Guidelines for Quantifying Benefits of Traffic Incident Management Strategies (2022)

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.

6 The Language of Traffic Incident Management Using common terminology and definitions when describing TIM­related activities is the first step toward uniform and consistent evaluation of TIM programs within and between agencies. This chapter presents considerations to encourage uniform use of terminology that is foundational to TIM and the evaluation of TIM programs. Figure 4 shows how TIM benefits depend on a common understanding of TIM terminology related to incidents, timeline, perfor­ mance measures, activities, outcomes, and quantification and monetization. The organization of this chapter loosely mirrors Figure 4, beginning with incidents, followed by TIM timeline and performance measures, TIM activities, and outcomes, and concluding with how outcomes are quantified and monetized. Definitions presented herein are from multiple sources, including the FHWA, previous National Cooperative Highway Research Pro­ gram (NCHRP) projects, peer­reviewed journals, and state TIM and intelligent transportation systems (ITS) plans. Definitions focus on providing support for calculating the FHWA core TIM performance measures of roadway clearance time (RCT), incident clearance time (ICT), and secondary incident counts. The focus of this chapter is on understanding terminology and explaining how variations in definitions and classification structures affect estimates of program benefit. What Are Incidents? How Are They Classified? While transportation providers and emergency responders tend to have different definitions for what constitutes an incident, the transportation definition better serves the purpose of estimating the value of TIM. Incidents are defined in the TIM Handbook as “any non­recurring event, of any origin, causing a reduction of roadway capacity or an abnormal increase in travel demand” (Farradyne, 2000). This includes traffic crashes, stopped or disabled vehicles, roadway obstructions or debris, and special events or road maintenance that are planned or unplanned. The Traffic Management Data Dictionary (TMDD), as published by Institute of Transportation Engineers and American Association of State Highway and Transportation Officials, offers a narrower definition of incidents to include only unplanned events. Figure 5 lists these incidents with “road work” and “special events” differentiated by color, reflecting that some agencies include planned events within their TIM, whereas others do not. The estimation of TIM benefits presented herein focuses on response to unplanned incidents; however, benefits of TIM for planned events may be estimated using a combination of methods, such as tools that focus on work zone analysis (FHWA Work Zone and Traffic Analysis Tools; [FHWA, n.d.(c)]). An incident may be reported within an agency’s database with numerous attributes. The most common fields include time and incident type, location of occurrence, duration, number/type of C H A P T E R 2

The Language of Trafc Incident Management 7   involved vehicles, number of lanes blocked, and whether primary or secondary. Other informa- tion oen collected includes: • Weather, light, and other environmental conditions. • Personnel and hours at the incident site by response agency. • Equipment and materials used and agency ownership. • Injury to persons involved and infrastructure damage. e data management structure for TIM programs should provide for elements that extend beyond the incident classication system. e guidance on TIM performance measures (Pecheux et al., 2014) proposes a database model with 34 elements along incident, roadway, vehicle, lanes, responder, and participant categories. e ability to design custom data types is present in many data soware as well as the schema. Many agencies also capture situational and impact information as incident data elements. For example, the Virginia Department of Transportation (VDOT) captures the number of miles trac is delayed and the degree to which trac is delayed. ese incident data elements, or attributes, should be aggregated through an incident classication system. An incident classication system allows the grouping of incidents for which average trac impact, TIM program performance, improvement needs, and benets may be addressed more simply than by examining individual incidents. It is the lens by which TIM performance statistics are presented and reviewed. A three-level approach for incident classication, shown in Figure 6, is recommended by the FHWA (Farradyne, 2000) and the Manual on Uniform Trac Control Devices [FHWA, n.d.(a)]. Duration is a meaningful basis for an incident classication In ci de nt s Crashes Vehicle Fires Disabled Vehicles Roadway Debris Spilled Cargo Loss of Load Hazardous Materials Road Construction or Maintenance Special Events May be planned or unplanned. Quantification and Monetization (variance in what and how to quantify and monetize) Outcomes and Performance Measures Mobility Environment E ciency Safety Traveler Satisfaction TIM Activities (difficult to measure independently) Strategic - Policy Tactical - Response Support - Systems TIM Performance Measures (slight variations in definition) Roadway Clearance Time Incident Clearance Time Secondary Incidents TIM Timeline (some consensus) Detection Ver cation Response Road Clearance Incident Clearance Incident Classification (FHWA classification, others also in use) Minor Intermediate Major Figure 4. Terminology within the context of TIM. Figure 5. Incident types. Minor Intermediate Major Roadway clearance time of at least 30 minutes but less than 2 hours Roadway clearance time of less than 30 minutes Roadway clearance time of 2 or more hours and/or involves fatality Figure 6. FHWA recommended incident classication structure.

8 Guidelines for Quantifying Benefits of Traffic Incident Management Strategies system because it is simple to apply and because time correlates with incident severity and number of secondary incidents (Karlaftis et al., 1999; Hirunyanitiwattana and Mattingly, 2006; Khattak et al., 2011). Agencies use different classification structures that meet their local requirements. For example, the Washington State DOT Gray Notebook (Washington State Department of Transportation) classifies and reports incidents as minor, intermediate, and major, but they define minor as less than 15 minutes, and major as more than 90 minutes. The Maryland Coordinated Highway Action Response Team (CHART) program captures a number of incident data elements and presents incident characteristics by regional traffic operations centers, operational and nonoperational hours, and incidents and disabled vehicles. They further use aggregation by incident type, including collision­fatality, collision­property damage, and collision­personal injury. They do not, however, classify incidents based on the FHWA­recommended structure. New York, Wisconsin, Virginia, and others use the FHWA­ recommended temporal classification, but with descriptors based on incident severity or slight deviations in duration. In creating a classification structure, an agency should consider how levels of detail align with data systems across the responder community. Categories should be sufficiently detailed to differentiate estimates within the TIM timeline and levels of response. The language to describe and categorize incidents should be, to the extent feasible, uniform across organiza­ tions involved in the TIM program. Greater consistency enables easier analysis and evaluation among stakeholder agencies and with peer agencies. What Is the TIM Timeline, and What Are Core TIM Performance Measures? Incident timestamps along the timeline of every incident, as presented in Figure 7 by T0 through T7, are the key components for measuring TIM performance. Transportation and other responder agencies document timestamps through a combination of automated and manual Why Incident Definition and Classification Matter • Incident counts differ based on definition; changing a key driver for TIM benefits estimates. • Localized classifications complicate comparison of performance and benefits with peer agencies or nationally. • Using TIM benefits estimates from a source with a different classification structure may result in inaccurate estimates of benefits. Figure 7. TIM timeline.

The Language of Traffic Incident Management 9   entry processes. The routine, organized recording of incident timestamps is necessary to compute TIM performance measures. The FHWA defines performance measures as “. . . the use of statistical evidence to determine progress toward specific defined organizational objectives.” The FHWA­recommended TIM performance measures include RCT, ICT, and secondary incident count. Linking TIM activities with changes in performance measures demonstrates the effects of investment in TIM. These measures within and along the timeline for TIM are discussed herein to establish common terminology. Figure 7 shows the timeline for an incident, which begins with detection time, followed by verification and response time. From the point of the incident report to open roadway lanes (T5–T1) is RCT. From the point of the incident report to the exit of all responders from the scene (T6–T1) is ICT. Each of these durations is defined and discussed in this section. Detection Time One of the primary goals for a TIM strategy is to decrease the detection time of incidents. Detection time is not typically reported because the actual time the incident occurred is often unknown. In today’s mobile communications environment, the time to detect that an inci­ dent has occurred on an urban freeway is typically fairly short. On the other hand, the time to detect that an incident has occurred in remote, rural areas can be significant (in such cases, the primary beneficiaries of the incident response are those immediately involved in the incident, since delays may not be occurring behind the incident). The TIM Handbook and other works (Farradyne, 2000; Owens et al., 2010) list commonly used methods to detect and verify incidents: • Social media and crowdsource data. • Mobile telephone calls from motorists. • Closed­circuit television cameras viewed by operators. • Roaming service patrols. • Police patrols. • Automatic vehicle identification (AVI) combined with detection software. • Electronic traffic measuring devices (e.g., video imaging, loop, or radar detectors). • DOT or public works crews. • Fleet vehicles (transit and trucking). • Others (motorist aid telephones or call boxes, aerial surveillance, traffic reporting services). Verification Time Incidents may be verified through a subset of sources listed for detection. Incidents can be verified through dispatch field units (e.g., police or service patrol), closed­circuit television (CCTV), or multiple corroborative cellular calls or crowdsourced reports. Response Time The response to an incident starts with a series of planned steps for an optimum deployment of personnel and equipment appropriate for each particular incident. Response time is the time between the incident being verified and the first responder arriving on scene. Each incident is unique, and the response sequence may vary, so it is vital that measurement information can be collected from traditional and specialized responders. Some agencies include verification time within the measure of response time (Chang et al., 2013). Incident response time is improved (or reduced) by strategic planning and positioning of resources, the establishment of processes, and the use of technology for quicker detection and verification. Table 1 highlights examples of improvements in response time for various TIM programs.

10 Guidelines for Quantifying Benefits of Traffic Incident Management Strategies Roadway Clearance Time RCT is one of the three TIM program performance measures identified by FHWA and is the period between the first recordable awareness of the incident by a responsible agency and the first confirmation that all lanes are available for traffic flow (Conklin et al., 2013). For those agencies without data to support the first recordable awareness of an incident, the first aware­ ness by the transportation agency may serve to compute a modified RCT. • Agencies may only have data for those incidents that are reported directly to them. • Incident data may also be limited to specific TIM activities (e.g., safety service patrol). • Incident data may be limited to TIM program hours of operation. Many TIM organizations have RCT goals and measure performance against those goals. The goals are specific to the incident classification defined by the program. Incident Clearance Time ICT is one of the three TIM program performance measures identified by the FHWA, and one of the core measures that drives estimation of TIM benefits. The FHWA defines ICT as the time from incident report to when the last responder has left the scene. The responder may be transportation, law enforcement, towing, or other responsible entity. Differences continue in the interpretation of “last responder leaving the scene,” with some agencies recording the departure of the transportation responder. In some instances, the reporting for incident clearance by law enforcement is specific to mainline clearance, wherein the responder may remain roadside (shoulder lane) while completing the incident logs. Both of these interpretations underestimate the true ICT, which in turn underestimates the delay attributable to the incident and the potential for improvements in TIM. Time to Return to Normal Flow Return to normal traffic flow is a difficult point in time to measure and requires signifi­ cant sensor or camera instrumentation. Often the full dissipation of queuing occurs several TIM Program Response Time Observations San Antonio TransGuide (Henk et al., 1997) 20% lower response time from the deployment of a combined communications network and CCTV Hudson Valley H.E.L.P. Safety Service (Haghani et al., 2006) Weekend 20-minute response and weekday evening 12-minute response reduced to 8 minutes with safety service patrol Maryland CHART Safety Service (Chang and Rochon, 2003) 13% lower response time from safety service patrol, despite worsening congestion and increasing incident counts Houston SAFEClear (Lomax and Simcic, 2016) 98% of incidents have 6-minute or less response time Virginia DOT Hampton Roads (Virginia Department of Transportation, 2013, 2014) Average response time decreased from 9.04 minutes (2011) to 8.3 minutes (2012), to 6.9 minutes (2013), and 4.3 minutes in the year 2014 Table 1. TIM improvements in response time.

The Language of Traffic Incident Management 11   miles upstream of the incident and may include multiple facilities. Consequently, the time duration to return to normal flow requires estimation. This represents the total duration of time that traffic is affected by the initial incident. This measure is typically not directly used in defining performance goals for TIM programs, given the complexity associated with measurement. Secondary Incidents The number of secondary incidents is one of the three TIM program performance measures identified by the FHWA. This measure is defined as “the number of unplanned incidents beginning with the time of detection of the primary incident where a collision occurs as a result of the original incident either within the incident scene or within the queue in either direction” (FHWA, 2010). While some agencies consider only crashes as secondary incidents, the FHWA includes engine stalls, overheating, and running out of fuel as secondary incidents (FHWA, 2004). The occurrence of secondary incidents is influenced by a number of primary incident characteristics, including primary incident duration, lanes affected, level of congestion, facility speed, vehicle types and count in the incident, and environmental conditions. Conceptually, incidents that occur within the time to return to normal flow from the primary incident, and within the spatial footprint of delayed flow, should be categorized as secondary. The challenge with this measurement primarily lies with the identification of incidents to a secondary status after the fact. Some agencies have moved to an on­site report via a checkbox in the incident report filed by police or other on­site personnel. This on­site reporting also has challenges asso­ ciated with the temporal and physical location of the on­site personnel and their awareness of activities up­ or downstream. Agencies that employ after­the­fact secondary incident identification use different temporal parameters such as 30 minutes, 60 minutes, or 120 minutes after the primary incident. Some agencies employ a dynamic temporal parameter equal to the primary ICT plus 15 minutes. The variance in the use of spatial parameters among agencies ranges from 0.5 to 3 miles upstream in the direction of traffic, with some also cap­ turing crashes that are within a 0.5­mile distance up­ and downstream for opposing traffic. Table 2 presents data points relating to secondary incident rates, recognizing that these estimates use somewhat different methods for identifying an incident as secondary. Analyses conducted as part of this study applied several methods and found that 16% of Maryland and 9% of Dallas incidents are secondary. Analyses conducted using data from the Chicago and the Hampton Roads regions (Raub, 1997; Khattaket al., 2011) further define rates of secondary incidents specific to the type of primary incident. These analyses suggest that among the set of all incidents: • 7.5% to 10% followed an earlier crash. • 1.5% to 9% followed a disabled vehicle. • 2.5% followed a police stop for a traffic violation. • 0.9% followed an abandoned vehicle. The sum of these percentages suggests that 12.4% to 22.4% of incidents in these regions are secondary in nature. The likelihood of occurrence of a secondary incident increases by 2.8% for every minute of increase in clearance time. This is based on analyses of Indiana’s Hoosier Helper freeway service patrol (Karlaftis et al., 1999). What Is a Secondary Incident? • A secondary incident occurs as a result of the original incident within the queue in either direction. • Secondary incidents occur more frequently during peak periods and on urban freeways. • Many agencies define secondary incidents as crashes only; the FHWA includes engine overheating and stalls and running out of fuel.

12 Guidelines for Quantifying Benefits of Traffic Incident Management Strategies According to analyses of incidents conducted across three regions in the development of this guide, the percentage of incidents that were prevented by implementing TIM programs ranged from 2.6% to 7.2%. This range reflects the application of three different methods for estimating the reduction in secondary incidents across three regions (North Dallas, Texas; Capital Beltway, Maryland; Seattle, Washington) and TIM programs, reducing incident duration by 24% in these regions. These data points can serve as a basis for comparison with local estimates. What Are Common TIM Activities? To achieve efficiencies throughout the timeline of incident response, TIM programs require strategic, tactical, and support activities. While strategic, tactical, and support TIM activities are equally critical to a strong TIM program, and the realization of TIM goals, estimating benefit from a single activity is difficult, particularly for strategic and support activities. Figure 8 provides a framework for studying TIM activities. Figure 8. TIM strategic, tactical, and support activities. TIM Region Percentage of Incidents Identified as Secondary Regional Emergency Action Coordination Team (Battelle Memorial Institute, 2002) 10% of crashes are secondary St. Louis (Sun et al., 2009) 10% of crashes are secondary Maryland CHART (Chang and Point-du-Jour, 2003) 6.8% of lane blocking incidents are secondary California Highway Patrol’s First Incident Response Service (Moore et al., 2004) 1.5% to 3.0% of crashes are secondary Chicago, Illinois (Raub, 1997) 5.4% of crashes are secondary Hampton Roads Area (Khattak et al., 2011) 2.0% of incidents are secondary FHWA Report (FHWA, 2004) 20% of incidents are secondary* *A foundational analysis could not be identified to reference this estimate. Table 2. Percentage of incidents identified as secondary.

The Language of Traffic Incident Management 13   Significant literature highlights best practices for TIM activities, including the TIM Handbook (Owens et al., 2010), Best Practices in TIM (Carson, 2010), the TIM Teams Best Practices Report (Delcan, 2010), and the National Traffic Incident Management Coalition’s Example Strategies for Building Stronger State TIM Programs (Haas, 2006). What is common among all literature is the criticality of establishing strong institutional relationships and multidisciplinary collabo­ ration for the sustainability of TIM programs and the realization of TIM benefits. Strategic TIM Activities Strategic activities focus on establishing TIM within the fabric of planning and operations among transportation, emergency, fire, law enforcement, recovery, and other stakeholders through formalized structures. This includes planning, preparing for, establishing, and formal­ izing the following: • Partnerships, policies, and legislation. • Multidisciplinary team training. • Public outreach and education. • Program resources and funding. • Methods for measuring progress. Tactical TIM Activities Tactical activities include surveillance and detection, mobilization and response, scene management, information dissemination, and clearance and recovery. Benefits from these activities have frequently been quantified, but typically in aggregate from multiple activities. For example, San Antonio’s TransGuide ITS system combines a communications network, CCTV, and loop detectors to improve incident detection and response. In its first year of deploy­ ment, TransGuide reduced incident response times by 20% (Haas, 2006). Support TIM Activities Support activities include technology implementation and upgrades, data integration, and cost or recovery management. Effective interoperable voice, data, and video communications and information exchange are vital to TIM. The most efficient way to accomplish real­time accurate information exchange is to develop interoperable systems that can electronically exchange data. Real­time communication and information exchange requires institutional, technical, and operational coordination among agencies, operational support centers, and systems (Owens et al., 2010). As TIM becomes integrated as part of the transportation agency operations, jurisdictions must understand the true cost of responding to traffic incidents and identify ways to offset or recover these recurring costs. Cost management and recovery aim to bridge this gap from transportation agency funding to a broader funding base. TIM Cost Management and Recovery Guidance developed by the FHWA (Rensel et al., 2012) suggests implementing a cost management roadmap and coordinating with legislators and private industry to identify and implement cost recovery methods. Recovery is defined as full or partial reimbursement from sources outside of the budget. The most common type of cost recovery used by state transportation agencies is the recovery of those costs related to infrastructure damage. Many states have enacted legislation that allows the transportation agency to submit claims to individuals or insurance companies for damages to infrastructure. Three primary elements also need to be managed: asset utilization,

14 Guidelines for Quantifying Benefits of Traffic Incident Management Strategies resource utilization, and performance measurement. More on cost recovery can be found in Rensel et al., 2012. Strategic, tactical, and support TIM activities target multiple parts of incident management, including detection and verification, traffic management, coordinated response, scene manage­ ment, and quick clearance. Figure 8 presents examples of TIM activities that affect incident management. What Are TIM Outcomes? Outcomes refer to the benefits from TIM activities. Outcomes may be organized into five broad categories: mobility, environment, safety, efficiency, and traveler satisfaction. Outcomes and their corresponding performance measures are increasing in importance in meeting broader safety and infrastructure goals as part of Moving Ahead for Progress in the 21st Century (MAP­21). MAP­21 establishes performance measures for the federal­aid highway program in several target areas that relate to TIM, including safety, reliability, and congestion reduction. State DOTs must set performance measures in support of those target areas (FHWA, 2013a). For outcomes of specific interest to an agency, performance measures designed to assess that outcome should be developed. For example, if increased responder safety is the desired outcome, the duration of time responders are on scene, and therefore exposed to traffic, may be measured by subtracting the response arrives on scene time from the response departs scene time. Another measure centered on responder safety may be the number of responders trained in TIM. Table 3 presents examples of TIM activities and outcomes. Note that this table does not specify an exhaustive list of outcomes from a specific TIM activity. TIM Activities Examples of TIM Outcomes* Increase operational hours of service patrol Mobility—Reduce delay from quicker clearance Safety—Reduce secondary incident counts Pass and/or publicize move over and/or slow down laws Safety—Reduce the number of secondary incidents, improve responder safety Partnerships with news media to notify drivers to avoid areas during an incident Mobility—Reduce demand and consequently delay on the facility with an incident Mobility—Expedite responder access to an incident, reducing incident duration and delay Environment—Reduce emissions and fuel consumption due to mobility outcomes Implement and use interoperable voice and data networks Safety—Increase public and responder safety Stage equipment and resources in areas able to easily access corridors and critical points Mobility—Reduce ICT Safety/Efficiency—Reduce responder’s time on scene *Note: Not all TIM outcomes are described for each activity. Table 3. Examples of activities and outcomes.

The Language of Traffic Incident Management 15   The common thread in evaluating the different types of outcomes is the need for data. All outcomes can be quantified, and some outcomes can be monetized. Often, outcomes of TIM programs may be inferred by the selection and implementation of specific activities. Among the outcomes in Table 3, mobility contributes the most to the overall monetized benefits, followed by safety benefits. Environmental and efficiency benefits are significant but far less than those of mobility and safety. Typically, traveler satisfaction benefits are not monetized. Mobility Outcomes TIM activities can improve mobility attributes related to quantity, quality, and utilization: • Quantity—vehicle, person, and truck miles traveled. • Quality—average travel speed, vehicle and person delay, and travel time reliability. • Capacity utilization—vehicles per lane mile, duration of congestion, percentage of miles or travel in congestion Among these activities, the most frequently estimated mobility outcome is the reduction in delay. This reduction in delay is achieved through TIM activities that decrease facility demand through traveler information and more quickly restoring facility capacity through efficient incident staging and clearance. For example, the Maryland CHART program measured inci­ dent duration at 23.3 minutes and 34.8 minutes with and without CHART assistance, respec­ tively, contributing to 36.5 million vehicle­hours reduction in delay. (Chang and Igbinosum, 2015). Safety Outcomes Safety outcomes from TIM activities include reduced numbers of secondary incidents, as well as reduced risk of serious injury or death for responders and motorists. Through TIM training activities and expedited incident clearance, the duration of time responders are exposed to traffic decreases significantly. Likewise, motorists who are assisted by TIM safety service patrol reduce their duration of time exposed to traffic. TIM activities also reduce the time from incident occurrence to arrival at trauma care, increasing the likelihood that those seriously injured from a crash arrive within the “golden hour.” The literature from the medical profession, however, does not find an association between EMS intervals, specifically the golden hour, and mortality among injured patients (Newgard et al., 2009). In 2011, Arizona Department of Public Safety (DPS) had a secondary crash rate of approx­ imately 6% and involved 54 first responders. Further, Arizona DPS lost 11 officers while involved with secondary crashes as of 2012 (Arizona Department of Transportation, 2012). The Maryland CHART 2014 estimate of secondary incidents is 930, with 458 secondary incidents prevented through reduced incident duration (Chang and Igbinosun, 2015). Environmental Outcomes Environmental outcomes from TIM activities include a reduction in vehicle emissions associated with the reduction in delay. These include hydrocarbons (HC), carbon monoxide (CO), nitrous oxides (NOx), and carbon dioxide (CO2). Other examples where TIM activities benefit the environment may include reduced use of products to treat hazardous materials spills or quicker extinguishing of vehicle fires. The Minnesota DOT Freeway Incident Response Safety Team (FIRST) program reduced incident duration from an average of 7.5 minutes for

16 Guidelines for Quantifying Benefits of Traffic Incident Management Strategies incidents less than 57 minutes, resulting in 662,000 gallons of fuel savings. This fuel savings translates to 270.4 tons of CO, 5.2 tons of HC, and 14.7 tons of NOx (Minnesota DOT, 2004). Efficiency Outcomes Efficiency outcomes can be achieved through TIM activities that streamline operations, ensuring that the right set of response personnel and vehicles is dispatched and that inci­ dent responder activities are executed more efficiently, leading to fewer equipment and staff resource­related expenditures. For example, a study in Maricopa County, Arizona, through the creation of the Regional Emergency Action Coordination Team Evaluation in 2002, high­ lighted the reduction in police and fire personnel responding to long­duration incidents on arterials by 20 to 40 hours per week (Battelle Memorial Institute, 2002). An assessment of personnel cost from a dedicated incident response team compared to response by maintenance crews identified a 24% reduction in personnel costs with the dedicated incident response program within the Oregon DOT Region 3, District 8 (Pecheux et al., 2016). The Highway Emergency Response Operator (HERO) program in Austin, Texas, increased the availability of Austin Police Department officers by 1,000 man­hours because 3,918 HERO responses did not require a police response (Central Texas Regional Mobility Authority, 2011). Traveler Satisfaction Outcomes Traveler satisfaction is an important outcome that may indirectly or directly factor into funding decisions for a TIM program. TIM programs have benefited the traveling public through educational outreach (e.g., move over laws), traveler information, and safety service patrol. The traveler satisfaction derived from these activities can be measured through surveys; however, this benefit is typically not monetized. Similarly, the traveler satisfaction benefit from safety service patrol is typically documented using motorist comment cards but not monetized. Tennessee, for example, has reported that of 725 comment cards regarding their Highway Emergency Local Patrol (HELP) service patrol in fiscal years 2013–2014, 99% rated the service as “good” or “excellent” (Tennessee Department of Transportation, 2014). How to Quantify TIM Outcomes? Quantification of TIM outcomes is the estimation of outcomes through the application of analysis techniques and tools. The most common quantified TIM outcomes include a reduction in the hours of delay, gallons of fuel consumed, and the number of secondary incidents. TIM agencies typically compute these three outcomes based on factors related to incident response performance (e.g., number of incidents and reduction in ICT), traffic operations, and roadway characteristics. In some cases, secondary incidents are identified by the responder in real time rather than estimated. This value is used as a basis to estimate the reduction in secondary incidents. The five general classes of methods to quantify the outcomes are: • Spreadsheet and Rule of Thumb—This approach is based on estimated characteristics of TIM programs without significant data and analysis. Rules of thumb often apply values, factors, and percentages to estimated outcomes. These computations are often straight­ forward, but the results are typically only intended as ballpark estimates and typically do not support detailed operational decision­making. Appendix A presents some of the most commonly cited rules of thumb for TIM and summarizes their applicability. Appendix B presents a simple­sketch spreadsheet method for estimating the mobility benefit from TIM in the absence of TIM data.

The Language of Traffic Incident Management 17   • Existing Tools—Some tools are available to quantify one or multiple outcomes. These tools offer an effective means to derive more precise estimates than spreadsheet or rule of thumb tools without requiring advanced mathematical, statistical, or simulation expertise. Often tools are based on empirical analysis or the application of simulation models to derive a relationship between incident and roadway characteristics and resultant delay. Tools may also provide quantified and monetized outcomes. • Mathematical and Statistical Modeling—This approach applies traffic flow and queuing theory, and HCM or local estimates of effective road capacity from lane blocking to estimate delay. The precision of outcomes is directly correlated to the types, quality, and completeness of data applied to define and calibrate the model. • Simulation—Methods in this category apply traffic simulation software to estimate delay and emissions effects from incidents of varying duration and capacity loss. Simulation requires significant expertise, effort, and data to calibrate and validate. This method for estimating outcomes, if conducted adequately, allows robust estimates of TIM outcomes. • Empirical Analysis—This approach measures incident frequencies, response times, delays, and resource use prior to and subsequent to the implementation of a TIM activity to measure changes in outcomes. While empirical analysis provides ground truth, the costs associated with direct measurement for a large system often prove untenable. Comparisons can be made between facilities with TIM and similar facilities in the region where TIM activities are not established. The following subsections summarize the key methods by which to quantify efficiency, traveler satisfaction, mobility, environmental, and safety outcomes. Quantifying Efficiency and Traveler Satisfaction To the extent feasible, traveler satisfaction and efficiency outcomes should be quantified using empirical analysis. This is done by tracking data before and after TIM implementation or by comparing data in regions with or without TIM. Traveler satisfaction metrics should include the number of motorists served through outreach and safety service patrol and the quality of service, which can be achieved through comment cards or surveys. Efficiency metrics can include the types, frequency, duration of equipment dispatched or used for incident response, as well as responder hours and other resources used. Quantifying Mobility Outcomes Quantifying delay reduction from TIM activities has the greatest variance in the method applied to estimate its value. Some agencies use rules of thumb and simple spreadsheet estimates, others develop mathematical models, and others develop and apply simulation. Empirical quantification of delay has been pursued, but mainly for a short time to validate the development of mathematical models. For example, a study conducted by the University of Washington for Washington State DOT used video extraction to empirically measure incident delay. This delay was compared with a deterministic queuing model that uses data from loop detectors prior to and subsequent to an incident to quantify delay (Wang et al., 2011). The three basic data points needed to estimate the change in delay using mathematical and statistical methods are duration of the incident, reduction in roadway capacity during the incident, and motorist demand. Many studies apply the capacity loss factors and delay esti­ mation process prescribed by the HCM. As a part of this study, the research team examined the capacity loss from freeway incidents with the varying number of lanes closed during peak traffic. The HCM capacity loss and proposed adjustments to these factors are presented in Appendix C.

18 Guidelines for Quantifying Benefits of Traffic Incident Management Strategies PARAMICS, VISSIM, and CORSIM are three commercially available simulation tools that have been applied to quantify delay and emissions outcomes from TIM activities. Many TIM programs have invested in the conduct of simulation models to derive measures of delay and other metrics. These include the Hoosier Helper, Georgia’s HERO, Maricopa County DOT’s Regional Emergency Action Coordination Team (REACT), New York State HELP, I­64 Missouri Arterial Service Patrol, Minnesota FIRST, and the Maryland CHART programs. Some of these programs have also conducted large sets of simulations to derive statistically significant relationships between incident, demand, and facility characteristics. These relationships, typically in the form of a simple equation, serve as the means for estimating delay by incident for their respective region. These simulation tools enable the capture of a robust set of delay estimates based on varying capacity loss from incidents, demand, and roadway characteristics. Simulation model calibra­ tion and validation require significant data, time, and expertise. Key inputs to each simulation effort are the estimate of capacity loss for each category or type of incident, the reduction in incident duration from TIM, and facility demand. Typically, the performing organization is a university research center, and the sponsoring organization is the state DOT. Most frequently, agencies apply mathematical and statistical models and simulations or use existing tools. The most recently developed tool, TIM Benefit­Cost (TIM­BC), estimates mobility, environmental, and safety outcomes for a number of TIM activities. This tool is assessed and compared with other mathematical and statistical methods assessed as a part of this study to estimate delay benefits from TIM. A description of the tool and comparative findings are presented in Appendix D. Below is a brief summary of tools that support TIM assessment: • Rutgers Incident Management System (RIMS) is an integrated incident management and traffic simulation tool developed in 2009 for the New Jersey Department of Transportation (NJDOT). It uses a cell transmission simulation component to estimate travel times, and has the capability to generate incidents to test various incident management activities and technologies. The key outcome captured by RIMS is delay savings. This tool has been applied to NJDOT facilities. • FREEVAL (FREeway EVALuation) 2010 is an Excel­based tool using VBA code to repre­ sent the HCM computation for travel time to compute delay (vehicle hours of delay). The tool requires segment­level data at 15­minute intervals, including segment type (onramp, mainline, etc.), length, lane count, speed, demand (vehicles per hour), percentage of trucks, and percentage of recreational vehicles. A revised input with changes in segment capacity, demand, and/or speed represents the incident scenario. A second revised input representing TIM may be reflected with fewer time durations exhibiting capacity loss. A comparison of these inputs provides an estimate of the delay reduction from TIM. This tool requires detailed temporal data. A complementary tool, FREEVAL­TR, estimates the reliability outcomes. • The Tool for Operations Benefit/Cost (TOPS-BC) is a spreadsheet­based tool made available in 2013 by the FHWA Office of Operations. Inputs to estimate TIM benefit include aver­ age volume, number of lanes, roadway capacity, free flow speed, link length, and percentage reduction in the incident response time. This is a sketch planning tool ideally used to estimate benefits and conduct a benefit­cost analysis during the visioning or long­range planning activities. The tool documentation advises that for more precise analysis during transportation improvement program or project development phase of activity, a more robust tool should be considered. • The Incident Management Assistance Patrols (IMAP) integrated cost­benefit estimation tool was developed by the Institute for Transportation Research and Education at North Carolina State University through funding from the North Carolina Department of Transportation

The Language of Traffic Incident Management 19   (NC DOT). This tool uses traffic volume, incident attributes, and IMAP incident log data inputs. The tool incorporates FREEVAL to estimate delay, fuel consumption, and crash reduction. The tool also includes a process to estimate TIM costs. This tool is geared for prioritizing and expanding freeway service patrol and has been applied to NC DOT facilities. • The Freeway Service Patrol Evaluation (FSPE) tool is an Excel–VBA tool developed by the University of California, Berkeley, to quantify the savings due to reductions in traffic delay, fuel, and emissions from safety service patrol. Inputs include service, roadway design, traffic, and incident characteristics. One limitation of FSPE is the inability to input incidents with multiple blocked lanes. The model uses nine incident type designations, which include accident, breakdown, and debris on the right shoulder, mainline, or left shoulder. The tool has been applied by multiple organizations, including California, Virginia, Florida, and Illinois. • The Traffic Incident Management Benefit-Cost (TIM-BC) tool is the newest and TIM­specific tool made available by the FHWA Office of Operations that estimates mobility, environmental, and safety benefits. TIM­BC incorporates the formerly published Safety Service Patrol Benefit­ Cost (SSP­BC). This tool has six independent modules that allow users to prepare a BCA for eight TIM activities: 1. Safety Service Patrol (SSP). 2. Driver Removal Laws (DRL). 3. Authority Removal Laws (ARL). 4. Shared Quick Clearance Goals (SQCG). 5. Pre­established Towing Service Agreements (PTSA). 6. Dispatch Colocation (DC). 7. TIM Task Forces (TIM­TF). 8. Second Strategic Highway Research Program (SHRP2) Training (TR). This tool is applied and compared with other mathematical and rule of thumb method applications to estimate delay benefits from TIM using data from the I­495 corridor in Maryland. A description of the tool and comparative findings are presented in Appendix D. Quantifying Environmental Outcomes Typically, environmental outcomes are quantified through spreadsheet analysis that relates pollutant tons to reduction in gallons of fuel use, without direct accounting for factors that affect emissions such as vehicle fleet type, altitude, ambient temperature, or travel speeds. Often, the translation from fuel use and emissions is specified by the state DOT so as to maintain consistency in emissions estimates between TIM, ITS, and other transportation programs. The report by Chang and Igbinosun (2015) estimates the emission of HC, CO, and NOx based on locally calibrated emission rates from the Maryland DOT and total delay. This estimate assumes a 24.18% increase in incident duration under non­TIM conditions. Table 4 summarizes the emission rates estimated through prior analyses. In addition, this strategy also considers engine estimates of CO2 emissions based on the engine type. Another emissions translation method, developed by Morris and Lee (1994), estimated the emission savings of HC, CO, and NOx based on the SSP response count performed by a TIM program during the analysis period. This estimate of emissions savings is also presented in Table 4. Though the Chang and Igbinosun strategy allows easy conversion from delay savings to emission savings, care should be taken in understanding the underlying assumption on the delay estimation. The Morris and Lee (1994) emission estimation method only requires the count of assists performed by the TIM program. Though this method is attractive due to the ease in application, it assumes that all assists have equal emission savings, and the analysis is based on data and analyses from over two decades prior.

20 Guidelines for Quantifying Benefits of Traffic Incident Management Strategies The other commonly assessed environmental measure that is included in TIM benefits analysis is fuel consumption. Recognizing the environmental and economic impact of fuel consumption, many researchers and government agencies have developed complex models for high­resolution estimates of fuel consumption in the transportation sector. The TIM agency may use tools such as MOBILE6 to estimate grams of emission per mile by vehicle type and travel conditions and then apply average fuel efficiency rates based on fleet composition. When an agency chooses to develop a simulation model to estimate TIM mobility outcomes, the same model may also deliver fuel savings and emissions outcomes. Figure 9 presents an example of the relationship between operating speed and emissions factors from the FSPE model. Similar to the method used to estimate emission savings, Chang and Igbinosun (2015) applied fuel consumption rates for cars and trucks based on delay. Here, the researchers used the rates of 0.156 gallon of gas per hour for passenger cars and 0.85 gallon per hour for trucks. These rates were gathered from the U.S. Census Bureau and the Energy Information Administration. The reduced fuel consumption was estimated by applying these rates to the delay savings estimates. Environmental benefits from more specific activities such as quicker extinguishing of vehicle fires have not been quantified by TIM agencies. Quantifying Safety Outcomes Reduction in secondary incidents is the core quantified safety outcome. As presented in the section “What Are Incidents,” secondary incidents are measured by on­site documentation and/or classification of incidents as secondary based on temporal and spatial parameters. Most TIM agencies apply static temporal and spatial parameters to identify an incident as secondary. This is one of the two items required to estimate the reduction in secondary inci­ dents from TIM activities. Pollutant Emissions (tons per hour of delay) Hydrocarbons (HC) 25.676/106 Carbon monoxide (CO) 338.69/106 Nitrous oxides (NOx) 36.064/106 Pollutant Emissions (pounds per gallon of fuel savings) Carbon dioxide (CO2) for gasoline 22.38 Carbon dioxide (CO2) for diesel 22.38 Pollutant Emissions (kg per safety assist) Hydrocarbons (HC) 3.51 Carbon monoxide (CO) 35.84 Nitrous oxides (NOx) 8.85 Table 4. Translation of pollutants by delay and fuel use estimated for Maryland TIM and Morris and Lee.

The Language of Traffic Incident Management 21   The second item is defining a relationship between reduction in incident duration and the reduction in the number of secondary incidents from TIM activities. The simplest method is to apply the same reduction in the percentage of ICT to the reduction in secondary incidents, that is, a reduction in incident duration of 25% equates to a 25% reduction in secondary inci­ dents. From this, one can reverse­engineer to compute the reduction in secondary incidents. For example: Assume that: • 75 incidents are categorized as secondary in the presence of a TIM activity. • TIM activity reduces ICT by 25%. Then: • 75% × (Number of secondary incidents without TIM) = 75 secondary incidents with TIM. • Number of secondary incidents without TIM = 100. • Number of secondary incidents avoided through TIM activity = 25. Some have offered an approach that derives dynamic parameters based on the duration of the incident, incident type, and volume–capacity ratio (Sun and Chilukuri, 2007). Using this relationship, the reduction in secondary incidents is not a perfect relationship with the reduc­ tion in incident duration, but rather in combination with incident type and volume–capacity ratio. A few studies have noted the percentage of incidents identified as secondary prior to and after TIM implementation based on on­site documentation or after­the­fact analysis. For example, the number of secondary crashes from 1987 through 1990 without motorist assistance is compared to the number of secondary crashes from 1993 to 1996 to estimate a 25.9% reduction in secondary crashes. Often agencies apply a rule of thumb for estimating the number of secondary incidents and the reduction in secondary incidents. A common rule of thumb frequently cited is that 20% of all incidents are secondary. As noted in the previous section, and detailed in Appendix A, this percentage is too high. While using temporal and spatial analysis to identify incidents as secondary is advised, a rule of thumb of 10% is more reasonable if the temporal and spatial Figure 9. Pollutants rates and operating speed from FSPE model.

22 Guidelines for Quantifying Benefits of Traffic Incident Management Strategies analysis is not feasible for identifying secondary incidents. This percentage is likely even lower for regions that do not experience heavy congestion. Table 5 provides estimates of secondary incidents found in the literature. Other safety outcomes center on reducing exposure to traffic for responders, including state, police, fire, EMS, and other responder personnel and motorists on scene. This may be quantified by measuring responder personnel­hours outside the vehicle while on the scene or at the scene prior to and after TIM activity to estimate the percent reduction in risk exposure hours. A simpler estimate is to use the reduction in ICT as an equivalent to the reduction in risk exposure. The second parameter is the historic national or program area struck by or fatality incidents. An example using national data follows: • Six law enforcement personnel were killed in 2014 by a vehicle while on duty (Federal Bureau of Investigation, 2015). This computes to a 0.00096 probability of a law enforcement personnel being struck. Program Reporting Year Secondary Incident Rates and Reductions in Incidents from TIM Percentage of Incidents Resulting from a Previous Incident FHWA 2004 Approximately 20% of all incidents are secondary (FHWA, 2004, TIM “Rule of Thumb”) Maryland CHART 1994 5% to 14% of accidents are secondary crashes (Chang and Point-du-Jour, 2003) Maryland CHART 1996 5.7% of incidents are secondary crashes Illinois 1996 5.4% of incidents are secondary crashes (Raub, 1997) • Among crash incident type, 10% are secondary • Among disabled incident type, 9% are secondary • Among traffic violation police stops, 2.5% are secondary REACT, Arizona 2002 10% of total crashes are secondary (Battelle Memorial Institute, 2002) St. Louis 2003 10% of total crashes are secondary (Sun et al., 2009) Maryland CHART 2003 6.8% of all incidents are secondary when one or more lanes are blocked (Chang and Point-du-Jour, 2003) California CHIP 2004 1.5% to 3.0% of incidents are secondary crashes (Moore et al., 2004) Hampton Roads 2011 2.0% of incidents are secondary (Khattak et al., 2011) • 7.5% of incidents had a secondary incident • 1.5% of disabled vehicles result in a secondary incident • 0.9% of abandoned vehicles resulted in a secondary incident Secondary Incident Avoided through TIM Hoosier Helper, Indiana 1999 18.5% reduction in secondary crashes during winter and 36.3% reduction in secondary crashes in other seasons if the primary crash duration is reduced by 10 minutes (Karlaftis et al., 1999) Maryland CHART 2013 24.2% reduction in secondary crashes assuming a linear relationship between secondary crash avoidance and ICT reduction (24.2%) through TIM activities (Chang and Raqib, 2013) Table 5. Secondary incident rates and incident reductions from TIM.

The Language of Traffic Incident Management 23   • By reducing the number of incidents and the duration of exposure, the probability will decrease accordingly. This is computed as 0.00096 × (1 – % reduction in incidents) × (1 – % reduction in time at incident). The difference between the computed risk probability and the 0.00096 is the reduction in risk from TIM activities for law enforcement personnel. The same computation can be performed for fire and motorist risk reduction if data are available at the national or state level. While this set of safety outcomes is typically noted anec­ dotally but not quantified or monetized, a few analyses have applied a process similar to the one above to estimate the reduction in responder risk (e.g., REACT program) (Battelle Memorial Institute, 2002). This quantification in reduction to risk exposure can subsequently be monetized by assigning a value to the cost of this risk, mainly the statistical value of life or the average cost for personal injury. Quantifying Traveler Satisfaction Outcomes Typically, traveler satisfaction outcomes are captured through comment cards distributed by SSP personnel at the completion of an assist. Public feedback can also be solicited during educational outreach with the public, whether at schools or community events. Figure 10 is the comment card from the NJDOT’s SSP program. Quantifying Efficiency Outcomes Efficiency outcomes are captured through detailed record keeping of time and resources applied for the response to incidents as well as post­incident activities among stakeholders, most notably the transportation, law enforcement, and fire entities. Ideally, this information is available prior to and after the implementation of TIM. It may include the duration of time personnel are on­site or completing the incident report, the reduction in unneeded resources through quick identification of response needs, or other areas where efficiency in incident management has been achieved through TIM. Figure 10. NJDOT comment card (NJDOT, 2016).

24 Guidelines for Quantifying Benefits of Traffic Incident Management Strategies How to Monetize TIM Outcomes? Monetization is applying financial values to quantified outcomes. Examples of financial values commonly used in monetization include the price per gallon of fuel, the environmental cost of emissions by ton or kilogram, the value of an hour of a person’s time, and the cost of a crash. Some outcomes are easier to monetize, while others are more difficult to monetize but are still important. For example, data are typically not available to support quantifying and monetizing a reduc­ tion in motorist time exposed to traffic from safety service support; however, an agency can apply national data to estimate a quantified and monetized benefit in reducing the risk of law enforcement­struck­by­vehicle. This outcome as a monetized benefit is typically far smaller than the mobility, environmental, and other safety benefits. Appendix E provides an example to translate the quantified outcomes to a benefit in dollars as well as parameters based on national economic values. The basic factors in this translation include the following: • Vehicle occupancy. • Percentage of traffic that is truck or commercial vehicle. • Value of time for motorists ($ per hour). • Value of time for commercial vehicles ($ per hour). • Cost of incident ($ per incident). • Cost of fatality ($ per fatality). • Cost of personal injury ($ per incident). • Fuel costs ($ per gallon). • Cost for emissions ($ per ton of HC, CO, CO2, and NOx). • Cost per hour by responder personnel types. • Costs of response vehicle. Many of these are defined by the state DOTs or the metropolitan planning organization. When this is not the case, comparable states or national parameters can be used to monetize TIM outcomes. For each of these factors, further differentiation can be defined for greater accuracy of benefits valuation. For example, a temporal variation in vehicle occupancy or percentage of commercial traffic can be defined for peak and off­peak, and the quantified outcomes can also be aggregated by peak and off­peak. Likewise, the cost of incidents can be aggregated by severity (fatality, personal injury, or property damage only). While the transformation of efficiency, mobility, and emissions outcomes is relatively straight­ forward and able to be articulated at more or fewer categories, the transformation of safety outcome has an additional consideration. In other words, is secondary incident severity the same as that of primary incidents? While some analyses suggest that the severity of secondary inci­ dents is greater than that of the primary incidents, others have found the opposite to be the case. If resources allow, TIM programs should examine local data to assess the severity of each type of incident that will be separately evaluated. Finally, in monetizing the safety benefit from reduced risk exposure to law enforcement personnel, the risk level reduction quantified can be multiplied by the value of a statistical life (year 2015 estimate of $9.4 million) and the size of the responder fleet. The number of total personnel should be considered given the statistical rate included all law enforcement, not specifically the subset of officers on duty to respond at any given time.