Decision-Making Guide for Traffic Signal Phasing (2020)

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

62 11 Combining Safety and Operations 11.1 Relationship When selecting the most appropriate phase mode and sequence, the user must consider both safety and operations, as shown in Figure 19. In some cases, promoting safety may lead to a degradation in operational performance and vice versa. Previously, the balance between promoting safety and promoting operational efficiency has been largely subjective and variable between jurisdictions. The guidance in this document provides an approach to determining the combined impacts on safety and operations at an intersection. The results of this chapter may be used as an additional factor for determining the most appropriate phase mode and sequence. Figure 19 – Components of Phasing Decision 11.2 Overall Approach The overall approach for selecting a phase mode and sequence that considers both safety performance and operational impacts entails equalizing the operational and safety impacts for two or more possible phase modes and sequences by converting to a consistent unit, dollars, and comparing the combined effects in a 3-step process.

63 11.2.1 Step 1. Estimate Annual Delay Cost The first step of the process is to estimate the annual delay costs, in dollars, of each phasing alternative under consideration. This step relies on converting the total annual intersection delay experienced into a dollar value using current wage rates. To estimate the annual delay, the user must first estimate the delay per vehicle during typical peak and non-peak hours. This delay may be estimated through use of a deterministic model or a microsimulation, as discussed in Chapter 10 of this Guide. From the movement delay per vehicle and the movement volumes, the total intersection delay experienced, in hours, should be calculated. The total intersection delay, in seconds, obtained from analysis must be converted to annual delay, using the hourly and annual volumes as a factor using the following equation: 𝐴𝑛𝑛𝑢𝑎𝑙 𝐷𝑒𝑙𝑎𝑦 (ℎ𝑜𝑢𝑟𝑠) = 𝑇𝑜𝑡𝑎𝑙 𝐷𝑒𝑙𝑎𝑦 (𝑠𝑒𝑐) ∗ 𝐴𝐴𝐷𝑇𝑇𝑜𝑡𝑎𝑙 𝑉𝑜𝑙𝑢𝑚𝑒 ∗ 3653600 where Total Delay (sec) and Total Volume refer to the sum of the delay for all analysis hours and the sum of the major road approach volume for all analysis hours, respectively. AADT is the average annual daily traffic of the major road. This equation may be applied for as many analysis hours as are available. The hourly value of travel time savings developed by the U.S. Department of Transportation (USDOT 2016) ($/hr) should be applied to the annual delay to calculate the annual delay costs in dollars. Agencies may choose to develop their own comprehensive unit delay costs, which may be used in lieu of the United States Department of Transportation’s (USDOT) hourly value of travel time savings. It is important to consider the same hours between all scenarios when converting to annual delay costs. For example, estimating a peak hour for one scenario and an off-peak hour for the other scenario is not an equivalent comparison and may exaggerate the peak-hour delay over the non-peak hour delay. Case Study 1b provides an example of this process. 11.2.2 Step 2. Estimate Annual Crash Cost Next, the annual crash costs must be calculated for each phasing alternative under consideration. At signalized intersections with existing crash history, the crash data for at least the three prior years should be retrieved, and the number of crashes correctable by the proposed alternative phasing should be determined. The CMFs below in Table 6 may be applied to estimate the annual crashes under each phasing alternative. For additional scenarios not covered by the recommended CMFs below, refer to the CMF Clearinghouse (FHWA 2019) for the most appropriate CMF. The unit crash costs below are used to convert the annual crashes into an average annual crash cost.

64 Table 6 - CMFs Derived from Research Treatment Crash Type CMFs Estimate Standard error Convert from permissive-only or protected-permissive left-turn phasing to protected-only left-turn phase on major road LTOPP1 0.507 0.070 LTOPP INJ2 0.443 0.073 Convert from permissive-only or protected-permissive left-turn phasing to protected-only left-turn phase on minor road LTOPP1 0.702 0.098 LTOPP INJ2 0.687 0.109 Convert from protected-only left-turn phasing to permissive-only or protected-permissive left-turn phasing on major road LTOPP1 1.973 0.273 LTOPP INJ2 2.259 0.374 Convert from protected-only left-turn phasing to permissive-only or protected-permissive left-turn phasing on minor road LTOPP1 1.424 0.198 LTOPP INJ2 1.457 0.231 1. Left-turn Opposing Crash, All Severities 2. Left-turn Opposing Crash Type, Injury/Fatality Only For new intersections or intersections that do not have representative crash history available, annual crashes may be estimated based on SPFs, further described in Chapter 9. Refer to Chapter 9 for additional instruction on applying the developed CMFs and SPFs. The following table, table 7, was developed from recommended FHWA unit crash cost per crash severity (FHWA 2018) and the ratio of crashes from the accompanying research study’s dataset. Agencies may choose to develop their own comprehensive unit crash costs, which may be used in lieu of the crash costs below. Table 7 - Recommended Unit Crash Costs (in 2016 dollars) Crash Severity Comprehensive Unit Crash Cost All Severities $ 105,900 Fatality + Injury $ 310,000 11.2.3 Step 3. Derive a Composite Cost The final step in the process is to combine the operational and safety costs calculated in steps 1 and 2 with a simple equation for each alternative. 𝐶𝑜𝑠𝑡 = 𝐷𝑒𝑙𝑎𝑦 + 𝐶𝑟𝑎𝑠ℎ CrashiDelayiCosti

65 where, Costi = Composite Estimated Annual Cost of phasing alternative i, in dollars ($) Delayi = Estimated Annual Delay Cost of phasing alternative i, in dollars ($) Crashi = Estimated Annual Crash Cost of phasing alternative i, in dollars ($) The phasing alternative with the lowest composite cost inherently has the lower combined impact on operations and safety. 11.3 Case Studies Two case studies are provided below to demonstrate the use of the concepts above. 11.3.1 Case Study 1a – New Traffic Signal Case Overview This case study focuses on a 4-leg, two-way stop-control intersection where a traffic signal is warranted. The future lane conditions are depicted in Figure 20. A previous study determined that a traffic signal is the preferred alternative rather than an alternative intersection configuration (e.g., roundabout). This case study focuses on determining the left-turn phase mode and sequence for the major street, which is functionally classified as a minor arterial. Permissive-only phasing is assumed for the minor approaches. The intersection is in a suburban setting on a commuter corridor. The two scenario alternatives are: 1. Permissive-only left-turn phasing 2. Protected-only left-turn phasing Data Collection Geometric Each major street approach has one shared through/right-turn lane and one dedicated left-turn lane, and each major street receiving approach has one lane. This case study confirmed that geometric features, such as sight distance or the ability to make concurrent opposing left turns, do not restrict the use of permissive left-turn movements. Safety Crash history data is not available because crashes at the existing stop-controlled intersection are not expected to be representative of conditions after a traffic signal is installed. Operations The major street has a speed limit of 35mph. An adjacent intersection, approximately 900 feet away, operates with permissive-only left-turn phase mode on one major street approach and protected- permissive-only left-turn phase mode on the opposite approach.

66 Volumes The major road approach has an AADT of 14,000 vehicles per day, and the minor road approach has an AADT of 3,600 vehicles per day, as obtained from the state AADT database. Turning movement counts for 24-hours are available for use in this analysis because data were collected for the earlier signal warrant study. The data includes separate pedestrian movement counts. The left-turning movements account for approximately 3% of the total approach volume on the major road approaches and 7% on the minor road approaches. Figure 20 - Diagram of Case Study 1

67 Estimate Annual Delay Cost This analysis employed a deterministic model to determine the delay per vehicle for a typical AM, PM, Midday, and off-peak hour, shown in Table 8. The deterministic model relied on software-optimized signal timings based on the lane configuration and volume inputs. Table 8 - Case Study 1a - Hourly Volume and Delay Major Road Volume Delay (sec) - Scenario 1 Delay (sec) - Scenario 2 AM 1,165 11,137 24,003 Midday 601 4,760 7,800 PM 1,262 17,117 24,286 Off-Peak 709 7,011 9,775 TOTAL 3,737 40,025 65,864 For each scenario, the delay from the four analysis hours was converted to total annual delay (in hours) using the 24-hour turning movement count ratios for the analysis hours’ volume to total daily volume with the following equation: 𝐴𝑛𝑛𝑢𝑎𝑙 𝐷𝑒𝑙𝑎𝑦 (ℎ𝑜𝑢𝑟𝑠) = 𝑇𝑜𝑡𝑎𝑙 𝐷𝑒𝑙𝑎𝑦 (𝑠𝑒𝑐) ∗ 𝐴𝐴𝐷𝑇𝑇𝑜𝑡𝑎𝑙 𝑉𝑜𝑙𝑢𝑚𝑒 ∗ 3653600 𝐴𝑛𝑛𝑢𝑎𝑙 𝐷𝑒𝑙𝑎𝑦 (ℎ𝑜𝑢𝑟𝑠) 𝑆𝑐𝑒𝑛𝑎𝑟𝑖𝑜 1 = 40,025 ∗ 14,0003,737 ∗ 3653600 𝐴𝑛𝑛𝑢𝑎𝑙 𝐷𝑒𝑙𝑎𝑦 (ℎ𝑜𝑢𝑟𝑠) 𝑆𝑐𝑒𝑛𝑎𝑟𝑖𝑜 2 = 65,864 ∗ 14,0003,737 ∗ 3653600 This equation may be applied for as many analysis hours as are available. The total estimated annual delay for each scenario is as follows: Table 9 - Case Study 1a - Annual Delay Scenario 1 Scenario 2 Annual Delay 15,203 hours 25,017 hours The estimated annual delay was converted into an annual delay cost by applying the most current recommended hourly value of travel time savings developed by the US Department of TransportationError! Bookmark not defined.. This analysis assumed an hourly rate of $14.10 (in 2016 dollars), which the USDOT recommended for local travel, all purposes. After accounting for inflation rates (Bureau of Labor Statistics, CPI Inflation Calculator), this value equals $14.98 in the current year (2019) dollars, and this value was applied. Converted into annual cost, the delay for each scenario is presented below.

68 Table 10 - Case Study 1a - Annual Delay Cost Scenario 1 Scenario 2 Annual Delay 15,203 hours 25,017 hours Annual Delay Cost, Delayi $ 227,741 $ 374,755 11.3.2 Estimate Annual Crash Cost Since representative crash data are not available, the safety analysis relied on the SPFs developed by the accompanying research effort. Chapter 9 provides more detail on the use of SPFs. The following SPF was chosen to evaluate crashes of all severities: 𝐿𝑇𝑂𝑃𝑃= 𝑒 . 𝑀𝐴𝐽𝐴𝐴𝐷𝑇 . 𝑀𝐼𝑁𝐴𝐴𝐷𝑇 . 𝑒( . . . . ) where, LTOPP = Left-turn opposing crashes (all severities) per year MAJAADT = major road AADT (sum of both approaches) MINAADT = minor road AADT (sum of both approaches) RATIOLTMAJ = Ratio of major road traffic turning left to total traffic on the major road in peak hours RATIOLTMIN = Ratio of minor road traffic turning left to total traffic on the major road in peak hours PROTMAJ = 1 if major road has protected-only left-turn phasing; 0 otherwise PROTMIN = 1 if minor road has protected-only left-turn phasing; 0 otherwise This SPF gives the following results for estimated annual crashes: Table 11 - Case Study 1a - Annual Crash Estimation Scenario 1 Scenario 2 Annual LTOPP Crashes (All Severities) 0.513 0.183 The estimated annual crashes were converted into an annual crash cost by applying the unit crash costs for injury crashes. This analysis used a unit crash cost of $105,900 per crash in 2016 dollars, which equals $112,513.72 in current year (2019) dollars. The annual crash costs for each scenario are presented below. Table 12 - Case Study 1a - Annual Crash Cost Scenario 1 Scenario 2 Annual Crashes (KABC Severities) 0.513 0.183 Annual Crash Cost, Crashi $ 57,720 $ 20,590 Derive a Composite Cost The final step is to derive a composite cost which will reveal the scenario with the lowest combined impact on safety and operations for this intersection. The following equation combines the annual delay cost and annual crash cost computed in the previous two steps to develop one comprehensive cost:

69 𝐶𝑜𝑠𝑡 = 𝐷𝑒𝑙𝑎𝑦 + 𝐶𝑟𝑎𝑠ℎ where, Costi = Composite Cost of phasing alternative I, in dollars ($) Delayi = Estimated Annual Delay Cost of phasing alternative, in dollars ($) Crashi = Estimated Annual Crash Cost of phasing alternative, in dollars ($) The composite cost for each scenario is: Table 13 - Case Study 1a - Annual Composite Cost Scenario 1 Scenario 2 Annual Delay Cost, Delayi $ 227,741 $ 374,755 Annual Crash Cost, Crashi $ 57,720 $ 20,590 Annual Composite Cost, Costi $ 285,461 $ 395,345 As Scenario 1 (permissive-only left-turn phasing) results in the lower composite cost, Scenario 1 inherently has the lower negative combined impact to the intersection, and, based on this analysis, is the more appropriate phasing strategy for this intersection. 11.3.3 Case Study 1b – New Traffic Signal Case Overview Assume the same characteristics as the intersection in 1a, but with the following volumes and delays: Table 14 - Case Study 1b - Hourly Volume and Delay Major Road Volume Delay (sec) - Scenario 1 Delay (sec) - Scenario 2 AM 1,700 45,574 87,846 Midday 1,056 20,173 35,536 PM 1,577 58,118 70,752 Off-Peak 829 9,395 15,815 TOTAL 5,162 133,260 209,949 AADT is 17,000 vehicles on the major road approach and 2,428 on the minor road approach. Left-turning vehicles account for approximately 8% of the major road traffic and 4% of the minor road traffic. The two scenario alternatives are: 1. Protected-Permissive left-turn phasing 2. Protected-only left-turn phasing The following tables provide the expected annual delay costs, annual crash estimation and cost, and the annual composite cost for each scenario.

70 Table 15 - Case Study 1b - Annual Delay Cost Scenario 1 Scenario 2 Annual Delay 44,496 hours 70,103 hours Annual Delay Cost, Delayi $ 666,550 $ 1,050,143 Table 16 - Case Study 1b - Annual Crash Cost Scenario 1 Scenario 2 Annual Crashes (KABC Severities) 0.287 0.146 Annual Crash Cost, Crashi $ 32,291 $ 16,427 Table 17 - Case Study 1b - Annual Composite Cost Scenario 1 Scenario 2 Annual Delay Cost, Delayi $ 666,550 $ 1,050,143 Annual Crash Cost, Crashi $ 32,291 $ 16,427 Annual Composite Cost, Costi $ 698,841 $ 1,066,570 At this intersection, Scenario 1 (protected-permissive left-turn phasing) results in the lowest composite cost and inherently has the lower negative combined impact to the intersection. Based on this analysis, Scenario 1 is the recommended phasing strategy. 11.3.4 Case Study 2 – Change in Existing Phasing Case Overview This case revisits the intersection in Case Study 1a. In this case, permissive-only left-turn phase mode has been in place for 5 years, and in that time, 12 LTOPP crashes occurred of all severities. Due to the crash occurrence, the local agency is contemplating a change to protected-only left-turn phasing. All other information remains the same as in Case Study 1. The two scenario alternatives remain as: 1. Permissive-Only left-turn phasing (existing) 2. Protected-only left-turn phasing (proposed) Data Collection The data collection portion remains the same as in Case Study 1, with the exception of the crash history. The previous 5 years of crash history is now available and reveals 12 LTOPP injury crashes, correctable by protected-only phasing, in those 5 years. Estimate Annual Delay Cost The estimated annual delay cost remains the same as in Case Study 1:

71 Table 18 - Case Study 2 - Annual Delay Cost Scenario 1 Scenario 2 Annual Delay 15,203 hours 25,017 hours Annual Delay Cost, Delayi $ 227,741 $ 374,755 Estimate Annual Crash Cost The annual crash frequency and crash cost for Scenario 2 remain the same as in Case Study 1a. However, the estimated annual crash frequency for Scenario 1 (the existing condition) requires inclusion of the known crash history. For Scenario 1 (the existing condition), a refined empirical Bayes estimate is applied as a weighted average between the known crash history and the SPF developed results in the estimated annual crash frequency. The empirical Bayes refined estimate, m, is a weighted average of the observed 12 crashes in 5 years (x = 2.4 crashes/year) and the SPF prediction (P), which remains the same as Case Study 1 (0.513 crashes/year). The overdispersion parameter, k, is defined in the model development report, Table 18 of Appendix B. Using the following equation, the empirical Bayes refined estimate is defined as m = α(P) + (1-α)(x) where, the weight α was estimated as follows: α = 1/(1 + NkP) P = SPF predicted crashes/year X = observed crashes/year N = number of years of observed crash data, and k = overdispersion parameter given for the SPF in the SPF parameter table in Appendix B (=0.528) Given the parameters defined above, for this case study, α is estimated as 0.425, and m is estimated as 1.598 crashes/year. Applying the unit comprehensive crash cost of $112,513.72, the new crash costs for this intersection are: Table 19 - Case Study 2 - Annual Crash Cost Scenario 1 Scenario 2 Annual Crashes (KABC Severities) 1.598 0.183 Annual Crash Cost, Crashi $ 179,797 $ 20,590 Derive a Composite Cost The final step is to derive a composite cost which will reveal the scenario with the lowest negative combined impact on safety and operations for this intersection. The following equation combines the annual delay cost and annual crash cost computer in the previous two steps to develop one comprehensive phasing cost: 𝐶𝑜𝑠𝑡 = 𝐷𝑒𝑙𝑎𝑦 + 𝐶𝑟𝑎𝑠ℎ

72 where, Costi = Composite Cost of phasing alternative I, in dollars ($) Delayi = Estimated Annual Delay Cost of phasing alternative, in dollars ($) Crashi = Estimated Annual Crash Cost of phasing alternative, in dollars ($) The composite cost for each scenario is: Table 20 - Case Study 2 - Annual Composite Cost Scenario 1 Scenario 2 Annual Delay Cost, Delayi $ 227,741 $ 374,755 Annual Crash Cost, Crashi $ 179,797* $ 20,590 Annual Composite Cost, Costi $ 407,538* $ 395,345 * Value to change from Case Study 1 In this case study, Scenario 2 (protected-only left-turn phase mode) results in the lowest composite cost, indicating that Scenario 2 inherently has the lower negative combined impact to the intersection, and, based on this analysis, is the most appropriate phasing option.