Prioritization Procedure for Proposed Road–Rail Grade Separation Projects Along Specific Rail Corridors (2019)

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.

5 Overview of Current Railroad–Highway Grade Crossing Separation Evaluation 2.1 USDOT Railroad–Highway Grade Crossing Handbook The traditional approach for making grade crossing invest- ment decisions in the United States has been guided by the 2nd revised edition of the Railroad–Highway Grade Crossing Handbook of the Federal Highway Administration (FHWA) of the U.S. Department of Transportation (USDOT) (FHWA, USDOT 2007) and its predecessor document. The handbook provides guidance for the consideration of grade separations for highway–rail crossings. The handbook addresses two sce- narios: (Case A) the highway–rail crossing must be grade sepa- rated or the crossing eliminated and (Case B) the highway–rail crossing should be considered for grade separation. Figure 2-1 provides a synopsis of the factors evaluated and considered for both Case A and Case B, with the primary factors being the highway facility type and the highway and railway traffic levels. 2.2 Issues and Trends Affecting Railroad–Highway Grade Crossings The 2013 Status of the Nation’s Highways, Bridges, and Transit: Conditions and Performance (FHWA USDOT 2013) estimated that U.S. rail freight tonnage would grow at a com- pound annual growth rate of 0.9 percent through 2040. Inter- modal freight (container-based freight that can use truck, rail, air, and water modes) is expected to increase at a 2.6 per- cent compound annual growth rate through 2040 (FHWA, USDOT 2013). To efficiently accommodate the demand for freight tonnage on the existing track system, trains have become longer and more frequent. In the 2004 Status of the Nation’s Highways, Bridges, and Transit: Conditions and Performance, FHWA dedicated an entire chapter to highway–rail grade crossings. In 2004, the Federal Railroad Administration (FRA) had identified more than 260,000 public and private grade crossings in the United States and noted, “Passive warning devices protect over 78 per- cent of the grade crossings” (FHWA, USDOT 2004). Data from that report showed that in areas with double tracks (two parallel mainline tracks) at-grade crossings were serving as many as 140 trains per day, and the number of crossings serving more than 100 trains per day was expected to double by 2024. The increased frequency of trains at crossings and the expanded length of trains will increase vehicle delay at blocked crossings. FHWA estimated that auto delay at crossings could increase by between 64.4 million and 86.6 million hours annu- ally by 2024. Truck delay over the same time could increase by between 9.9 million and 10.7 million hours, depending on the time of day that crossings are blocked (FHWA, USDOT 2004). Following the enactment of the Passenger Rail Investment and Improvement Act of 2008 (PRIIA), USDOT has more aggressively supported improvements to the intercity passen- ger rail system. Annual trips by public transportation reached 10.1 billion in 2006 because of a resurgence in demand for and use of public modes; this was the highest total since 1949 (FRA 2009). Most passenger rail systems in the United States share track with freight rail. Capacity constraints can lead to poor on-time performance, slower train speeds, and more congestion. Safety at grade crossings has been greatly increased over the past 25 years. According to information obtained from FRA’s Office of Safety, accidents at grade crossings reached their lowest point in 2007 and have remained nearly static since, with minor increases occurring in 2013 and 2014 before a subsequent reduction in 2015. The overall trend for accidents at grade crossings in shown in Figure 2-2. In the case of multiple grade crossings in a short corri- dor, crossing consolidation is another alternative that must be considered. Crossing consolidation requires a more com- prehensive decision-making approach than that required for a single grade separation. Multiple crossings must be evalu- ated simultaneously to optimize safety and mobility of the passengers and goods using the crossing. A corridor-focused C H A P T E R 2

6 Case A. Highway–rail grade crossings should be considered for separation or eliminated whenever one or more of the following exist: Case B. Grade crossings separation considered when the cost can be economically justified based on life-cycle costs and one or more of the following: 1. The facility is an Interstate highway. 1. Facility is part of the National Highway System. 2. The design is for full controlled access. 2. The design is for partial controlled access. 3. Posted highway speed equals/exceeds 70 mph . 3. Posted highway speed equals/exceeds 55 mph. 4. AADT >100,000 in urban areas or >50,000 rural. 4. AADT >50,000 in urban areas or >25,000 rural. 5. Maximum authorized train speed exceeds 110 mph. 5. Maximum authorized train speed exceeds 100 mph. 6. An average is 150 or more trains per day or 300 million gross tons per year. 6. An average is 75 or more trains per day or 150 million gross tons per year. 7. The average is 75+ passenger trains/day in urban areas or 30+ passenger trains/day in rural areas. 7. The average is 50+ passenger trains/ day in urban areas or 12+ passenger trains/ day in rural areas. 8. Train exposure (product of trains/day & AADT) >1 million in urban or >250,000 in rural areas. 8. Train exposure (product of trains/day & AADT) >500,000 in urban or >125,000 in rural areas. 9. Passenger train exposure >800,000 in urban areas or 200,000 in rural areas. 9. Passenger train exposure >400,000 in urban areas or 100,000 in rural areas. 10. Expected accident frequency as calculated by the USDOT accident prediction formula, including 5-year accident history, exceeds 0.5. 10. Expected accident frequency as calculated by the USDOT accident prediction formula, including 5-year accident history, exceeds 0.2. 11. Vehicle delay exceeds 40 vehicle hours per day. 11. Vehicle delay exceeds 30 vehicle hours per day. 12. Engineering study indicates that lack of a grade separation structure will result in the highway performing at a level of service below intended minimum design level 10% or more of the time. Source: FHWA, USDOT August 2007. Figure 2-1. Railroad–highway grade separation criteria (AADT = average annual daily traffic). (USDOT, Report FHWA -SA-07-010) Source: FRA 2016. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 1980 1985 1990 1995 2000 2005 2010 2015 N um be r o f C ol lis io ns /F at al iti es /I nj ur ie s Year Accident Trend in the United States Total no. of accidents No. of accidents on public crossing Injuries on public crossing Fatalities on public crossing Figure 2-2. U.S. trends in accidents at road–rail at-grade crossings (1980–2015). (The difference between the total number of accidents and the number of accidents on public crossings is accidents on private crossings.)

7 approach may also use enhanced public involvement to engage the community in the decision-making process. Consider- ing multiple closures during a corridor-based crossing con- solidation may also reduce the administrative burden on the involved parties by allowing a single review for multiple loca- tions (FRA 2009). 2.3 Road–Rail Grade Separation Decision Process: Literature Review Grade separation projects are costly and can have long- term effects on a neighborhood or community. Locations under consideration for grade separation must be carefully analyzed so that scarce resources provide the greatest benefits to citizens and businesses. While the traditional approach for prioritizing grade crossing separation in the U.S. has largely focused on improving safety, the literature review identified a number of evaluation efforts that involved non-safety criteria when grade separations are prioritized. Examples follow: • A 1997 Israeli study found that vehicle delay due to blocked crossings was a primary contributor to economic loss (Hakkert and Gitelman 1998). • Grade separation prioritization in the Peel Region of Canada is focused on maximizing the efficiency of goods movement. The Peel Goods Movement Task Force identi- fied prioritization of grade crossing separation decisions as a key action item to help maximize freight efficiency (Peel Regional Council 2014b). • Prioritization of crossings in Kern County, California, focused on allowing the county to allocate financial resources to projects that would provide the greatest benefit to traffic flow improvements, freight movement, passenger movement, and safety (Wilbur Smith Associates 2011). The first step in the research process was to identify the factors that others had used in planning for or making deci- sions about grade separation across the United States and in other countries. A review of existing literature was under- taken to help identify those factors. Forty-three factors were identified from the literature. The research team then orga- nized the factors into six broad categories: A. Safety-related factors 1. Accident prediction values 2. Accident history (FRA safety history) 3. Near misses B. Traffic- and delay-related factors 1. Current road delay 2. Future road delay 3. Rail delay 4. Posted highway speed 5. Speed reduction 6. AADT 7. Average annual train traffic (AATT) 8. Train distribution 9. Passenger train count 10. Train speed 11. Train length 12. Exposure 13. Traffic growth 14. Duration of crossing closure 15. Level of service (LOS) 16. Queue length C. Location and geometry related factors 1. Existing land use 2. Distribution of crossings 3. Geometry of crossing/sight distance/clearance time for road vehicles 4. Number of highway lanes 5. Number of rail tracks 6. Adjacent grade separations 7. Warning devices 8. Constructability of the grade separation D. Environmental factors 1. Noise 2. Air quality/emissions/fuel savings 3. Sites of environmental significance E. Community considerations 1. Population/population density (near the crossing) 2. Vulnerable population 3. Transit/emergency/school bus routes 4. Social significance 5. Community cohesion/accessibility/connectivity 6. Quiet zone 7. Strategic fit 8. Local agency priority/isolated location 9. Visual amenities F. Economic considerations 1. Vehicle operating cost/delay and accident cost 2. Crossing operating cost/life-cycle cost 3. Construction cost 4. Economic losses Appendix A lists each of the more than 25 grade crossing studies examined as part of the literature review, arrayed in six tables by category. From the review of previous studies, safety and delay variables are the two most common factors applied to grade separation decisions. Another aspect of the literature review examined the processes or models employed for applying these factors to the decision process.

8 Techniques and Models Applied to Grade Separation Decisions Many evaluation techniques that have been applied to grade crossing decisions were identified during the literature review. While researchers, state agencies, and other govern- ment organizations often use similar factors, a variety of techniques are used for analyzing and prioritizing those factors when investments in grade separations are being made. Through the literature review, the research team iden- tified the following approaches for evaluating grade separa- tion decisions: 1. Multicriteria analysis (MCA), 2. Benefit–cost analysis (BCA), 3. Trade-off analysis, 4. Transport modeling, and 5. Other techniques [network modeling, integrated corri- dor management (ICM), and economic impact of vehicle delay, and resource allocation]. Each of these techniques is reviewed and discussed in the following section. Multicriteria Analysis MCA was the most common approach cited in literature for making assessment and prioritization decisions about grade crossings. The general approach used in MCA is to assign a consolidated score for each crossing in the analysis. The most frequently used criteria when MCA is applied are road traffic, train traffic, and the number of accidents. The criteria are typically variables in a standardized formula used to calculate an index or score for each crossing location. Con- solidated scores for each crossing are based on a weighted average score from each variable. The variables and weights differed across the studies examined in the literature review. An overview of several representative examples of study approaches using MCA identified in the literature are dis- cussed in the following section. Caltrans Section 190 The California DOT (Caltrans) Section 190 Grade Separa- tion Program authorizes approximately $15 million per year for grade separation projects. Funding decisions are based on a priority list of grade separation projects prepared with the use of two formulas. The first formula is used for the cross- ings nominated for separation or elimination. The second formula is used to evaluate the existing grade separations that are in need of alteration or renovation. These formulas are shown below: For crossings nominated for separation or elimination: 0.1 LRT AH 1 SCF (1)P V T C ( ) ( ) = + + +   For existing separations nominated for alterations or reconstruction: 0.1 LRT SF (2)P V T C ( ) = + +  where P = priority index number, V = average daily vehicle traffic, T = average daily freight/commuter train traffic, LRT = average daily light rail train, C = project cost share to be allocated from grade separa- tion fund, AH = accident history, SCF = special conditions factor, and SF = separation factor. As these formulas demonstrate, the Caltrans application of MCA is used to develop a priority rank for grade separation. The formula weights vehicular and train volumes at crossings along with project costs, accident history, and special condi- tions or separation factors (used to consider factors such as sightlines along crossing approaches, the angle of the tracks to the roadway, and traffic delays caused by trains traveling through the crossing). Riverside County, California In 2012, Riverside County applied an MCA approach using nine criteria as inputs for prioritization, shown in Table 2-1. The weights that Riverside assigned to the variables were based on the feedback from a technical steering committee consisting of staff from affected jurisdictions. Weights were assigned to the first eight criteria and the criterion of isolated CRITERION WEIGHT 1. Safety 25 2. Existing vehicle delay 15 3. Future vehicle delay 15 4. Emissions 10 5. Residential noise 10 6. Adjacent grade separation 10 7. Local priority 10 8. Project readiness 5 9. Isolated location 25 Table 2-1. MCA factors and weights applied to grade separation decisions in Riverside, California.

9 location was considered as a bonus criterion contributing an additional 5% to the overall score (InfraConsult 2012). Minnesota DOT Increases in the number of crude-by-rail trains rolling through the state prompted the Minnesota DOT to under- take an evaluation of at-grade crossings on oil train routes across the state in 2014. The MCA process used to evaluate crude-by-rail corridor grade crossings had a multipart com- parative score involving three index numbers: • The first index score was a public risk assessment based on population density within 1/2 mile of each crossing. Spe- cial consideration was given to fixed vulnerable popula- tions, such as hospitals, nursing homes, and prisons, and to transient vulnerable populations such as schools. The pres- ence of public service facilities, including fire and police stations, was also quantified in this score. • The second score involved the use of the established FRA Safety Index, a predictive index of possible grade crossing accidents. • The third index provided a current evaluation of the physi- cal conditions of crossings, which may influence accident risk and movements over the crossing. This score, developed from site visits, evaluated sight lines, grades, and approaches to the crossing; the crossing itself; and road surface con- ditions. The condition score could also include unusual situations, such as proximity to refineries, truck terminals, power plants, special event venues, casinos, and chemical or fuel storage (Minnesota DOT 2014). The maximum values possible are 19 points for risk, 15 points for safety, and 10 points for condition. MnDOT used the sum of the three scores to prioritize crossing decisions. Texas A&M Transportation Institute MCA may also be applied in multiple stages, with each stage providing the opportunity to assign additional screen- ing factors than the previous stage to filter crossings to be selected for grade separation. A study report from the Texas A&M Transportation Institute (TTI) (Nichelson, Rex, and Reed 1999) used a multistage MCA for grade separation projects. The first step began with a preliminary feasibility determi- nation. Feasible projects were then scrutinized by a detailed technical analysis. An initial analysis ruled out projects that could not be implemented from a physical, practical or com- monsense basis. The preliminary analysis included six crite- ria. After the feasibility determination, a second step applied five additional factors to determine if the project should be built. The criteria applied in the initial feasibility and second stage are listed in Table 2-2. International Approaches to MCA Studies for both the Peel Region of Canada (Southern Ontario) and Melbourne, Australia, used MCA to evaluate and prioritize grade crossings for separation. The Peel Region used existing and future train and vehicle volumes, develop- ment impacts from future projects, and other region-specific criteria. The method did not consider a combined score, but each criterion was used in the elimination process (Peel Regional Council 2014a). The Melbourne study used a four-stage progressive priori- tization process with each stage eliminating crossings as crite- ria were added (Taylor and Crawford 2009). Stage 1 identified any crossing that could be considered for closure (i.e., local roads with traffic less than 5,000 vehicles per day, not a bus route, and an available alternate route). Stage 2 assessed con- gestion and safety (the study used an exposure rate based on train and car volumes) to evaluate and rank crossings. Stage 3 grouped criteria into four categories: economic, social, envi- ronmental, and strategic fit. Stage 4 applied more reliable traffic data and refined cost estimates. Benefit–Cost Analysis BCA monetizes the impacts of a project as either costs or benefits, which are then discounted over a defined period (e.g., 20 or 30 years) to evaluate a proposed improvement. A benefit–cost ratio can be calculated to compare projects of dissimilar size. A benefit–cost ratio greater than 1.0 indicates a project that is expected to return a net benefit to the facil- ity or area; a ratio lower than 1.0 has a net cost system. The authors note, “A cost–benefit analysis aspires to estimate the Table 2-2. Criteria applied in the MCA grade separation study by TTI. Initial Feasibility Criterion Second Stage Analysis Criterion Physical feasibility Traffic projections Surrounding land development User cost elements Highway traffic/rail traffic Project costs Crash experience or exposure/ crash frequency Social and environmental impacts Cost of improvements / life-cycle cost analysis Traffic projections Excessive cost/project economics

10 profitability of a project from the whole community point of view by quantifying the willingness-to-pay or the willingness- to-accept” (Aoun, El Koursi, and Lemaire 2010). A problem when BCA was applied in early studies related to appropriate monetary values that should be applied to various factors (accident cost, time savings, fuel savings) as well as determination of an appropriate discount rate. J. S. Dodgson conducted early BCA research related to grade separation decisions. Dodgson identified the following key benefits from a grade separation project (Dodgson 1984): • Reduction in accidents, • Reduction in highway and rail traffic delay, • Reduction in grade crossing protection equipment, and • Reduction in surface maintenance costs. In 1993, TTI developed MicroBENCOST software for an NCHRP project. Roper and Keltner discussed the soft- ware developed to perform BCA for highway improvement programs. Grade separation projects are one of the specific project types MicroBENCOST can evaluate. The program compares existing user costs (cost of vehicle delay, vehicle operation costs, accident costs, etc.) to the user costs if the improvements are to be completed. MicroBENCOST requires data be input for construction costs, route segments, value of travel time, vehicle operation costs, accident costs, traffic volumes, traffic distributions, type of vehicles, railroad data including train distribution, train length, and speed of train (Roper and Keltner 1999). Trade-Off Analysis NCHRP Report 288: Evaluating Grade Separated Rail and Highway Crossing Alternatives examines trade-off analysis (Taggart et al. 1987). The research sought to develop a system- atic and credible tool for making decisions regarding alterna- tives for improving deteriorated bridges separating highways and railroads. The alternatives considered included • Replacement of the structure, • Rehabilitation of the structure, • Relocation of the structure, • Construction of an at-grade crossing in place of the structure, • Closure of the road, and • Closure of the rail line. Researchers cited five factors that influence the decision- making process: (1) safety, (2) cost, (3) rail and highway operations, (4) land use, and (5) environmental concerns/ institutional issues. The decision framework presented in NCHRP Report 288 uses a layered approach of four steps. The first step is to structure the problem—including defining the alternatives in engineering terms, followed by three levels of engineering analysis. The first level of engineering analy- sis uses 11 factors, the second includes 20 factors, and the third level of analysis has 33 factors. The project is reviewed after each level of analysis. A problem that is found to be unreasonable, impractical, or inferior is not carried forward into subsequent phases of engineering analysis. Transportation Modeling Various transportation modeling techniques have also been used to assess multiple impacts associated with railroad cross- ings. The Illinois Commerce Commission (ICC) developed a model to determine grade crossing delay (ICC 2002). ICC maintains a database describing the physical, administrative, and operational characteristics of all public, private, and pedes- trian at-grade crossings and grade-separated structures in Illi- nois. The grade crossing inventory and statistical information system (CRISIS) provided the basic data used in the analysis. CRISIS also contains information describing the maximum, minimum, and timetable authorized train speeds and infor- mation about AADT for highway vehicles. These data items form the building blocks to model grade crossing delay. The model also takes into account the presence of a Metra station (Chicago’s commuter rail provider) or a yard near the railway crossing to determine the gate down time. Twenty grade cross- ings were selected as control crossings to validate the model. The number of vehicles per minute was then multiplied by the total number of gate down minutes to estimate the num- ber of vehicles impacted (ICC 2002). Network Modeling Modeling of railroad and highway traffic networks can be useful in understanding traffic behavior. If conducted at the appropriate level of detail, modeling can be used to assess the changes in network flow due to closing or separating rail- road grade crossings. To consider effects of closing a crossing or grade separation on highway and railroad networks, one needs to model the interaction of the two networks. Various models are identified in the literature for trans- portation network simulation. A model describing the inter- action between the highway network and the railroad network should address the following questions: 1. What are the impacts at the network level when a grade crossing is separated? 2. What are the impacts at the adjacent grade crossings due to a grade separation?

11 3. What are the impacts to the rail network and the train operations due to a grade separation? While network modeling offers a more comprehensive view of road and rail interactions, given the focus of this research on rail crossings within a corridor it was determined network modeling approaches were beyond the scope of this project. Grade Crossing Safety Evaluation Tools During the literature review and since the start of the research for NCHRP Project 25-50, a number of tools, sev- eral internet-based and some only recently available, were identified as focusing on assessing grade crossing safety risks. Generally, many of these tools apply similar models using physical and operational characteristics as inputs to estimate accident frequency. Since the purpose of the research for NCHRP Project 25-50 was to expand factors used for grade separation decisions beyond safety, the research team did not expend significant effort evaluating these tools. However, a brief description of these tools follows. GradeDec.NET (USA-FRA) GradeDec.NET is a web-based application and decision- support tool for the identification and evaluation of high- way–rail grade crossing upgrades, separations, and closures. Designed for the needs of federal, state and local authority decision-makers, GradeDec.NET employs benefit–cost meth- odologies to assess grade crossing investment alternatives at the corridor level or in a region (FRA 2018). The modeling framework in GradeDec.NET was developed for FRA based on research findings from the Volpe National Transporta- tion Systems Center, and the National Cooperative Highway Research Program. GradDec.NET is a web-based decision- support tool to maximize the safety return on public invest- ment. The primary uses of GradeDec.NET are for safety analysis and investment analysis. GradeX (Canada - Transport Canada) GradeX (http://gradex.ca/) is an online decision-support tool that identifies crossing hot spots and evaluates safety countermeasures to prioritize safety improvements. Devel- oped by engineering professors at the University of Waterloo, L. Fu and F. F. Saccomanno, GradeX incorporates collision history and two collision prediction models based on nega- tive binomial equations coupled with an empirical Bayes technique (Saccomanno, Fu, and Miranda 2004). All colli- sion history and crossing data are stored within the model and provide users with the ability to do consequence model- ing and ranking of all crossings. GradeX focuses on groups of crossings and provides little adaptability to consider the risks and characteristics of individual crossings. Australian Level Crossing Assessment Model (Australia and New Zealand) The Australian Level Crossing Assessment Model (ALCAM) is an assessment tool used to identify potential risks at level crossings (i.e., at-grade) and to assist in prioritizing crossings for upgrades. The risk model is used to support a decision- making process for both road and pedestrian level crossings and to help determine the most cost-effective treatments. The ALCAM model consists of three separate components: (1) infrastructure module, (2) exposure module, and (3) con- sequence module. When combined, these three components produce a unique risk score for each level crossing All Level Crossing Risk Model (United Kingdom - RSSB) The All Level Crossing Risk Model (ALCRM) was devel- oped by the Rail Safety and Standards Board (RSSB), an independent agency in the United Kingdom charged with conducting research and setting standards for the rail indus- try. The ALCRM was launched in January 2007 after 5 years of research to develop an assessment tool capable of pre- dicting risk at all types of level crossings on the national rail network. In 2016, RSSB contracted for enhancements to the ALCRM algorithms. The ALCRM predicts the risk of colli- sions between trains and all road users, including pedestrians. The model uses decision fault trees to calculate accident fre- quency by taking into account types of hazards, user types, and user behavior. A distinguishing characteristic of this model is that it applies algorithms that account for decreas- ing road traffic speeds as a result of higher road volumes (RSSB 2010). Integrated Corridor Management At the request of the project panel, the research team extended the literature review to explore whether applica- tions of ICM could benefit the research. ICM is an attempt to make optimal use of underutilized infrastructure capacity in urban corridors, primarily for highways (Cronin, Mortensen, and Thompson 2008). Research by the Minnesota DOT con- cluded that ICM is an effective way to reduce congestion and enhance safety by appropriately diverting the traffic to parallel routes containing unused capacity (Minnesota DOT 2012).

12 USDOT partnered with eight large metropolitan trans- portation agencies (Pioneer Sites) to apply a systems engi- neering approach to determine the needs for ICM in key corridors. ICM seeks to improve corridor mobility by inte- grating existing ITS devices and systems, including assets operated by different agencies, into a proactive solution designed to manage demand and capacity across all travel modes (USDOT 2012). ICM is achieved with the help of the integrated corridor management system (ICMS), a suite of tools being developed for USDOT to perform analysis and assist in making operational decisions to utilize capacity in parallel corridors or modes. ICM implementation guidance uses seven phases, starting with conception through retirement of an ICM project. The guidance provides lessons learned from each phase. ICMS implementation is a complicated procedure that by its very nature involves bringing together multiple agencies that per- form different operations. Because of the complexity of ICM, the literature recommends a systems engineering approach be used for implementation, similar to the VEE Development Model used for ITS projects (USDOT, FHWA California Division 2009). To date, most of the published ICM studies center on freeways, and parallel highways or streets. Some efforts have examined parallel passenger rail corridors, but have not focused rail freight corridors. Resource Allocation Members of the research panel guiding NCHRP Proj- ect 25-50 also suggested that the research team examine resource allocation as a potential approach for grade separa- tion prioritization within corridors. Resource allocation gen- erally suggests available resources should be used to maximize the benefits, because the resources often are not sufficient to fund all projects. Resource allocation approaches discussed in the literature include • Ranking by safety-related measures, • Incremental BCA, • Linear programming methods, • Dynamic programming methods, and • Multiobjective resource allocation. The approaches for resource allocation procedures have been summarized in NCHRP Report 486: Statewide Impact of Safety and Traffic Operations Design Decisions for 3R Projects (Harwood et al. 2003) and the Highway Safety Manual of the American Association of State Highway and Transporta- tion Officials (AASHTO) (AASHTO 2010a). The research for NCHRP Project 25-50 aims to develop a prioritization method for grade separation that can benefit from the tech- niques employed in resource allocation studies. Resource allocation approaches were considered in determining priori- tization procedures across the evaluation modules for ranking grade separation projects. 2.4 Survey of State and Local Program Managers Regarding Highway–Railroad Grade Separation Decisions The project team developed a web-based survey for gather- ing information about the current state-of-the-practice in the United States for making investment decisions related to road– rail grade separations. Approximately 240 stakeholders— representing state DOTs, public utility commissions, metro- politan planning organizations, cities, counties, transit agencies, port authorities, associations, and other entities involved in grade separation decisions—were contacted to participate voluntarily in an online survey. The purpose of the survey was to determine the types of issues respondents encounter when making decisions about road–rail grade separation projects. The survey was conducted during March and April of 2016. Of those contacted, 47 stakeholders at least partially completed the survey. The responding 47 stake- holders were located in 16 states, as shown in Figure 2-3. Of the 47 respondents, 81 percent represented state DOTs. Responses were received from 13 DOTs across the United States. Other responding entities included three state-level public utility agencies, a local unit of government, and a state rail agency. Survey Summary of Who Responded and Current At-Grade Crossing Investment Policies • When asked about their familiarity with the topic of grade separation investments, 80 percent of those who responded to the survey indicated that they had an expertise or knowledge of decisions about investment in railroad grade crossings. • Forty-six percent of respondents reported that their orga- nizations had a formal grade separation policy in place. Five of the respondents were able to provide electronic links to their organization’s grade separation policy: – State of Washington, – State of Michigan, – Georgia DOT, – Oregon DOT, and – California Public Utilities Commission. • Forty-six percent of respondents also indicated that their organization had dedicated funding programs to support grade separation projects. Specific funding programs were identified for Georgia, Michigan, Oregon, and California.

13 Survey Summary of At-Grade Crossing Decision Factors and Data Sources • Thirty-eight percent of respondents provided information relating to the various factors or resources that they had previously used to evaluate railroad grade separation proj- ects. Factors were grouped into 12 categories and ranked by frequency of response. The categories and ranking are shown in Figure 2-4. When prompted for information about data that they found most useful for evaluating highway/rail at-grade crossing Figure 2-4. Factors most often identified for making grade separation decisions. Figure 2-3. Geographic distribution of grade separation survey respondents.

14 separations, respondents said cost–benefit information and collision data were the most useful. The hierarchy for use of the remaining data cited is shown in Figure 2-5. Funding availability was used by more than one-half of respondents as a key data source for project evaluation. All other data types were used by less than half the respondents with only 8 per- cent utilizing an internal point system for evaluation. One respondent used political support from city councils and legislators as a data source for the decision process. When evaluating rail grade separation projects, respon- dents used a variety of data sources to develop cost informa- tion to support benefit–cost calculations. The most common cost–benefit inputs included the roadway AADT and train counts at a crossing. More than 90 percent of respondents indicated using these two data sources. More than 77 percent of respondents also used historical accident costs as an input to cost–benefit analysis. More than half the respondents also used hours of vehicle delay and predicted accidents (FRA pre- dicted collisions) when developing cost estimates. • Only 17 percent of respondents currently maintain or had previously used a specific data collection program or tool to support at-grade crossing separation improvement decisions. Examples of data programs include crossing inventories, roadway traffic data at crossings, statewide grade separation studies, and the FRA Web Accident Pre- diction System (WBAPS) (FRA 2016). • Thirty-four percent of the respondents identified spe- cific data gaps and/or limitations with current evaluation methods. Concern with crossing data quality, accuracy, and consistency was the most frequent limitation. Respon- dents would like to see higher quality data pertaining to train volumes (including mainline versus switching traf- fic), school bus volumes, hazardous vehicle counts, and accident history information at crossing locations. • Several respondents commented that funding grade sepa- rations was challenging when competing with traditional roadway capacity projects. One noted, “Political leadership does not necessarily see the value of grade separation proj- ects.” Building local support for grade separation projects was also identified as a challenge. Respondents were asked to identify potential negative effects of a grade separation project (other than construction costs). • More than 80 percent of respondents identified the avail- ability and acquisition right-of-way as a negative effect. Revenue loss for businesses near the crossing was identi- fied by 43 percent of participants. The potential increase of the speed limits on new routes was identified as a nega- tive by 19 percent of respondents. Other negative effects included community severance, environmental impacts, utility impacts, higher maintenance costs, and conflicting user demands between modes. Participants were given a list of potential criteria for decision- making related to grade separations and asked to rate the cri- teria on the basis of their perceived importance for ranking projects. Each criterion was to be rated on a five-point scale with five being the most important and one being least impor- tant. The weighted results from this question are shown in Figure 2-6. Accident prediction information and historical safety records were the two highest rated criteria for decision- 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% What type(s) of data have you found most useful when evaluating highway/rail at-grade crossing separation decisions? Figure 2-5. Data sources identified as most useful for making grade separation decisions.

15 making. Of the given criteria, the emergency evacuation routes and consideration of the transient population rated lowest. • When asked if their organization used a weighting system to rank at-grade crossing separation decisions, 32 percent of respondents said that they did apply factor weights. • When asked to indicate how they weigh the factors used in their evaluation process, number of trains per day and AADT were the factors given the greatest weight. Respondents were asked to identify the causes for devel- oping a grade separation project. Participants were able to select all factors that applied to their organizational decision-making process. The responses are summarized in Figure 2-7. • Of the respondents who selected other, 63 percent were directly related to public/motorist safety; 13 percent iden- tified blocked crossings as a motivating factor. Weighted scores for importance of decision-making criteria Figure 2-6. Scores by importance of decision making. Figure 2-7. Causal factors that contribute to developing grade separation projects.

16 When given an open opportunity to provide additional topics for this research effort, 33 percent of respondents dis- cussed the lack of a dedicated funding program or source and cited needs greater than can be met with existing programs. Additionally, a respondent discussed a desire to explore the efficacy of traffic queue cutter loops in urban areas with sig- nal preemption to prevent traffic from backing up onto the railroad tracks. 2.5 Conclusions from Existing Literature and Stakeholder Input 1. Existing literature about when to invest in rail–road grade separations suggest that there may be better returns from expenditures made for improving a large number of exist- ing grade crossings rather than replacing a select few with grade separations. This issue also needs to be considered before prioritization of grade crossings for grade separation. 2. The cost–benefit analysis reviewed in the literature consid- ered only those costs that could be easily monetized. Social costs due to environmental impacts and noise impacts were not considered in the majority of the reviewed studies. This finding was also collaborated with the responses from state and local officials when asked about cost–benefit data sources. 3. With MCA, determining the weights to be used for each criterion to calculate a scoring for each crossing appears dependent on organizational and/or regional considerations. 4. Much of the literature emphasizes regional conditions when a methodology is being developed for prioritizing the grade crossings. Determining broad criteria for use in corridors for a national study is a more challenging task. This factor is consistent with survey findings that motorist complaints were the primary motivation in making grade crossing improvements. 5. Little has been mentioned in the literature about the change in the economic value of land after the completion of the grade separation project. 6. GradeDec, the FRA application for analysis of investment in highway–rail grade crossings, was developed as part of the Next-Generation High-Speed Rail Program, but none of the state or local decision-makers surveyed as part of this research effort mentioned the application as a tool that they used in making decisions about grade-crossing investment. 7. More than one-third of the survey respondents indicated issues with data quality (e.g., FRA crossing inventory) or lack of accurate data (e.g., trains per day). 8. While survey respondents weighed safety and accident data as the most important for making grade separation decisions, current and future delay to motorists also influ- enced project decisions. 9. One-third of the respondents indicated that the lack of funding for grade separations was an issue.