Guide for Pavement Friction (2009)

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.

53 CHAPTER 5. PAVEMENT FRICTION MANAGEMENT INTRODUCTION Highway safety management in the U.S. essentially began in 1966 with the passage of the Highway Safety Act. This Act created a unique partnership among federal, state and local governments to improve and expand the nation's highway safety activities. The Act established the State and Community Highway Safety Grant Program (U.S.C. Title 23, Section 402), commonly known as the "402" program. State Highway Safety Offices were created as a result of the legislation and were funded mainly with 402 funds. A typical highway safety management program encompasses all phases of highway life (from roadway design to maintenance), driver attitudes and performance capabilities, environmental conditions, and their influence on the driving task. The aspects of a safety management program of interest to pavement engineers are the design and maintenance of roadways surfaces that enhance highway safety by reducing skid-related crashes (i.e., ensuring there is adequate friction at the pavement–tire interface throughout a pavement service life). To accomplish this goal, highways agencies in the U.S. and worldwide are increasingly interested in setting up or improving pavement friction management (PFM) programs that help ensure adequate levels of surface friction and texture to minimize the risk of skid- related crashes (FHWA, 1980). To be effective, a PFM program must be an integral part of a comprehensive highway safety management program. This chapter presents an overview of the typical PFM program. It describes in detail all key components or elements required in setting up and managing a program, and it provides examples of PFM practices and policies. FEDERAL MANDATES Since 1966, the U.S. Congress has approved several Acts concerning highway safety. A chronological summary of these Acts and associated directives from federal agencies are summarized in the following sections. Highway Safety Act of 1966 (23 USC Chapter 4) The main objective of the Highway Safety Act of 1966 (revised in 1998) was for each state to establish a highway safety program designed to reduce traffic crashes and deaths, injuries, and property damage. The Highway Safety Act specifically mentioned the need for the provision of the following (McNeal, 1995): • An effective record system of crashes (including injuries and deaths resulting from the crashes). • Crash investigations to determine the probable causes of crashes, injuries, and deaths.

54 • Highway design and maintenance (including lighting, markings, and surface treatment). • Surveillance of traffic for detection and correction of high or potentially high crash locations. Highway Safety Program Standard 12 (HSPS No. 12) of 1967 As a result of the Highway Safety Act, the FHWA issued Highway Safety Program Standard (HSPS) 12, “Highway Design, Construction, and Maintenance” (McNeal, 1995). The general objectives of this directive were to ensure “that existing streets and highways are maintained in a condition that promotes safety, and that capital improvements either to modernize existing roads or to provide new facilities meet approved safety standards.” It was further required that each state develop special provisions for high skid-resistant qualities in pavement design and construction and for correction of locations with low skid resistance by providing improved surface characteristics (McNeal, 1995). FHWA Instructional Memorandum 21-2-73 In 1973, the FHWA issued Instructional Memorandum 21-2-73, “Skid-Accident Reduction.” This document changed the federal emphasis from establishing skid-accident reduction programs to evaluating existing programs. The memorandum required every state program to include an evaluation of current pavement design, construction, and maintenance practices to ensure that skid-resistance properties were suitable for the needs of traffic. It also required a systematic procedure to identify and correct hazardous skid- prone locations (McNeal, 1995). 1975 Federal-Aid Highway Program Manual In 1975, the FHWA issued the Federal-Aid Highway Program Manual. A directive, titled “Skid Measurement Guidelines for the Skid-Accident Reduction Program,” suggested that each state’s program consist of the following three basic activities (McNeal, 1995): • The evaluation of pavement design, construction, and maintenance to ensure that only pavements with good skid resistance characteristics are used in construction and resurfacing. • The detection of locations with a high incidence of wet-pavement accidents by utilizing the state accident record system and local accident record system, where applicable, and the development of priorities for correction of the locations. • The analysis of skid resistance for all roads with a speed limit of 40 mi/hr (64 km/hr) or greater, so that skid resistance can be given consideration in development of priorities for resurfacing and maintenance programs. 1980 FHWA Technical Advisory T 5040.17 In 1980, the FHWA issued Technical Advisory T 5040.17, “Skid Accident Reduction Program.” This advisory was a comprehensive guide for state and local highway agencies in conducting skid-accident reduction programs. The purpose of the Skid Accident Reduction Program was to minimize wet-weather skidding accidents through:

55 • Identifying and correcting sections of roadway with a high or potentially high incidence of skid-accidents. • Ensuring that the new surfaces have adequate and durable skid resistance properties. • Utilizing resources available for accident reduction in a cost effective manner. A model for the process was developed and proposed, as shown in figure 23. Intermodal Surface Transportation Efficiency Act (ISTEA) of 1991 The 1991 Intermodal Surface Transportation Efficiency Act (ISTEA) required each state to develop and implement a Safety Management System (SMS) by October 6, 1996. National Highway System (NHS) Designation Act of 1995 The National Highway System (NHS) Designation Act of 1995 removed the ISTEA mandate for states to implement the management systems. States could elect to adopt the systems in whole or in part. Transportation Equity Act for the 21st Century of 1998 (TEA-21) In 1998, the Transportation Equity Act for the 21st Century (TEA-21) was signed into law. TEA-21 basically determined funding levels and formulas to distribute federal transportation trust fund revenues and identified local “high priority” projects. As with past federal transportation Acts, TEA-21 placed considerable emphasis on improving public safety by dedicating over $2 billion to safety programs. TEA-21 placed more emphasis on incentives to improve safety, rather than federal mandates. Specifically, TEA-21 provided the following: • State Highway Safety Data Improvement Incentive Grants to encourage states to take effective actions to improve the timeliness, accuracy, completeness, uniformity, and accessibility of their highway safety data. • Highway Safety Research and Development Program that specifies several new categories of research, including training in work zone safety management; measures that may deter alcohol- or drug-impaired driving; and programs to train law enforcement officers on motor vehicle pursuits. • Infrastructure Safety: TEA-21 designates “the safety and security of the transportation system for motorized and non-motorized users” as one of the seven newly established areas to be considered in the overall planning process, both at the metropolitan and statewide level.

56 Select Sites Representing New and Typical Design Mixes Identify and List High Wet Weather Accident Sites Develop Representative Sampling Plan with Stratification by Highway Type, Area, and ADT Develop Wet and Dry Pavement Times for Highway Location Sample Analyze Wet Pavement Accident Rates Collect Skid Resistance Data List Selected Sites in Sample Calibrate Skid Tester at Test Center Collect Auxiliary Pavement Data as Needed Prepare Skid Number Distribution by Highway Type, Area, and ADT for Representative Sample Prepare Listing of Hazardous Sites by Priority Order Evaluate New and Typical Pavement Mixes Establish Performance of Mixes Conduct Cost-Effectiveness Analysis of Treatments for High- Priority Sites Schedule Highway Projects for Resurfacing and Other Remedial Treatments (within constraints of funds) Provide Feedback to Design, Operations, and Research Implement Projects in Coordination with Safety Improvements, 3R** Pavement Management, Maintenance, and Other Applicable Programs Prepare Annual Report on Program Implementation Prepare Next Year’s Test and Sample Plans * ADT: Average Daily Traffic *** 3R: Resurfacing, Restoration, and Rehabilitation Figure 23. Model skid crash reduction program (FHWA, 1980).

57 2005 FHWA Technical Advisory T 5040.36 In 2005, the FHWA issued Technical Advisory T 5040.36, “Surface Texture for Asphalt and Concrete Pavements.” This advisory issued (a) information on state-of-the-practice for providing friction and surface texture on pavements and (b) guidance for selecting techniques that will provide adequate wet pavement friction and low pavement–tire noise characteristics. The implementation of the various summarized Acts and directives has led to the development of various forms of highway safety management plans and PFM programs. The survey conducted under this study (see appendix C) indicates that most agencies do have some form of PFM. The forms of these programs range from the rudimentary (i.e., periodic friction and texture testing resulting in friction restoration) to the more sophisticated programs that involve routine testing, design and construction guidelines, and research relating friction to skid crashes. The framework of the typical PFM program, along with key elements of the program, is presented in the sections below. PAVEMENT FRICTION MANAGEMENT FRAMEWORK Federal mandates and directives generally allow SHAs some flexibility in developing and implementing a PFM program. However, there are three basic elements that are considered vital to any successful PFM program: • System for evaluating in-service pavements for friction. ¾ Collect and analyze friction data of representative pavement sections within a network to develop an understanding of how effective pavement design, construction, and maintenance practices are in providing good friction characteristics. • System for correlating available friction with wet-weather crashes. ¾ Develop an understanding of how pavement friction properties impact crash risk. • Guidance on the design, construction, and maintenance of pavement surfaces with adequate surface friction throughout the pavement design life. ¾ Utilize pavement design, construction, and maintenance practices that result in good friction characteristics to minimize wet weather crashes. Examples of typical PFM programs are presented in figures 24 through 26. The steps shown in these figures are representative of successful strategies for managing pavement friction. Agencies may vary in the emphasis placed on each of the basic elements of the programs, depending on their current level of understanding of their pavement properties, their access to complete and timely crash data, their ability to collect network friction data, and considerations of the best use of available funds to meet the safety objective.

58 Figure 24. Procedure for identification and prioritization of sites (Highways Agency, 2004). Categorize sites (Table 4.1) and assign Investigatory Levels Define/review local criteria for setting Investigatory Levels Network changes or > 3 years since Investigatory Level review? Review/revise Investigatory Level CSC at or Below Investigatory Level? No Yes No No further action until next CSC measurement Carry out site investigation in prioritized order Carry out SCRIM survey(s) and calculate CSC for site Other indication of increased skidding accident risk Yes Treatment needed? Yes No Consider revising Investigatory Level No further action until next CSC measurement Erect warning signs if required Identify and cost suitable treatment strategy Prioritize and treat sites, taking account of budget and program considerations Site treated? Yes No Add to next years program Remove warning signs No further action until next CSC measurement

59 Figure 25. Overview of a proactive strategy to manage friction on road networks (Austroads, 2005). Figure 26. Simplified example of a friction strategy currently operated by a state road authority in Australia utilizing SCRIM (Austroads, 2005).

60 ESTABLISHING THE PAVEMENT FRICTION MANAGEMENT PROGRAM Defining Network Pavement Sections In a traditional PMS, pavement sections within a network are defined for evaluation based on the consistency of structural capacity defined using characteristics such as structural composition (surface type, layer thicknesses, and so on), construction history, and traffic. Sample units within the sections are then identified for the purpose of inspection, field testing, and evaluation. Defining pavement sections for a PFM program is based on similar principles. The pavement sections must have similar friction demand levels. Therefore, pavement section definition must consider all or any combination of the following factors that influence friction demand: • Number of crashes, crash rate, or crash severity (or a combination of these). • Traffic levels (e.g., a criterion based on a set number of vehicles or trucks per lane or per day). • Highway functional class (e.g., all sites at, or above, a certain functional class in the network will be tested). • Climatic zones (e.g., all sites with a specified range of annual rainfall or number of wet days per year). • High risk locations (e.g., curves, intersections, signalized intersections, railway crossings, sites where guardrail is installed on curves). • Age of surfacing. For both pavement management and friction management programs, the location of pavement sections are identified using information such as route name or number, direction, county, nearby city or town, milepost limits, and/or station limits. The locations of mostly permanent benchmarks (e.g., bridges, underpasses, and interchanges) closest to the pavement section should also be noted as they provide an important reference point. A PFM program can be easily integrated into an existing PMS if the pavement sections are defined such that they match fairly closely. Overlapping of pavement management and friction management sections makes merging the two difficult. However, the PFM program pavement section can be within a pavement management section. Matching these sections not only makes data storage and retrieval less confusing, but also makes it easier to coordinate field inspection/testing needs for both programs. Some agencies categorize and prioritize sites based on engineering judgment and/or local concerns and knowledge. However, such criteria should be defined and applied carefully. The criteria ultimately selected are applied to the network to determine the sites that will be subject to testing. The sections can then be programmed with the aim of ensuring geographic efficiency of testing. Allowing other sites (e.g., those identified for research purposes) to be added to the identified sites should be encouraged. Identifying the level of friction needed by the driving public is the important first step in a PFM program. Although there are no universal criteria for determining the exact level of

61 friction needed for a project, a rational estimate can be developed by evaluating the array of factors comprising three broad categories—highway alignment, highway features/ environment, and highway traffic characteristics. A fourth category, driver/vehicle characteristics, which covers driver skills and age, vehicle tire characteristics, and vehicle steering capabilities, is difficult to assess in terms of friction demand. Discussions of the specific factors comprising the first three categories are provided below. Highway Alignment Friction demand is significantly influenced by both the horizontal and vertical alignment of a highway. Descriptions of each and their impacts on demand are summarized below. Horizontal Alignment The horizontal alignment of a highway is defined by tangents and curves. The typical curves encountered are simple, compound, and spiral. A horizontal curve is used whenever there is a change in highway direction of sufficient length to avoid the appearance of a kink in the highway horizontal alignment. Small changes in horizontal alignment that will not be noticed by drivers usually do not require horizontal curves. The amount of friction required increases with increasing complexity of the highway horizontal alignment (i.e., as the alignment changes from a tangent to a horizontal curve). To counter increasing friction demand in horizontal curves, highway designers increase the horizontal radius of curvature and super-elevate the highway cross-section. However, this does not eliminate the need for additional friction. Figure 27 illustrates the change in highway cross-section as the horizontal alignment transitions from a tangent to a horizontal curve, while figure 28 shows the lateral forces that act on a vehicle as it travels along a curve. As can be seen, the lateral friction developed at the pavement–tire interface is directly related to the square of its speed. As the speed increases, the force required to maintain a circular path eventually exceeds the force that can be developed at the pavement–tire interface and super-elevation. At this point, the vehicle begins to slide in a straight line tangential to the highway alignment, as shown in figure 29 (Farber et al., 1974; Page and Butas, 1986).

62 Figure 27. Change in highway cross-section as the horizontal alignment transitions from a tangent to a curve. Figure 28. Lateral forces that act on a vehicle as it travels along a curve. Figure 29. Lateral sliding. A A B B Section A-A normal crown g g Section B-B full super-elevation g F = (m×V2) / R Point at which Maximum Friction is Exceeded

63 The relationship between side-force friction for horizontal curves (the most critical horizontal alignment), vehicle speed, radius of curvature, and highway cross-section (super- elevation) is defined using the following AASHTO Green Book equation (AASHTO, 2001): Eq. 23 where: FS = Side-force friction demand. e = Super-elevation rate, ft/ft. V = Speed, mi/hr. R = Radius, ft. FS is also a function of climate, tire condition, and driver comfort while performing maneuvers (i.e., braking, making sudden lane changes, and making lateral movements within a lane). Super-elevation is governed by climate (amount of ice and snow), terrain (flat, rolling, mountainous), and frequency of occurrence of slow moving vehicles (that may slip when encountering high super-elevation rates). The maximum super-elevation rate is typically 8 percent, although rates of 10 to 12 percent are common for low-volume roadways in climates with no ice and snow. The side-force friction computed using this equation is universally accepted as representing the level of friction required for safe driving maneuvers and is recommended in the AASHTO Green Book (2001). Vertical Alignment Vertical alignment consists of a series of gradients (grades) connected by vertical curves. It controls how the highway follows existing terrain and its properties are mainly controlled by terrain, horizontal alignment, sight distance, and other factors. The amount of friction required increases with increasing complexity of the vertical alignment (e.g., grade, stopping sight distance). AASHTO (2001) defines stopping sight distance (SSD) as the distance required for a driver (with a 3.5-ft [1-m] eye height) to clearly see an object 0.5 ft (0.15 m) or more in height on the highway with enough distance to perceive, react, and brake to a stop on a poor wet pavement. It quantifies SSD as the sum of two distances—(a) the distance traveled between the time the driver sees an object and strikes the brakes and (b) the distance traveled after braking commences until the time the vehicle stops. SSD is determined using the equation below: Eq. 24 where: SSD = Stopping sight distance, ft. V = Vehicle speed, mi/hr. t = Driver reaction time, sec. G = Longitudinal grade, percent. µ = Coefficient of friction at the pavement–tire interface. ( ) )30( 1.47 2 G VtVSSD ±+××= μ e R VFS −×= 15 2

64 While the first part of this equation is determined based on driver skill, experience, reaction time, and perception, the second part depends, to some extent, on the highway geometry (longitudinal grade) and available surface friction. Highway Features/Environment Highway features/environment is an important but hard-to-measure characteristic of traffic flow that can significantly influence pavement friction. This characteristic of traffic flow is largely defined by the level of interacting traffic situations (e.g., access drives, intersections, entrance/exit ramps), the presence of specially designated lanes (e.g., separate turn lanes at intersections, center left-turn lanes, through versus local traffic lanes), the presence and type of median barriers, and the setting (urban versus rural) of the roadway facility. In general, as the highway environment becomes more difficult and complex, significantly higher levels of friction are required to help drivers perform the necessary maneuvers (e.g., sudden braking). Table 12 provides an example of how criteria can be established for individual highway features/environment factors for different levels of friction demand. Clearly, for a given factor, the greater the difficulty of driving that is imposed, the higher the demand for pavement friction. Table 12. Example criteria for highway features/environment factors corresponding to different friction demand levels (modified from TXDOT, 2004). Example Criteria for Different Levels of Friction Demand Facility Type Highway Features/Environment Factors Mild Moderate Severe Frequency of entrance/exit ramps (number per 1-mi segment) 0 to 2 3 to 4 >4 Designated lanes Full-length (interchange to interchange) entrance/exit lanes Partial-length entrance/exit lanes None Setting Rural Rural/Urban Urban Controlled Access Lateral Clearance (adequacy of median and inside and outside shoulders Unrestricted Partially Restricted Severely Restricted Frequency of access drives (number per 1-mi segment) ≤ 10 11 to 30 > 30 Frequency of signed/signalized intersections (number per 1-mi segment) 0 1 to 3 > 3 Designated lanes Separate turn lanes or turning not permitted Center lane left turn Turn lanes from through traffic Setting Primarily Residential Residential/ commercial Commercial Uncontrolled Access Median Type Wide median (> 20 ft) Narrow median (≤ 20 ft) No median 1 mi = 1.61 km 1 ft = 0.305 m

65 Highway Traffic Characteristics Traffic characteristics that influence friction demand include traffic volume, composition, and speed. Discussions of each are provided below. Traffic Volume As traffic volume increases, the number of driving maneuvers taking place along any given segment increases. The risk associated with these increased maneuvers is elevated, especially in high-speed areas. When traffic volume is increased to the point that congestion occurs (ADT > 7,500 veh/day per lane), the possibility of crashes is aggravated if a highway facility is undivided and traffic speed is high (Page and Butas, 1986; Mahone and Runkle, 1972). Traffic Composition For the same traffic volume, the composition of traffic vehicles (i.e., the percentage of trucks in the traffic stream) can significantly affect friction demand. There are three primary reasons for this phenomenon. They are: • Stopping distances of trucks are significantly longer than stopping distances of passenger cars (AASHTO, 2001). • Trucks have inferior steering capability compared to passenger cars. • Truck tires produce less friction than passenger car tires. Hence, for highway segments where a high percentage of trucks is anticipated, friction demand will typically be higher than a corresponding highway having predominantly passenger cars or lower percentage of trucks. Traffic Speed Vehicle speed is the most important variable influencing friction demand. For wet pavement surfaces, for instance, an increase in truck speed on tangents from 20 to 70 mi/hr (32 to 113 km/hr) results in an increase in truck stopping distance from 50 to 1,200 ft (15 to 366 m) (Radlinski and Williams, 1985). Such an increase in stopping distance significantly increases the risk of a crash. Figure 30 shows a conceptual relationship between friction demand and friction availability for wet pavements. As can be seen, an increase in speed results in an increase in friction demand and a decrease in available surface friction (Glennon, 1996). Speed also contributes to the severity of impact when a collision occurs. For passenger cars colliding with an impact speed of 65 mi/hr (105 km/hr), the likelihood of death is 20 times greater than that associated with an impact speed of 20 mi/hr (30 km/hr) (WHO, 2004). Finally, increasing speed (above 40 mi/hr [64 km/hr]) increases the likelihood of

66 Figure 30. Conceptual relationship between friction demand, speed, and friction availability. hydroplaning, which is a major cause of wet-weather crashes (Glennon, 1996). The speed of vehicles on the highway must therefore be considered in determining friction demand. Highways with higher posted speed limits and overall travel speeds (85th percentile of vehicle speed) require higher levels of pavement surface friction. Establishing Friction Demand Categories Examples of how various agencies categorize friction demand are presented in table 13. Based on the information presented in this table, pavement friction demand categories should be established logically and systematically using the highway alignment, highway features/environment, and highway traffic characteristics factors described above. Ideally, friction demand categories should be established for individual highway classes, facility types, or access types. Also, the number of demand categories should be kept reasonably small (say, 3 to 5 per highway class, facility type, or access type), so that a sufficient number of PFM sections are available for each category from which to define investigatory and intervention friction levels. Pavement-Tire Frictional Capability Friction Demand of Vehicle Speed of Impending Skid Pa ve m en t-T ire F ric tio n Vehicle Speed

67 Table 13. Typical friction demand categories. Site Description Site Category VicRoads/RTA 1996 New Zealand (Transit New Zealand, 2002) Main Roads Queensland, Australia Transport South Australia 2001 United Kingdom (Viner et al., 2004) Maryland SHA (Chelliah et al., 2003) Texas DOT (TXDOT, 2004) 1 • Traffic light- controlled intersections • Pedestrian/ school crossings • Railway level crossings • Roundabout approaches • Approaches to railway level crossings, traffic lights, pedestrian crossings, roundabouts, stop- and-give way controlled intersections (state highway only), one lane bridges (including bridge deck) (High) • Curves with radius ≤100 m) • Roundabouts • Traffic light- controlled intersections • Pedestrian/ school crossings • Roundabout approaches • Difficult sites (steep grades, traffic light approaches, tight bends, roundabouts) • (Q) Approaches to and across minor and major junctions, approaches to roundabouts • Approach railroad crossing, traffic lights, pedestrian crossing, roundabouts, stop-and-give way controlled intersections Rainfall, in/yr >40 ADT, veh/day >15,000 Speed, mi/hr >60 Trucks, % >15 Vertical grade, % >5 Horizontal curve, ° >7 Driveways, #/mi >10 ADT of intersecting roadways, veh/day >750 Cross slope, in/ft >1/4 Design life, years >7 2 • Curves with radius ≤250 m • Gradients ≥5% and ≥50 m long • Freeway/ highway on/off ramps • Curve <250 m radius • Down gradient >10% (Intermediate) • Curves with radius ≤250 m • Gradients ≥5% and ≥50 m long • Freeway and highway on/off ramps • Intersections • Urban arterial roads • (K) Approaches to pedestrian crossings and other high risk situations • Curves with radius < 250 m, downhill gradients >10 percent and ≥50 m long freeway/highway on/off ramp Rainfall, in/yr >20 and ≤40 ADT, veh/day >5000 and ≤15,000 Speed, mi/hr >35 and ≤60 Trucks, % >8 and ≤15 Vertical grade, % >2 and ≤5 Horizontal curve, ° >3 and ≤7 Driveways, #/mi >5 and ≤10 ADT of intersecting roadways, veh/day >500 and ≤750 Cross slope, in/ft 1/4 to 3/8 Design life, years >3 and < 7 1 ft = 0.305 m 1 in = 25.4 mm 1 mi = 1.61 km

68 Table 13. Typical friction demand categories (continued). Site Description Site Category VicRoads/RTA 1996 New Zealand (Transit New Zealand, 2002) Main Roads Queensland, Australia Transport South Australia 2001 United Kingdom (Viner et al., 2004) Maryland SHA (Chelliah et al., 2003) Texas DOT (TXDOT, 2004) 3 • Intersections • Approaches to road junctions • Down gradient 5-10% • Motorway junction area including On/Off Ramps (Low) • Maneuvers— free areas of undivided roads • Maneuver— free areas of divided roads • Rural arterial roads • (R) Roundabout • Approach to intersections, downhill gradients 5 to 10 percent Rainfall, in/yr ≤20 ADT, veh/day ≤5000 Speed, mi/hr ≤35 Trucks, % ≤8 Vertical grade, % ≤2 Horizontal curve, ° ≤3 Driveways, #/mi ≤5 ADT of intersecting roadways, veh/day ≤500 Cross slope, in/ft 3/8 to 1/2 Design life, years < 3 4 • Maneuver-free areas of undivided roads • Undivided carriageway (event-free) • Urban lightly trafficked • (G1) Gradient 5-10% and ≥50 m • Undivided highways without any other geometrical constraints which influences frictional demand 5 • Maneuver-free areas of divided roads • Divided carriageway (event-free) • (G2) Gradient >10% and ≥ 50 m • Divided highways without any other geometrical constraints which influences frictional demand 6 • Curves with radius ≤100 m • (S1) Bend radius <500 m – dual carriageway 7 • Roundabouts • (S2) Bend radius <500 m – single carriageway 1 ft = 0.305 m 1 in = 25.4 mm 1 mi = 1.61 km

69 Data Collection Three key data inputs are required for an effective PFM program: pavement friction, pavement texture, and crash rates. Procedures for collecting these data are presented in this section. Pavement Friction and Texture Following are the key issues in setting policy for a routine friction and texture testing program: • Selection of testing protocol. • Determination of testing frequency. • Standardizing testing conditions. • Test equipment acquisition. • Equipment calibration and maintenance. Testing Protocol At the network level, the locked-wheel friction tester (ASTM E 274) is the most appropriate method of testing. The method is standardized (e.g., test speed, water flow rate), can be performed quickly and at high speeds, and is generally quite repeatable. The method can assess friction and texture by performing tests with both smooth and ribbed tires or with a properly mounted texture laser. Frequency of Testing For a network-level evaluation, it is desirable to test all pavement sections annually because of the year-to-year variation in pavement friction. The testing frequency is determined by the length of network to be tested and available resources. A practical approach is a rolling or cyclical testing regime, whereby portions of the network are tested once every few years (e.g., for a rolling 3-year program, one-third of the network is tested each year). A maximum frequency of 4 years is generally desired. Statistical sampling of pavement sections for network level analysis is an acceptable option, as many agencies cannot test 100 percent of their pavement network due to budgetary and/or other constraints. Testing Conditions Because pavement friction is influenced by various factors, such as pavement surface temperature, test speed, and ambient weather conditions, testing should be performed under standardized conditions to control the effect of these factors on test results. Controlling testing conditions will minimize variability in test results and produce repeatable measurements. The factors presented in table 14 should be considered along with other relevant factors in establishing testing conditions (Highways Agency, 2004).

70 Table 14. Summary of key issues to be considered in standardizing test conditions. Factors Consideration Season for testing Because significant variations in measured friction may occur across seasons within a given year, friction testing should be limited to a specific season or time of year when friction is typically lowest (Highways Agency, 2004). This will help maintain some consistency in year-to-year measurements and reduce variability in measured data. For agencies that cannot perform all testing requirements within a given season, the following can be considered to reduce test variability: • Develop correction factors, as needed, to normalize raw friction test data to a common baseline season. • For a given pavement section, initial and subsequent testing must be done within a specific season (e.g., pavement sections originally tested in fall should subsequently be tested in fall). Test speed The standard speed recommended by ASTM E-274 for pavement friction tests is 40 mi/hr (64 km/hr). However, since most agencies conduct friction tests without traffic control and because posted or operational speeds vary dramatically throughout a network, it is very difficult for the operator to conduct testing at just this speed. For such situations, the operator typically adjusts test speeds to suit traffic conditions and to assure a safe operation. Thus, it is recommended that friction values corresponding to testing done at speeds other than 40 mi/hr (64 km/hr) be adjusted to the baseline 40- mi/hr (64-km/hr) value to make friction measurements comparable and useful. To do this requires the establishment of correlations between friction measurements taken at 40 mi/hr (64 km/hr) and those taken at other speeds (i.e., speed gradient curves). The following equation can be used to adjust friction measurements to FN40: where: FN(S) = Adjusted value of friction for a speed s. FNV = Measured friction value at speed V. SP = Speed number. In order to produce accurate estimates of FN(S), SP must be established for a broad range of pavement macro-textures and texture measuring devices. Test lane and line Friction measurements must be done in the most heavily trafficked lane, as this lane usually carries the heaviest traffic and is, therefore, expected to show the highest rate of friction loss (worst case scenario). For 2-lane highways with a near 50-50 directional distribution of traffic, testing a single lane will suffice; otherwise, the lane in the direction with heavier traffic should be tested. For multi- lane highways, the outermost lane in both directions is typically the most heavily trafficked and should be tested. Where the outermost lane is not the most heavily trafficked, a different lane or more than one lane should be tested. Test measurements must be carried out within the wheelpath, as this is the location where friction loss is greatest. Note that it is important to test along the same lane and wheelpath to maintain some consistency between test results and to reduce variability. If it is necessary to deviate from the test lane and wheelpath (e.g., to avoid a physical obstruction or surface contamination), the test data should be marked accordingly. Ambient conditions Because ambient conditions can have an effect on pavement friction, it is important to standardize ambient test conditions to the extent possible and document ambient test conditions so the measurements can be corrected as needed. The following should be noted when setting ambient conditions for testing: • Testing in extremely strong side winds must be avoided because these can affect the measurements by creating turbulence under the vehicle that causes the water jet to be diverted from the correct line. • Testing must be avoided in heavy rainfall or where there is standing water on the pavement surface. Excess water on the surface can affect the drag forces at the pavement–tire interface and influence the measurements. • Measurements shall not be undertaken where the air temperature is below 41°F (5°C) (Highways Agency, 2004). Contamination Contamination of the pavement surface by mud, oil, grit, or other contaminants must be avoided. pS VS V eFNSFN −− ×=)(

71 Equipment Calibration and Maintenance Proper calibration and maintenance of the friction testing equipment is essential to the collection of reliable friction data. To this end, agencies should follow the manufacturer- specified regime or guidance for calibration and maintenance. Crash Data Crash data form the basis for analyzing pavement friction and texture data; therefore, an efficient system for collecting and analyzing crash data is critical to a successful PFM program. The quality of the crash data must be as high as possible to be useable in analysis. Crash data are generally available from an agency’s crash database or from other sources, such as law enforcement agencies and statistical bureaus. Although the specific information required for the PFM program depends on the program’s objectives, key inputs required to classify and describe crashes include (1) the location (route, milepost, direction) of each crash, (2) vehicles involved along with their characteristics, (3) drivers and passengers involved along with their characteristics, (4) ambient weather conditions at the time of the crash, and (5) injury levels and property damage as a result of the crash. To get the type of crash information needed to monitor safety throughout a highway network, the crash data must be processed into useable statistics, reportable across defined segments of roadway. Examples of useable statistics are presented in table 15. Table 15. Examples of useable crash statistics. Control Data for a Given Friction Demand Category Crash Statistics Project Specific Local Agency Regional/State National No. Crashes in X Years of Analysis Period Crashes Per 100 mi Crashes Per 106 veh-mi Serious Injury Ratio Wet Crash Rate (WCR), percent Wet Skidding Rate (WSR), percent Skid Crash Rate, percent ƒ WSR, which was developed for the national highway network in the UK in the early 1980s, has been found to give a high degree of correlation with changes in skid resistance, and is defined as: WSR (%) = (no. of skidding crashes in wet conditions/no. of crashes in wet conditions) × 100. ƒ WCR (%) = (no. of crashes in wet conditions / total no. of crashes) × 100. ƒ Crashes per 100 mi = (average no. of crashes per year / site length in mi) × 100. ƒ Crashes per 106 vehicle mi = average no. of crashes per year / (site length in mi × vehicles per day × 365 / 106). ƒ Serious injury ratio = crashes where a person was killed or seriously injured / total no. of crashes. ƒ Skid crash rate (in percent) = (no. of skidding crashes / total number of crashes) × 100.

72 Crash data must be stored in a structured databank, so that each individual crash location can be related to a unique PFM pavement section. The databank must be compatible with an agency’s PFM program and must contain sufficient amounts of data for meaningful analysis (i.e., contains crash data for a minimum of 10 years). It is essential to establish protocols that describe which institutions are responsible for different crash-related data within the highway agency jurisdiction and how the data can be collected. The amount and quality of data vary from institution to institution. For example, the number of crashes reported by the police could be much less than the number reported by hospitals, which could be seen as low compared with the data from insurance companies. Hence, there should be protocols established to cross-check data from all sources and to determine the best sources and the most accurate data to be included in the database. Figure 31 provides an illustration of Iowa’s highway safety data integration and analysis system, and shows how crash-related data collected from various agencies are integrated, analyzed, and used in safety management. Figure 31. Illustration of Iowa’s highway safety data integration and analysis system (Iowa DOT, 2005).

73 Data Analysis Establishing Investigatory and Intervention Friction Threshold Levels Because conditions and circumstances along a highway change, there is no one friction level that defines the threshold between “safe” and “potentially unsafe.” Although the ideal situation is to have friction supply meet or exceed friction demand over the entire system, such a practice would be prohibitively expensive (as well as largely unnecessary) and would not generate the cost-benefits associated with a better-targeted strategy. A more practical approach, therefore, is to maintain an appropriate level of pavement friction for all pavement sections within the highway network, based on each section’s friction demand. This approach ensures the provision of adequate friction levels for a variety of roadway (intersections, approaches to traffic signals, tight curves) and traffic conditions. In a PFM program, the adequacy of friction is assessed using the two distinct threshold levels defined earlier in this chapter—investigatory and intervention. Pavement sections with measured friction values at or below an assigned investigatory level are subject to a detailed site investigation to determine the need for warning or remedial action, such as erecting warning signs, performing more frequent testing and analysis of friction data and crash data, or applying a short-term restoration treatment. For pavement sections with friction values at or below the intervention level, remedial action may consist of either immediately applying a restoration treatment or programming a treatment into the maintenance or construction work plan and erecting temporary warning signs at the site of interest. The establishment of investigatory and intervention levels requires detailed analyses of micro-texture and macro-texture data, and crash data, if available. Such analyses must be carried out separately for each friction demand category established by the agency. Presented in the sections below are three feasible methods for setting investigatory or intervention friction levels, either in terms of FN or in terms of IFI(F(60),SP). These methods are derived from many years of discussions at national and international meetings and workshops on pavement friction (e.g., ASTM E 17, TRB AFD90, PIARC TC 1 [now T4.2], and the NASA Wallops Friction Workshops). It is recommended that one of these methods be used in identifying deficient or potentially deficient PFM sections. Establishing Thresholds Using Historical Pavement Friction Data Only (Method 1) This method uses historical trends of friction loss determined by plotting friction loss against pavement age or time for a specific friction demand category. The investigatory level is set at the pavement friction value where friction loss begins to increase at a significantly faster rate. The intervention level is then set at a certain amount (e.g., five F(60),SP or five FN points) or percentage (e.g., 10 percent) below the investigatory level. The friction value at which friction loss begins to increase rapidly can be determined graphically or through the use of analytical/statistical methods. An example graphical based method includes the following steps:

74 • Step 1—Plot pavement friction versus age/time for a given friction demand category (figure 32). • Step 2—Develop a friction loss deterioration curve based on the measured data. • Step 3—Graphically determine the slopes of the three stages of the S-shaped friction loss versus pavement age/time relationship. • Step 4—Set the investigatory level as the friction value where there is a significant increase in the pavement friction loss. • Step 5—Set intervention level at a certain value or percentage below the investigatory level. Figure 32. Setting of investigatory and intervention levels for a specific friction demand category using time history of pavement friction. Establishing Thresholds Using Both Historical Pavement Friction Data and Crash Data (Method 2) This method compares historical pavement friction and crash data for the given friction demand category for which levels are being set. Figure 33 shows a plot of friction and wet- to-dry crash trends for a specific friction demand category. The investigatory level is set corresponding to a large change in friction loss rate while the intervention level is set where there is a significant increase in crashes. Pavement Age, years F( 60 ),S p o r FN Investigatory Level Intervention Level Friction F( 60 ),S p o r FN

75 Figure 33. Setting of investigatory and intervention levels for a specific friction demand category using time history of friction and crash rate history. Establishing Thresholds Using Pavement Friction Distribution and Crash Rate–Friction Trend (Method 3) This method uses the distribution of friction data versus the crash rates that correspond with the friction for the category of roadway for which the levels are being set. An example of using this method includes the following steps: • Step 1—Plot a histogram of pavement friction for a given friction demand category, based on current history. On the same graph, plot the current wet-to-dry crash ratio for the same sections as the friction frequency distribution (figure 34). • Step 2—Determine the mean pavement friction and standard deviation for the pavement friction frequency distribution. • Step 3—Set the investigatory level as the mean friction value minus “X” standard deviations (say, 1.5 or 2.0) of the distribution of sections and adjust to where wet-to- dry crashes begin to increase considerably. • Step 4—Set intervention level as the mean friction value minus “Y” standard deviations (say, 2.5 or 3.0) of the distribution of sections and adjust the level to a minimum satisfactory wet-to-dry crash rate or by the point where the amount of money is available to repair that many roadway sections. Pavement Age, years F( 60 ), Sp o r FN Investigatory Level Intervention Level C ra sh R at es Friction Crash Rates F( 60 ), Sp o r FN C ra sh R at es

76 Figure 34. Setting of investigatory and intervention levels for a specific friction demand category using pavement friction distribution and crash rate–friction trend. Method 3 is the most robust approach. It has the advantage of allowing one to discern the number of roadway sections below a certain level and to make adjustments to the level to accommodate a highway agency’s needs and budget. As in any engineering decision, one must weigh the financial implications of maintaining highway safety through managing pavement friction levels. Thus, an agency should examine the effects of using different investigatory and intervention levels in terms of the improvement in safety and the cost to achieve the level. The levels can then be adjusted to optimize the increase in safety to the agency’s budget. Regardless of the method used, the investigatory and intervention levels selected should be reviewed periodically and revised as needed. Improvements in highway safety standards may require changes in the levels set by an agency. Examples of recommended investigatory friction levels developed by selected agencies are presented in tables 16 through 19. F(60) ,Sp or FN N um be r o f S ite s W et to D ry C ra sh es , % Intervention Level Investigatory Level Wet to Dry Crashes Mean – X*(Std Dev) Mean – Y*(Std Dev) N um be r o f S ite s W et to D ry C ra sh es , %

77 Table 16. Levels of pavement friction required for various friction demand categories (VicRoads/RTA, 1996). Investigatory Level (SFC50) Site Category Site Description Primary Roads and Secondary Roads >2,500 vehicles per lane per day Secondary Roads <2,500 vehicles per lane per day 1 Traffic light controlled intersections Pedestrian/school crossings Railway level crossings Roundabout approaches 0.55 0.50 2 Curves with radius ≤250 m Gradients ≥5% and ≥50 m long Freeway/highway on/off ramps 0.50 0.45 3 Intersections 0.45 0.40 4 Maneuver-free areas of undivided roads 0.40 0.35 5 Maneuver-free areas of divided roads 0.35 0.30 Investigatory Level (SFC20) 6 Curves with radius ≤100 m 0.60 0.55 7 Roundabouts 0.55 0.50 1 ft = 0.305 m Table 17. Levels of pavement friction required for various friction demand categories (Transit New Zealand, 2002). Category Site Definition Investigatory Level (SFC) 1 Approaches to railway level crossings, traffic lights, pedestrian crossings, roundabouts, stop and give way controlled intersections (state highway only), one lane bridges (including bridge deck). 0.55 2 Curve <250 m radius. Down gradient >10% 0.50 3 Approaches to road junctions. Down gradient 5 to 10%. Motorway junction area including on/off Ramps. 0.45 4 Undivided carriageway (event-free). 0.40 5 Divided carriageway (event-free). 0.35 1 ft = 0.305 m

78 Table 18. Friction demand categories used by Maryland SHA (Chelliah et al., 2003). Site Category Site Description Design FN Required Demand Category 1 Approach railroad crossing, traffic lights, pedestrian crossing, roundabouts, stop-and-give way controlled intersections. 55 High 2 Curves with radius <250 m, downhill gradients >10 percent, and ≥50 m long freeway/highway on/off ramp. 50 High 3 Approach to intersections, downhill gradients 5 to 10 percent 45 High 4 Undivided highways without any other geometrical constraints which influences frictional demand. 40 Low 5 Divided highways without any other geometrical constraints which influences frictional demand. 35 Low Note: The Maryland SHA procedures for determining friction demand are based on a procedure originally developed by VicRoads in Australia. The VicRoads procedure was modified and calibrated for U.S. traffic conditions and aggregate testing methods. Friction demand is categorized for this procedure based on how much shear stress the pavement surfacing attracts from vehicles performing evasive traffic actions. The nature and complexity of the evasive actions is directly related to the level of pavement surface friction that would be required to ensure its success. Sites without any geometrical constraints are categorized as low frictional demand sites, while sites with geometrical constraints, such as railroad crossings, traffic lights, pedestrian crossings, roundabouts, stop and yield controlled intersections, curves, and freeway entrance/exit ramps are categorized as high frictional demand sites. Maryland also uses the following model to determine friction demand: where: FNDESIGN = Required friction at anticipated maximum speed. D = Stopping distance, ft. V = Anticipated speed, mi/hr. G = Gravitational acceleration, ft/sec2. 5280 525.242 2 ×⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ ×××= GD VFN DESIGN

79 Table 19. U.K. site categories and investigatory levels (Viner et al., 2004). Investigatory Level at 50 km/hr Site Category Definition 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 A Motorway B Dual carriageway non-event C Single carriageway non-event Q Approaches to and across minor and major junctions, approaches to roundabouts K Approaches to pedestrian crossings and other high-risk situations R Roundabout G1 Gradient 5 to 10% longer than 50 m G2 Gradient >10% longer than 50 m S1 Bend radius <500 m – dual carriageway S2 Bend radius <500 m – single carriageway 1 ft = 0.305 m Notes: 1. Investigatory levels are for the mean skidding resistance within the appropriate averaging length. 2. Investigatory levels for site categories A, B, and C are based on 100 m averaging lengths (50 m lengths for some Overseeing Organizations) or the length of the feature if it is shorter. 3. Investigatory levels and averaging lengths for site categories Q, K, G and S are based on the 50 m approach to the feature but this shall be extended when justified by local site characteristics. 4. Investigatory levels for site category R are based on 10 m lengths. 5. Residual lengths less than 50% of a complete averaging length may be attached to the penultimate full averaging length, providing the site category is the same. 6. As part of site investigation, individual values within each averaging length should be examined and the significance of any values which are substantially lower than the mean value assessed. Identification of Pavement Sections Requiring Detailed Site Investigation or Intervention Once a section has been identified as being at or below a friction threshold level, steps must be taken to identify the cause(s) of the deficiency. If FN is being used, then the agency must caution highway users by installing appropriate signs (e.g., slippery when wet, reduced speed) and then proceed with plans for a detailed investigation of the section. If the IFI is being used, a quick assessment can be made of the friction and texture measurements to determine if micro-texture or macro-texture, or both, are inadequate and in need of improvement. A graph similar to figure 35 can be developed and used, not only as an aid to the detailed investigation, but to select the type of warning that should be posted.

80 Figure 35. Determination of friction and/or texture deficiencies using the IFI. Detailed Site Investigation A detailed site investigation of all pavement sections at or below the investigatory or intervention level is necessary to (a) identify all other factors besides friction that are adversely impacting safety and (b) determine the specific causes of inadequate micro- texture and/or macro-texture. The detailed investigation involves two steps, as described below. Step 1—Conduct Visual/Video Survey Each deficient section should first be evaluated for features or characteristics of the roadway that may be compounding the friction problem, both in terms of available friction and friction demand. Such items include the horizontal and vertical alignment, the layout of lanes, intersections, and traffic control devices, the presence, amount, and severity of pavement distresses (e.g., potholes, rutting, bleeding, deteriorated patches), longitudinal pavement smoothness, and transverse pavement profile. Also of importance in the detailed investigation are the issues of glare (as caused by the pavement or the lack of appropriate traffic aids), splash and spray, and hydroplaning potential (often linked to rutting or inadequate cross-slope). Detailed discussions of these latter two items are provided below. Marginal Friction/Micro-texture Inadequate Macro-texture Marginal Macro-texture Adequate Macro-texture LEGEND Texture Investigatory Level Texture Intervention Level Friction Investigatory Level Friction Intervention Level Pavement Macro-Texture Pa ve m en t F ric tio n Inadequate Friction/Micro-texture Adequate Friction/Micro-texture Pa ve m en t F ric tio n

81 Splash and Spray While there is currently very little data/information on the relationship between highway crashes and splash and spray, it is obvious that these occurrences can reduce a driver’s vision and increase the risk of crashes. Splash and spray from passing and/or leading vehicles make seeing ahead, behind, and to the sides more difficult, particularly at night. The fog-like phenomenon associated with spray typically results in greater loss of visibility as compared to splash, due to the propensity of small water droplets to remain airborne longer than large droplets. This “fog” can linger as long as it is being replenished by the interaction of the three elements that produce it—standing water; a hard, smooth, non- porous, surface; and turbulent air flow that picks up and carries the water (NHTSA, 1998). The occurrence of splash/spray is influenced by the drainage condition at the pavement surface. Providing positive drainage that quickly removes standing water from the pavement surface will reduce the occurrence of splash/spray significantly. Pavement surface drainage is enhanced by providing adequate amounts of macro-texture and cross- slope. Hydroplaning Potential As discussed earlier, hydroplaning refers to the separation of the tire contact from the pavement surface by a layer of water. It is a complex phenomenon that is affected by (1) the water film thickness (WFT) on the pavement surface, (2) pavement macro-texture, (3) tire tread depth, (4) tire inflation pressure, (5) tire contact area, and (6) vehicle speed. For a vehicle to experience hydroplaning, two things must occur simultaneously: there must be a sufficient buildup of water on the pavement surface and the vehicle must be traveling at a speed high enough to cause hydroplaning. Thus, the potential for hydroplaning for a given highway segment can be assessed by determining (1) the frequency of water buildup from precipitation (rainfall only) on the pavement surface and (2) whether the traveling speeds of vehicles is high enough to result in hydroplaning for the water buildup conditions. A three-step procedure for determining hydroplaning potential is presented below. • Step 1—Estimate Critical Hydroplaning Speed (HPS): An approximate relationship between the vehicle speed (in mi/hr) at which hydroplaning for both asphalt and concrete pavements will occur and the tire inflation pressure (in lb/in2) is as follows (Ong and Fwa, 2006): HPS = 10.35 pressuretire Eq. 25 This equation assumes that WFT on the pavement surface exceeds the combined capability of the surface macro-texture and tire design (i.e., tread depth) to remove water from the pavement surface.

82 • Step 2—Compute WFT using agency-established models or procedures or the WFT prediction models (and accompanying software) developed in NCHRP Project 1-29 (Anderson et al., 1998). • Step 3—Determine Hydroplaning Potential: As shown in table 20, hydroplaning potential is categorized as none, low, moderate, or high. Table 20. Assessment of hydroplaning potential based on vehicle speed and water film thickness. WFT, in Average Vehicle Speed (85th Percentile of Traveling Speed) minus Critical Hydroplaning Speed (HPS), mi/hr a < 0.02 0.02 to 0.06 > 0.06 Less than –5 None None None Between +5 None Low Moderate Greater than 5 None Moderate High 1 mi/hr = 1.61 km/hr 1 in = 25.4 mm a Guidelines for determining design speed based on highway functional classification, location (i.e., rural versus urban), and terrain type (i.e., level, rolling, and mountainous) can be found in the AASHTO Green Book (AASHTO, 2001). Step 2—Evaluate Micro-Texture and Macro-Texture The second step in the detailed site investigation involves testing the pavement surface for micro-texture and macro-texture. These two properties can be evaluated using various types of equipment, including: • Micro-texture, which can be evaluated using any of the following: ¾ Locked-wheel friction tester. ¾ British Pendulum Tester (BPT). ¾ Dynamic Friction Tester (DFT). • Macro-texture, which can be evaluated using any of the following: ¾ High-speed laser. ¾ Circular Texture Meter (CTM). ¾ Sand Patch Method (SPM). Testing must be done in a manner that produces results that are representative of the entire pavement section. In addition to the micro-texture and macro-texture data, the following information must be obtained from the records or through field testing: • Traffic applications, including truck percentages. • Pavement surface age. • Surface material type and/or finishing method. • Data on all materials used in the surface pavement (e.g., fine/coarse aggregate type), including polishing/wear characteristics, structure, hardness, and so on, if available. • Other information, such as data from laboratory tests.

83 Using the micro-texture and macro-texture results and the data listed above, the exact cause of friction loss can be determined. Common causes of friction loss include polishing of coarse aggregates and excessive wearing of the pavement surface resulting in a loss of macro-texture. Table 21 lists the many specific actions recommended when conducting a detailed site investigation. Table 21. Recommended actions for detailed site investigations (Chelliah et al., 2003; TXDOT, 2004; Austroads, 2005; Viner et al., 2004). Step Description Recommended Action 1 Site location 1. What is the friction demand for this location? 2. What are the current investigatory and intervention levels? 3. Has there been any substantial change in the amount or type of traffic applied or highway features to warrant a change in friction demand category and associated changes in investigatory and intervention levels? If so, reclassify the friction demand as appropriate. 4. Document recent weather and traffic conditions at the site location. Has there been any unusually bad weather (excessive rainfall, snow, blizzards, etc.)? Document unusual weather occurrences and investigate if they can be a possible reason for spikes in crash rates. 2 Pavement condition 1. What is the current friction levels? 2. By how much is the current friction level below the investigatory level and over what length? 3. Is pavement friction uniform along the site or are there significant variations? If there are significant variations, perform a detailed visual assessment and testing as needed to describe this situation in detail. 4. Is the minimum pavement friction measurement below the intervention level? If so what percentage of the site is below the intervention level? 3 Crash history 1. What is the location of crashes in relation to the observed variability in measured pavement friction? 2. Are crashes generally located in localized areas with low friction? 3. If not, is there any other pattern apparent in the location or type of crashes that would warrant more crash investigation? 4. Have there been any significant changes to the site or the traffic using it in the analysis period, which could have affected the number of crashes? 4 Visual assessment 1. Is a visual inspection of surface condition consistent with the available survey data? 2. Friction is generally measured in the nearside wheel track in the outside lane. Is the rest of the area of the maintained pavement surface visually consistent with the measured path, or are there any localized areas of polished surfacing, low texture depth, patching or areas otherwise likely to give rise to uneven friction (i.e., is it likely that the friction of other lanes could be lower than the lane tested)? 3. If there is a lack of uniformity in friction measurements across the site, is it likely to increase the risk of crashes occurring?

84 Selection and Prioritization of Restoration Treatments The final step in a PFM program is to analyze the collected data to identify sites requiring more frequent monitoring or forensic investigation, and sites requiring friction restoration. Highway agencies normally use pavement friction and other condition data to identify and prioritize sites to be included in a program for: • Short-term remedial (maintenance) works. • Comprehensive restoration treatment (e.g., diamond grinding, cold milling, thin overlays, chip seals) aimed directly at improving friction. In analyzing pavement friction data, the minimum desirable outcome is to ensure that the most “deficient” sites are detected and given reasonable priority, because they are likely to have more impact on highway user safety. The extent of the analysis and use of pavement friction and other data is determined locally by the agency. However, in its simplest possible form, analysis can be restricted to identifying all sites where the measured pavement friction is at or below any investigatory or intervention level that has been set. This is followed up by a detailed site investigation to identify required actions that include: 1. Continue to monitor the site: Such a decision typically would be reached where (a) current crash rates are sufficiently low and an increase is not expected to significantly impact safety and (b) the pavement surface does not require maintenance because of other factors. 2. Listing the site for remedial action to improve pavement friction (e.g., resurface, retexture): This usually would apply where an increase in crash rate might occur if friction remains the same or continues to decrease, and such an increase would significantly impact safety. Deficient sites requiring restoration are prioritized so that sites urgently requiring attention are dealt with first. In practice, a cut-off is likely to be reached when the available funding is exhausted, after which it is common for the remaining sites in the list to be considered together with other sites requiring short-term remedial works. Pavement Friction Management Approach and Framework To develop PFM policies, an agency must identify an overall approach for managing pavement friction and a process for implementing it. The comprehensive PFM program shown in figure 36 may be used. It is comprised of the following key components: • Network Definition—Subdivide the highway network into distinct pavement sections and group the sections according to levels of friction need. ¾ Define pavement sections. ¾ Establish friction demand categories. • Network-Level Data Collection—Gather all the necessary information. ¾ Establish field testing protocols (methods, equipment, frequency, conditions, etc.) for measuring pavement friction and texture. ¾ Collect friction and texture data and determine overall friction of each section. ¾ Collect crash data.

85 Figure 36. Example PFM program. Process Crash Data for all Sections at or Below Investigatory Level Define Pavement Network & Identify Sites (Re-assess Site Categories Periodically) Review Pavement Friction Testing Frequency Perform Routine Friction Testing and Collect Crash Data No Yes No No No Yes Yes Perform Detailed Site Investigation Yes Does Site Need Restoration Yes No Higher Lower Are Wet Crash Rates High? Are Wet Crash Rates High? Is Friction At or Below Investigatory Level? Is Friction At or Below Intervention Level? Assess Risk 1. Shortlist Sites Requiring Restoration in Order of Priority 2. Perform Short-Term Remedial Works, if Needed 3. Identify Preferred Restoration Design Strategy 4. Develop Schedules for Restoration Activities For all Sections Above Investigatory Level, Process and Evaluate Crash Data. Conduct Detailed Investigation of Sections with High Crash Rates to Determine if High Crash Rates are Due to (1) Setting of Inadequate Investigatory Levels or (2) Non-Friction Related Causes. Develop Appropriate Recommendations Based on Investigation Results

86 • Network-Level Data Analysis—Analyze friction and/or crash data to assess overall network condition and identify friction deficiencies. ¾ Establish investigatory and intervention levels for friction. Investigatory and intervention levels are defined, respectively, as levels that prompt the need for a detailed site investigation or the application of a friction restoration treatment. ¾ Identify pavement sections requiring detailed site investigation or intervention. • Detailed Site Investigation—Evaluate and test deficient pavement sections to determine causes and remedies. ¾ Evaluate non-friction-related items, such as alignment, the layout of lanes, intersections, and traffic control devices, the presence, amount, and severity of pavement distresses, and longitudinal and transverse pavement profiles. ¾ Assess current pavement friction characteristics, both in terms of micro-texture and macro-texture. ¾ Identify deficiencies that must be addressed by restoration. ¾ Identify uniform sections for restoration design over the project length. • Selection and Prioritization of Short- and Long-Term Restoration Treatments—Plan and schedule friction restoration activities as part of overall pavement management process. ¾ Identify candidate restoration techniques best suited to correct existing pavement deficiencies. ¾ Compare costs and benefits of the different restoration alternatives over a defined analysis period. ¾ Consider monetary and non-monetary factors and select one pavement rehabilitation strategy.