adjacent segments and information provided by the crash reports reviewed, several conditions within the tunnel segment appear to be contributing factors to the high crash frequency observed in the 2012 and 2013 crash data. The data suggests that two lighting issues are occurring. First, the difference between the day-time ambient illumination level outside of the tunnel and Figure 18. Westbound Tunnel in the Afternoon the illumination level at the tunnel portals, or threshold zones, is affecting drivers vision as their eyes adjust. Second, as noted in the Deck Park Tunnel Energy Efficiency Study (Ref. 9) the illumination levels within the tunnel do not meet current IES recommendations for tunnel lighting. At the tunnel threshold zones (entrances), the measured illumination levels are below the original design levels. The effects of age and environment on the Kittelson & Associates, Inc 60 I-10 Phoenix Corridor Safety Study Final Report luminaires reduce lumen output and resulting illumination levels achieved. To reduce the safety effects of these two issues, the tunnel lighting system will need to be upgraded. A range of options exist, however the primary improvement would be to upgrade the existing system to meet current IES recommended practice for tunnel lighting levels. The upgrade would replace the current HPS fixtures with fixtures that provide greater lumen output, either HPS or LED. The LED option offers greater efficiency (i.e. more lumens per watt) and the ability to dim the lighting in the tunnel transition areas depending on the ambient illumination levels. With LED, the lighting could also be efficiently dimmed based on changes in daytime ambient light levels and variable speed. The IES standard changes based on posted speed limit. Several traffic issues within the tunnel are also likely contributing to the high expected crash frequency. These include lane changing associated with the 7th St and 7th Ave entry and exit ramps in close proximity to the tunnel portals and unexpected queueing that typically occurs within the tunnel during peak periods. Reducing the lane changing activity during peak traffic periods could be achieved by metering the eastbound 7th Ave and westbound 7th St entry ramps. The potential benefit of metering is described as a separate countermeasure. Finally, the horizontal curve within the tunnel is likely also contributing to the crash frequency. A potential countermeasure for this geometric condition is to reduce vehicle operating speeds through the tunnel. The benefit of reduced operating speed was previously discussed. Potential Crash Reduction CMFs for tunnel lighting are not available and no studies were identified that might provide information to estimate the potential crash reduction associated with tunnel lighting. To quantify the potential effect of tunnel presence (with the existing illumination design) on safety performance, the expected crash frequency within the tunnel was compared with the frequency in the adjacent freeway segments. As these segments have essentially the same roadway characteristics (lane and shoulder width, HOV lanes, median and shoulder barrier, roadway curvature, grade) and traffic conditions, a significant difference in crash frequency could be attributed to the five issues previously identified. However, rather than simply compare the expected crash frequencies, the“Excess”expected crash frequencies within these segments were compared. When the expected crash frequency exceeds the predicted crash frequency, the difference is called excess expected crashes. These excess crashes reflect the effect of unknown factors or a combination of factors that are not included in the predictive methodology. The tunnel and adjacent segments all exhibit high excess expected crash values. A comparison of the “Excess” statistics for the segments before and after the tunnel exhibit shows a similar large number of excess crashes. The high excess values extending from MP 144.42 to 145.82 are provided in Table 37. Note that the lower crash rate in the segment beginning at MP 144.92 is likely a crash reporting error, with crashes incorrectly located in the adjacent segment beginning at MP 144.80. Kittelson & Associates, Inc 61 I-10 Phoenix Corridor Safety Study Final Report Table 37. Excess Expected Average Crashes in Tunnel and Adjacent Segments Segment Begin MP Excess Expected Average Crashes per mile per year 144.42 144.80 144.92 144.96 (tunnel) 145.42 145.52 145.78 219.9 368.4 36.8 269.2 161.6 215.5 136.9 The average excess for the segments adjacent to the tunnel is 220 total crashes/mile (excluding the tunnel segment). The tunnel has an excess of 269 total crashes/mile. Thus, the tunnel segment has about 22% (= 100 [269/220 -1], more excess crashes than the average segment in this area. This data infers that the tunnel with the existing illumination design may be contributing to a 22% increase in crashes, relative to the neighboring segments. This amounts to nearly 49 crashes per year or 38 crashes if only daytime crashes are considered. Table 38 lists the crash reduction that could potentially occur over 20 years with upgraded tunnel lighting. Table 38. Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from Upgraded Tunnel Lighting Fatal Crashes (K) All Excess Tunnel Crashes (49/yr) Daylight Excess Tunnel Crashes (38/yr) Serious Injury Crashes (A) Non-serious Injury Crashes (B) Possible Injury Crashes (C) PDO Crashes Total Reduced Total Reduced Total Reduced Total Reduced Total Reduced 21.7 1.9 52.2 14.2 374.4 96.0 676.8 169.4 3400.6 850.9 21.7 1.4 52.2 10.8 374.4 73.0 676.8 128.7 3400.6 646.7 Potential Crash Reduction Benefit/Cost The estimated benefit value of the reduction in all crashes and in fatal and serious injury crashes only is presented in Table 39. Note that these values assume that all of the excess expected crashes will be affected by improved tunnel lighting. Practically, it is more likely that some portion of the excess expected crashes would be eliminated. The Deck Park Tunnel Energy Efficiency Study offered several lighting upgrade options, however the LED option is preferred. Several LED lighting design alternatives can be considered, including designing for a constant speed limit (55, 60, 65 mph) or for a variable speed limit. Since it is unknown if the crash reduction effects of the upgrade tunnel lighting and the VSL system can be combined, only the tunnel lighting upgrade to was included in the B/C analysis. The lighting upgrade was assumed to provide illumination levels required for a 65 mph posted speed limit and will have dimming capabilities. Kittelson & Associates, Inc 62 I-10 Phoenix Corridor Safety Study Final Report Table 39. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from Upgraded Tunnel Lighting All Excess Tunnel Crashes (49/yr) Daylight Excess Tunnel Crashes (38/yr) Fatal Crashes (K) Serious Non-serious Possible Injury Injury Injury Crashes PDO Crashes Crashes (C) Crashes (A) (B) $10,957,418 $5,699,100 $7,682,524 $7,112,876 $8,327,638 $4,331,316 $5,838,718 $5,405,786 Total All Crash Severities K+A $3,403,571 $34,855,489 $16,656,518 $2,586,714 $26,490,172 $12,658,954 Crash cost factors from ADOT Highway Safety Improvement Program Manual, March 2010 The tunnel energy efficiency study estimated the construction cost for an LED upgrade at $20,055,000. Annual maintenance costs are estimated at $43,000 and electrical costs are estimated to be $327,000, approximately $8,000 lower than current electrical costs with the HPS system. Table 40 presents the benefit/cost ratios for the tunnel lighting upgrade. Note that the annual electrical savings is included as a benefit with the value of the crash reduction benefit. Table 40. 20-yr B/C Analysis of Upgraded Tunnel Lighting Crash Redcuction All Excess Tunnel Crashes (49/yr) Daylight Excess Tunnel Crashes (38/yr) 1. Total Const. Cost Annual Const. Cost (CRF x Total Const Cost) (1) Change in Annual O&M Cost Total Annual Cost Annual Benefit $20,055,840 $2,043,690 $35,298 $20,055,840 $2,043,690 $35,298 B/C All Crashes K+A All Crashes K+A $2,078,988 $1,742,774 $832,826 0.84 0.40 $2,078,988 $1,324,509 $632,948 0.64 0.30 CRF =0.1019; 20 yrs at 8% interest; ADOT HSIP Manual, 2010 ELIMINATE LANE DROPS Description There are three lane drops within the study section, EB at 7th St, WB at 7th Ave, and EB at Washington St. Review of the crash data and sample reports for each segment indicated that the lane drops at 7th St and 7th Ave contribute to the observed crash frequency, particularly during peak traffic periods. In the 7th St lane drop segment, 146 EB crashes occurred in 2012 and 2013, while 188 WB crashes occurred in the 7th Ave lane drop segment. In the Washington St lane drop segment, 37 crashes occurred over Kittelson & Associates, Inc Figure 19. Westbound Lane Drop West of 7th Ave Exit; Note queuing in right two lanes 63 I-10 Phoenix Corridor Safety Study Final Report this 2-year period. In addition to contributing to crashes due to added lane changing, the lane drops also contribute to congestion which is a primary factor in crashes occurring within the study section. The estimated expected crash frequencies in both the 7th St and 7th Ave lane drop segments are well above the value for a typical urban freeway segment in Phoenix and corresponding high Z-scores indicate that there is substantial potential to reduce crashes in these segments. The potential for crash reduction in the Washington St lane drop section is much lower. Eliminating the 7th St and 7th Ave lane drops can be achieved by either extending the lane to the downstream entry ramp or dropping the lane at the upstream exit ramp. In both cases, extending the lane will change a lane add configuration into a merge configuration with a speed-change lane at the downstream entry ramp. An important advantage of extending the lane over dropping it at the upstream exit is that this improvement would provide additional through lane capacity in addition to eliminating a forced merge. The potential capacity impacts associated with changing the downstream lane add to a merge will need to be evaluated for each location. Potential Crash Reduction A CMF for eliminating a lane drop is not available. The potential crash reduction of eliminating the 7th St and 7th Ave lane drops was estimated using the crash prediction methodology and the ISATe analysis tool developed for the study. Extending each lane will affect three conditions that influence crash potential and which can be analyzed using the predictive methodology: 1) eliminate the forced merge, 2) change the downstream lane add to a speed-change lane, and 3) add a through lane to the downstream segment. Forcing the lane off at the upstream exit can also be evaluated using the predictive method. While both options can be evaluated using the predictive method the potential effect that either option will have on reducing congestion related crashes is not readily captured Table 41 summarizes the estimated potential effect on crashes of each lane drop option. Note that at 7 th Ave, dropping the lane at the upstream exit will potentially result in higher expected crashes (a negative reduction in crashes). However, at 7th St, the lane drop option produces a higher reduction in crashes than extending the lane. Potential Crash Reduction Benefit/Cost Table 42 provides the estimated benefit value of eliminating the 7th St and 7th Ave lane drops. For 7th St, both options are provided. The benefit/cost analysis is summarized in Table 43. Detailed cost estimates to extend the lanes to the downstream on-ramp are included in Appendix H. These estimates assume that the roadway will need to be widened to extend the lanes and convert the lane add configuration at the downstream ramp to a speed-change lane. Kittelson & Associates, Inc 64 I-10 Phoenix Corridor Safety Study Final Report Table 41. Estimated Reduced Crashes over 20 years on I-10 Study Section Resulting from Eliminating Lane Drops Serious Injury Crashes (A) Fatal Crashes (K) Total Reduced Total Reduced Non-serious Injury Crashes (B) Total Reduced Possible Injury Crashes (C) PDO Crashes Total Reduced Total Reduced Extend Lane to Downstream Entry Ramp th 6.9 0.2 48.2 1.5 305.4 9.5 298.0 7.6 1871.3 46.5 th 3.5 0.1 25.8 0.8 166.9 4.9 257.9 7.6 1390.5 29.8 WB Lane Drop at 7 Ave EB Lane Drop at 7 St Drop Lane at Upstream Exit Ramp WB Lane Drop at 7th Ave 6.9 0.0 48.2 -0.6 305.4 -4.3 298.0 -1.8 1871.3 -7.2 EB Lane Drop at 7th St 3.5 0.2 25.8 1.2 166.9 8.1 257.9 14.8 1390.5 28.0 Restriping of the mainline to utilize existing inside and outside shoulders to add the lane would be a substantially lower cost and would produce higher B/C ratios, however it is not preferred since it would significantly reduce the shoulder width and impact the left-side HOV ramp merge gore areas. Reducing the shoulder width will increase crash potential since there is a continuous median barrier and would eliminate shoulder width used by disabled vehicles and enforcement. The estimated construction cost for dropping the 7th St lane at the upstream exit is roughly assumed to be $500,000, although it could be lower. While the benefit/cost results based on estimated crash reduction do not necessarily support this countermeasure, it is again important to note that additional crash reduction will likely be achieved due to reduced congestion. Table 42. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from Eliminating Lane Drops Fatal Crashes (K) Serious NonPossible Injury serious Injury Crashes Injury Crashes (C) (A) Crashes (B) Extend Lane to Downstream On-Ramp Total PDO Crashes All Crash Severities K+A WB Lane Drop at 7th Ave $1,227,888 $596,748 $759,458 $320,220 $185,800 $3,090,114 $1,824,636 EB Lane Drop at 7th St $576,606 $302,318 $390,286 $320,220 $119,284 $1,708,714 $878,924 $111,974 $2,751,532 $1,370,282 Drop Lane at Upstream Exit Ramp EB Lane Drop at 7th St $899,654 $470,628 $648,014 $621,262 Crash cost factors from ADOT Highway Safety Improvement Program Manual, March 2010 Kittelson & Associates, Inc 65 I-10 Phoenix Corridor Safety Study Final Report Table 43. 20-yr B/C Analysis of Eliminating Lane Drops Annual Benefit Annual Const. Cost Change in Total Const. Total Annual (CRF x Total Const Annual All Crash Cost Cost K+A Cost) (1) O & M Cost Severities Extend Lane to Downstream On-Ramp WB Lane Drop at 7th Ave EB Lane Drop at 7th St B/C All Crash Severities K+A $1,500,000 $152,850 $1,000 $153,850 $158,067 $91,232 1.03 0.59 $1,400,000 $142,660 $1,000 $143,660 $85,436 $43,946 0.59 0.31 $137,577 $68,514 2.65 1.32 Drop Lane at Upstream Exit Ramp EB Lane Drop at 7th St 1. $500,000 $50,950 $1,000 $51,950 CRF =0.1019; 20 yrs at 8% interest; ADOT HSIP Manual, 2010 Meter EB 7th Ave and WB 7th Street Entry Ramps Description The EB 7th Ave and WB 7th St entry ramps are unmetered. Although auxiliary lanes are in place, heavy traffic volumes entering the tunnel from these ramps creates substantial lane changing within the weaving section in the Deck Park Tunnel which has a fairly high observed and expected crash frequency relative to typical freeway segments in the urban Phoenix area. Based on 2012 and 2013 crash data, 116 crashes/yr occurred in the eastbound direction and 123 crashes/yr in the westbound direction within the tunnel. Metering of the heavy on-ramp traffic during peak periods could reduce crashes within the tunnel. Potential Crash Reduction A CMF of moderate quality is available from the FHWA Clearinghouse. The CMF is based on a study conducted in 2013 which looked at the effect of installing ramp metering on 19 ramps in Northern California (Ref.8). The study found that freeway collisions in the vicinity of the ramp speed-change lane were 36% lower with ramp metering in place. The study noted that PDO crashes were primarily affected. The results of the safety performance evaluation conducted for this study using the calibrated crash prediction model and diagnostic analysis indicates that lane changing is a primary factor contributing to the observed and expected crash frequency within the tunnel. The calculated CMF that reflects the effect of lane changing on crash frequency within the tunnel is nearly 1.7 for fatal and injury crashes and 1.5 for PDO crashes, indicating a substantial effect. While we are unable to estimate the specific effect that metering of these two ramps will have on crash reduction, it is reasonable to consider modest reductions in crashes (i.e. 5 to 15%) in order to assess the potential benefit. Based on current observations 56% of crashes within the tunnel occur during the morning (7:00 to 9:00 am)and evening (3:00 – 6:00 pm) peak periods during which metering would operate. The estimated reduction in expected peak period crashes over 20 yrs with metering of these two ramps is presented in Table 44 for several scenarios. Kittelson & Associates, Inc 66 I-10 Phoenix Corridor Safety Study Final Report Table 44. Estimated Reduced Crashes over 20 years Resulting from Metering the EB 7th Ave and WB 7th St Onramps Peak Period Crash Reduction (1) Total Reduced 1. Serious Injury Crashes (A) Fatal Crashes (K) Non-serious Injury Crashes (B) Possible Injury Crashes (C) PDO Crashes Total Reduced Total Reduced Total Reduced Total Reduced 5% 4.2 0.2 31.9 1.6 214.9 10.7 379.0 19.0 1904.3 95.2 10% 4.2 0.4 31.9 3.2 214.9 21.5 379.0 37.9 1904.3 190.4 15% 4.2 0.6 31.9 4.8 214.9 32.2 379.0 56.9 1904.3 285.6 Based on 56% of crashes occurring during peak periods when metering typically operates. The estimated benefit value of the reduction in all crashes and in fatal and serious injury crashes only is presented in Table 45. Table 46 presents the benefit/cost ratios for metering these ramps. The construction cost for adding meters is estimated at $250,000 per ramp, which does not include improvements that might be required at the cross street ramp terminals to add storage for additional queueing. Table 45. Estimated 20-yr Benefit Value of Reduced Crashes Resulting from Metering the EB 7th Ave and WB 7th St On-ramps Peak Period Non-serious Fatal Crashes Serious Injury Possible Injury Crash Reduction Injury Crashes (K) Crashes (A) Crashes (C) (1) (B) Total PDO Crashes All Crash Severities K+A 5% $1,226,158 $637,740 $859,690 $795,946 $380,866 $3,900,400 $1,863,898 10% $2,452,316 $1,275,482 $1,719,380 $1,591,892 $761,734 $7,800,804 $3,727,798 15% $3,678,474 $1,913,222 $2,579,070 $2,387,836 $1,142,600 $11,701,202 $5,591,696 1. Based on 56% of crashes occurring during peak periods when metering typically operates. 2. Crash cost factors from ADOT Highway Safety Improvement Program Manual, March 2010 Table 46. 20-yr B/C Analysis of Metering the EB 7th Ave and WB 7th St On-ramps Annual Const. Change in Cost Total Annual Annual (CRF x Total Cost O & M Cost Const Cost) (1) Peak Period Crash Reduction (1) Total Const(2). Cost 5% $500,000 $50,950 $1,000 10% $500,000 $50,950 15% $500,000 $50,950 1. 2. Annual Benefit B/C All Crash Severities K+A All Crash Severities K+A $51,950 $195,020 $93,195 3.75 1.79 $1,000 $51,950 $390,040 $186,390 7.51 3.59 $1,000 $51,950 $585,060 $279,585 11.26 5.38 CRF =0.1019; 20 yrs at 8% interest; ADOT HSIP Manual, 2010 Construction cost reflects the cost for installing two ramp meters with connection to the FMS. It does not include costs for improvements to the cross streets that may be needed to provide queue storage. Kittelson & Associates, Inc 67 I-10 Phoenix Corridor Safety Study Final Report RESTRICT HOV LANE ACCESS Description The crash analysis revealed that approximately 60% of the crashes that occurred between lane 1 and the HOV lane in 2012 and 2013 involved single occupancy vehicles entered the HOV lane, typically when the general purpose lanes were congested. Crashes in the HOV lanes typically involved higher speed vehicles traveling in the HOV lanes colliding with lower speed vehicles entering the HOV lane. Several countermeasures to address crashes occurring between the interaction of the HOV lane and Lane 1 can be considered. Increased enforcement of the HOV lanes, particularly during peak traffic periods would reduce the potential of single occupant vehicles jumping into the HOV lane to avoid congestion. The increased enforcement could be achieved using additional DPS patrols within the corridor during peak traffic periods. Increased fines for HOV violations with appropriate signing placed along the HOV lanes may also lower violations. Restricting HOV lane access to all vehicles using striping (double yellow lines) and signing, as is done in California, may also reduce HOV related crashes. The restrictions could be placed on a limited section of the HOV lanes within the study section, between 7th St and 7th Ave for example, where the highest crash frequencies were observed in the 2012-2013 data and are expected based on the safety performance evaluation of the study section. Potential Crash Reduction The potential crash reduction associated with increased enforcement and fines is unknown. Limited restriction of HOV lane access will have to be carefully assessed, as research conducted in California (Ref. 11) suggests that access restrictions may actually result in higher crash frequency than with unrestricted access control. Kittelson & Associates, Inc 68 I-10 Phoenix Corridor Safety Study Final Report HIGH FRICTION SURFACE TREATMENT Description High-Friction Surface Treatment (HFST) is a specialty pavement treatment used specifically to enhance friction. It is commonly used as a safety treatment on curves, at intersections, on ramps, and on steep grades to reduce stopping distance and the potential of skidding off the road, particularly under wet conditions. HFST is comprised of a thin a resin binder sprayed over the pavement surface followed by broadcasting of abrasion and polish-resistant aggregate. A recent study conducted by the FHWA that evaluated pavement safety performance (Ref.12) reported that while initial evaluation of applications on curves and ramps around the U.S. indicates potentially significant crash reduction benefits, additional research is needed to develop a CMF that can be provided to practitioners. Since rear-end crashes comprise the highest proportion of crashes on the I-10 study section, HFST may be an effective countermeasure for this crash type, particularly in instances when higher speed vehicles encounter stopped or slowed traffic. Testing HFST on a segment of the I-10 study section would be an appropriate application. A possible candidate segment would be in the westbound direction of the Deck Park Tunnel. This segment includes a downgrade and horizontal curve. Westbound traffic entering the tunnels often encounter slowed or stopped traffic, either within the tunnel or at the tunnel exit. Since the application of HFST has primarily been on lower volume highways and freeways, its survivability and service life on a high volume, high speed freeway is unknown and will need to be assessed. The New Mexico DOT recently installed HFST on a bridge deck on I-10 in Albuquerque with a 65 mph posted speed limit. This site may provide information on the service life of the treatment. Manufacturers’ have suggested a 7 to 10 year service life for the I-10 study section traffic conditions. The potential effect of HFST on traffic generated noise levels will also need to be considered. The cost of applying HSFT is $21 to $26/sq yd. For the westbound direction of the tunnel segment, the estimated installation cost, excluding traffic control, would be in the range of $550,000. OTHER POTENTIAL COUNTERMEASURES TO CONSIDER As noted in the diagnostic review, congestion and high speed during peak traffic periods are primary factors associated with the high observed and expected crash frequncy within the study section. Lane changing associated with merging and weaving are also contributing factors. As such, improvements that will reduce these conditions can be considered. Several additional cost effective countermeasures for consideration include the following:   Install freeway lane markings in advance of the I-17 and SR 51/SR 202 interchanges where two or more lanes exit I-10. Turn-on the ramp meter on the westbound Sky Harbor Blvd on-ramp during morning and evening peak hours to reduce the impact of merging traffic on the congested mainline. Kittelson & Associates, Inc 69 I-10 Phoenix Corridor Safety Study   Final Report Install a high friction pavement surface treatment to reduce rear end crashes, which are the primary crash type within the study section. Additional information on the potential benefit and cost of this countermeasure is needed. Non-engineering countermeasures intended to reduce speed and increase driver awareness during peak traffic periods can also be considered. Potential countermeasures could include: o Heightened and visible DPS presence within the corridor. For example, station DPS vehicles at highly visible locations, such as at the entrances to the tunnel. o Provide messages using existing dynamic message signs reminding drivers to be prepared for upcoming congestion and to drive at a safe speed. These messages can be alternated with the current travel time information being provided. o Develop and implement a public outreach campaign designating this section of I-10 as a “safety corridor”. The campaign would inform the public of the need to improve safety on this freeway section, steps being taken by ADOT and DPS to do so, and actions that the public can take to reduce crash potential. o Regional and statewide efforts to reduce distracted driving. Study Results and Recommendations I-10 STUDY SECTION SAFETY PERFORMANCE The safety performance of I-10, between 35th Ave and Sky Harbor Blvd was evaluated using the predictive methods prescribed in the Highway Safety Manual. Thirty basic freeway segments and 29 speed change lane segments within the study section were evaluated. Several performance measures were applied to determine the magnitude of the safety problem within each segment and identify segments with high potential for reducing crash frequency and severity. Safety performance on the I-10 study section is described as follows:    4900 reported crashes occurred on the I-10 study section over the 3-year period (2011 to 2013). The crash rate in the I-10 study section from 2011 to 2013 was 3.10 crashes per million vehicle miles compared to 1.47 crashes per million vehicle miles on a representative sample of other Phoenix area freeways. Referring to Figure 7, 18 of the study section segments have expected crash frequencies exceeding 150 crashes per mile per year. Six segments have crash frequencies exceeding 250 crashes per mile per year. The average for 40 other Phoenix area freeway segments is 82 crashes/mile per year, considering all crashes. When considering only fatal and injury crashes, 22 segments have a crash frequency of 40 crashes per mile per year compared to an average of 25 crashes per mile per year for the other Phoenix freeways. These results indicate that the I-10 study section is a priority in the region for safety improvement. Kittelson & Associates, Inc 70 I-10 Phoenix Corridor Safety Study  Final Report 6 fatal crashes (0.2%) and 40 incapacitating injury crashes (1.2%) occurred over the 2-year period 2012 to 2013. Rear-end crashes are the predominant crash type. Referring to Figure 6, the level of service of safety (LOSS) results for all crashes, or only considering fatal and injury crashes, indicates moderate to high potential to reduce crash frequency on a large majority of the study segments. When considering the expected crash frequency and LOSS results, it is clear that the I-10 study section is a strong candidate for implementing appropriate crash reduction countermeasures.    CONTRIBUTING FACTORS       Focusing on the high crash locations identified from the safety performance evaluation, detailed crash diagnosis was conducted to identify roadway, operations, and traffic factors that are contributing to the high expected crash frequencies. The diagnosis included assessment of crash trends and patterns, review of 756 crash reports of 3,255 crashes reported in 2012 and 2013, and field reviews. The crash diagnosis produced the following results:Speed and congestion were contributing factors in approximately 57% of all crashes occurring within the study section in 2012 and 2013. Congestion is a primary contributing factor, with 70% of the reported crashes occurring during the morning peak period (6:00 to 9:00) and afternoon peak period (3:00 to 7:00). Speed is also a primary factor. The average speed appears to range from 65 to 70 mph during off peak periods. During peak periods, crashes associated with high speed vehicles encountering slowed or stopped vehicles is common. Of the 40 incapacitating injury crashes, half noted congested conditions and speed as factors. During peak periods, differential speeds between stop-and-go traffic in the general lanes and vehicles in the HOV lanes is a factor. Crashes associated with vehicles entering/exiting the HOV lanes in combination with the noted speed differential during peak periods is a factor. The number of reported crashes at the left-side HOV entry ramps (3 locations) are low. As such this roadway condition is not a contributing factor. A combination of several factors contribute to the exceptionally high crash frequency in the Deck Park Tunnel. These factors include: o the illumination levels at the tunnel threshold zones are lower than the original design, likely due to the effect of age and environment on the luminaires, o the difference between the day-time ambient illumination level outside of the tunnel and the illumination level at the tunnel portals which can affect drivers as their eyes adjust when they enter/exit the tunnel, o lane changing within the tunnel associated with the eastbound and westbound weaving sections between 7th St and 7th Ave, o the horizontal curve within the tunnel, and o congestion occurring within the tunnel and just outside the exit portals. Kittelson & Associates, Inc 71 I-10 Phoenix Corridor Safety Study   Final Report Heavy lane changing associated with the major weaving sections at the I-17 and L202/SR51 system interchanges creates congestion at the exit ramps, resulting in rear-end and sideswipe crashes. Mainline right-side lane drops at two locations, WB at 7th Ave and EB at 7th St are a factor in observed crashes. The lane drops contribute to congestion and produced forced lane changes. The EB lane drop at Washington St, is not as much a factor in crash frequency as at the other two lane drop locations. POTENTIAL CRASH COUNTERMEASURES Countermeasures to mitigate the contributing factors identified in the diagnostics process, resulting in reduced crash frequency within the study section were identified and analyzed. The analysis estimated the potential reduction in crashes using crash modification factors available from the HSM and the FHWA CMF Clearinghouse, and estimated using the ISATe crash prediction modeling tool. When possible, a benefit/cost analysis of each countermeasure was prepared following the guidelines from the ADOT HSIP manual. The following summarizes the results of the countermeasure evaluation and recommendations for potential implementation. Lower Speed Limit: Lowering the speed limit in conjunction with heightened and continued enforcement is expected to lower the average operating speed on the freeway, potentially reducing crash frequency substantially at a high benefit/cost. Lowering the speed limit from 65 to 55 mph with focused enforcement is expected to lower average operating speeds from 2 to 4 mph. A 2 mph reduction in operating speed is estimated to result in 54 fewer fatal and injury crashes per year, producing an annual benefit/cost of 2.14 to 7.14 when considering K+A+B crashes and 1.62 to 5.41 for K+A crashes only. The benefit/cost range reflects the varying effect that a lowered speed limit will have on crashes during peak periods when congestion is present. PDO crash frequency would also drop, however there is no CMF available for estimating the amount. Lowering the operating speed by 3 or 4 mph would produce substantially higher safety benefits. Recommendation: ADOT and DPS should evaluate this countermeasure for implementation. Variable Speed Limit: VSL systems are effective in lowering primary and secondary crashes, and increasing capacity and throughput on high speed access controlled roadways. VSL systems are typically part of an active traffic management system that may also include advance warning of congestion, crashes, and lane closures. Installing a variable speed limit system that operates primarily during weekday peak traffic periods to lower operating speeds on I-10 is estimated to result in potentially 100 fewer total crashes and 30 fatal and injury crashes each year. The estimated benefit/cost, based on a system implementation cost of $8.2 million and additional enforcement is 2.3 for all crashes and 1.6 for K+A crashes only. Recommendation: Prepare a VSL system design concept to define system limits, operations, system requirements, and construction, operating, and enforcement methods and costs. Kittelson & Associates, Inc 72 I-10 Phoenix Corridor Safety Study Final Report Upgrade Deck Park Tunnel Lighting: It is not possible to definitively quantify the effect of existing illumination levels on the high expected crash frequency within the tunnel, particularly since there are other contributing factors (weaving sections, horizontal curve, and congestion). However, based on the review of crash reports and field review, and the results of the illumination study conducted as part of the Deck Park Tunnel Energy Efficiency Study, it is reasonable to conclude that upgrading the tunnel illumination to meet current IES recommendations and to reduce the blinding effect of the daytime ambient light level at the tunnel portals will reduce crashes. Considering the high cost of a lighting upgrade ($20 million) and the uncertainty of the improvement on reducing crash frequency, developing a reasonably sound benefit/cost based is not possible. Recommendation: It is understood that ADOT is moving forward on a possible upgrade to tunnel lighting to reduce energy costs. If implemented, a study of the effect of tunnel illumination improvements on crash frequency should be conducted. Eliminate Lane Drops: Although the right-side lane drops at 7th Ave and 7th St are a factor in crash frequency at these locations, analysis of the benefit of extending the lanes indicates a fairly low reduction in crashes and low benefit/cost. The option of dropping each lane at the upstream exit results in similar low reduction in crashes for the 7th St lane drop, but higher crash frequency at the 7th Ave lane drop. Recommendation: Countermeasure not recommended for implementation Meter EB 7th Ave and WB 7th St Entry Ramps: Lane changing is contributing to the high crash frequency in the tunnel. During peak periods, heavy ramp volumes are entering each weaving section in the tunnel, often at higher speed than the mainline traffic. Although studies on the effect of ramp metering on safety limited and high quality CMFs are not available, a nominal 5% crash reduction during peak periods is estimated to produce a benefit/cost of 3.75 for all crashes and 1.8 for K + A crashes. Recommendation: Conduct further study of the effect of ramp metering on traffic operations on the mainline, at the ramp terminal intersections, and on the cross streets. Consider metering both ramps. Restrict HOV Lane Access: Review of crash records indicates that vehicles, often single occupant, jumping from the congested mainline lanes to the HOV lane are causing crashes due to the higher speeds of vehicles traveling in the HOV lanes. Based on a recent California study, it is unlikely that restricting continuous HOV lane access, as currently is the case on Phoenix area freeways, will reduce HOV lane related crashes. Reducing the speed differential between congested mainline lanes and HOV lanes, could be achieved from a VSL system. Recommendation: Restricting HOV lane access is not recommended. High Friction Surface Treatment: Since rear end-crashes are predominant, application of HFST may prove to be an effective countermeasure for this crash type. Although a crash modification factor is Kittelson & Associates, Inc 73 I-10 Phoenix Corridor Safety Study Final Report not available, a recent FHWA study concludes that this treatment potentially offers a significant crash reduction benefit. Recommendation: Evaluate the service life of this treatment for application on the I-10 study section. Consider applying HFST on a high crash segment to test it’s effectiveness, possibly westbound I-10 within the Deck Park Tunnel. If installed, evaluate the safety benefit of this treatment. Automated Speed Enforcement: The Highway Safety Manual identifies automated speed enforcement as an effective method of lowering average operating speeds, potentially lowering crashes by as much as 17%. Recommendation: This enforcement method could be considered with lowering the speed limit or a variable speed limit system. Other Recommended Countermeasures    Install freeway lane markings in advance of the I-17 and SR 51/SR 202 interchanges where two or more lanes exit I-10. Turn-on the ramp meter on the westbound Sky Harbor Blvd on-ramp during morning and evening peak hours to reduce the impact of merging traffic on the congested mainline. Implement non-engineering countermeasures intended to reduce speed and increase driver awareness during peak traffic periods. Potential countermeasures could include: o Heightened and visible DPS presence within the corridor. For example, station DPS vehicles at highly visible locations, such as at the entrances to the tunnel. o Provide messages using existing dynamic message signs reminding drivers to be prepared for upcoming congestion and to drive at a safe speed. These messages can be alternated with the current travel time information being provided. o Develop and implement a public outreach campaign designating this section of I-10 as a “safety corridor”. The campaign would inform the public of the need to improve safety on this freeway section, steps being taken by ADOT and DPS to do so, and actions that the public can take to reduce crash potential. O Regional and statewide efforts to reduce distracted driving. APPLICATION OF STUDY RESULTS AND TOOLS FOR FUTURE SAFETY EVALUATIONS Not only has the I-10 safety study successfully demonstrated the application of the HSM quantitative safety performance process, but the results and tools produced can directly be applied to safety evaluations of other urban freeway sections in the Phoenix area, as well as for evaluating the safety performance of projects intended to add freeway capacity or improve operations and safety. The training provided to ADOT staff as part of the study, which included segmentation of a freeway section, gathering roadway condition data using Google Earth, and application of the ISATe evaluation tool will Kittelson & Associates, Inc 74 I-10 Phoenix Corridor Safety Study Final Report support future safety evaluations. Application of the study results and tools, as well as limitations on their use, is described below. CALIBRATED CRASH PREDICTION MODELS Eight freeway crash prediction models, listed in Table 47 were calibrated to reflect local conditions so that the models will more accurately assess the safety performance on freeways in the Phoenix metropolitan area. Table 47. Freeway Predictive Models Calibrated for Phoenix Urban Area Urban Basic Freeway Segments Urban Freeway Speed-change Lanes  Multiple vehicle crashes – FI  Multiple vehicle crashes – FI  Multiple vehicle crashes – PDO  Multiple vehicle crashes – PDO  Single vehicle crashes – FI  Single vehicle crashes – FI  Single vehicle crashes – PDO  Single vehicle crashes - PDO Each model predicts average crash frequency using Safety Performance Functions (SPFs), Crash Modification Factors (CMFs), and a local calibration factor (C). The calibration factors provided in Table 48 were developed using randomly selected Phoenix urban freeway segments with generally similar operational characteristics as exists on the I-10 study section. Included in the calibration database were segments with 6, 7, 8, 9, 10, and 11 through lanes, all with a posted speed limit of 65 mph. Two particularly key features included in all of the calibration segments are ramp metering and HOV lanes. Since both are common on Phoenix freeways and are present in the study section, the effect of these features are represented in the calibration factors. Table 48. Predictive Model Calibration Factors for Phoenix Area Urban Freeways Model Location F+I Calibration Factor, C PDO Calibration Factor, C Multiple vehicle Freeway Segment 1.63 1.37 Single vehicle Freeway Segment 0.94 1.12 Multiple vehicle Speed-change lane 2.27 2.29 Single vehicle Speed-change lane 1.58 2.14 The crash severity distribution function (SDF), which is used to predict the proportion of fatal and injury crashes by severity, (K, A, B, and C), was also calibrated. The procedure for quantifying this calibration factor is described in the Highway Safety Manual First Edition Supplement (Ref 6). This factor adjusts the SDF such that the predicted frequency of K-fatal, A-incapacitating injury, B- nonKittelson & Associates, Inc 75 I-10 Phoenix Corridor Safety Study Final Report incapacitating injury, and C-possible injury crashes matches more closely to that for the Phoenix metro area. The calculated SDF calibration factor is 2.14 and applies to the predictive models for both basic freeway segments and speed-change lanes. The eight calibrated crash prediction models for basic freeway segments and speed-change lanes can be used to evaluate safety on Phoenix urban freeways with the following limitations. First, since the model calibration sections included continuous access HOV lanes and ramp meters, the model should not be applied to freeway sections without HOV lanes as well as freeways without ramp metering, without additional calibration for these conditions. Second, care should be used in applying the model to freeway sections with posted speed limits below 65 mph, in which case the free-flow speed should be approximately 65 mph. Again, additional calibration for segments with posted speeds less than 65 mph may be desirable, depending on the operating speeds on the segments under study. SAFETY PERFORMANCE EVALUATION TOOLS ISATe The ISATe tool, with calibrated models for Phoenix urban freeways is available. The calibrated tools can also be used to evaluate safety performance on existing freeways, assess freeway design alternatives, aide in crash diagnostics, evaluate crash mitigation measures, and analyze design exceptions/variances. IHSDM Design Support Software The Interactive Highway Safety Design Model (IHSDM) was developed by FHWA as a suite of software analysis tools used to evaluate operational effects of geometric design decisions on highways. The IHSDM was originally designed to provide “decision-support” in the highway design process – comparing existing or proposed roadway designs against relevant design and operations policy values. A crash prediction module, which incorporates the HSM methodology was added to estimate the safety impacts of design decisions. One advantage that the IHSDM software has over ISATe is the ability to import CAD files which can substantially simplify both segmenting of the freeway section being evaluating and gathering the necessary roadway geometry data. However, if CAD files are not available, manual segmentation and gathering of roadway geometry is required. Performance Measure Spreadsheet The spreadsheet developed for the study to calculate safety performance measures can be used by ADOT for other safety studies. The spreadsheet is linked to the ISATe segment output data, automating the calculation of performance measures. Kittelson & Associates, Inc 76 I-10 Phoenix Corridor Safety Study Final Report GIS Crash Analysis Tool: The GIS-based crash analysis tool developed for the I-10 study section provides an efficient means of performing crash diagnostics. A similar tool can be created for other freeway sections to be studied. DATA COLLECTION SOURCES AND METHODS Roadway Geometry and Physical Characteristics The study results have provided valuable information regarding the availability and quality of data on roadway geometry and characteristics, traffic volume data, and crash data. Roadway geometry data provided in the ADOT RCID is appropriate for use in safety studies. Additional roadway characteristic data will need to be collected manually, preferably using recent aerial photographs. Use of EarthTools in conjunction with Google Earth offers an accurate and efficient method of extracting the roadway condition information. Given the relatively good correlation between the RCID data and the data collected manually, as noted in Table 49, it would be more expedient on future safety studies to use the RCID data for these data elements and use the manual method to fill the gaps. Table 49. Recommended Data Sources for Safety Performance Evaluations Feature ADOT RCID Basic segment length Speed-change lane length No. through lanes Width of lanes, shoulders, median Horizontal curve radius Horizontal curve length Manual Method (1)         Feature Barrier length Barrier offset Continuous median barrier width Clear zone width ADOT RCID  Manual Method (1)     Type B weaving section  Distance to upstream    ramp Rumble strips; inside & Distance to downstream   outside shoulders ramp 1. Roadway condition data can be extracted manually from aerial photos, design drawings, as-built plans, and using Google Earth. Traffic Volume Data Mainline AADT data is available from both the ADOT Traffic Data Management System (TDMS) and the FMS. It is recommended that the TDMS data be used to determine mainline segment volumes for freeway segments. Segment volumes can be calculated based on the AADT from the nearest TDMS mainline count station, adjusted to reflect traffic entering or exiting from adjacent ramps. Ramp AADT’s are available from the biannual traffic count conducted by ADOT. Level of congestion, described as the proportion of high volume hours occurring during a typical day can be determined from hourly volume data available from the FMS. Using Microsoft Access, the FMS Kittelson & Associates, Inc 77 I-10 Phoenix Corridor Safety Study Final Report database can be queried to determine the average annual proportion of high volume hours for each segment. A high volume hour is defined as when the average lane volume exceeds 1,000 vphpl. The proportion is calculated by summing the volume during each hour where the average lane volume exceeds 1,000 vphpl and then dividing this sum by the AADT. Crash Data Multiple years, typically 3 to 5, of crash data should be used to evaluate freeway safety performance. Review of the crash data provided from the ADOT Safety Data Mart for calibration of the crash prediction models and evaluation of safety performance on the I-10 study section provided several important results. First, the new GPS-based Traffic and Citation (TraCS) crash reporting system implemented by the Arizona Department of Public Safety in 2010 and 2011 provides more accurate crash location data, certainly to the tenth of a mile and possibly to the hundredth of a mile. This improved level of accuracy was observed consistently beginning with the 2012 crash data. Crash records from earlier years show a substantial proportion of crashes located at the mile or half-mile. Therefore, it is recommend that safety evaluations only use crash data beginning in 2012. Second, crash data needs to be screened to remove duplicate crash records, crashes that occurred during mainline or ramp construction or during an unusual event (i.e. freeway flooding or dust storm), and crashes that may have been incorrectly coded to the mainline, ramp, or frontage road. Periods when work zones were present or a freeway closure was in effect can be identified using the ADOT Highway Condition Reporting System (HCRS) database. Crash data during these periods would be removed from the safety study database. In addition, the RoadCondition attribute in each crash record can indicate the presence of a work zone. Several attributes in each crash record can be reviewed to confirm the freeway facility where the crash occurred. TrafficWayType identifies the type of roadway on which the crash occurred and JunctionRelation indicates if a crash was located on a ramp or frontage road. Kittelson & Associates, Inc 78 I-10 Phoenix Corridor Safety Study Final Report EXPANSION OF ARIZONA FREEWAY CRASH PREDICTION MODELS The I-10 study has provided a solid foundation for conducting safety performance evaluations on Arizona freeways using the HSM predictive methods. In order to expand the predictive models to cover all freeways throughout the state, it is recommended that the following additional model calibration be conducted.     Develop calibration factors for urban freeways without HOV lanes and without ramp metering. Develop calibrated crash predictive models for 4 and 6-lane rural freeways. Evaluate the need for calibration factors for urban freeway with posted speeds of 55 mph. Evaluate the applicability of calibration factors based strictly on MAG region data to other urban areas in the state. Should ADOT desire to expand the predictive models to cover ramps and ramp terminals to support the evaluation of interchange safety performance, calibration factors for these facilities can also be developed. FHWA should also consider conducting research on the following issues that were identified in the I-10 safety study. The effects of HOV lanes with unlimited access and ramp metering on crash frequency is unknown. These effects are reflected in the calibration factors and not as a separate CMF. Closer examination of HOV lane crashes in the diagnostics phase of the study found a notable trend in crashes associated with vehicles entering or exiting the HOV lane. Quantifying this effect on Arizona’s HOV lanes in comparison to other HOV facilities in the U.S. which do not restrict access location, will enhance the crash prediction models, provide information on ways to mitigate these crashes, and better inform future design of HOV facilities. Kittelson & Associates, Inc 79 I-10 Phoenix Corridor Safety Study Final Report REFERENCES 1. Zegeer, C., and C. Williams. “Calculation of Accident Rates by Roadway Class for HSIS States.” UNC Highway Safety Research Center, June 1994. Accessed 8/16/2014. http://www.hsrc.unc.edu/research_library/PDFs/Calculation1994.pdf 2. 2011 State Highway Crash Rate Tables. Transportation Data Section, Oregon Dept. of Transportation. August 2012. Accessed 8/16/2014. http://www.oregon.gov/ODOT/TD/TDATA/car/docs/2011_Crash_Rate_Book.pdf 3. Dille, J., C. Fountain, and S. Jackson. “NCTCOG Regional Crash Rate Development.” Presentation made to the Regional Safety Working Group. September 23, 2011. Access 8/16/2014. http://www.nctcog.org/trans/safety/CrashRateDevProcess_092311.pdf 4. J.A. Bonneson, S. Geedipally, M. Pratt, and D. Lord. Safety Prediction Methodology and Analysis Tool for Freeways and Interchanges. Final Report. NCHRP Project 17-45. Texas Transportation Institute, College Station, Texas, 2012. 5. J.A. Bonneson, S. Geedipally, M. Pratt, and D. Lord. Enhanced Interchange Safety Analysis Tool – Users Manual. Texas Transportation Institute, College Station, Texas, 2013. 6. Highway Safety Manual, First Edition Supplement, AASHTO, 2014 7. Bham, G. H., et.al., “Evaluation of Variable Speed Limits on I-270/I-255 in St. Louis.” RI08-025, Missouri University of Science and Technology, Rolla, MO., (2010). 8. Liu, C.and Wang, Z., "Ramp Metering Influence on Freeway Operational Safety near On-ramp Exits." International Journal of Transportation Science and Technology, Vol.2, No.2, Multi Science Publishing, (2013) pp.87-94. 9. Deck Park Tunnel Energy Efficiency Study, Final Report, DH Lighting Solutions and Energy Systems Design, October, 2011. 10. Liu, C.and Wang, Z., "Ramp Metering Influence on Freeway Operational Safety near On-ramp Exits." International Journal of Transportation Science and Technology, Vol.2, No.2, Multi Science Publishing, (2013) pp.87-94. 11. Jang, K., K. Chung, D. Ragland, and C.Y. Chan, "Safety Performance of High-Occupancy-Vehicle Facilities: Evaluation of HOV Lane Configurations in California," Transportation Research Record: Journal of the Transportation Research Board, No. 2099, Transportation Research Board of the National Academies, Washington, D.C., 2009, pp. 132–140 12. Merrit, D., C.,Lyon, and B.Persaud, , Evaluation of Pavement Safety Performance , Federal Highway Administration, FHWA-HRT-14-065, Feb. 2015. 13. Kittelson & Associates, Inc 80 I-10 Phoenix Corridor Safety Study Final Report APPENDIX A – DISCUSSION OF SAFETY PERFORMANCE MEASURES PROBABILITY OF SPECIFIC CRASH TYPES EXCEEDING THRESHOLD PROPORTIONS Description: The purpose of this performance measure is to identify segments with a higher proportion of a specific crash type or severity that is of interest to an agency. For example, an agency may want to focus on roadway departure crashes or fatal and incapacitating injury crashes. The agency would establish a threshold proportion of the specific crash type or severity that is considered “acceptable”. The threshold could be the average proportion across similar sites or a percentile proportion (e.g., 85th percentile), or a targeted desired proportion (e.g., goal of less than 15% of speed related crashes). For each roadway segment, the probability that the long-term predicted proportion of crashes at the site is greater than the threshold proportion is calculated. Segments where the long-term prediction proportion is greater than the threshold proportion are identified. Strengths: Considers variance in data; not affected by RTM bias; can also be used as a diagnostic tool. Limitations: Does not account for traffic volume; segments may be flagged for review due to unusually low frequency of non-target crash types. Application: For the initial application of this performance measure to the I-10 study, the evaluation focused on severe crashes (K, fatal; A, incapacitating injury; B, non-incapacitating injury). The threshold proportions applied were 0.144 (or14.4 % of total crashes) for basic segments and 0.108 for speedchange lanes. Note that these proportions represent what was found in the freeway segments used for the safety prediction model calibration. Therefore, the intent of this evaluation is to identify segments in the I-10 study section where the predicted long term proportion of severe crashes is greater than the proportions observed on typical Phoenix freeways. Results: Each of the six calculated performance measures is listed in Table 50 and Table 51 for basic segments and speed-change lanes, respectively. Column 3 lists the probability that the proportion of severe (KAB) crashes within each segment of the I-10 study section will exceed the typical proportion on similar Phoenix freeways. Based on the rankings provided in Column 4, the top 10 basic segments and speed change lanes are fairly well spread out within the I-10 study section. It is worth noting that all of the top 10 speed-change lanes have a probability 70% or higher, while only two basic segments have a probability above this section. This suggests that focusing improvements on the speed change lanes may provide the greatest potential for crash reduction. It also suggests that the severity of crashes on the basic segments in the study section are not more severe relative to typical Phoenix freeways. However, crash severity may be worse on the speed-change lanes . Kittelson & Associates, Inc 81 I-10 Phoenix Corridor Safety Study Final Report Table 50. Calculated Performance Measures and Ranking - Basic Segments 1 2 3 4 1. Probability of Specific Crash Type Exceeding Threshold Proportion 6 7 8 9 2. Excess Proportion 3. Expected of Specific Crash Type Average Crash Frequency with EB Adjustment 10 11 12 13 4. Equivalent PDO Average Crash Frequency with EB Adjustment 14 15 16 17 5. Excess Expected Crash Frequency with EB Adjustment 18 19 20 21 22 23 24 6. Modified Level of Service of Safety (MLOSS) Basic Segments LOSS I II III IV Crash Severity: Crash Severity: Avg. Crash Frequency for Calibraion Avg EPDO for Calibration Sites: Severe (KAB) Severe (KAB) Sites: FI = 51.2 FI+PDO = 164.8 Economic: 551 Comp.: 885 Threshold Proportion: Threshold Proportion: EPDO CRASHES PER MILE / 2 YEARS p* = 0.144 p* = 0.144 Basic Segments + Speed-Change Lanes Begin Expected Crashes per mile/2yr PDO FI PDO+FI PDO FI PDO+FI Z SV FI Z 25 85.9 12 318.4 11 1227 17 2284 17 49.2 12.6 58.4 12 16 11 0.80 III 0.15 III 3.12 IV 1.57 IV 11 16 106.0 7 394.6 8 1482 14 2698 14 71.3 16.7 86.3 10 11 10 0.66 III -0.01 II 6.14 IV 2.62 IV 8 142.43 0.88 1 0.07 2 92.1 11 282.4 13 1304 15 2375 15 36.8 18.5 54.7 15 10 12 0.20 III 0.34 III 3.03 IV 2.27 IV 10 142.71 0.83 2 0.13 1 69.2 16 186.2 19 816 22 1508 22 17.4 15.3 33.6 21 13 17 -0.33 II 0.03 III 1.73 IV 1.80 IV 17 142.88 0.53 7 0.06 3 68.9 17 176.7 21 804 23 1495 23 12.8 16.3 27.5 23 12 20 0.32 III -0.40 II 0.79 III 1.64 IV 20 I-17 143.00 0.02 28 -0.06 22 48.3 21 179.9 20 638 26 1146 26 25.5 7.8 32.5 17 20 18 0.41 III 0.47 III 4.03 IV 2.25 IV 9 19th Ave 143.67 0.65 3 0.04 5 99.7 9 276.8 14 1257 16 2365 16 -2.1 10.3 10.5 26 17 24 -0.66 II 0.07 III 0.23 III 1.23 III 22 143.95 0.36 12 0.00 13 85.8 13 267.3 15 1930 11 3648 11 18.5 8.5 28.3 20 18 19 -0.65 II 0.20 III 2.18 IV 1.30 III 18 144.36 0.24 17 -0.14 27 2.5 29 64.7 28 2455 4 4503 4 20.5 -3.2 18.1 19 24 23 -0.56 II -0.16 II 1.41 III -1.77 I 23 144.42 0.29 15 -0.01 15 155.6 4 583.3 5 2346 5 4445 5 167.2 56.0 219.9 3 3 3 144.80 0.62 4 0.02 7 251.2 1 887.5 1 3615 1 6823 1 263.3 102.8 368.4 1 1 1 -0.50 144.92 0.18 22 -0.14 27 10.9 27 194.1 17 2023 9 3805 8 52.6 -12.9 36.8 11 27 16 144.96 0.00 29 -0.05 21 214.2 2 835.9 2 2783 2 4874 2 214.4 57.8 269.2 2 2 2 145.42 0.53 6 0.03 6 147.5 6 524.4 6 2019 10 3741 10 116.0 42.9 161.6 6 5 145.52 0.34 14 0.00 14 156.7 3 620.2 3 2098 7 3759 9 166.5 49.4 215.5 4 4 4 0.14 III 145.78 0.36 11 0.01 9 97.9 10 402.6 7 2592 3 4538 3 104.6 30.0 136.9 8 7 7 -0.73 II 145.82 0.22 19 -0.07 23 28.7 25 114.9 25 1649 13 2995 12 1.7 -3.0 0.2 25 23 26 -0.97 145.94 0.05 26 -0.03 19 154.3 5 616.3 4 2320 6 4194 6 124.8 25.9 149.8 5 8 6 0.39 III -1.14 146.44 0.38 10 0.00 12 102.0 8 353.2 9 2093 8 3889 7 81.5 30.9 110.6 9 6 9 0.93 III 146.85 0.06 25 -0.11 26 57.6 19 341.4 10 1687 12 2953 13 112.8 13.1 122.1 7 15 8 0.72 III -0.38 146.93 0.21 21 -0.02 17 39.1 23 164.6 22 653 25 1201 25 37.5 7.8 40.9 14 19 14 1.60 IV 0.94 147.51 NA NA -0.14 27 NA NA NA NA 551 29 1024 29 NA NA NA NA NA NA NA NA NA 147.67 0.35 13 0.00 11 36.7 24 103.6 27 588 27 1115 27 -6.1 2.0 0.0 28 22 27 -2.63 I -0.57 147.96 0.54 5 0.06 4 71.0 15 199.4 16 955 19 1830 19 32.1 22.5 54.5 16 9 Z LOSS 148.07 0.40 9 0.01 10 60.4 18 191.3 18 863 21 1630 21 21.0 13.5 39.4 18 14 15 -2.71 148.31 0.22 19 -0.07 23 22.1 26 114.9 26 923 20 1656 20 8.0 -5.3 4.3 24 25 25 -1.01 148.43 0.40 8 0.02 8 56.0 20 159.1 23 694 24 1214 24 -49.0 -26.8 -74.2 29 29 29 -0.69 148.66 0.18 23 -0.03 20 41.5 22 149.0 24 561 28 1026 28 16.1 4.8 20.8 22 21 22 0.06 III 149.14 0.27 16 -0.14 27 9.6 28 59.6 29 337 30 571 30 -4.2 -15.3 -24.0 27 28 28 0.51 III -0.16 149.17 0.23 18 -0.03 18 80.7 14 313.3 12 1050 18 1857 18 40.6 -12.8 21.3 13 26 21 1.06 III IV 5.89 IV 1 II 7.59 IV 4.74 IV 5 0.31 III -0.32 II 1.51 IV -2.81 I 25 0.94 III -0.32 II 11.80 IV 5.76 IV 2 5 -0.67 II -0.63 II 4.14 IV 2.50 IV 12 0.79 III 8.18 IV 4.29 IV 3 0.74 III 2.56 IV 1.17 III 16 II -0.28 II 0.38 III -0.42 II 24 0.00 0.39 III 10.12 LOSS II -0.05 13 1.30 Z Rank -0.02 III LOSS MV FI -0.08 1.04 LOSS MV PDO 24 Kittelson & Associates, Inc Rank Comprehensive Rank SV PDO 27 Sky Harbor Blvd Rank Economic Crash Reduction Potential Low Low to Moderate Moderate to High High 0.18 Van Buren St FI+PDO 30 0.02 SR 51 Rank 29 141.92 7th St FI Rank 28 141.67 Deck Park Tunnel Rank Excess Crashes per mile/1yr 27 Probaility 7th Ave p - p* EPDO 26 Milepost 35th Ave Rank EPDO 25 II 8.44 IV 3.33 IV 6 III 6.63 IV 4.21 IV 4 II 3.68 IV 1.13 III 13 III 4.85 IV 2.39 IV 7 NA NA NA NA . II 0.34 III 0.54 III 27 IV 2.01 IV 14 III 0.56 III 2.19 I 0.21 III 2.95 IV 2.03 IV 19 II -0.27 II 0.88 III -0.96 II 26 II -0.61 0.08 0.05 II -5.01 III 2.52 II -0.98 III 1.47 I -4.01 I 29 1.33 III 15 II -3.16 I 28 III -1.20 II 21 IV 82 I-10 Phoenix Corridor Safety Study Final Report Table 51. Calculated Performance Measures and Ranking - Speed Change Lanes 1 2 Speed-Change Lanes 3 4 1. Probability of Specific Crash Type Exceeding Threshold Proportion 6 7 8 9 2. Excess Proportion 3. Expected of Specific Crash Type Average Crash Frequency with EB Adjustment 10 11 12 13 4. Equivalent PDO Average Crash Frequency with EB Adjustment 14 15 16 17 5. Excess Expected Crash Frequency with EB Adjustment Crash Severity: Crash Severity: Avg. Crash Frequency for Calibraion Severe (KAB) Severe (KAB) Sites: FI = 2.1 FI+PDO = 6.8 Threshold Proportion: Threshold Proportion: p* = 0.108 p* = 0.108 Expected Crashes per mile/2yr Type Probaility Rank p - p* Rank FI Rank FI+PDO 18 19 Excess Crashes per mile/1yr EN EN EX EX Rank PDO FI PDO FI 20 21 22 23 24 6. Modified Level of Service of Safety (MLOSS) Rank EN/EX EN/EX PDO FI LOSS I II III IV Entry PDO Z 25 26 27 Crash Reduction Potential Low Low to Moderate Moderate to High High Entry FI Exit PDO LOSS Z LOSS Z 28 29 Exit FI LOSS Z 30 Rank LOSS 142.38 ex 0.53 18 0.00 17 2.0 21 7.3 22 . . 0.1 -0.1 16 22 0.20 III -0.52 II 20 142.65 ex 0.40 22 -0.02 20 2.7 18 18.2 7 . . 0.7 -0.1 11 20 0.90 III -0.36 II 14 143.95 ex 0.89 6 0.29 2 2.9 15 9.5 16 . -0.1 11 20 0.90 III -0.36 II 14 143.95 en 0.73 9 0.03 12 10.0 3 23.5 4 28 28 -7.83 . 29 144.30 en 1.00 1 0.56 1 10.0 3 23.5 4 . 0.6 24 10 144.31 ex 0.62 13 0.03 13 2.9 15 9.5 16 . -0.1 19 24 144.36 en 0.53 17 0.00 17 3.8 11 17.6 8 23 18 -1.18 144.36 ex 0.56 16 0.01 16 5.9 7 20.5 6 15 23 Deck Park Tunnel 144.92 ex 0.99 2 0.20 5 4.7 9 13.7 11 0.0 -0.2 . . 17 25 -0.04 145.38 ex 0.50 19 0.00 19 2.7 19 13.3 12 0.0 -0.2 . . 17 25 -0.04 II -0.34 7th St 145.78 en 0.64 12 0.03 11 4.8 8 14.0 10 3.8 0.4 . . 3 12 2.48 IV 0.81 145.82 en 0.56 14 0.02 14 1.4 24 8.4 20 1.2 0.1 . . 8 19 1.03 III 0.08 III 145.82 ex 0.90 4 0.09 7 11.4 2 28.5 3 7 11 146.38 ex 0.04 29 -0.08 21 6.8 6 29.5 2 I -3.01 I 146.64 en 0.88 7 0.07 8 17.7 1 48.0 1 . . -2.6 1.8 27 7 -1.80 I 146.76 ex 0.89 5 0.11 6 7.4 5 17.0 9 . . 8.7 2.5 1 2 3.80 IV 146.85 en 0.19 28 -0.11 22 3.0 14 9.2 18 146.89 ex 0.70 10 0.06 9 2.1 20 8.2 21 147.45 en 0.56 15 0.01 15 1.8 23 12.2 13 147.51 en 0.27 27 -0.11 22 3.4 12 10.4 14 . . 3.6 0.8 4 8 2.39 IV 147.51 en 0.67 11 0.05 10 3.4 12 10.4 14 . . 2.9 1.8 6 6 2.13 IV 147.67 en 0.40 23 -0.11 22 1.9 22 3.4 26 9 17 147.91 ex 0.40 23 -0.11 22 0.5 28 3.1 27 5 5 1.94 IV 148.07 ex 0.44 20 -0.11 22 0.5 27 2.7 28 13 3 0.23 III 1.70 IV 148.31 en 0.87 8 0.23 4 3.9 10 8.5 19 13 3 0.23 III 1.70 IV 148.34 ex 0.94 3 0.27 3 2.8 17 7.1 23 . . -0.3 0.4 21 13 -0.30 148.43 ex 0.33 26 -0.11 22 0.7 25 4.1 24 . . -0.3 0.4 21 13 148.64 ex 0.36 25 -0.11 22 0.7 25 4.1 24 . . 5.3 0.6 2 9 149.14 ex 0.44 20 -0.11 22 0.5 29 2.3 29 . . 0.8 0.4 10 15 I-17 SR 51 Van Buren St Sky Harbor Kittelson & Associates, Inc . -6.9 0.7 -1.5 . . -0.6 . -0.5 . -0.1 0.1 . . . . -0.7 0.4 3.4 2.1 . 2.5 . 25 26 . 3.5 2.5 . 0.4 20 . 0.2 . . 29 . 0.3 . 1.0 0.5 . -0.7 . -1.2 -0.1 1.7 -5.8 . -0.2 . . 0.3 . -15.4 . . . I -1.92 II I -0.62 II 0.88 III 18 -0.17 II -0.68 II 25 . 26 0.18 III II 19 II -0.34 II . 22 II . 22 III . 6 . 12 0.39 1.53 29 -5.90 27 -0.13 II -1.48 -0.37 0.59 II III 0.32 0.24 IV 1.01 III 7 . 28 1.43 III 21 2.18 IV 1 . 27 IV 10 . 24 1.19 III 5 1.98 IV 2 . 13 1.93 IV 3 . 8 II 1 16 -0.82 III -0.62 II 2.17 III III . 8 II 0.57 III 16 -0.30 II 0.57 III 16 2.86 IV 0.88 III 4 0.84 III 0.70 III 11 83 I-10 Phoenix Corridor Safety Study EXCESS PROPORTIONS OF SPECIFIC CRASH TYPES Description: The purpose of this performance measure is to identify segments with an overrepresentation of a specific crash type of interest relative to other crash types. This measure overlaps with the previous measure (probability of specific crash types exceeding threshold proportions), identifying the degree that the proportion of a specific crash type exceeds the threshold proportion established by an agency and aiding in the ranking of segments relative to potential for crash reduction. Strengths: Considers variance in data; not affected by RTM bias; can also be used as a diagnostic tool. Limitations: Does not account for traffic volume; segments may be flagged for review due to unusually low frequency of non-target crash types. Application: For the initial application of this performance measure to the I-10 study, the evaluation focused on severe crashes (K, fatal; A, incapacitating injury; B, non-incapacitating injury). The threshold proportions applied were 0.144 for basic segments and 0.108 for speed-change lanes, representing what was found in the freeway segments used for the safety prediction model calibration. Results: Column 6 in Table 50 and Table 51 lists the excess proportion of severe crashes within the I10 section that exceeds the typical proportion on similar Phoenix freeways. A large positive value indicates greater potential for improvement of crashes associated with the subject crash type. As would be expected, the top 10 basic segments and speed-change lanes are nearly identical to those produced by the probability performance measure. With the exception of several speed change lanes, the excess proportion values for both basic segments and speed-change lanes are fairly low when compared with the threshold proportion that represents typical Phoenix freeways. The highest excess proportion value for the basic segments is only 0.13 higher than the threshold, while the value on five speed-change lanes range from 0.20 to 0.56 above the threshold. While these results suggest that speed-change lanes offer the greatest potential for crash reduction, they also suggest that the safety performance, based on severe crashes, on much of the I-10 section is not much worse, or perhaps better than on similar Phoenix freeways. The crashes are not really more severe—just much more frequent. EXPECTED AVERAGE CRASH FREQUENCY WITH EB ADJUSTMENT Description: This performance measure estimates the EB adjusted expected average crash frequency for each segment, applying the calibrated safety performance functions and observed crash data to calculate expected crash frequency for each segment. The measure can be calculated for specific crash types or severity categories, if desired. Segments can then be ranked based on expected long-term average crashes to compare relative safety performance and prioritize segments. Strengths: Accounts for RTM bias; accounts for influence of traffic volume; accounts for influence of roadway geometry and characteristics. Limitations: Requires SPFs calibrated to local conditions (not a limitation for this study). Application: EB adjusted expected average crash frequencies are calculated directly from the ISATe spreadsheet. Expected average FI crash frequencies were calculated for both basic segments and speed change lanes. Total (FI + PDO) crash frequencies were also calculated for basic segments. Expected crash frequencies for basic segments were calculated on a “per mile” basis to normalize the segments to allow for more meaningful comparison and ranking among segments. Kittelson & Associates, Inc 84 I-10 Phoenix Corridor Safety Study Results: The performance measure was calculated for FI and FI+PDO crashes. The results are provided in Columns 8 and 10 in Table 50 and Table 51. A large positive value indicates greater potential for improvement. The expected average crash frequencies per mile for the calibration segments are also shown in each table. These calibration segment crash frequencies can be considered as reference values, representing similar Phoenix freeways. Comparing performance measure values, the expected crash frequencies per mile on a majority of the basic segments and speed-change lanes within the I-10 study section exceed the reference values, often substantially. This finding is consistent with that found in the examination of crash rates (as shown in Table 7. Calibration Segments Crash Summary and). An examination of the top 10 basic segments and speed-change lanes indicates that the greatest potential for crash reduction is on the section of I-10 between I-17 and SR 51. EQUIVALENT PDO AVERAGE CRASH FREQUENCY WITH EB ADJUSTMENT Description: This performance measure weights crashes by severity to produce a single combined frequency and severity score for each segment. The measure incorporates calibrated SPFs and EB to address random fluctuations in crashes. The weighting factors are calculated relative to PDO crashes and are based on societal costs each crash severity. This measure allows segment ranking, highest to lowest score, based on expected long-term average crash frequency and severity. Strengths: Accounts for RTM bias; accounts for influence of traffic volume; accounts for influence of roadway geometry and characteristics; provides a single ranking that considers crash frequency and severity. Limitations: Requires SPFs calibrated to local conditions (not a limitation for this study). Application: EB adjusted expected average crash frequencies are calculated directly from the ISATe spreadsheet. Equivalent PDO crash frequencies for basic segments were calculated per mile. The total segment equivalent PDO crash frequency, including basic segments and speed change lanes were calculated and ranked. PDO weighting factors were determined using two types of crash costs: economic and comprehensive. The economic costs include medical costs, lost wages, and property damage. Comprehensive costs include economic costs PLUS cost associated with lost quality of life and productivity. The crash costs and PDO weighting factors are provided in Table 52. Table 52. Crash Cost Factors and PDO Weighting Factors PDO PDO Economic Weighting Comprehensive Weighting Crash Severity Cost (1) Factor Cost (2) Factor K, Fatal $1,804,479 241.4 $5,800,000 1450.0 A, Incapacitating Injury $205,862 61.78 $400,000 100.0 B, Non-incapacitating injury $40,570 12.18 $80,000 20.0 C, Possible injury $24,922 7.48 $42,000 10.5 PDO $3,332 1.00 $4,000 1.00 Source: Crash Cost Estimates by Maximum Police-Reported Injury Severity within Selected Crash Categories. FHWA-HRT-05-051, October 2005 Source: ADOT Highway Safety Improvement Program Manual, March 2010 Kittelson & Associates, Inc 85 I-10 Phoenix Corridor Safety Study Results: The equivalent PDO performance measure results are provided in Columns 12 and 14 in Table 50 and Table 51. A large positive value indicates greater potential for improvement. The rankings produced using economic cost and comprehensive cost weighting factors are nearly identical. The ranking of the top 10 segments suggests that the focus for safety improvements should focus on the section inclusive of the 7th Ave interchange to the 7th St interchange. EXCESS EXPECTED AVERAGE CRASH FREQUENCY WITH EB ADJUSTMENT Description: This performance measure identifies segments where the expected average crash frequency is higher than the predicted average crash frequency for a segment. The predicted crash frequency is calculated using calibrated SPFs (without the EB adjustment). Thus, the predicted crash frequency represents the long-run crash frequency for typical segment having the same geometry and traffic volume as the subject segment. The expected crash frequency is calculated by weighting the predicted crashes with observed crash data for the subject segment. Thus, the expected crash frequency represents the long-run crash frequency for the subject segment. The difference between predicted and expected crashes is called “Excess Expected Crashes”. It represents the amount by which the safety of the subject segment deviates from that of the typical segment, given the same geometry and traffic volume. Segments are then ranked based on the highest excess expected crashes. Strengths: Accounts for RTM bias; accounts for influence of traffic volume; accounts for influence of roadway geometry and characteristics; provides a single ranking that considers crash frequency and severity. Limitations: Requires SPFs calibrated to local conditions (not a limitation for this study). Application: Excess expected crashes were calculated for the following categories:          Basic segment, single-vehicle crash, PDO Basic segment, single-vehicle crash, FI Basic segment, multiple-vehicle crash, PDO Basic segment, multiple-vehicle crash, FI Basic segment, FI+PDO Entrance speed-change lane, PDO Entrance speed-change lane, FI Exit speed-change lane, PDO Exit speed-change lane, FI For basic segments, “excess expected crash frequency” was calculated on a “per mile” basis to normalize the segments to allow for more meaningful comparison and ranking among segments. Results: The excess average crash frequency performance measure results are provided in Columns 16 thru 18 in Table 50 and Table 51. The top 10 basic segments identified using this performance measure are consistent with those produced by two other EB adjusted performance measures. However, the same results are not observed with the speed-change lanes, where there is greater spread. Kittelson & Associates, Inc 86 I-10 Phoenix Corridor Safety Study MODIFIED LEVEL OF SERVICE OF SAFETY (MLOSS) Description: The Level of Service of Safety (LOSS) has been effectively used to qualitatively describe the safety performance of a roadway and identify segments with greater potential for crash reduction. The methodology ranks segments by comparing their average crash frequency to the predicted average crash frequency for all of the segments being evaluated. The degree of standard deviation from the predicted average crash frequency defines the LOSS levels, I, II, III, and IV. The higher the level, the poorer the LOSS. One weakness in the original methodology is that it does not control for RTM bias, however this is in the process of being remedied by the developers. For the purposes of this study, a modified version of the LOSS method was applied to add an EB adjustment to account for RTM bias. Specifically, the MLOSS is based on the expected average crash frequency whereas the LOSS is based on the observed crash frequency. The LOSS categories are defined as follows: LOSS I II III IV Potential for Crash Reduction Low Low to Moderate Moderate to High High Strengths: Performance measure is normalized by the variance in the crash data allowing for meaningful comparison of segments; accounts for volume; establishes a threshold for measuring crash frequency. Limitations: Requires SPFs calibrated to local conditions (not a limitation for this study); 3-5 years of crash data are recommended. Application: EB-based expected average crash frequency was substituted for the observed crash frequency when computing the LOSS using the criteria provided in the HSM (Table 4-11). LOSS for FI and PDO crashes was computed for basic segments and speed-change lanes. The basis for determining the LOSS was the z-score (z-score = (predicted - reported)/(reported)0.5). This measure is a "relative measure". It "normalizes" the difference between the expected and predicted crash frequencies (where this difference relates the subject segment to the typical segment, in a manner similar to the preceding "excess expected average crash frequency with EB adjustment" measure). The z-score conveys the magnitude of the difference. Results: The LOSS values, listed in Columns 22 thru 29 in Table 50 and Table 51, indicate that the crash reduction potential on nearly all of the basic segments and nearly all of the speed-change lanes within the I-10 study section is either moderate to high (LOSS III) or High (LOSS IV), relative to typical Phoenix freeways. Multi-vehicle PDO crashes (e.g. rear-ends) offer the highest potential for reduction, followed by multi-vehicle FI crashes. While LOSS levels don’t necessarily aide in focusing further investigation on specific segments, they may suggest that there may be a system-wide factor that is contributing to the relatively high crash frequencies. Potential system-wide factors could include high speed, or perhaps the continuous access HOV lanes. These factors will be investigated further in subsequent project tasks. While the LOSS levels themselves do not differentiate the segments within the study section, the zscores do. These values are also provided in Table 50 and Table 51. Segments were ranked based on Kittelson & Associates, Inc 87 I-10 Phoenix Corridor Safety Study summing all z-scores. The top 10 basic segments are very similar to the rankings using the excess expected average crash frequency with EB performance measure. The segments of highest potential for crash reduction are between 7th Ave and SR 51, although several segments east of 35th Ave are also of interest. The top 10 speed-change lanes are also very similar, with the speed-change lanes of highest accident potential located between 7th St and Sky Harbor Blvd. Kittelson & Associates, Inc 88 I-10 Phoenix Corridor Safety Study APPENDIX B – ISATE OUTPUT Kittelson & Associates, Inc 89 Output Summary General Information Project description: I-10 Study Section - 35th Ave to Sky Harbor Blvd 4/3/2015 Urban Analyst: Area type: 0 Date: First year of analysis: 2012 Last year of analysis: 2013 Crash Data Description Segment crash data available? Yes First year of crash data: 2012 Freeway segments Project-level crash data available? No Last year of crash data: 2013 Segment crash data available? No First year of crash data: Ramp segments Project-level crash data available? No Last year of crash data: Segment crash data available? No First year of crash data: Ramp terminals Project-level crash data available? No Last year of crash data: Estimated Crash Statistics Crashes for Entire Facility Total K A B C PDO Estimated number of crashes during Study Period, crashes: 2450.2 16.2 38.9 269.0 360.9 1765.3 Estimated average crash freq. during Study Period, crashes/yr: 1225.1 8.1 19.4 134.5 180.5 882.6 Crashes by Facility Component Nbr. Sites Total K A B C PDO Freeway segments, crashes: 20 2450.2 16.2 38.9 269.0 360.9 1765.3 Ramp segments, crashes: 0 0.0 0.0 0.0 0.0 0.0 0.0 Crossroad ramp terminals, crashes: 0 0.0 0.0 0.0 0.0 0.0 0.0 Crashes for Entire Facility by Year Year Total K A B C PDO Estimated number of crashes during 2012 1225.1 8.1 19.4 134.5 180.5 882.6 the Study Period, crashes: 2013 1225.1 8.1 19.4 134.5 180.5 882.6 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 Distribution of Crashes for Entire Facility Estimated Number of Crashes During the Study Period Crash Type Crash Type Category Total K A B C PDO Multiple vehicle Head-on crashes: 8.1 0.1 0.3 1.9 2.6 3.2 Right-angle crashes: 47.5 0.4 1.1 7.3 9.8 28.9 Rear-end crashes: 1577.8 10.9 26.3 181.9 244.9 1113.8 Sideswipe crashes: 536.0 2.7 6.4 44.1 59.4 423.5 Other multiple-vehicle crashes: 56.9 0.4 1.1 7.3 9.9 38.2 Total multiple-vehicle crashes: 2226.3 14.6 35.0 242.6 326.6 1607.5 Single vehicle Crashes with animal: 3.7 0.0 0.0 0.1 0.1 3.5 Crashes with fixed object: 162.2 1.1 2.8 19.1 24.8 114.4 Crashes with other object: 25.4 0.1 0.2 1.4 1.9 21.7 Crashes with parked vehicle: 2.9 0.0 0.0 0.3 0.4 2.1 Other single-vehicle crashes 29.8 0.3 0.8 5.5 7.1 16.1 Total single-vehicle crashes: 223.9 1.6 3.8 26.4 34.3 157.8 Total crashes: 2450.2 16.2 38.9 269.0 360.9 1765.3 Output Summary General Information Project description: I-10 Study Section - 35th Ave to Sky Harbor Blvd Analyst: Area type: 0 Date: 4/3/2015 Urban First year of analysis: 2012 Last year of analysis: 2013 Crash Data Description Segment crash data available? Yes First year of crash data: 2012 Freeway segments Project-level crash data available? No Last year of crash data: 2013 Segment crash data available? No First year of crash data: Ramp segments Project-level crash data available? No Last year of crash data: Segment crash data available? No First year of crash data: Ramp terminals Project-level crash data available? No Last year of crash data: Estimated Crash Statistics Crashes for Entire Facility Total K A B C PDO Estimated number of crashes during Study Period, crashes: 986.6 6.8 16.4 112.6 153.5 697.3 Estimated average crash freq. during Study Period, crashes/yr: 493.3 3.4 8.2 56.3 76.8 348.7 Crashes by Facility Component Nbr. Sites Total K A B C PDO Freeway segments, crashes: 14 986.6 6.8 16.4 112.6 153.5 697.3 Ramp segments, crashes: 0 0.0 0.0 0.0 0.0 0.0 0.0 Crossroad ramp terminals, crashes: 0 0.0 0.0 0.0 0.0 0.0 0.0 Crashes for Entire Facility by Year Year Total K A B C PDO Estimated number of crashes during 2012 493.3 3.4 8.2 56.3 76.8 348.7 the Study Period, crashes: 2013 493.3 3.4 8.2 56.3 76.8 348.7 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 Distribution of Crashes for Entire Facility Estimated Number of Crashes During the Study Period Crash Type Crash Type Category Total K A B C PDO Multiple vehicle Head-on crashes: 3.1 0.0 0.1 0.8 1.0 1.2 Right-angle crashes: 18.6 0.2 0.4 2.9 4.0 11.1 Rear-end crashes: 613.0 4.4 10.7 73.5 100.7 423.8 Sideswipe crashes: 209.8 1.1 2.6 17.8 24.4 163.9 Other multiple-vehicle crashes: 21.9 0.2 0.4 2.9 4.0 14.4 Total multiple-vehicle crashes: 866.6 5.9 14.2 97.9 134.1 614.4 Single vehicle Crashes with animal: 1.8 0.0 0.0 0.0 0.0 1.7 Crashes with fixed object: 86.1 0.6 1.5 10.5 13.8 59.6 Crashes with other object: 14.1 0.0 0.1 0.8 1.1 12.0 Crashes with parked vehicle: 1.7 0.0 0.0 0.2 0.3 1.2 Other single-vehicle crashes 16.3 0.2 0.5 3.1 4.1 8.4 Total single-vehicle crashes: 120.0 0.9 2.1 14.7 19.4 82.9 Total crashes: 986.6 6.8 16.4 112.6 153.5 697.3 I-10 Phoenix Corridor Safety Study APPENDIX C – SEGMENT PERFORMANCE MEASURES Kittelson & Associates, Inc 90 I-10 Phoenix Corridor Safety Study APPENDIX D – SPEED AND CONGESTION PROFILES Crash Frequency Distribution by Time of Day (Segment: I-10 WB, MP 145.777) Speed Profile: I-10 WB, MP 146.823, 16th St 80 70 Speed (mph) 60 50 40 30 20 10 41338 mph lane1 41338 mph lane2 0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 6:00 6:30 7:00 7:30 8:00 8:30 8:59 9:29 9:59 10:29 10:59 11:29 11:59 12:29 12:59 13:29 13:59 14:29 14:59 15:29 15:59 16:29 16:59 17:29 17:59 18:29 18:59 19:29 19:59 20:29 20:59 21:29 21:59 22:29 22:59 23:29 23:59 41338 mph HOV 0 Time of Day Speed Profile for westbound I-10, MP 146.823 Speed Profile: I-10 WB, MP 145.681, 7th St 80 70 Speed (mph) 60 50 40 30 20 10 0 41338 mph HOV 41338 mph lane1 41338 mph lane2 41338 mph lane3 41338 mph lane4 0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 6:00 6:30 7:00 7:30 8:00 8:30 9:00 9:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30 16:00 16:30 17:00 17:30 18:00 18:30 19:00 19:30 20:00 20:30 21:00 21:30 22:00 22:30 23:00 23:30 0:00 Time of Day Frequency Percent 0:00 0 0.00% 1:00 0 0.00% 2:00 0 0.00% 3:00 0 0.00% 4:00 0 0.00% 5:00 0 0.00% 6:00 0 0.00% 7:00 0 0.00% 8:00 0 0.00% 9:00 0 0.00% 10:00 0 0.00% 11:00 0 0.00% 12:00 0 0.00% 13:00 2 9.09% 14:00 1 4.55% 15:00 3 13.64% 16:00 3 13.64% 17:00 4 18.18% 18:00 6 27.27% 19:00 1 4.55% 20:00 0 0.00% 21:00 0 0.00% 22:00 1 4.55% 23:00 1 4.55% Total 22 100.00% Time of Day Speed Profile for westbound I-10, MP 145.681 Collision Type Distribution (Segment: I-10 WB, MP 145.777) Collision Type REAR_END SIDESWIPE_SAME_DIRECTION SINGLE VEHICLE Total Kittelson & Associates, Inc Frequency 19 1 2 22 Percent 86.36% 4.55% 9.09% 100.00% 91 I-10 Phoenix Corridor Safety Study Drivers’ Maneuvering/Action (Segment: I-10 WB, MP 145.777) Drivers Action Frequency Percent AVOIDING_VEHICLE_OBJECT_PEDESTRIAN 1 1.96% CHANGING_LANES 2 3.92% GOING_STRAIGHT_AHEAD 14 27.45% SLOWING_IN_TRAFFICWAY 19 37.25% STOPPED_IN_TRAFFICWAY 13 25.49% UNKNOWN 2 3.92% Total 51 100.00% Distribution of Slowing/Stopped in the Traffic Way Distribution by Time of Day (Segment: I-10 EB, MP 145.777) Time of Day 0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 Total Frequency 0 0 0 0 0 0 0 0 0 0 0 0 0 4 1 6 3 8 9 0 0 0 0 1 32 Kittelson & Associates, Inc Percent 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 12.50% 3.13% 18.75% 9.38% 25.00% 28.13% 0.00% 0.00% 0.00% 0.00% 3.13% 100.00% 92 I-10 Phoenix Corridor Safety Study APPENDIX E – 2012-2013 CRASH DATA SUMMARY BY SEGMENT Kittelson & Associates, Inc 93 I-10 Phoenix Corridor Safety Study Kittelson & Associates, Inc 94 I-10 Phoenix Corridor Safety Study Kittelson & Associates, Inc 95 I-10 Phoenix Corridor Safety Study APPENDIX F – FATAL AND SERIOUS INJURY CRASH DETAILS Kittelson & Associates, Inc 96 I-10 Phoenix Corridor Safety Study Final Report Fatal Crash Locations: Kittelson & Associates, Inc 97 I-10 Phoenix Corridor Safety Study  Fatal Crash Summaries ID: 2595077 Date: 03/03/2012 (Saturday) Time: 0:12 am Total Injuries: 0 Total Fatalities: 1 Units involved: 3 Conditions: Dark-lighted, Clear, Dry, Level Manner of Impact: Other Travel direction: EB / V2, V3 - WB Lane: HOV Description: None V1: Pedestrian D1: DOB 1984 Action: Walking against traffic Officer Estimated Speed: Unknown Violation: Walked on wrong side of road / Other Note: Drug test results = 1 Reviewer comments: Possible roadway factors: None Possible driver factors: Yes – Inattention, drugs? Possible vehicle factors: None Possible environmental factor: None Congestion related: No V2: Unknown D2: Unknown Action: Unknown Officer Estimated Speed: Unknown Violation: Unknown Note: None V3: 2005 Ford Sedan D2: DOB: 1957 Action: Going straight Officer Estimated Speed: 65 mph Violation: None Note: None Kittelson & Associates, Inc 98 I-10 Phoenix Corridor Safety Study ID: 2599442 Date: 04/12/2012 (Thursday) Time: 5:08 pm Total Injuries: 4 Total Fatalities: 1 Units involved: 4 Conditions: Daylight, Cloudy, Dry, Level Manner of Impact: Rear-End Travel direction: WB Lane: HOV Description: None V1: 2000 Dodge Sedan D1: DOB 1927 Action: Going straight Officer Estimated Speed: 65 mph Violation: Speed too fast Reviewer comments: Possible Roadway factors: None for conditions Possible driver factors: Yes – Inattention, Speed, Age related Note: Condition: Other capability (visual recognition) to drive Possible vehicle factors: None V2: 2008 Infinity Sedan Possible environmental factor: None D2: DOB 1976 Congestion related: Yes Action: Stopped in Traffic Officer Estimated Speed: Unknown Violation: None Note: None V3: 2012 Dodge Caravan D2: DOB 1959 Action: Stopped in Traffic Officer Estimated Speed: Unknown Violation: None Note: None V4: 2008 Toyota Sedan D2: DOB 1975 Action: Going straight Officer Estimated Speed: 50 mph Violation: None Note: None Kittelson & Associates, Inc 99 I-10 Phoenix Corridor Safety Study ID: 2660116 (MP:140.1) Date: 10/24/2012 (Wednesday) Time: 10:30 am Total Injuries: 0 Total Fatalities: 1 Units involved: 4 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Other Travel direction: EB (D1) / WB Lane: 3 / 4 Description: None Reviewer comments: Possible Roadway factors: None Possible driver factors: None Possible vehicle factors: Yes Possible environmental factor: None Congestion related: No V1: 2008 Freightliner Tractor Trailer D1: DOB 1971 Action: Going straight Officer Estimated Speed: 65 mph Violation: None Note: Motor Vehicle: Wheels V2: 1999 Jeep Sedan D2: DOB 1990 Action: Going straight Officer Estimated Speed: 65 mph Violation: None Note: None (Fatal Injury) V3: 2011 International Tractor Trailer D2: DOB 1959 Action: Going straight Officer Estimated Speed: 45 mph Violation: None Note: None V4: 2012 Kia Sedan D2: DOB 1972 Action: Going straight Officer Estimated Speed: 65 mph Violation: None Note: None Kittelson & Associates, Inc 100 I-10 Phoenix Corridor Safety Study ID: 2676135 (MP: 145.6) Date: 11/15/2012 (Thursday) Time: 4:47 am Total Injuries: 0 Total Fatalities: 1 Units involved: 2 Conditions: Dark lighted, Clear, Dry, Level Manner of Impact: Other Travel direction: WB Lane: 3 / 2 Description: None Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes Possible vehicle factors: None Possible environmental factor: None Congestion related: No V1: Pedestrian D1: DOB 1985 Action: Crossing road Officer Estimated Speed: Unknown Violation: Walking on wrong side of road Note: Alcohol V1: 1997 Mitsubishi Truck D1: DOB 1985 Action: Going straight Officer Estimated Speed: 60 mph Violation: None Note: None Kittelson & Associates, Inc 101 I-10 Phoenix Corridor Safety Study ID: 2689544 (MP: 146.65) Date: 12/15/2012 (Saturday) Time: 12:11 pm Total Injuries: 0 Total Fatalities: 1 Units involved: 1 Conditions: Daylight, Rain, Wet, Level Manner of Impact: Single vehicle Travel direction: WB Lane: 3 Description: None Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes (speeding) Possible vehicle factors: None Possible environmental factor: Yes (wet) Congestion related: No V1: 2000 GMC Sedan D1: DOB 1983 Action: Negotiating a curve Officer Estimated Speed: 70 mph Violation: Speed too fast for conditions Note: Drug/alcohol test given ID: 2735405 (MP: 142.5) Date: 06/13/2013 (Friday) Time: 9:29 pm Total Injuries: 0 Total Fatalities: 1 Units involved: 1 Conditions: Dark lighted, Clear, Dry, Uphill Manner of Impact: Single vehicle Travel direction: EB Lane: 4 Description: None Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (drugs, inattention) Possible vehicle factors: None Possible environmental factor: None Congestion related: No V1: 2001 Dodge Pickup D1: DOB 1967 Action: Unknown Officer Estimated Speed: Unknown Violation: Unknown Note: Drug blood test given Kittelson & Associates, Inc 102 I-10 Phoenix Corridor Safety Study Final Report Serious Injury Crash Locations Kittelson & Associates, Inc 103 I-10 Phoenix Corridor Safety Study  Section A  Section B Kittelson & Associates, Inc Final Report 104 I-10 Phoenix Corridor Safety Study  Final Report Section C Kittelson & Associates, Inc 105 I-10 Phoenix Corridor Safety Study  Section D Kittelson & Associates, Inc Final Report  Section E 106 I-10 Phoenix Corridor Safety Study Final Report Serious Injury Crash Summaries ID: 2579668 Date: 01/25/2012 (Wednesday) Time: 8:00 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 2 Conditions: Dark – Lighted, Clear, Dry, Level Manner of Impact: Unknown Travel direction: EB Lane: HOV Description: None V1: Unknown D1: Unknown Action: Unknown Officer Estimated Unknown Violation: Unknown Note: None Speed: V2: 1996 Dodge Pickup D2: DOB 1948 Action: Going Straight Officer Estimated Speed: 40 mph Violation: Unknown Note: None Kittelson & Associates, Inc Reviewer comments: Possible roadway factors: None Possible driver factors: Unknown Possible vehicle factors: None Possible environmental factor: None Congestion related: No 107 I-10 Phoenix Corridor Safety Study Final Report ID: 2583826 (MP: 143.8) Date: 01/19/2012 (Thursday) Time: 4:33 pm Total Injuries: 2 Total Fatalities: 0 Units involved: 2 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-End Travel direction: WB Lane: 2 Description: D1 left the scene. Few days later reported that D1 was distracted by police vehicle. V1: 1993 Hyundai Sedan D1: DOB 1968 Action: Going straight Officer Estimated Speed: Unknown Violation: Inattention / Distraction Note: D1 left the scene. Few days later D1 reported he was distracted by police vehicle to his right, which stopped another vehicle. V2: 2008 Chevrolet Sedan D2: DOB 1980 Action: Stop in traffic Officer Estimated Unknown Violation: None Note: None Kittelson & Associates, Inc Reviewer comments: Possible roadway factors: None Possible driver factors: Yes - distraction Possible vehicle factors: None Possible environmental factor: None Congestion related: Yes Speed: 108 I-10 Phoenix Corridor Safety Study Final Report ID: 2585018 Date: 02/22/2012 (Wednesday) Time: 5:58 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 3 Conditions: Dusk, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: EB Lane: 1 Description: None V1: 2001 Mitsubishi sedan D1: 1994 Action: Going straight Officer Estimated Speed: 5 mph Violation: Speed too fast for conditions Note: None V2: 1994 Chevrolet Pickup D2: DOB 1962 Reviewer comments: Action: Going straight Officer Estimated Speed: 50 Possible Roadway factors: None Possible driver factors: Yes - Inattention mph Violation: Speed too fast for Possible vehicle factors: None Possible environmental factor: None conditions Congestion related: Yes Note: None V3: 2001 Ford VAN D2: DOB 1971 Action: Going straight Officer Estimated Speed: 5 mph Violation: None Note: None Kittelson & Associates, Inc 109 I-10 Phoenix Corridor Safety Study Final Report ID: 2585023 Date: 02/22/2012 (Wednesday) Time: 8:35 am Total Injuries: 2 Total Fatalities: 0 Units involved: 2 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: EB Lane: 1 Description: D1 tried to change lanes possible to avoid collision V1: 2007 Mazda Sedan D1: DOB 1989 Action: Going straight Officer Estimated Speed: 65 mph Violation: Speed too fast for Reviewer comments: conditions Possible Roadway factors: None Note: None Possible driver factors: Yes - Inattention Possible vehicle factors: None V2: 2011 Mercedes Sedan Possible environmental factor: None D2: DOB 1981 Congestion related: Yes Action: Slowing in traffic Officer Estimated Speed: 15 mph Violation: None Note: None Kittelson & Associates, Inc 110 I-10 Phoenix Corridor Safety Study Final Report ID: 2585064 Date: 02/20/2012 (Monday) Time: 3:24 pm Total Injuries: 2 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: EB Lane: 1 Description: None V1: 2010 Chevrolet Sedan D1: 1964 Action: Going straight Officer Estimated Speed: 55 mph Violation: Speed too fast for conditions Note: None V2: 2005 Kia Sedan D2: DOB 1976 Action: Stopped in traffic Officer Estimated Speed: None Violation: None Note: None Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes - Inattention Possible vehicle factors: None Possible environmental factor: None Congestion related: Yes V3: 2004 Chevrolet Pickup D2: DOB 1965 Action: Stopped in traffic Officer Estimated Speed: None Violation: None Note: None Kittelson & Associates, Inc 111 I-10 Phoenix Corridor Safety Study Final Report ID: 2585385 (MP: 145) Date: 02/24/2012 (Friday) Time: 9:30 am Total Injuries: 2 Total Fatalities: 0 Units involved: 2 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: WB Lane: 1 Description: None V1: 2007 Dodge Pickup D1: DOB 1938 Action: Changing Lanes Officer Estimated Speed: 60 mph Violation: Speed too fast for conditions Reviewer comments: Note: None Possible Roadway factors: None Possible driver factors: Yes - Inattention V2: 2012 Toyota Camry Possible vehicle factors: None D2: DOB 1960 Action: Slowing in traffic Possible environmental factor: None (crash location: Tunnel) Officer Estimated Congestion related: Yes Speed: 55 mph Violation: None Note: None Kittelson & Associates, Inc 112 I-10 Phoenix Corridor Safety Study Final Report ID: 2593799 (MP: 144.5) Date: 03/14/2012 (Saturday) Time: 5:11 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 1 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Single vehicle Travel direction: EB Lane: HOV Description: None V1: 2011 Harley Motorcycle D1: DOB 1963 Action: Going straight Officer Estimated Speed: 65 mph Reviewer comments: Violation: Speed too fast for Possible Roadway factors: None conditions Possible driver factors: Yes – Inattention Note: None Possible vehicle factors: None Possible environmental factor: None Congestion related: No Kittelson & Associates, Inc 113 I-10 Phoenix Corridor Safety Study Final Report ID: 2595374 Date: 03/27/2012 (Tuesday) Time: 7:50 am Total Injuries: 1 Total Fatalities: 0 Units involved: 2 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Other Travel direction: EB Lane: HOV / 1 Description: Motorcyclist followed too close. D2 applied brakes due to traffic. D1 lost control V1: 2011 Suzuki Motorcycle D1: DOB 1990 Action: Going straight Officer Estimated Speed: 40 mph Violation: Speed too fast for conditions Note: None Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes-Inattention Possible vehicle factors: None Possible environmental factor: None Congestion related: No V2: 2009 Nissan Sedan D2: DOB 1969 Action: Slowing in Traffic Officer Estimated Speed: 20 mph Violation: None Note: None Kittelson & Associates, Inc 114 I-10 Phoenix Corridor Safety Study Final Report ID: 2595773 (MP:142.8) Date: 03/30/2012 (Friday) Time: 2:53 pm Total Injuries: 2 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Clear, Dry, Level Manner of Impact: RearEnd Travel direction: EB Lane: 6 Description: None V1: 2008 Dodge Pickup D1: DOB 1983 Action: Changing Lanes Officer Estimated Speed: 30 mph Violation: Speed too fast for Reviewer comments: conditions Possible Roadway factors: None Note: None Possible driver factors: Yes-Inattention Possible vehicle factors: None V2: 2002 Dodge Sedan Possible environmental factor: None D2: DOB 1957 Congestion related: Yes Action: Stopped in Traffic Officer Estimated Speed: Unknown Violation: None Note: None V3: 2002 Chrysler Sedan D2: DOB 1993 Action: Stopped in Traffic Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc 115 I-10 Phoenix Corridor Safety Study Final Report ID: 2595789 Date: 03/28/2012 (Wednesday) Time: 4:03 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-End Travel direction: EB Lane: 4 Description: None V1: 2000 GMC Pickup D1: DOB 1993 Action: Going straight Officer Estimated Speed: 40 mph Violation: Speed too fast for conditions Note: None V2: 2004 Dodge Sedan D2: DOB 1985 Action: Stopped in Traffic Officer Estimated Speed: Unknown Violation: None Note: None Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes-Inattention Possible vehicle factors: None Possible environmental factor: None Congestion related: Yes V3: 2004 Acura Sedan D2: DOB 1988 Action: Stopped in Traffic Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc 116 I-10 Phoenix Corridor Safety Study Final Report ID: 2596216 Date: 03/27/2012 (Tuesday) Time: 9:43 pm Total Injuries: 3 Total Fatalities: 0 Units involved: 2 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-End Travel direction: WB Lane: 3 Description: None V1: Hyundai Sedan D1: DOB 1987 Action: Going straight Officer Estimated Speed: 40 mph Violation: Speed too fast for conditions Reviewer comments: Note: None Possible Roadway factors: None Possible driver factors: Yes-Inattention V2: 2008 Ford Sedan Possible vehicle factors: None D2: DOB 1957 Action: Slowing in Traffic Possible environmental factor: None Officer Estimated Speed: Congestion related: Yes 15 mph Violation: None Note: None Kittelson & Associates, Inc 117 I-10 Phoenix Corridor Safety Study ID: 2596917 Date: 03/27/2012 (Tuesday) Time: 4:05 pm Total Injuries: 2 Total Fatalities: 0 Units involved: 1 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Single vehicle Travel direction: WB Lane: 3 Description: None Final Report Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes-Inattention, speed Possible vehicle factors: None Possible environmental factor: None Congestion related: Yes (?) V1: 2010 Kawasaki Motorcycle D1: DOB 1992 Action: Going straight Officer Estimated Speed: 40 mph Violation: Speed too fast for conditions Note: None Kittelson & Associates, Inc 118 I-10 Phoenix Corridor Safety Study Final Report ID: 2622201 Date: 06/21/2012 (Thursday) Time: 8:24 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Cloudy, Dry, Level Manner of Impact: Rear-End Travel direction: EB Lane: 1 Description: None V1: 2008 Nissan Sedan D1: DOB 1956 Action: Going straight Officer Estimated Speed: 55 mph Violation: Speed too fast for conditions Reviewer comments: Note: Condition: Other Possible Roadway factors: None Possible driver factors: Yes – Inattention, speed V2: 2002 Honda Sedan Possible vehicle factors: None D2: DOB 1983 Possible environmental factor: None Action: Stopped in Traffic Congestion related: Yes Officer Estimated Speed: Unknown Violation: None Note: None V3: 2012 Toyota Sedan D2: DOB 1990 Action: Stopped in Traffic Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc 119 I-10 Phoenix Corridor Safety Study Final Report ID: 2624458 (MP: 145) Date: 07/09/2012 (Monday) Time: 4:23 pm Total Injuries: 3 Total Fatalities: 0 Units involved: 5 Conditions: Daylight, Cloudy, Dry, Level Manner of Impact: Rear-End Travel direction: EB Lane: 4 / 3 Description: None V1: 2001 Buick Sedan D1: DOB 1957 Action: Going straight Officer Estimated Speed: 65 mph Violation: Speed too fast for Reviewer comments: conditions Note: Condition: Other Possible Roadway factors: None Possible driver factors: Yes – Inattention, speed V2: 1993 Honda Sedan Possible vehicle factors: None D2: DOB 1961 Possible environmental factor: None Action: Going straight Congestion related: Yes Officer Estimated Speed: 65 mph Violation: Speed too fast for conditions Note: None V3: 2010 Chrysler Sedan D2: DOB 1972 Action: Going straight Officer Estimated Speed: 65 mph Violation: Speed too fast for conditions Note: None V4: 2001 Jeep Sedan D2: DOB 1987 Action: Stopped in traffic Officer Estimated Unknown Violation: None Note: None Speed: V5: 2007 Toyota Sedan D2: DOB 1984 Action: Stopped in traffic Officer Estimated Unknown Violation: None Note: None Speed: Kittelson & Associates, Inc 120 I-10 Phoenix Corridor Safety Study Final Report ID: 2638496 Date: 08/10/2012 (Friday) Time: 11:40 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 1 Conditions: Dark - lighted, Clear, Dry, Level Manner of Impact: Single vehicle Travel direction: WB Lane: 3 Description: Roadway curve left. V1: 2012 Yamaha motorcycle D1: DOB 1976 Action: Avoiding vehicle/object/ped/cyclist Officer Estimated Speed: 75 mph Violation: Speed too fast for Reviewer comments: conditions Possible Roadway factors: Yes (Curve) Note: None Possible driver factors: Yes – Inattention, speed Possible vehicle factors: None Possible environmental factor: None (report: moving vehicle) Congestion related: No Kittelson & Associates, Inc 121 I-10 Phoenix Corridor Safety Study ID: 2642347 Date: 07/19/2012 (Thursday) Time: 6:57 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 1 Conditions: Dusk, Cloudy, Dry, Downhill Manner of Impact: Single vehicle Travel direction: EB Lane: Non-roadway Description: None V1: 2003 Suzuki Motorcycle D1: DOB 1979 Action: Negotiating a curve Officer Estimated Speed: 45 mph Violation: Speed too fast for conditions Note: None Kittelson & Associates, Inc Final Report Reviewer comments: Possible Roadway factors: None (alignment – curve right) Possible driver factors: Yes – Inattention, speed Possible vehicle factors: None Possible environmental factor: None Congestion related: No 122 I-10 Phoenix Corridor Safety Study Final Report ID: 2651125 Date: 09/25/2012 (Tuesday) Time: 7:16 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 2 Conditions: Dark-lighted, Clear, Dry, Uphill Manner of Impact: Rear-End Travel direction: WB Lane: HOV Description: D2 applied brakes to stop. D1 tried to stop but did not have enough time. V1: 2006 Chevrolet 3500 D1: DOB 1988 Action: Going straight Officer Estimated Speed: 40 mph Violation: Speed too fast for conditions Note: None V2: 2012 Chevrolet Sedan D2: DOB 1974 Action: Stopped in traffic Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes – Inattention, speed Possible vehicle factors: None Possible environmental factor: None Congestion related: Yes 123 I-10 Phoenix Corridor Safety Study Final Report ID: 264810 Date: 09/17/2012 (Monday) Time: 8:09 am Total Injuries: 3 Total Fatalities: 0 Units involved: 4 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Angle Travel direction: EB Lane: 1 / HOV / 1 Description: None V1: 2003 Toyota Sedan D1: DOB 1963 Action: Slowing in traffic Officer Estimated Speed: 40 mph Reviewer comments: Violation: Speed too fast for Possible Roadway factors: None conditions Possible driver factors: Yes – Inattention, speed Note: None Possible vehicle factors: None Possible environmental factor: None V2: 2001 Dodge Van Congestion related: Yes D2: DOB 1984 Action: Going straight Officer Estimated Speed: 55 mph Violation: None Note: None V3: 2005 Honda Sedan D2: DOB 1994 Action: Stopped in traffic Officer Estimated Speed: 40 mph Violation: None Note: None V4: 2012 Ford Sedan D2: DOB 1948 Action: Going straight Officer Estimated Speed: 55 mph Violation: None Note: None Kittelson & Associates, Inc 124 I-10 Phoenix Corridor Safety Study Final Report ID: 2661954 (MP: 144.1) Date: 10/29/2012 (Monday) Time: 6:49 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 2 Conditions: Dark Lighted, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: WB Lane: HOV / 1 Description: Changed lanes from HOV into 1. Traffic stopped in lane 1. D1 did not have enough time to stop. V1: 2006 Yamaha Motorcycle D1: DOB 1993 Action: Changing lanes Officer Estimated Speed: Unknown Violation: Speed too fast for conditions Note: None Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes (inattention) Possible vehicle factors: None Possible environmental factor: None Congestion related: Yes V2: 2010 Lexus Sedan D2: DOB 1959 Action: Slowing in traffic Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc 125 I-10 Phoenix Corridor Safety Study Final Report ID: 2662051 (MP: 146.7) Date: 10/31/2012 (Monday) Time: 4:54 pm Total Injuries: 3 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: EB Lane: 4 Description: Traffic abruptly stopped in front of D1. D1 applied brake as well as gas. V1: 2002 Mitsubishi sedan D1: DOB 1976 Action: Slowing in traffic Officer Estimated Speed: 35 mph Violation: Other Reviewer comments: Note: None Possible Roadway factors: None Possible driver factors: Yes V2: 2007 Ford Sedan Possible vehicle factors: None D2: DOB 1962 Possible environmental factor: None Action: Stopped in Congestion related: Yes traffic Officer Estimated Speed: Unknown Violation: None Note: None V3: 2006 Volkswagen Sedan D2: DOB 1973 Action: Stopped in traffic Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc 126 I-10 Phoenix Corridor Safety Study Final Report ID: 2665404 Date: 11/15/2012 (Thursday) Time: 5:40 pm Total Injuries: 2 Total Fatalities: 0 Units involved: 3 Conditions: Dusk, Cloudy, Dry, Level Manner of Impact: Sideswipe, same direction Travel direction: EB Lane: HOV Description: Vehicle in front of D1 in the HOV swerved into the shoulder. D1 observed other vehicle stopped in HOV lane. D1 tried to avoid impact, sideswiped D2 and hit V3. V1: 2009 Harley Motorcycle D1: DOB 1961 Action: Going straight Officer Estimated Speed: Unknown Violation: Speed too fast for condition, inattention/distraction Note: None Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes Possible vehicle factors: None Possible environmental factor: None Congestion related: Yes V2: 2009 Harley Motorcycle D2: DOB 1957 Action: Slowing in traffic Officer Estimated Speed: Unknown Violation: None Note: None V3: 1993 Chevrolet Sedan D2: DOB 1962 Action: Stopped in traffic Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc 127 I-10 Phoenix Corridor Safety Study Final Report ID: 2665932 (MP: 143.38) Date: 11/11/2012 (Sunday) Time: 2:55 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 1 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Single vehicle Travel direction: WB Lane: 3 Description: None V1: 1996 Acura Sedan D1: DOB 1992 Action: Going straight Officer Estimated Speed: 65 mph Violation: None Note: None Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes Possible vehicle factors: None Possible environmental factor: None Congestion related: No Kittelson & Associates, Inc 128 I-10 Phoenix Corridor Safety Study Final Report ID: 2671728 Date: 11/23/2012 (Friday) Time: 4:15 am Total Injuries: 1 Total Fatalities: 0 Units involved: 1 Conditions: Dark lighted, Clear, Dry, Level Manner of Impact: Single vehicle Travel direction: WB Lane: HOV Description: None V1: 2006 Pontiac Sedan D1: DOB 1992 Action: Going straight Officer Estimated Speed: 85 mph Violation: Speed too fast for Reviewer comments: condition Possible Roadway factors: None Note: Fell asleep/fatigued Possible driver factors: Yes Possible vehicle factors: None Possible environmental factor: None Congestion related: No Kittelson & Associates, Inc 129 I-10 Phoenix Corridor Safety Study Final Report ID: 2689512 (MP:143.87) Date: 01/19/2013 (Saturday) Time: 9:35 am Total Injuries: 3 Total Fatalities: 0 Units involved: 5 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: WB Lane: 4 / Unknown / 1 Description: D3 (in lane 5) observed D1 passing by at higher rate of speed (in lane 4). D3 observed D1 brake hard and swerved toward lane 5 in front of D3 and collided with D2. V1: 2005 Toyota Sedan D1: DOB 1973 Action: Going straight Officer Estimated Speed: 50 mph Violation: Speed too fast for conditions Note: D1 does not remember the collision V2: Unknown D1: Unknown Action: Slowing in traffic Officer Estimated Unknown Violation: Unknown Note: fled the scene Reviewer comments: Possible Roadway factors: None Possible driver factors: Yes Possible vehicle factors: None Possible environmental factor: None Congestion related: No/Yes Speed: V3: 2011 Chevrolet Pickup D1: DOB 1960 Action: Slowing in traffic Officer Estimated Speed: 50 mph Violation: None Note: None V4: 2012 Toyota Pickup D1: DOB 1958 Action: Going straight Officer Estimated Speed: 40 mph Violation: None Note: None Kittelson & Associates, Inc 130 I-10 Phoenix Corridor Safety Study Final Report V5: 2003 Chevrolet Pickup D1: DOB: 1983 Action: Going straight Officer Estimated Speed: 65 mph Violation: None Note: None Kittelson & Associates, Inc 131 I-10 Phoenix Corridor Safety Study Final Report ID: 2696715 Date: 02/23/2013 (Saturday) Time: 03:30 am Total Injuries: 2 Total Fatalities: 0 Units involved: 2 Conditions: Dark lighted, Clear, Dry, Level Manner of Impact: Sideswipe same direction Travel direction: EB Lane: 1 / 2 Description: D1 turn right and sideswipe V2. V1: 2003 Chevrolet Pickup D1: DOB 1990 Action: Changing lanes Officer Estimated Speed: 60 Reviewer comments: mph Possible Roadway factors: None Violation: Unsafe passing Possible driver factors: Yes (inattention) Note: None Possible vehicle factors: None Possible environmental factor: None V2: 2010 Kia Sedan Congestion related: No D1: DOB 1956 Action: Going Straight Officer Estimated Speed: 60 mph Violation: None Note: None Kittelson & Associates, Inc 132 I-10 Phoenix Corridor Safety Study Final Report ID: 2706271 Date: 03/21/2013 (Thursday) Time: 03:29 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: WB Lane: 3 / 4 / 3 Description: Changing Road Width. Warning Sign present. D1 changed lanes from 1 into 2 and then into 3. When changing lanes into lane 4,traffic stopped abruptly in front of D1. D1 attempt to brake but struck V2 at 55 mph. Before the impact, D2 applied brakes to slow down due to traffic in front. Heavy traffic. Lane Reviewer comments: drop. Possible Roadway factors: Yes (lane drop) Possible driver factors: Yes (inattention) V1: 1997 Chevrolet Pickup Possible vehicle factors: None D1: DOB 1981 Possible environmental factor: None Action: Changing lanes Congestion related: Yes Officer Estimated Speed: 55 mph Violation: Speed too fast for conditions Note: None V2: 2007 Hyundai SUV D1: DOB 1968 Action: Slowing in traffic Officer Estimated Speed: 10 mph Violation: None Note: None V2: 2003 Hyundai Sedan D1: DOB 1977 Action: Slowing in traffic Officer Estimated Speed: 5 mph Violation: None Note: None Kittelson & Associates, Inc 133 I-10 Phoenix Corridor Safety Study Final Report ID: 2715241 (MP: 143) Date: 04/19/2013 (Friday) Time: 03:13 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 2 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Sideswipe same direction Travel direction: EB Lane: HOV Description: D2 stated that he was in HOV. D1 changed from lane 1 into HOV (to avoid impacting another vehicle in lane 1) and collided with V2. V1: 2003 Volkswagen Sedan D1: DOB 1993 Action: Changing lanes Officer Estimated Speed: 50 mph Violation: Speed too fast for conditions Note: None Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (inattention) Possible vehicle factors: None Possible environmental factor: None Congestion related: Yes V2: 2013 Toyota Sedan D1: DOB 1958 Action: Going straight Officer Estimated Speed: 60 mph Violation: None Note: None Kittelson & Associates, Inc 134 I-10 Phoenix Corridor Safety Study Final Report ID: 2715800 (MP: 147.8) Date: 04/25/2013 (Thursday) Time: 6:33 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 2 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: WB Lane: 3 Description: Traffic in front of D1 slowed down. D1 applied brakes but could not stop. V1: 2012 Chrysler Sedan D1: DOB 1956 Action: Going straight Officer Estimated Speed: 50 mph Violation: Reviewer comments: Inattention/Distraction Possible Roadway factors: No Note: None Possible driver factors: Yes (inattention) Possible vehicle factors: None V2: 2009 Honda Sedan Possible environmental factor: None D1: DOB 1975 Congestion related: Yes Action: Stopped in traffic Officer Estimated Speed: 30 mph Violation: None Note: None Kittelson & Associates, Inc 135 I-10 Phoenix Corridor Safety Study Final Report ID: 2719776 (MP: 147.15) Date: 04/19/2013 (Friday) Time: 2:58 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 1 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Single vehicle Travel direction: WB Lane: HOV Description: None V1: 2006 Yamaha Motorcycle D1: DOB 1954 Action: Going straight Officer Estimated Speed: 45 mph Violation: Speed too fast for conditions Note: None Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (inattention, speed) Possible vehicle factors: None Possible environmental factor: None Congestion related: No Kittelson & Associates, Inc 136 I-10 Phoenix Corridor Safety Study Final Report ID: 2719871 Date: 03/29/2013 (Friday) Time: 10:35 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 1 Conditions: Dark lighted, Clear, Dry, Level Manner of Impact: Single vehicle Travel direction: WB Lane: 5 Description: D1 lost control, veered left and hit concrete wall. V1: 2002 Chevrolet Sedan D1: DOB 1988 Action: Going straight Officer Estimated Speed: 65 mph Violation: Unknown Note: Alcohol Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (inattention, speed, alcohol) Possible vehicle factors: None Possible environmental factor: None Congestion related: No Kittelson & Associates, Inc 137 I-10 Phoenix Corridor Safety Study Final Report ID: 2738892 Date: 07/18/2013 (Thursday) Time: 3:17 pm Total Injuries: 3 Total Fatalities: 0 Units involved: 2 Conditions: Daylight, Cloudy, Dry, Level Manner of Impact: Rear-end Travel direction: WB Lane: 1 Description: Environmental factor: stopped/parked vehicle V1: 1996 Ford Sedan D1: DOB 1992 Action: Slowing in traffic Officer Estimated Speed: 50 mph Violation: Speed too fast for conditions Note: None Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (inattention) Possible vehicle factors: None Possible environmental factor: Yes (stopped/parked vehicle) Congestion related: Yes V2: 2006 Chevrolet Pickup D1: DOB 1971 Action: Slowing in traffic Officer Estimated Speed: 20 mph Violation: None Note: None Kittelson & Associates, Inc 138 I-10 Phoenix Corridor Safety Study ID: 2744192 Date: 07/22/2013 (Monday) Time: 7:20 am Total Injuries: 1 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Clear, Uphill Manner of Impact: Rear-end Travel direction: EB Lane: 1 Description: None Final Report Dry, V1: 1995 Ford Sedan D1: DOB 1990 Action: Going straight Officer Estimated Speed: 50 mph Violation: Speed too fast for conditions Reviewer comments: Note: None Possible Roadway factors: No Possible driver factors: Yes (inattention) V2: 2003 Chevrolet Sedan Possible vehicle factors: None D1: DOB 1956 Possible environmental factor: None Action: Slowing in traffic Congestion related: Yes Officer Estimated Speed: 10 mph Violation: None Note: None V3: 2010 Hyundai Sedan D1: DOB 1953 Action: Slowing in traffic Officer Estimated Speed: 10 mph Violation: None Note: None Kittelson & Associates, Inc 139 I-10 Phoenix Corridor Safety Study Final Report ID: 2752349 (MP: 144.88) Date: 08/16/2013 (Friday) Time: 5:20 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: EB Lane: 3 Description: Traffic stopped suddenly in front of D1. D1 tried to brake but could not stop. V1: 1998 Ford Pickup D1: DOB 1981 Action: Going straight Officer Estimated Speed: 65 mph Violation: Speed too fast Reviewer comments: for conditions Possible Roadway factors: No Note: None Possible driver factors: Yes (inattention, speeding) Possible vehicle factors: None V2: 2000 Ford Sedan Possible environmental factor: None D1: DOB 1990 Congestion related: Yes Action: Stopped in traffic Officer Estimated Speed: Unknown Violation: None Note: None V3: 2002 Saturn Sedan D1: DOB 1976 Action: Stopped in traffic Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc 140 I-10 Phoenix Corridor Safety Study Final Report ID: 2752470 Date: 09/03/2013 (Tuesday) Time: 6:39 am Total Injuries: 3 Total Fatalities: 0 Units involved: 1 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Single vehicle Travel direction: EB Lane: 5 Description: D1 tried to avoid piece of tire/debris in the roadway. Lost control. V1: 1998 Ford SUV D1: DOB 1966 Action: Avoiding vehicle/object/ped/cyclist Officer Estimated Speed: 50 mph Violation: Speed too fast for conditions Note: None Kittelson & Associates, Inc Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (inattention) Possible vehicle factors: None Possible environmental factor: Yes (debris) Congestion related: No 141 I-10 Phoenix Corridor Safety Study Final Report ID: 2757122 Date: 08/29/2013 (Thursday) Time: 8:51 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 1 Conditions: Dark lighted, Clear, Dry, Level Manner of Impact: Single vehicle Travel direction: WB Lane: HOV Description: None V1: 2005 Chrysler Van D1: DOB 1949 Action: Unknown Officer Estimated Speed: 50 mph Violation: Speed too fast for conditions Reviewer comments: Note: Fell Asleep/Fatigue Possible Roadway factors: No Possible driver factors: Yes (inattention, fatigue) Possible vehicle factors: None Possible environmental factor: No Congestion related: No Kittelson & Associates, Inc 142 I-10 Phoenix Corridor Safety Study Final Report ID: 2762880 Date: 09/26/2013 (Thursday) Time: 4:11 pm Total Injuries: 3 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Angle Travel direction: EB Lane: 3 / HOV /1 Description: None V1: 1995 BMW Sedan D1: DOB 1990 Action: Avoiding vehicle/object/ped/cyclist Officer Estimated Speed: 50 mph Violation: Speed too fast for conditions Note: None Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (inattention, speed) Possible vehicle factors: None Possible environmental factor: No Congestion related: No/Yes V1: 2007 Chevrolet Pickup D1: DOB 1984 Action: Going straight Officer Estimated Speed: 60 mph Violation: None Note: Note V1: 2007 Toyota SUV D1: DOB 1984 Action: Going straight Officer Estimated Speed: 65 mph Violation: None Note: None Kittelson & Associates, Inc 143 I-10 Phoenix Corridor Safety Study Final Report ID: 2768280 Date: 10/16/2013 (Wednesday) Time: 7:02 am Total Injuries: 2 Total Fatalities: 0 Units involved: 3 Conditions: Daylight, Clear, Dry, Level Manner of Impact: Angle Travel direction: EB Lane: 5 /2 / 2 Description: None V1: 2002 Honda Sedan D1: DOB 1991 Action: Avoiding vehicle/object/ped/cyclist Officer Estimated Speed: 65 mph Violation: Speed too fast for conditions Reviewer comments: Note: None Possible Roadway factors: No Possible driver factors: Yes (inattention, speed) V1: 2005 Hyundai Sedan Possible vehicle factors: None D1: DOB 1968 Possible environmental factor: No Action: Going straight Congestion related: No/Yes Officer Estimated Speed: 65 mph Violation: None Note: Note V1: 2005 Hyundai Sedan D1: DOB 1971 Action: Going straight Officer Estimated Speed: 65 mph Violation: None Note: None Kittelson & Associates, Inc 144 I-10 Phoenix Corridor Safety Study Final Report ID: 2778115 Date: 11/18/2013 (Monday) Time: 7:37 pm Total Injuries: 1 Total Fatalities: 0 Units involved: 3 Conditions: Dark-Lighted, Clear, Dry, Uphill Manner of Impact: Rear-end Travel direction: WB Lane: 4 Description: D2 stated that traffic suddenly stopped. Applied brakes and was hit from behind by V1. V1: 2006 Yamaha Motorcycle D1: DOB 1981 Action: Going straight ahead Officer Estimated Speed: 65 mph Violation: Speed too fast for conditions Note: None V2: 2010 Kia Sedan D1: DOB 1940 Action: Slowing in traffic Officer Estimated Speed: 25 mph Violation: None Note: Note Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (inattention, speed) Possible vehicle factors: None Possible environmental factor: No Congestion related: Yes V3: 2002 Nissan Pickup D1: DOB 1974 Action: Stopped in traffic Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc 145 I-10 Phoenix Corridor Safety Study ID: 2782249 (MP: 145.27) Date: 11/06/2013 (Wednesday) Time: 10:54 am Total Injuries: 2 Total Fatalities: 0 Units involved: 4 Conditions: Daytime, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: EB Lane: Unknown Description: None Final Report Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (inattention, speed, illness) Possible vehicle factors: None Possible environmental factor: Unknown Congestion related: No V1: 2013 Dodge Pickup D1: DOB 1948 Action: Other Officer Estimated Speed: Unknown Violation: Speed too fast for conditions Note: Illness V2: 2006 Pontiac Sedan D1: DOB Unknown Action: Properly Parked Officer Estimated Speed: Unknown Violation: None Note: Note V3: 2009 Mazda Sedan D1: DOB Unknown Action: Properly Parked Officer Estimated Speed: Unknown Violation: None Note: None V4: Pedestrian D1: DOB 1985 Action: Standing Officer Estimated Speed: Unknown Violation: None Note: None Kittelson & Associates, Inc 146 I-10 Phoenix Corridor Safety Study ID: 2788793 (MP: 145.24) Date: 11/28/2013 (Thursday) Time: 05:20 am Total Injuries: 3 Total Fatalities: 0 Units involved: 2 Conditions: Dark lighted, Clear, Dry, Level Manner of Impact: Rear-end Travel direction: EB Lane: 3 / 5 Description: None Final Report Reviewer comments: Possible Roadway factors: No Possible driver factors: Yes (inattention, speed) Possible vehicle factors: None Possible environmental factor: None Congestion related: No V1: 2003 Chevrolet Sedan D1: DOB 1980 Action: Going Straight Officer Estimated Speed: 65 mph Violation: Speed too fast for conditions, failed to stay in proper lane Note: None V2: 2006 Isuzu Truck D1: DOB 1962 Action: Going straight Officer Estimated Speed: 65 mph Violation: None Note: Note Kittelson & Associates, Inc 147 I-10 Phoenix Corridor Safety Study Final Report APPENDIX G – VARIABLE PPEED LIMIT SYSTEM LITERATURE REVIEW INTRODUCTION Variable speed limit (VSL) systems have been widely used to improve safety and operational efficiency around the world. VSL is also known as Dynamic Speed Limit or Dynamic speed display. The principle behind VSL systems is to post a speed limit that is appropriate for current conditions considering time dependent freeway traffic demand, speed profile and/or special conditions like adverse weather and incidents. This provides an opportunity to warn drivers of downstream conditions and decrease headways and encourage more uniform flow. VSL has the capability of increasing safety by reducing both primary and secondary crashes. It also reduces travel time, congestion and emission, and to increase mobility. VSL systems have been one of the most heavily researched ATM techniques, and a number of deployments have occurred, especially at United Kingdom, Germany, Netherland, France and Denmark in the European continent. The purposes of the practice were congestion management, incident management, and or weather condition. The practice could be mandatory or advisory. The following benefits have been reported in different published research documents from their evaluation studies: • Increase average congested period throughput – 3% to 7% • Increase overall capacity – 3% to 22% • Decrease primary incidents – 3% to 30% • Decrease secondary incidents – 40% to 50% Recently several state DOTs in the United States including Florida DOT, Caltrans, Washington DOT, Minnesota DOT, Virginia DOT, Missouri DOT, Utah DOT, Colorado DOT, and also in Australia have deployed the concept of VSL in their practice. The purpose of this document is to summarize the practices of VSL around the world, the challenges they have faced and the benefits they have gained in terms of safety and operational capacity improvement. Overseas Experience: Practices and Benefits Several countries in Europe including Germany, United Kingdom, Netherland, and Denmark, and Australia have deployed VSL systems on their highway network to improve safety and mobility. Germany’s Experience In Germany, VSLs have been used since the 1970s. It is estimated that VSL systems are installed on more than 800 km (497 mi) of road in Germany. The German VSL systems use gantries placed over the road to display the VSLs, lane control messages, and pictographs representing congestion, when Kittelson & Associates, Inc 148 I-10 Phoenix Corridor Safety Study Final Report present. Spacing of overhead gantries varies depending on the roadway. Autobahn A5 uses a gantry spacing of 1 km (0.62 mi), and congestion pictographs are provided on either side of the structure for queue warning. On an 18-km (11.2 mi) stretch of Autobahn A9 near Munich, overhead gantries were placed at an average spacing of 1.8 km (1.12 mi). Inductive loop detectors spaced between 340 and 1750 m (1115 to 5741 ft) on this road were also used to provide detection. Effects on Traffic Flow and Safety have been found in Germany. Several studies evaluated VSL operational and safety impacts on the German Autobahn. On the A5 Autobahn, crash rates fell by 20% after VSL systems were installed and increased by 10% at a comparable site with no VSL system. There was also a 67 percent decline in secondary crashes. Secondary crashes are generally defined as crashes that occur as a result of congestion caused by an initial primary crash, although the researchers did not specify any time or distance thresholds for identifying secondary crashes in this case. Reduced travel times, decreased fuel consumption, and lower emissions were also cited as benefits of the system. The A5 Autobahn gained several other significant safety improvements after VSLs were installed. A 3% reduction in property damage only (PDO) crashes with light damage and a 27% reduction in PDO crashes with heavy damage occurred. A 30% reduction in injury crashes also occurred. For the A9 Autobahn, researchers found that the VSL system responded well to traffic but congestion and shockwaves were still present. One set of researchers used available detector data to examine the flow-speed-density relationships on the German Autobahn when VSLs were in use. They found that VSLs decreased the slope of the flowoccupancy diagram at undercritical conditions, shifted occupancy to higher values, and enabled higher flows at the same occupancy in overcritical conditions. The speed-flow diagram showed that a 50 mph VSL clearly had a higher critical flow rate than when no VSL was posted, indicating that heavy flow could be sustained for a longer period before breakdown occurred. Although there was significant stochastic variation in flow and speed, the critical occupancy was about 5% higher with the VSLs active than when they were not. U.K’s Experience In the United Kingdom, VSL systems have been installed on the M25 and M42 motorways. The M25 systems were installed in 1995. The M25 is a freeway with four lanes in each direction, and VSLs were placed on overhead gantries spaced at 1-km (0.62 mi) intervals. Inductive loops were placed at 500-m (1,640 ft) spacings to monitor traffic and provide data used by the VSL system to determine the appropriate speed limits. The other U.K. VSL system is on a 17-km stretch of the M42. This road has an Kittelson & Associates, Inc 149 I-10 Phoenix Corridor Safety Study Final Report average annual daily traffic (AADT) of 120,000 vehicles, and a total of 50 gantries holding 250 signs were installed. Gantries were spaced every 0.5 to 1 km (0.31 to 0.62 mi). A number of studies of the operational and safety effects of VSLs have also been performed in the United Kingdom. A 2-year study on the M25 found that the VSL system produced more even headways. Results from the first year of operation showed a 28% reduction in injuries and a 25% reduction in PDO crashes. A 25% to 30% reduction in rear-end crashes was also observed. Data from the second year of operation showed that these results had been maintained. It was also estimated that the system increased capacity by 5% to 10%. A subsequent study in 2005 also examined the M25 which reported that the VSLs produced the following impacts: • Neutral impact on travel time and travel time reliability • 15% reduction in injury crashes • Estimated 2% to 8% reduction in emissions • Estimated fuel consumption reduction of 10% • 1.5% increase in throughput • 5% improvement in speed limit compliance. Another analysis conducted using 7 years of data after the M25 VSL deployment began showed a 10% to 20% reduction in injuries. A subsequent expansion of M25 VSLs by 8 km (4.97 mi) found that travel times did not change significantly but injuries fell by 10% on the new section. Netherland’s Experience In the Netherlands, VSLs have been used since 1981. Overhead VSLs and lane control signals are deployed every 500 m (1640 ft). Studies of VSL systems in the Netherlands showed that throughput increased between 3% and 5%. Collisions were also reduced by about 16%. A study at 4 test locations in the Netherlands found a 20% to 30% reduction in NOx and a 10% reduction in particulate matter below 10 microns (PM10) when VSLs were implemented. Denmark’s Experience Several other results were reported from deployments in other European countries. A work zone VSL system was installed on M3 around Copenhagen, Denmark. Incidents did not increase during construction despite reduced lane widths at that site. Australia’s Experience Kittelson & Associates, Inc 150 I-10 Phoenix Corridor Safety Study Final Report In Australia, a VSL system was developed for the Western Ring Road, which has an AADT of approximately 100,000 vehicles, with 15% trucks. The system is implemented on a 26-km (16.16 mi) section of road that has a base speed limit of a 100 km/h (62.14 mph). Loop detectors were placed using an 0.5-km (0.31 mi) spacing. VSL system has also been deployed on M4, F3, and the Adelaide-Crafers Highway. The project objective were reduction of rear-end crashes, incident, queue management, and safe traffic operation in adverse weather conditions. The deployment has been able to reduce crash rates by 11-24% on F3 in the first month of installation. U.S. Experience: Practices and Benefits There have been several recent successful VSL deployments in the United States in the state of Florida, Washington, Minnesota, Missouri, Virginia, Maine, Colorado and Utah. Florida’s Experience A VSL system was deployed on a 10-mile section of I-4 in Orlando, Florida, in 2008. The section had an AADT of approximately 200,000 vehicles. A total of 20 VSL signs were installed at 16 locations, and inductive loops were used to measure speed, volume, and occupancy at 30-sec intervals. The Orlando VSL system was evaluated by looking at speed data from 4 P.M.to 6 P.M. for 1 month before VSL activation as compared to 1 month after VSL activation. The data showed that speed changes were more strongly correlated with changes in occupancy than changes in the posted speed limit. The evaluators concluded that the VSL had no significant impact on speed compliance or mean travel speed. A crash analysis was also conducted, but no conclusions could be drawn because of limited data. Washington’s Experience In August 2010, Washington DOT installed VSL systems on 7 miles of I-5 northbound as it approaches downtown Seattle. Similar systems were installed on 8 miles of S.R. 520 eastbound and westbound in November 2010 and on I-90 eastbound and westbound in June 2011. In Washington, evaluation results were more limited. There was a 6 month time lag between when a crash actually occurred and when it was entered into the DOT crash database, so WSDOT was unable to make definitive assessments of the safety impact of the system as of early 2012. Preliminary analysis examined the ATM segment of I-5, a segment immediately downstream, and 3 other urban segments further removed from the ATM segment. The preliminary 2011 trends showed that collisions at the ATM segment and the segment immediately downstream declined, whereas crashes at the other 3 segments increased. These are preliminary data, however, and no firm conclusions can be drawn. The WSDOT deployment did identify some safety-related benefits in terms of work zone and incident management since speed limits and lane control signs could be used to supplement traditional traffic control. Kittelson & Associates, Inc 151 I-10 Phoenix Corridor Safety Study Final Report Minnesota’s Experience The Minnesota DOT is also operating VSL systems, lane control signs on a 10-mile segment of I35W in the Minneapolis–St. Paul area. Signs are spaced 0.5 mile apart. An extension of the system is planned on an 8-mile section of I-94 between downtown St. Paul and downtown Minneapolis in summer 2012. Minnesota conducted a preliminary evaluation of the safety and operational effects of their system. Measures were compared for 3 months after the VSL system was activated to the same 3 months during the year before the VSL system was installed. Analysis of the detector data showed that the average maximum deceleration declined by 19.6% with the VSL, indicating smoother transitions between flow regimes. Travel times did increase by 13.3% with the VSL, however, because of posting slower speeds while transitioning from free flow to congested flow. It was also estimated that throughput increased by 6.1% at a known bottleneck because of reduced shock wave impacts. Crashes were not evaluated in this study. Missouri’s Experience The Missouri DOT installed 65 VSL signs along 38 miles of I-270 and I-255 in St. Louis. Data from the St. Louis deployment were evaluated using 150 days of data before and after the deployment. Conditions on typical weekdays were examined using 3 point sensors, and the speed-occupancy-flow relationships were examined before and after the VSLs were activated. The results indicated that the speed-flow-occupancy curve changed after the VSLs were activated, although direction of the change was not consistent at the three sites. Capacity increased at one site, declined at another, and remained the same at the third. The same trends were observed in mean speed. Speed variance did decline at all sites, however. Virginia’s Experience As of early 2012, VDOT had installed VSL systems on several bridge and tunnel facilities. These VSLs are used to reduce speeds primarily because of incidents and weather conditions and are reduced manually by operators. Two VSL systems are currently in development to mitigate safety issues related to foggy conditions on I-64 at Afton Mountain and I-77 at Fancy Gap, but they have not yet been deployed. Maine’s Experience Analysis of compliance with the Maine VSLs during poor weather showed low compliance to the 45 mph limit. The researchers did note, however, that the system was often left active when it was not warranted which may have eroded confidence in the system. A small survey of drivers was also conducted to assess driver perceptions of the system. Of 62 drivers surveyed, only 56% found the Kittelson & Associates, Inc 152 I-10 Phoenix Corridor Safety Study Final Report system to be useful and only 45% said they altered their speed in response to the VSLs. The researchers recommended that speed limits in the future be set based on available stopping sight distance and surface conditions. State of Utah and Colorado have also successfully deployed VSL systems on their state and interstate highway system very recently and the study on the effectiveness is in process now. Kittelson & Associates, Inc 153 I-10 Phoenix Corridor Safety Study Final Report APPENDIX H – COST ESTIMATES Variable Speed Limit System for the I-10 Study Section Item No. Unit Est. Qty Unit Price 6060036 BRIDGE SIGN STRUCTURE (SD9.52, TYPE 3F, DMS) EACH 4 $115,000.00 $460,000 6060037 BRIDGE SIGN STRUCTURE (SD9.52, TYPE 4F, DMS) EACH 8 $135,000.00 $1,080,000 6060080 FOUNDATION FOR BRIDGE SIGN STRUCTURE (SD9.20, TYPE 3F, DMS) EACH 6 $12,500.00 $75,000 6060083 FOUNDATION FOR BRIDGE SIGN STRUCTURE (SD9.20, TYPE 4F, DMS) EACH 14 $13,000.00 $182,000 6070055 SIGN POST (PERFORATED) (2 1/2 S) L. FT. 540 $15.00 $8,100 6070060 FOUNDATION FOR SIGN POST (CONCRETE) EACH 36 $170.00 $6,120 SQ. FT. 270 $25.00 $6,750 7020011 IMPACT ATTENUATION DEVICE (SAND BARRELL CRASH CUSHION, TYPE A) EACH 3 $6,000.00 $18,000 7320070 ELECTRICAL CONDUIT (3") (PVC) L. FT. 1,200 $13.00 $15,600 7320073 ELECTRICAL CONDUIT (2-3") (PVC) L. FT. 2,800 $15.00 $42,000 7320420 PULL BOX (NO.7) EACH 32 $625.00 $20,000 7320455 PULL BOX (NO.9) EACH 10 $2,400.00 $24,000 7320540 CONDUCTOR (NO. 4) L. FT. 30,000 $1.30 $39,000 7320585 CONDUCTOR (INSULATED BOND) (NO.8 GREEN BOND) L. FT. 15,000 $0.80 $12,000 7320765 SINGLE MODE FIBER OPTIC CABLE (12 FIBERS) L. FT. 8,400 $2.25 $18,900 7320794 FIBER OPTIC SPLICE CLOSURE (FMS) EACH 14 $1,800.00 $25,200 7340103 CONTROL CABINET EACH 40 $8,000.00 $320,000 7340251 CONTROLLER (MODEL 2070) EACH 40 $3,000.00 $120,000 7340304 CONTROL CABINET FOUNDATION (CABINET & TRANSFORMER) EACH 40 $1,200.00 $48,000 7350030 LOOP DETECTOR FOR TRAFFIC SURVEILLANCE (6'X6') EACH 66 $550.00 $36,300 7350051 DETECTOR CARD EACH 33 $170.00 $5,610 7350165 LOOP DETECTOR LEAD-IN CABLE L. FT. 6,600 7360250 MODIFY LOAD CENTER EACH 7370430 TRANSFORMER (CABINET ASSEMBLY) (3 KVA) 7370431 TRANSFORMER (CABINET ASSEMBLY) (7.5 KVA) Description 6080005 WARNING, MARKER OR REGULATORY SIGN PANEL Extended Price $0.60 $3,960 4 $2,500.00 $10,000 EACH 26 $2,000.00 $52,000 EACH 14 $2,800.00 $39,200 7379111 VARIABLE MESSAGE SIGN ASSEMBLY EACH 83 $25,000.00 $2,075,000 9240121 MISCELLANEOUS WORK (Microwave/non intrusive detector) EACH 16 $6,000.00 $96,000 9240122 MISCELLANEOUS WORK (GigE SWITCH) EACH 40 $2,000.00 $80,000 9240133 MISCELLANEOUS WORK (In Tunnel mounting signs) EACH 2 $75,000.00 $150,000 9240133 MISCELLANEOUS WORK (MEDIAN BARRIER TRANSITION) EACH 7 $30,000.00 $210,000 9240133 MISCELLANEOUS WORK (Special Foundation for Sign Structure Elevated area) EACH 4 $40,000.00 $160,000 9240133 MISCELLANEOUS WORK (Cable BARRIER relocatation) EACH 2 $15,000.00 SUBTOTAL Including Design, System integration, Contingency, Communications and other Miscellaneous (1.5*Subtotal) Per mile cost would be ((G36)/3) Per mile Variable Speed Limit system capital cost would be Kittelson & Associates, Inc $30,000 $5,468,740 $8,203,110.00 $2,734,370 2.73 million dollars 154 I-10 Phoenix Corridor Safety Study Final Report Extend the WB Lane Drop at 7th Ave. ENGINEER'S ESTIMATE Item No. Description Unit 2010001 CLEARING AND GRUBBING 2020021 REMOVAL OF CONCRETE CURB AND GUTTER 2020027 REMOVAL OF CONCRETE BARRIER 2020031 REMOVAL OF PORTLAND CEMENT CONCRETE PAVEMENT Est. Qty L.SUM 1 L.FT. 2,300 Unit Price Extended Price $2,000.00 $2,000 $10.00 $23,000 L.FT. 20 $50.00 $1,000 SQ.YD. 195 $100.00 $19,500 2020041 REMOVAL OF PIPE L.FT. 60 $30.00 $1,800 2020072 REMOVE AND SALVAGE GUARD RAIL L.FT. 100 $20.00 $2,000 2020156 REMOVE (CATCH BASIN) EACH 3 $300.00 $900 2030301 ROADWAY EXCAVATION CU.YD. 1,100 $20.00 $22,000 3030022 AGGREGATE BASE, CLASS 2 CU.YD. 1,100 $30.00 $33,000 4010012 PORTLAND CEMENT CONCRETE PAVEMENT (12") SQ.YD. 3,200 $100.00 $320,000 4150040 ASPHALTIC CONCRETE (ASPHALT- RUBBER) (END PRODUCT) TON 350 $70.00 $24,500 4150042 ASPHALT RUBBER MATERIAL (FOR AR-AC) (END PRODUCT) TON 30 $550.00 $16,500 4150044 MINERAL ADMIXTURE (FOR AR-AC) (END PRODUCT) TON 4 $100.00 $400 5010011 PIPE, CORRUGATED METAL, 24" L.FT. 45 $100.00 $4,500 5010111 PIPE, CORRUGATED METAL, SLOTTED, 24" L.FT. 60 $150.00 $9,000 5030080 CONCRETE CATCH BASIN (C-15.30) SINGLE, H =8' OR LESS (Type F Barrier) EACH 3 $4,000.00 $12,000 5030702 JUNCTION STRUCTURE 7015052 OBLITERATE PAVEMENT MARKING (STRIPE) EACH 3 $3,000.00 $9,000 L.FT. 6,000 $1.00 $6,000 7042001 PAVEMENT MARKING (PCCP WHITE SPRAYED THERMOPLASTIC)(0.060') L.FT. 6,000 $0.50 $3,000 7042031 PRIMER-SEALER FOR PCCP THERMOPLASTIC STRIPING L.FT. 6,000 $0.30 $1,800 7042051 REMOVAL OF CURING COMPOUND FROM PCCP STRIPING L.FT. 6,000 $1.00 $6,000 L.SUM 8 9100037 CONCRETE BARRIER (SPECIAL HALF) (32" w/GUTTER) L.FT. 9100038 CONCRETE BARRIER (SPECIAL HALF) (32" w/CATCH BASIN) L.FT. 7320480 RELOCATE EXISTING PULL BOXES Misc Items (20%) $1,000.00 $8,000 1,800 $125.00 $225,000 30 $150.00 $4,500 $151,080 SUBTOTAL DESIGN $135,972 2% $18,130 3% $27,194 CONSTRUCTION TRAFFIC CONTROL 10% $90,648 MOBILIZATION 10% $90,648 EROSION CONTROL QUALITY CONTROL CONSTRUCTION SURVEY CONSTRUCTION ENGINEERING AND CONTINGENCIES TOTAL Kittelson & Associates, Inc $906,480 15% 2% $18,130 20% $181,296 $1,468,498 155 I-10 Phoenix Corridor Safety Study Final Report Extend EB Lane Drop at 7th St ENGINEER'S ESTIMATE Item No. Description Unit 2010001 CLEARING AND GRUBBING 2020021 REMOVAL OF CONCRETE CURB AND GUTTER 2020027 REMOVAL OF CONCRETE BARRIER Est. Qty L.SUM 1 L.FT. 1,900 Unit Price Extended Price $2,000.00 $2,000 $10.00 $19,000 L.FT. 20 $50.00 $1,000 SQ.YD. 165 $100.00 $16,500 2020041 REMOVAL OF PIPE L.FT. 100 $30.00 $3,000 2020072 REMOVE AND SALVAGE GUARD RAIL L.FT. 2020031 REMOVAL OF PORTLAND CEMENT CONCRETE PAVEMENT $20.00 2020156 REMOVE (CATCH BASIN) EACH 5 $300.00 $1,500 2030301 ROADWAY EXCAVATION CU.YD. 1,200 $20.00 $24,000 3030022 AGGREGATE BASE, CLASS 2 CU.YD. 850 $30.00 $25,500 4010012 PORTLAND CEMENT CONCRETE PAVEMENT (12") SQ.YD. 2,600 $100.00 $260,000 4150040 ASPHALTIC CONCRETE (ASPHALT- RUBBER) (END PRODUCT) TON 280 $70.00 $19,600 4150042 ASPHALT RUBBER MATERIAL (FOR AR-AC) (END PRODUCT) TON 25 $550.00 $13,750 4150044 MINERAL ADMIXTURE (FOR AR-AC) (END PRODUCT) TON 3 $100.00 $300 5010011 PIPE, CORRUGATED METAL, 24" L.FT. 200 $100.00 $20,000 5010111 PIPE, CORRUGATED METAL, SLOTTED, 24" L.FT. 100 $150.00 $15,000 5030080 CONCRETE CATCH BASIN (C-15.30) SINGLE, H =8' OR LESS (Type F Barrier) EACH 5 $4,000.00 $20,000 5030702 JUNCTION STRUCTURE 7015052 OBLITERATE PAVEMENT MARKING (STRIPE) EACH 5 $3,000.00 $15,000 L.FT. 5,000 $1.00 $5,000 7042001 PAVEMENT MARKING (PCCP WHITE SPRAYED THERMOPLASTIC)(0.060') L.FT. 5,000 $0.50 $2,500 7042031 PRIMER-SEALER FOR PCCP THERMOPLASTIC STRIPING L.FT. 5,000 $0.30 $1,500 7042051 REMOVAL OF CURING COMPOUND FROM PCCP STRIPING L.FT. 5,000 $1.00 $5,000 L.SUM 8 9100037 CONCRETE BARRIER (SPECIAL HALF) (32" w/GUTTER) L.FT. 9100038 CONCRETE BARRIER (SPECIAL HALF) (32" w/CATCH BASIN) L.FT. 7320480 RELOCATE EXISTING PULL BOXES Misc Items (20%) $1,000.00 $8,000 1,700 $125.00 $212,500 50 $150.00 $7,500 $139,630 SUBTOTAL DESIGN $125,667 2% $16,756 3% $25,133 CONSTRUCTION TRAFFIC CONTROL 10% $83,778 MOBILIZATION 10% $83,778 EROSION CONTROL QUALITY CONTROL CONSTRUCTION SURVEY CONSTRUCTION ENGINEERING AND CONTINGENCIES TOTAL Kittelson & Associates, Inc $837,780 15% 2% $16,756 20% $167,556 $1,357,204 156 I-10 Phoenix Corridor Safety Study Final Report Deck Park Tunnel Lighting Upgrade Kittelson & Associates, Inc 157