SAFETY IMPACT OF HIGH FRICTION SURFACE TREATMENT INSTALLATIONS IN PENNSYLVANIA

Transcription

1 0 0 0 SAFETY IMPACT OF HIGH FRICTION SURFACE TREATMENT INSTALLATIONS IN PENNSYLVANIA Kimberley Musey Graduate Research Assistant Department of Civil and Environmental Engineering Villanova University 00 Lancaster Avenue Villanova, PA kmusey@villanova.edu Seri Park, Ph.D., PTP * Assistant Professor Department of Civil and Environmental Engineering Villanova Center for the Advancement of Sustainability in Engineering (VCASE) Villanova University 00 Lancaster Avenue Villanova, PA seri.park@villanova.edu Monica Kares Undergraduate Research Assistant Department of Civil and Environmental Engineering Villanova University 00 Lancaster Avenue Villanova, PA mkares@villanova.edu Date of Submission: November, 0 Paper Submitted to the th Annual Meeting of the Transportation Research Board January -, 0, Washington, D.C. Number of words:,0 Total number of words:, (Max: 00 words) Number of tables: (x 0 = 00) Number of figures: ( x 0 = 0) * Corresponding Author

2 Musey et al 0 0 ABSTRACT Each year, thousands of drivers in the United States are involved in motor vehicle crashes. In order to address this issue, transportation professionals have continued to investigate countermeasures to improve roadway safety including high friction surface treatments (HFSTs). This treatment maximizes the existing infrastructure, and provides exceptional skid resistance in spot locations where friction demand is critical, such as intersection approaches or horizontal curves. Since the early 000s, there has been an increase in state HFST installation projects. This research seeks to evaluate the performance of HFST installation projects in the state of Pennsylvania from both a safety and economic perspective. Using project construction and crash data provided by the Pennsylvania Department of Transportation (PennDOT), it reviews how effective the installations were in reducing both crash rates and crash severity through a beforeafter and benefit-cost study of over 0 sites. The results of these two investigations show that Pennsylvania received the greatest reduction in crash number and severity as well as the greatest return on investment for intersections on horizontal curves that are located in an urban environment. The results, along with further statistical analysis, will be used in future study to develop crash modification factors to quantify the expected crash reduction that can be expected at a particular location by installing HFSTs, with the ultimate goal of helping DOTs maximize return on investment and to better anticipate the safety benefits of HFSTs when programming projects. Keywords: High Friction Surface Treatment, Crash Severity, Benefit-Cost Ratio, Return on Investment

3 Musey et al INTRODUCTION AND BACKGROUND According to the National Highway Traffic Safety Administration (NHTSA), a total of,0 people died in motor vehicle crashes on the U.S. highway system in the year 0 (). According to the Pennsylvania State Transportation Innovation Council, Pennsylvania experienced nearly 00 fatalities and 00 major injuries each year involving crashes on slippery or wet pavement over a recent period of five years (). From statistics such as these, it is evident that transportation professionals need to investigate cost efficient and sustainable methods of improving roadway safety. Since the early 000s, there has been an increase in state HFST installations projects as a safety countermeasure. According to the FHWA, noteworthy practices have been seen in a number of states including Kentucky, South Carolina, California, Florida, New York, and others. Now that HFSTs are continuing to be implemented on such a wide scale, it is important, to evaluate how they are performing, and determine the extent of their effectiveness in reducing crash rates and crash severity. In addition, studies should be conducted to evaluate the long-term benefits versus the project costs. This paper seeks to review the performance of several HFST installations from a safety and economic perspective through an analysis of projects in the State of Pennsylvania. It includes a before-after analysis and benefit-cost study of over sites. The results along with further statistical analysis were also used to develop a number of crash modification factors (e.g. for an urban, rural, intersection, segment, curve, or tangent facility) to calculate the expected crash reduction when deciding whether or not to proceed with a HFST project. This study uses data provided by the Pennsylvania Department of Transportation (PennDOT). The completed research from the state of Pennsylvania will serve as a proof of concept, but can be expanded in the future to include other states. The goal of this study is use results to enable DOTs to better anticipate the safety benefits of HFSTs prior to implementing new projects. The findings can also be used by DOTs to determine the top locations to invest in HFST projects for the maximum safety impact and best return on investment. The findings would greatly benefit transportation agencies when programming projects and carefully investing government funds to meet traffic safety goals. RESEARCH OBJECTIVES The objectives of this research are to: Conduct a comprehensive review and analysis of crash data to further explore the impact of HFST installation on crash rates and crash severity for projects in Pennsylvania; Perform before-and after analyses of the crash data; Conduct a benefit-cost study of HFTS installations; and Establish a basis for future research that includes the development of a crash modification factor (CMF) of pavement friction factor along with other site specific features (e.g. degree of horizontal curvatures, rural vs urban environments, and other roadway features. This paper is organized into major sections that include: a background and reviews of HFST installation projects; detailed crash data description and analysis methodology; resulting dataset findings; and the future direction of this research. LITERATURE REVIEW Pavement Friction

4 Musey et al When dealing with roadways, friction factor is the amount of resistive force between a vehicle tires and the road. The degree of roadway friction is generally quantified using what is known as a skid number. Pavement friction plays a very important role in driver safety. Overall, the less friction that is available, the less a driver can control his vehicle. Friction becomes a problem particularly under wet, icy, or slippery conditions. Even small amounts of water on the road surface can reduce friction by 0% to 0% (). Therefore, friction plays a critical role in pavement and geometric roadway design. High Friction Surface Treatments Horizontal curves, steep grades, and intersection approaches are all locations where drivers tend to brake excessively. As a result, the roadway tends to lose its friction more quickly than other locations. This reduction in pavement friction causes vehicles to skid, depart from the lane on curves, or rear-end leading vehicles when approaching an intersection. HFST are pavement treatment systems that can be applied at these spot locations to provide drivers with exceptional skid resistance and for a much longer period of time than traditional paving (). The greater pavement friction causes the loss of microtexture friction due to the wet weather to be less critical (). Calcined bauxite is the recommended aggregate for use in HFST applications due to their high resistance to polishing (). Other materials that have also been evaluated for this purpose include flint, granite, and slags, which are all commercially available. The binder materials can include Bitumen-extended epoxy resins, epoxy resin, polyester-resin, polyurethane-resin, or acrylic-resin. The installation process can be done with either manual or machine automated mixing (). HFST have a moderate cost compared to other alternatives when considering the life cycle of the pavement. Construction is also relatively short, meaning that construction is likely to have a minimal impact on the general public. It can therefore be a sustainable and cost-effective option to preventing roadway crashes and fatalities. Continuing to incorporate additional pavement friction in design may be a key element in ensuring roadway safety (). Case Studies HFST arrived in Pennsylvania in the year 00. The roadway chosen for this pilot program was a horizontal curve in which vehicles regularly slid on the pavement into either guiderail, or even more dangerously, into opposing traffic. After many attempts at a solution, HST was decided to be applied. The district reported that the results were immediate, with wet-pavement-related crashes at the spot drop from 0 in the years prior to the treatment to zero in the seven years since it has been installed (). These positive results lead to the deployment of several similar projects, which continue to experience similar success in reducing crashes. Pennsylvania, however, is not the only state to experience the benefits of HFSTs. Since the early 000s, there has been an increase in installation projects across the United States. In fact,. As of July 0, 0% of states have utilized HFSTs in at least one location (). In the state of Kentucky, some rural areas are characterized by very mountainous terrain, which presents a challenge to drivers. sites were chosen for the installation of high friction surfaces from the year 00 through 0. The treatments comprised of a compound comprising a two-part, highly modified epoxy resin binder and a specially graded, high-friction bauxite aggregate. According to FHWA s Everyday Counts initiative, a preliminary evaluation of projects shows a % reduction in crashes (). One particular project focused on Oldham County Bridge Hill. This

5 Musey et al particular area had experienced crashes from August, 00 through August, 00, resulting in injuries and fatality. HFST was installed for a total cost of $,00. As of April 0, only crashes were reported. These results among others in this state, led the state to develop an official HFST program that selects sites based on the Empirical Bayes methodology (). The state of South Carolina also experienced positive results from HFST installation. A one-mile section of US was traversed a rural and mountainous terrain, contained sharp horizontal and vertical curves, and experienced high operating speeds. When HFST was installed in 00, wet crashes were reduced by percent and total crashes by percent. Other locations in South Carolina also experienced benefits. One study that considered seven HFST installations, reported that there was an percent reduction in wet crashes and percent reduction in total crashes (). In addition to looking purely at before and after data, some studies have also been conducted that look into the benefit-cost ratios associated with HFST projects. A recent before and after study from the SCDOT performed a similar analysis for a series of horizontal curves. The study revealed benefit-cost ratios of about to. In contract, the State of Kentucky placed HFST on curves, and saw values from. to. at those locations (). The State of Florida also conducted a study identified HFST sites in Florida. These were each analyzed based upon its bidding records, roadway geometry, and crash history. The cost of installing each project was determined by the average HFST unit cost, and was scaled by the size of the application. The savings were calculated based on the average cost per crash severity from the Department of Transportation s KABCO scale. On average, HFST applications on tight curves reduced the total crash rate by % and the wet weather crash rate by %. The average BC ratio on tight curve sections was between and. Wide curve and tangents sections had few accidents initially, and HFST had negligible impact (). HFST Life Expectancy Just like pavement performance, HFST wear is dependent on many things such as initial construction quality, friction demand, and traffic volume as well as the severity of the climate and the weight and number of heavy truck axle (). As a result, life expectancy of is difficult to generalize since it will vary with type of roadway, geometry, traffic volume and type. International data indicates that if it is correctly applied, the expected service life is at least - years. Some data from the United States indicate a service life of over years when applied to bridge deck. For approximately,000 vehicles per day, vendors have reported that the HFST lasted anywhere from - years. For a greater amount of 0,000 vehicles per day, the HFST lasted up to years (). Some states have been experiencing issues with the durability of HFSTs. In Kansas for example, four locations were chosen on both existing asphalt and concrete pavement to evaluate the long term effectiveness and durability of the high friction surface materials. In general, total surfaces are performing poorly, with one of the lower trafficked surfaces failing in less than three years. The surfaces on concrete are peeling off and skid resistance numbers are dropping. One way to improve the durability is through appropriate site preparation and a review of the application specifications (). Sand blasting the surface to remove contaminants and filling any existing cracks is crucial in order to get a strong bond between the asphalt and the HFST that will not separate and create spalls and patching with use over time ().

6 Musey et al 0 0 Risk Analysis Traditionally, engineers have used a nominal approach to safety, and often left to use good engineering judgment to make design choices. However, recent risk assessment methodologies, such as the one proposed by Ray and Carrigan () demonstrate that a more quantitative approach for measuring the risk associated with design alternatives can help to identify where the greatest safety benefit can be realized, which is directly in line with the goals of this research. STUDY LOCATION AND DATA COLLECTION This research utilized crash data provided by the Pennsylvania Department of Transportation. (PennDOT). The crash database includes crash data from PennDOT Districts,,,,, and. A variety of sites were included from rural and urban facilities, intersections and segments, horizontal curves and tangents. The number of lanes ranged from two to six, and speed limits ranged from mile-per-hour (mph) to over mph. Some sites included features like rumble stripes, medians. The data file used in this research contained a total of 0 crashes from the years 00 through 0. Each crash data entry recorded by PennDOT included detailed information on the crashes, including the county, state route, segment and offset where the crashes took place, the date and time of the crash, lighting, roadway surface conditions, weather, crash severity, environmental roadway factors, vehicle events, and driver actions. This information was crucial to analyze how each of these features correlated to the occurrence of crash events. In addition to the crash database, PennDOT also provided a database of its HFST projects throughout the state, both planned and completed. As of March 0, this included 0 locations where HFST projects were completed, and 0 that were planned or in construction. These projects are in nearly all counties in the state. In this database, each project site has a Multi-modal Project Management System (MPMS) number, district, county, state route, segment, offset, targeted crash type, actual/anticipated completion date, funding source, and amount of funding. METHODOLOGY This paper seeks to review the performance of HFSTs from a safety and economic perspective through an analysis of HFST installation projects in the State of Pennsylvania. Only projects with a completion date of 0 or earlier were included in this study in order to ensure at least one year of crash data after HFST installation necessary to compare to the crash history prior to the installation. In addition, only locations where geometric features can be found were included. This reduced the dataset to approximately locations from the original 0 locations, and are shown in Figure below.

7 Musey et al 0 FIGURE Map of Study Site Locations The first step of was the crash data compiling and database construction process. This involved collecting the crash records provided by PennDOT, followed by a review to the ensure accuracy of the data transfer. Then, a second database was constructed for the details regarding HFST projects throughout the state. This process involved the use of PennDOT s multimodal project management system (MPMS) query, Engineering and Construction Management System (ECMS), Internet Traffic Monitoring System (itms) and google earth to obtain roadway features for each site. This information includes lane number, shoulder information, clearance zones, median type or other barriers, segment length, the presence of rumble strips, speed limit, AADT, rural/urban, functional classification, whether it was an intersection or segment, tangent or curve, and curve degree if applicable. During this step, each project site was assigned a unique site identification number. After this data was also filtered and reviewed, steps and of this research involved the before and after analysis. This began by going through all each of the,0 crashes individually, and correlating them to its respective site identification number based on the county, state route, segment, and offset. The site ID and the HFST installation date from the second database to determine whether the crash occurred before or after the HFST was installed, and then to calculate the number of years between these dates. From here, two before-and-after analyses were performed, one limiting the crash data to the same number of years before and after the HFST installation, and one analysis where the data is averaged based on the years of data available before and after installation. This is performed for each project site when the data is available. Step was a benefit-cost analysis. This analysis used the results of the two before-after analyses along with project cost information from ECMS and the MPMS Query. Costs were obtained by dividing total project cost proportionally to the site segment lengths within each project to determine the cost of each site. The benefits were determined based upon the reduction in crashes by severity levels.

8 Musey et al The last step focuses on the development of Crash Modification Factors for the effect of HFST. A summary of this methodology is presented in Figure. It should be noted that step is still under investigation and hence, this paper presents findings through step. STEP : Database Building STEP : Same Period Before-After Study PennDOT Crash Data PennDOT ECMS, itms, and MPMS Query Construction Data STEP : Averaged Before-After Study STEP : Economic Analysis STEP : Crash Modification Factor 0 FIGURE Research Process Flow Chart DATA ANALYSIS AND RESULTS Before-After Analysis As defined in the methodology portion of this report, the before-after analysis was divided into two comparisons. One analysis involved limiting crashes per site to include only those with the same before period of time before and after the HFST installation. The second comparison considered the average number of crashes per year before and after HFST installation. Both of these were to ensure that the comparison of crashes before and after was fair. This section will describe the results of each investigation. Total Data Analysis The top portion of Table shows the results of the simple before-after analysis. For this investigation crash data was limited to crashes within the same period before and after the HFST installation. For more recent projects, data was often limited by the available crash data after the project was completed. For example, if HFST was installed on January st 0, a crash database ending in June 0 would mean. years of after crash data. Therefore, the before crash data

9 Musey et al 0 was limited to only include crashes. years prior to the January 0 install. For the sites, the periods used varied from less than a year to four years. TABLE Before-After Analysis Before After Crash Reduction % Reduction Simple Before-After Analysis Fatal Major Injury -. Moderate Injury -. Minor Injury - 0. PDO -.0 Unknown Injury -. Total -. Average Crashes Per Year Fatal Major Injury -. Moderate Injury -. Minor Injury PDO -0.0 Unknown Injury -. Total - 0. As a result of the simple before-after analysis, the total percent reduction in crashes across all severity levels was.%. The installation of HFST was able to eliminate fatalities by 0% at all of the sites. The severity level with the least percent crash reduction was major injury, however, installing HFST still reduced crashes here by.%. Overall, the total data showed that HFST improved roadway safety by reducing both the number and the highest severity of crashes. The bottom portion of Table shows the before-after analysis for the averaged data. As mentioned previously, the number of crashes before and after are divided by the total years of before and after crash data respectively. Therefore, the results are represented in average crashes per year. The total percent reduction in crashes across all severity levels was 0.%. Once again, the installation of HFST was able to eliminate fatalities by 0% at all of the sites. The severity level with the least percent crash reduction was major injury, however, installing HFST still reduced crashes here by.%. Overall, the total data in both analyses showed that HFST improved roadway safety by reducing both the number and the highest severity of crashes. Individual Site Analysis After the analysis of the data as a whole, the sites were analyzed on an individual basis. First examined was in the framework of the simple before-after analysis. Of the total sites, were able to eliminate total crashes among all severity levels by 0%. sites actually saw an increase in total crashes; however, the maximum increase was only by crashes, and the severity of crashes after installation tended to be much lower with the majority being PDO crashes and none of the crashes being fatal. Locations with the highest crash reductions were also investigated to identify any similarities between the geometric and operational features of these sites. From this examination, it is of note that 0% were in urban areas, % were intersection related crashes, and

10 Musey et al % were associated with a horizontal curve, with low and medium curvature sites experiencing the greatest reduction. After the simple before-after analysis was the individual site investigation for the averaged data. Of the total sites, were able to reduce total crashes among all severity levels by 0%. sites actually saw an increase in one of their crash categories; however, the maximum increase was only by. average crashes per year, and the severity of crashes after installation were much lower with the majority being PDO crashes, a few minor injuries, and none of the crashes being fatal. Locations with the highest crash reductions were also investigated to identify any similarities between the geometric and operational features of these sites. From this examination, it is of note that % were in urban areas, % were intersection related crashes, and % were associated with a degree of horizontal curvature, with low curvature sites experiencing the greatest reduction. Impact of Curvature The next area that the dataset examined was the impact of horizontal curve degree on crashes in locations where HSFT were considered. Examples of high, medium, and low curvature from the PennDOT installation sites can be found in Figure below. These categories were defined based on the angle, θ, between the two legs of the curve along with Google imagery. Low curvature was had an angle θ from about 0-, medium curvature from 0-0, and high curvature from FIGURE Examples of Low, Medium, and High Curvature Roadway Sites Table shows the simple before-after analysis when the sites were broken down by curvature. For this dataset, the greatest reduction in crashes, however, was seen by medium curvature roadways, followed by high curvature. However, it is noteworthy that these values are relatively close. Also, it is worth noting is crash severity. All of the fatalities prior to HFST occurred on low curvature roadways, and the installation was able to reduce these fatalities to zero during the analysis period. Therefore, HFST appear to be a good solution for all degrees of curvature.

11 Musey et al TABLE Simple Before-After Analysis by Curvature Crash Severity Before After Reduction in Crashes % Reduction Death 0 0% Major Injury 0 0% Moderate Injury % Minor Injury 0 0% PDO % Unknown Injury % Total % Death Major Injury 0 0% Moderate Injury % Minor Injury 0 0% PDO 0 0% Unknown Injury 0 0% Total % Low Curvature Med Curvature High Curvature Death Major Injury 0 0% Moderate Injury 0% Minor Injury 0 0% PDO % Unknown Injury 0 0% Total 0% Table shows the same curvature investigation for the averaged data. The same guidelines were used in defining curvature categories as in the previous simple before-after analysis. The greatest reduction in crashes, was seen by medium curvature roadways, followed by high curvature. Similar to the previous investigation, however, when considering crash severity, all fatalities occurred on low curvature roadways (an average of per year). Therefore, the ability of HFST to eliminate fatalities until the end of the data collection period shows that such projects may be a good solution on all curvature roadways.

12 Musey et al TABLE Average Crashes Per Year Before-After Analysis by Curvature Crash Severity Before After Reduction in Crashes Death 0 0% Major Injury 0 0% Moderate Injury % Minor Injury 0 0% PDO % Unknown Injury 0 0% Total % Low Curvature Med Curvature High Curvature % Reduction Death % Major Injury 0 0% Moderate Injury 0 % Minor Injury 0 0% PDO % Unknown Injury 0 0% Total % Death % Major Injury 0 0% Moderate Injury % Minor Injury 0 0% PDO % Unknown Injury 0 0% Total % BENEFIT-COST RATIO While a before-after analysis suggests practitioners the types of roadway facilities that experienced the greatest percent reduction in crashes, a benefit-cost analysis will indicate which types of locations will provide the greatest return on investment (ROI). A benefit cost ratio summarizes the overall value for money of a project or proposal. As shown in equation, the ratio is calculated by dividing the expected benefit of proceeding with a project by the costs. B C Ratio = Total Benefit from Crash Reduction Total Installation Costs (Equation ) Having a B/C ratio greater than.0, indicates that the benefit outweighs the cost, and often provides evidence that the project is recommended to proceed. Based on this formula, the greater B/C Ratio, the better the argument that the project was successful and fulfilled its objective. This is particularly important because limited project budgets, so transportation professionals must ensure that resources are used in the most efficient way possible. For this analysis, the cost of the project (per site) was determined based on the material and construction costs to install the HFST. This information was obtained on the PennDOT ECMS website. In general, HFSTs are considered to be an economical approach to roadway safety. When compared with transportation related construction projects with budgets in the million-dollar

13 Musey et al 0 0 range, HFSTs cost between tens or hundreds of thousand dollars. For this dataset, the prices of installation per site ranged from about $,000 to $,000. The benefit of installing HFST was measured based on the lives and injuries saved, which was measured in the reduction in crashes by severity. The average cost associated with each injury level can be seen in Table below, and are based on data provided by PennDOT from the years 0 to 0. These were multiplied by the reduction in crashes for each corresponding severity level, to calculate the total benefit that was derived from the project. TABLE Costs Per Crash Severity Crash Severity Average Cost Fatal $,,.0 Major Injury $,,.0 Moderate Injury $,.0 Minor Injury $,.00 PDO $,.00 Unknown $,.00 For the B/C analysis, the averaged data was used in order to define the average benefit cost ratio per year of installing HFSTs. The benefit-cost ratio for the combined cost and benefit of all projects was.. When the sites were examined individually, the median B/C ratio was 0. and the mean was.. The minimum B/C ratio observed was -., and the maximum was at a value of.. (%) of the sites where HFST was installed resulted in a B/C ratio greater than.0. All of these results indicate that while individual projects may vary quite significantly, the overall value of deploying multiple HFST projects may provide desirable ROI costs. The greatest returns on investment for this dataset were experienced by -lane roadways (which represented the majority of the project sites) with the following ranks and features:. Site # SR 0, Rural, minor arterial, intersection, tangent. Site # SR 0, Rural, minor collector, segment, horizontal curve. Site # SR 0, Urban, minor arterial, segment, horizontal curve. Site # SR 0, Urban, Urban collector, intersection, horizontal curve. Site # SR 0, Urban, minor arterial, intersection, horizontal curve. Site # SR 0, Urban, minor arterial, intersection, horizontal curve The least return on investment for this dataset also occurred on -lane roadways with the following ranks and features:. Site # SR 0, Rural, minor collector, segment, horizontal curve. Site # SR 0, Urban, minor arterial, segment, horizontal curve. Site # SR 00, Rural, major collector, intersection, horizontal curve. Site # SR 00, Rural, principal arterial, segment, horizontal curve. Site # SR 0, Rural, local road, intersection, horizontal curve. Site # SR 00, Rural, major collector, intersection, horizontal curve In terms of ROI, of the top sites were in an urban environment, while of the bottom were in a rural setting. In addition, of the top sites were impacted by an intersection, while of the bottom sites were associated with a roadway segment. Based on these two lists, it may be determined that PennDOT received the best return on investment in urban environments and at

14 Musey et al intersections. Therefore, when taking cost into consideration, these results are in agreement with the two previous before-after analyses. Although horizontal curves were represented in both the top and bottom ranked B/C ratios for the dataset, this may simply be due to the fact that curves represent the vast majority of HFST projects. Therefore, in line with the benefit cost analyses, it can be deduced that in addition to an urban intersection, horizontal curves also serve as critical locations for HFST. CONCLUSIONS AND NEXT STEPS Conclusions This research sought to perform a comprehensive review and analysis of PennDOT crash data to determine the efficiency of HFST projects throughout the state of Pennsylvania. It evaluated their ability to reduce not only crash occurrences, but also crash severity. This was performed by means of two before-and after analyses, one in which the same crash history period was used before and after the HFST installation. The second analysis averaged that data and therefore results were reported in average crashes per year. The two before-after analyses were conducted on a total of sites. For this particular dataset, the HFST was able to reduce the number of crashes by at least % for each degree of curvature and each crash severity. Most importantly, fatalities at all sites were reduced by 0%. Sites with the greatest reduction in total crashes were urban environments, intersections, and horizontal curves. The results of the benefit cost analysis showed very similar results in the selection of roadway facilities. However, comparing the individual sites indicated that while overall return on investment is high when deploying a large number of HFST projects throughout the state, the return on investment for individual projects may not show the same promising results. Overall this paper shows that HFST installations throughout the state of Pennsylvania have been effective in their goal of reducing both crash rates and severity. For this dataset, PennDOT received the greatest results at intersections involving horizontal curvature that are located in urban environments. Therefore, these may potentially be the types of facilities that departments of transportation could target first in order to most efficiently improve roadway safety. Future Research The next phase of this research will involve regression modeling in order to develop crash modification factor that could be used to help predict the safety impact of installing HFST at a site in the future. The study will also be expanded to include a complete crash dataset, and any other HFST project sites that have been completed since output of this report. Afterwards, future research may include crash data and studies in other states and an analysis to determine the trend of crashes after HFST and determine if crash reductions persist or if a learning curve is present. ACKNOWLEDGEMENTS The authors would like to acknowledge the PennDOT for providing the data which was used as a part of this research. The contents of this paper do not necessarily reflect the official views or policies of the State of Pennsylvania.

15 Musey et al REFERENCES. National Highway Traffic Safety Administration. 0 Motor Vehicle Crashes: Overview. August 0. US Department of Transportation. DOT HS. Accessed July, 0.October, 0.. State Transportation Innovation Council, PennDOT. High Friction Surface Treatment. FHWA. Accessed July, 0.. Albin, R., V. Brinkly, J. Cheung, et al. Low-Cost Treatments for Horizontal Curve Safety 0. Federal Highway Administration. FHWA-SA--0. January 0. Accessed July, 0.. Julian, F. High Friction Surface Treatments. 0. Everyday Counts : NACE Conference. 0Fric%0Treatments%0Julian.pdf. Accessed May 0.. Cheung, Joseph. Federal Highway Administration. High Friction Surface Treatments. USDOT. 0_faqs_hfst_mar0_0.pdf. Accessed June0.. Pulugurtha, S.S., P.R. Kusam, K.J. Patel. Assessing the Role of Pavement Macrotexture in Preventing Crashes on Highways. Traffic Injury Prevention, Vol., Issue, 0, pp. -.. Federal Highway Administration. High Friction Surface Treatments (HFST). September, 0. United States Department of Transportation. Accessed October, 0. Federal Highway Administration. Case Study: Kentucky Transportation Cabinet s High Friction Surface Treatment and Field Installation Program. FHWA-SA ytc/ky_hfst 0.pdf. Accessed October, 0.. Federal Highway Administration. South Carolina Case Study: A Cost-Effective and Time- Sensitive Safety Solution. FHWA-SA c/sc.pdf. Accessed October, 0.. Federal Highway Administration. Frequently Asked Questions about High Friction Surface Treatments (HFST). Accessed October, 0.. Wilson, B. T., Brimley, B. K., Mills, J., et al. Benefit Cost Analysis of Florida High-Friction Surface Treatments. In Transportation Research Record: Journal of the Transportation Research Board, Volume 0. DOI:./0-0. Transportation Research Board of the National Academies, Washington, D.C.. Meggers, D. Evaluation of High Friction Surface Locations in Kansas. Paper No. -0 submitted for the th Annual Meeting Transportation Research Board, Kansas Department of Transportation, Topeka, KS, 0, pp..

16 Musey et al. Hall, J. W., Smith, K.L., Titus-Glover, L. Guide-for-Pavement-Friction. In National Cooperative Highway Research Program. Project 0-, Transportation Research Board of the National Academies, Washington, D.C., 00. Ray, M. and C. Carrigan. Use of Risk Analysis to Minimize Adverse Consequences in Nonstandard Designs. In Transportation Research Record: Journal of the Transportation Research Board, Volume. DOI:./-. Transportation Research Board of the National Academies, Washington, D.C.. Harkey, D. L., R. Srinivasan, J. Back, et al. Accident Modification Factors for Traffic Engineering and ITS Improvements. In Transportation Research Record: National Cooperative Highway Research Program, Report, Transportation Research Board of the National Academies, Washington, D.C., 00, pp. -.