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.

9 CHAPTER 3. PAVEMENT FRICTION AND HIGHWAY SAFETY HIGHWAY SAFETY Safety, as a general term, is often defined in two ways—the quality or condition of being safe (i.e., freedom from danger, injury, or damage) or any of certain devices or actions designed to prevent a crash from happening. Thus, it stands to reason that highway safety can be characterized as a driving environment free from danger or, more appropriately, one that is operated with rules and features designed to minimize crashes and the associated consequences (fatalities, injuries, economic loss). Since the early years of motor vehicle transportation, governmental agencies and industry groups have worked continuously to institute highway safety measures. Over the past few decades, however, the societal demand for mobility and economic growth has increased substantially, resulting in spiraling rates of vehicle travel and unprecedented levels of risk for highway users. Between 1990 and 2003, an average of 6.4 million highway crashes (all vehicle types) occurred annually on the nation’s highways, resulting in 3 million injuries, 42,000 fatalities, and countless amounts of pain and suffering. This rate of fatality equates to 115 fatalities per day, or 1 death every 12 minutes (Noyce et al., 2005; National Highway Traffic Safety Administration [NHTSA], 2004). Crashes occur at significant cost to the nation’s economy. In 2000, the cost of highway crashes was estimated at $230.6 billion (Noyce et al., 2005; NHTSA, 2004). This figure continues to increase year after year, taking up resources that could be used to improve the highway infrastructure. Figures 2 and 3 present summaries of total crashes and resulting fatalities in the U.S. between 1990 and 2003. According to the National Transportation Safety Board (NTSB) and the FHWA, approximately 13.5 percent of fatal crashes and 25 percent of all crashes occur when pavements are wet (Kuemmel et al., 2000). Highway crashes are complex events that are the result of one or more contributing factors. Such factors fall under three main categories—driver-related, vehicle-related, and highway condition-related (Noyce et al., 2005). Of these three categories, highway agencies can control only highway conditions. This can be done by developing and administering effective design, construction, maintenance, and management practices and policies.

10 Figure 2. Total crashes (from all vehicles types) on U.S. highways from 1990 to 2003 (NHTSA, 2004). Figure 3. Total fatalities (from all vehicles types) on U.S. highways from 1990 to 2003 (NHTSA, 2004). 36 37 38 39 40 41 42 43 44 45 46 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 To ta l F at al iti es , t ho us an ds 5.6 5.8 6.0 6.2 6.4 6.6 6.8 7.0 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 To ta l c ra sh es , m illi on s 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 To ta l c ra sh es , m illi on s

11 RELATIONSHIP BETWEEN WET-WEATHER CRASHES AND HIGHWAY PAVEMENT SURFACE CONDITIONS Although most highway crashes involve multiple causative factors, crash investigations have consistently shown a link between crashes and pavement surface conditions/ characteristics, such as friction and texture. Thus, there is a need for in-depth knowledge and understanding of the relationship between the two so that engineers can develop effective solutions to potentially hazardous situations. Wet-Weather Crashes and Pavement Friction While the exact relationship between wet-weather crashes and pavement friction is difficult to quantify, much empirical research has been done that shows that the number of wet crashes increases as pavement friction decreases (all other factors, such as speed and traffic volume, remaining the same). A summary of the many published research findings is presented below. • Rizenbergs et al., 1972—In this study, crash and measured pavement friction data obtained from mostly rural interstates and parkway roadways in Kentucky were analyzed. The results of the analysis showed increased wet crash rates at pavement friction values (SN40R, skid/friction number determined with a locked-wheel friction tester operated at 40 mi/hr [64 km/hr]) less than 40 for low and moderate traffic levels (see figure 4). Similar trends were observed when analyzing wet-to-dry crash ratios as a function of pavement friction, as shown in figure 5. • Giles et al., 1962; Cairney, 1997—Pavement friction was evaluated at 120 sites where a skid-related crash had occurred along with a 100 randomly chosen control sites on highways of similar functional class and traffic volumes. The relative risk of a site being a skid-related crash site was computed by dividing the proportion of skid-related crash sites by the proportion of control sites for different pavement friction categories. The risk of a skid-related crash was small for friction values (SN) above 60, but increased rapidly for friction values below 50. • McCullough and Hankins, 1966—In a study of the relationship between pavement friction and crashes from 571 sites in Texas, it was found that a large proportion of crashes occurred with low pavement friction and relatively few occurred with high pavement friction. A minimum desirable friction coefficient of 0.40 measured at 30 mi/hr (48 km/hr) was recommended. This value was obtained as a convenient value close to the point where the slope of crash rate versus friction decreased significantly.

12 Figure 4. Relationship between wet-weather crash rates and pavement friction for Kentucky highways (Rizenbergs et al., 1973). Figure 5. Ratio of wet-to-dry pavement crashes versus pavement friction for Kentucky highways (Rizenbergs et al., 1973).

13 • Miller and Johnson, 1973; Cairney, 1997—Available friction before and after resurfacing was determined on the M4 highway in England (i.e., resurfacing increased the average friction from 0.40 to 0.55, at 50 mi/hr [80 km/hr]). Crashes were recorded for 2 years before and 2 years after resurfacing, for a total of over 500 incidents. Data from this study showed that pavement resurfacing (increased pavement friction) resulted in a 28 percent reduction in dry pavement crashes and a 63 percent reduction in wet pavement crashes. Total crashes for the study area were reduced by 45 percent. • Kamel and Gartshore, 1982—Selected hazardous sections (i.e., sections with low friction levels experiencing a high rate of wet pavement crashes) on the highway network in Ontario, Canada were resurfaced to increase pavement friction. For intersections, crashes were reduced by 46 percent overall, 21 percent in dry conditions and 71 percent in wet conditions. For freeways, total crashes were reduced by 29 percent, 16 percent in dry conditions and 54 percent in wet conditions. • Gothie, 1996—Three separate studies were performed to define the cause-and-effect relationships between highway surface properties. The following was concluded: (1) wet crash rates increased by at least 50 percent when moving from a section with a Sideway Force Coefficient (SFC) greater than 0.60 to a section with an SFC less than 0.50, (2) for a reduction in SFC of 0.05, the risk and severity of crashes increased by approximately 50 percent. • Bray, 2002—In this study, 40 pavement sections experiencing unusually high amounts of wet crashes were identified. Before and after hot mix asphalt (HMA) resurfacing (increased pavement friction) crash analyses showed significant reductions in the 740 recurring crashes (from which 540 were wet surface crashes) after rehabilitation. • McLean, 1995—Before and after evaluation of resurfacing projects in England indicated that crash rates on rural highways can increase even when pavement friction is improved significantly. This finding implies that the gains from improved friction can be offset by the increased risk caused by improved ride quality (i.e., drivers tend to use smooth roads more often then rough ones, and they tend to travel at higher speeds). • Organization for Economic Cooperation and Development (OECD), 1984—The OECD’s International Scientific Expert Group on Optimizing Road Surface Characteristics revealed a linear crash-friction relationship in the U.S. (i.e., reduction in friction was associated with a linear increase in crashes). This behavioral function differs from other relationships obtained in Europe, where research suggests a non-linear relationship between pavement friction and crashes. • Wallman and Astrom, 2001—In this research, a comprehensive evaluation of friction measurements and crash rates revealed that increasing pavement friction does reduce crash rates significantly, as summarized below.

14 Friction Interval Crash Rate (injuries per million vehicle km) < 0.15 0.80 0.15 – 0.24 0.55 0.25 – 0.34 0.25 0.35 – 0.44 0.20 • Gandhi et al., 1991—In the early 1990s, a study conducted in Puerto Rico found a statistically significant relationship between the minimum Mu-Meter skid number and the ratio of wet-to-dry crashes. Using linear regression, an R-squared value of 0.55 was obtained for those two variables. Other dependent variables considered included the ratio of wet crashes to the total number of crashes. The average friction coefficient in a section was found to be related less to crash rates than to the minimum friction coefficient. • Craus et al., 1991—This study was conducted by the Israeli Public Works Department to examine the relationship between pavement frictional condition measured by a Mu-Meter and highway crashes. It was found that average Mu- Meter readings greater than 37 for the network could reduce the total number of crashes by 7.5 percent. • Larson, 1999—French research reported in 1996 found a five-fold increase in the wet crash rates on the Bordeaux Ring Road when the SFC decreased from greater than 0.60 to less than 0.50. This study also found that the risk of wet crashes increases greatly for surfaces with an estimated texture depth less than 0.016 in (0.40 mm). • Xiao et al., 2000—Researchers at the Pennsylvania Transportation Institute (PTI) developed two fuzzy logic models to predict wet-pavement crashes. The skid number, posted speed, average daily traffic (ADT), pavement wet time, and driving difficulty were the variables selected as having the greatest effect on the risk of skidding crashes at a site. These models were used to calculate the improvement in safety expected from improvements in each of the input variables. It was shown that the safety condition, measured by the percent reduction in wet pavement crashes, could be improved nearly 60 percent if the skid number increased from 33.4 to 48. • Schulze et al., 1976—The effect of wet climate on safety was further demonstrated by a study conducted in Germany, where the proportion of wet crashes was compared to pavement surface friction, as shown in figure 6. Friction number for this study was measured at 50 mi/hr (80 km/hr). Although there was a large scatter in the data, this figure clearly shows there is a significant increase in wet pavement crashes as the pavement friction decreases.

15 Locked wheel braking (force coefficient at 80 km/h) 0.40.5 0.20.3 0.1 A cc id en ts (r at io o f w et to ta l [ w et a nd d ry ]) A cc id en ts (r at io o f w et to ta l [ w et a nd d ry ]) Figure 6. Relationship between wet crashes and pavement surface friction (Schulze et al., 1976). Empirical evidence from these research studies shows that vehicle crashes are more likely to occur on wet pavements (with lower friction levels) and that, as pavement friction levels decrease, there is a corresponding increase in crash rates. Research also shows that when pavement friction falls below a site-specific threshold value, the risk of wet crashes increases significantly (Kuttesch, 2004). The exact nature of the relationship between pavement friction and wet crashes is site- specific, as it defined by not only pavement friction but many others factors. Thus, pavement friction and wet crashes relationships must be developed for the sites that are typically present in a given pavement network. An example of such a relationship developed for single carriageways in the U.K. shows that crash risk approximately halves as pavement friction doubles over normal ranges, as shown in figure 7 (Viner et al., 2004).

16 Figure 7. Relationship between pavement friction and crash risk (Viner et al., 2004). Wet-Weather Crashes and Splash/Spray “Splash” and “spray” describe vehicle-induced water droplets and mist that adversely affect driver visibility on wet highways. Splash consists of very large liquid droplets that fall ballistically to the ground, while spray consists of very small liquid droplets that remain in the air for a long time in the form of a fog cloud before falling to the ground (NHTSA, 1998). Conditions that most favor the spray and must be present for it to occur are: (1) standing water, (2) a hard or smooth surface struck by the water, and (3) a turbulent airflow to pick up and carry the water (NHTSA, 1998). Splash does not significantly influence driver visibility, as the splashed droplets typically remain close to the ground and out of the line of the driver's vision. However, because spray can remain airborne for a long time, it can surprise, confuse, and disorientate drivers (NHTSA, 1998), leading potentially to the hazards listed in table 1. Although it is accepted that splash and spray increase crash risk, there is very limited information on how many of the crashes that occur on the nation’s highways are a direct result of splash and spray. A review of federal and state crash-related databases indicates that most agencies do not collect sufficiently detailed information to assign specific cause to splash and/or spray. Skid Resistance 0.3 0.4 0.5 0.6 5 0 10 15 20 “Double the skid resistance and halve the accidents” A cc id en t R is k (a cc id en ts p er 1 0^ 8 ve h- km )

17 Table 1. List of hazards that can potentially be caused by splash and spray (NHTSA, 1998). Potential Hazard Target Description Lead vehicle Spray obscures following vehicles. Follower Spray from lead vehicles obscures visibility of lead vehicles, signs, edge lines, other traffic at intersections, traffic signals. Follower Doused during passing, lane change. Spray obscures visibility of vehicles in passing lane; spray obscures lead vehicles, signs, edge lines, other traffic at intersections, traffic signals (in own lane). Opposing vehicles (2 lanes are adjacent) Doused during encounter causes loss of control. Opposing vehicles (divided highways) Spray drift from vehicle in opposing lanes obscures visibility of lead vehicles, signs, edge lines, other traffic at intersections, traffic signals. Numerous visibility problems. Knowing they cannot be seen, they navigate differently. Motorcycles, cyclists, and pedestrians Doused when vehicle passes. Upon reviewing data from the Fatality Analysis Reporting System (FARS) and the National Automotive Sampling System (NASS) General Estimates System (GES), the University of Michigan Transportation Research Institute (UMTRI) reported the following (NHTSA, 1998; NHTSA, 2005): • FARS identified 29 splash- and spray-related crashes out of a total of 255,928 reported from 1991 through 1997 (i.e., approximately 0.011 percent of all crashes were identified as having splash or spray as a contributing factor). This rate is somewhat higher than the rate from the FARS for 1982 through 1987, as reported in the 1994 Report to Congress. Nevertheless, the percentage of crashes attributed to splash and spray is too small to have any significance. Of the 29 crashes reported in FARS, only one involved a truck. • In the GES records, a total of 17 crashes were recorded between 1991 and 1997. This translates into a weighted estimated total of 1,622 splash/spray-related crashes occurring for this period (i.e., 0.0036 percent of the total of 45,024,000 crashes estimated to have occurred over those 8 years). Thus, the information available in both FARS and GES indicates that the number of recorded splash/spray crashes is extremely small and may not be a significant contributor to highway crashes. However, when it occurs, it is not limited to high-speed highways with a high percentage of truck traffic, and there is sufficient empirical information to suggest that a puddle in the middle of a highway can initiate the process that results in a splash- or spray-related crash.

This page intentionally left blank.