Use of Pavement Markings to Reduce Excessive Traffic Speeds on Hazardous Curves

Transcription

1 Use of Pavement Markings to Reduce Excessive Traffic Speeds on Hazardous Curves EFFORTS TO REDUCE DESPITE PROGRESS IN REDUCing motor vehicle crash deaths and TRAFFIC SPEEDS ON injuries in the United States, sharp horizontal curves situated on narrow twolane highways a common roadway CURVES SHOULD design feature pose a significant safety CONCENTRATE ON THE threat. According to data published by the National Highway Traffic Safety TANGENT SECTIONS Administration, nearly 80 percent of all fatal crashes in the United States occur PRECEDING THE POINTS on two-lane roads. 1 Motor vehicle crashes on curved roadway sections occur OF CURVATURE. more frequently and tend to be more severe than those on straight sections. About 40 percent of fatal roadside crashes occur on curves. 2 Troxel et al. analyzed data from the Fatality Analysis Reporting System and National Automotive Sampling System and reported that more than twice as many occupants were involved in crashes per kilometer of curved road compared with straight sections. 3 Crashes on curved roads resulted in nearly three times as many fatalities per kilometer as crashes on straight sections. Crash types that occur disproportionately on curves include collisions with fixed objects, head-on collisions and crashes involving rollovers. 4 Viner reported that run-off-road crash risk on rural two-lane roads increases with degree of curvature. 5 A study of 300 fatal roadside object crashes found a significant overrepresentation of road locations with curvatures greater than 6 degrees and negative gradients of 2 percent or steeper. 6 Excessive traffic BY RICHARD A. RETTING AND CHARLES M. FARMER speed is a significant factor in crashes on curves and in the severity of these crashes. 4 Vehicles traveling at high speeds on curved road sections not only risk departure to the outside of the curve but also frequently transgress into the opposite lane a major cause of crashes. 7 The speed at which a vehicle enters a curve is related more to the speed of the vehicle as it approaches the curve than to the sharpness of the curve. 7 Therefore, efforts to reduce traffic speeds on curves should concentrate on the tangent sections preceding the points of curvature. Limited research has been published on the feasibility of installing pavement markings in advance of hazardous curves to reduce excessive traffic speeds. In two separate studies, experimental transverse pavement markings, which were intended to create an illusion of acceleration and prompt drivers to slow down, were placed across roads in advance of curves. 8,9 These marking schemes were associated with significant reductions in traffic speeds. For example, Agent reported that average traffic speeds were reduced from 41.3 to 33.9 miles per hour (mph) one week after markings were installed; six months after treatment, the average speed was 34.8 mph 16 percent less than that observed during the baseline period. Rockwell et al. also reported reductions in traffic speeds, most notably high speeds, resulting from pavement markings designed to make the roadways appear narrower at the beginning of the curves. 10 Significant reductions in crashes were associated with yellow bar pavement markings installed perpendicular to the direction of travel at 42 sites in advance of roundabouts and termination of high-speed roads; 52 percent fewer crashes were reported in two years following treatment, and crash benefits were sustained four years after treatm ent at seven sites for which data were available. 11 This feature reports on a research experiment to examine the effectiveness of special pavement markings intended to reduce excessive traffic speeds at rural and suburban two-lane roadway locations with sharp horizontal curvatures. 30 ITE JOURNAL / SEPTEMBER 1998

2 the hot thermoplastic to enhance nighttime retroreflectivity. Figure 1. Indications of run-off-road crashes at curve. METHODS Setting The experiment was conducted at a single location on a suburban two-lane secondary road in Northern Virginia (USA) that includes a sharp left curve (approximately 90 degrees) preceded by a long tangent section. The road is approximately 20 feet wide, has one lane in each direction and is marked with a double yellow centerline. Land use in the area is mostly residential, and the speed limit is 35 mph. The sharp left curve shows evidence (i.e., skid marks) of numerous prior run-off-road crashes (Figure 1). An advisory, or recommended safe, speed of 15 mph is posted approximately 500 feet before the curve using a standard reverseturn sign and advisory speed plate. 12 The experimental pavement marking was intended to reinforce this existing advisory. A separate left curve (approximately 45 degrees) in the opposite (northbound) direction of the same highway, approximately one-quarter mile away, was chosen as a control site. Average daily traffic on the roadway was approximately 5,000 vehicles. Experimental Design A before/after design with control was employed. At the experimental site, traffic speeds were measured on the tangent section 90 feet prior to the point of curvature before and after installation of the pavement marking (Figure 2). At the upstream site, traffic speeds were measured approximately 650 feet prior to the point of curvature. The upstream counter allowed for analyses of general changes in speed associated with the pavement marking. However, it was not possible to match specific speed measurements at the experimental site with corresponding measurements at the upstream site. Thus, the number of measurements at the upstream and experimental sites are different. At the control site, traffic speeds were measured approximately 100 feet prior to the point of curvature. The control site controlled for factors that might affect speed such as weather and seasonal variability in travel patterns. Installed approximately 220 feet before the curve, the experimental pavement marking consisted of the word SLOW in eightfoot-high white letters, a white eight-foot-high left curve arrow and an 18-inch-wide white line perpendicular to the road at both the beginning and end of the text/symbol message (Figure 3). Thermoplastic marking material was used, and glass beads were spread over Instrumentation Speed measurements were made using TimeMark Delta Traffic Counters connected to pneumatic road tubes. Two road tubes were installed at each site and spaced 20 feet apart. Baseline measurements were made in August 1996, and the marking was installed in January Speed measurements were collected two weeks after installation of the marking, allowing time for local drivers to adapt to the intervention. Janus software was used to process raw data from the traffic counters, including vehicle classification, gap and speed. Measurement periods were free of precipitation. Data for all three sites during both periods were recorded beginning Saturdays at 10:30 a.m. and ending Sundays at 3 a.m. Data Analyses Analyses were limited to passenger vehicles because there were insufficient numbers of trucks observed at the study sites. A minimum gap of three seconds Figure 2. Pavement marking and speed measurement locations. ITE JOURNAL / SEPTEMBER

3 between vehicles was employed. Trucks with speed data were analyzed for three time periods: daytime (10:30 a.m. to 5 p.m.), evening (9 p.m. to midnight) and late night (midnight to 3 a.m.). Summary measures included the mean speed, 90th percentile speed and percentage of vehicles exceeding given speed thresholds. Logistic regression models were used to measure the effect of the pavement marking on the likelihood that a vehicle exceeded 40 mph. RESULTS Table 1 summarizes daytime speed measurements. At the experimental site, the mean traffic speed dropped from 34.3 mph during the baseline to 33.2 mph after marking. The 90th percentile speed dropped from 39.7 to 38.1 mph. In contrast, speeds increased at both the upstream and control sites. The 3 percent decrease in mean speed at the experimental site along with the 4 percent increase at the upstream site results in an estimated 7 percent decrease in mean speed associated with the pavement marking. The percentage of vehicles exceeding 40 mph on the experimental curve dropped by more than half, from 9.1 percent to 3.5 percent. This was significantly different (p < 0.001) from the upstream site, where the percentage of vehicles exceeding 40 mph rose from 54 percent to 66 percent. Figure 3. Experimental pavement marking. Table 2 summarizes evening speed measurements. At the experimental site, the mean traffic speed dropped from 33.5 mph during the baseline to 31.9 mph after marking. The 90th percentile speed dropped from 39.9 to 37.4 mph. Speeds also decreased at the upstream site, but by a lesser amount. The 5 percent decrease in mean speed at the experimental site along with the 3 percent decrease at the upstream site results in an estimated 2 percent decrease in mean speed associated with the pavement marking. The percentage of vehicles exceeding 40 mph on the experimental curve dropped from 9.4 percent to 3.3 percent. This was not significantly different (p = 0.094) from the upstream site, where the percentage of vehicles exceeding 40 mph dropped from 49 percent to 39 percent. Table 3 summarizes late night speed measurements. At the experimental site, the mean traffic speed dropped from 35.1 mph during the baseline to 31.7 mph after marking. The 90th percentile speed dropped from 41.9 to 37.6 mph. Speeds also decreased at the upstream site, but by a lesser amount. The 10 percent decrease in mean speed at the experimental site along with the 3 percent decrease at the upstream site results in an estimated 7 percent decrease in mean speed associated with the pavement marking. The percentage of vehicles exceeding 40 mph on the experimental curve dropped from 18.5 percent to 1.6 percent. This was significantly different (p = 0.035) from the upstream site, where the percentage of vehicles exceeding 40 mph dropped from 48 percent to 40 percent. DISCUSSION Motor vehicle crashes on curved twolane roads are a significant and costly problem. Many factors contribute to runoff-road, head-on and rollover collisions on curves, including driver impairment, fatigue, inattention, visual deficits and excessive vehicle speed. While most driver factors are outside the direct influence of transportation engineers, it is apparent from the present study and prior research Table 1. Daytime traffic speeds: Saturday from 10:30 a.m. until 5 p.m. Standard 90th Percentage Percentage Percentage Site/Period Number Mean Deviation Percentile > 35 mph > 40 mph > 45 mph Experimental Site Baseline After * Upstream Site Baseline After * Control Site Baseline After * *Mean speed in the after period significantly different from the baseline. 32 ITE JOURNAL / SEPTEMBER 1998

4 Table 2. Evening traffic speeds: Saturday from 9 p.m. until midnight. Standard 90th Percentage Percentage Percentage Site/Period Number Mean Deviation Percentile > 35 mph > 40 mph > 45 mph Experimental Site Baseline After * Upstream Site Baseline After Control Site Baseline After *Mean speed in the after period significantly different from the baseline. that strategically placed pavement markings can influence driver speed selection upon entering sharp horizontal curves. The pavement marking used in this study was associated with a decrease in vehicle speed of approximately 6 percent overall and 7 percent during daytime and late night periods. This was in addition to any effect of the existing advisory signs. Despite these findings, relatively little use has been made of such innovative pavement markings in the United States. The present research was limited by the use of only one experimental site and the inability to distinguish short-term or novelty effects from long-term effects. However, prior research by Agent 8 did report sustained reductions in traffic speeds associated with installation of similar pavement markings on the approach to a sharp curve. It is important to note that even short-term effects are useful since warning devices of the type studied here generally are designed for drivers unfamiliar with road dangers such as sharp horizontal curves. Reductions in average traffic speeds, and in the proportion of high traffic speeds reported in this study, are highly significant given the exponential relationship between fatality risk and change in velocity (delta V) during collisions. 13 Therefore, even seemingly small reductions in mean traffic speeds are likely to result in significant safety benefits. Late night Saturday (midnight to 3 a.m Sunday) is an especially high-risk time for run-off-road crashes. 1,2 The finding that high traffic speeds appear to be signifi- Table 3. Late night traffic speeds: Sunday from midnight until 3 a.m. Standard 90th Percentage Percentage Percentage Site/Period Number Mean Deviation Percentile > 35 mph > 40 mph > 45 mph Experimental Site Baseline After * Upstream Site Baseline After Control Site Baseline After *Mean speed in the after period significantly different from the baseline. ITE JOURNAL / SEPTEMBER

5 cantly reduced during this particularly dangerous period suggests that markings were even effective, perhaps especially so at times when drivers are more likely to be impaired by alcohol or fatigue. Because traffic speeds on curves are closely correlated to approach speeds, efforts to reduce speeds should concentrate on deployment of effective devices on tangent sections preceding curvature points. Advance curve warning signs and advisory delineation signs frequently are employed. However, a large body of research suggests that such traffic signs may be largely ineffective. For example, Zador et al. reported that use of postmounted delineators and chevrons tends to increase nighttime speed distributions. 14 Lyles, and Agent and Creasey, studied various sign treatments for reducing traffic speeds in the vicinity of horizontal curves and found they were generally ineffective. 15,16 There is some evidence that common warning signs are poorly understood by drivers. 17 Automated speed enforcement, used extensively outside the United States, has been found to reduce speeding and injury crashes on two-lane rural roads. 18,19 Though not generally effective for speed control, rumble strips may be useful for alerting drivers to certain roadway features such as sharp horizontal curves at which they would normally reduce speed. 20 In designing traffic control devices, it is important to recognize and understand what drivers expect. 21 For example, if several moderately sharp curves are preceded by curve warning signs, it is reasonable for drivers to expect that any additional curve warning signs they encounter downstream would be followed by similar curves. If an unusually sharp downstream curve is not differentiated by some special warning, driver expectancy is violated, and drivers may not slow sufficiently. The most effective method for reducing crashes and crash severity on hazardous curves is roadway realignment. Geometric improvements such as curve flattening, roadway widening and super-elevation have been associated with significant reductions in crashes and crash severity. 4,22,23 Increasing pavement friction also may be beneficial. However, these measures often are costly and are not always feasible due to financial, environmental and right-of-way restrictions. Special pavement markings that can reduce excessive traffic speeds should be considered as an interim measure at dangerous curves prior to road reconstruction, and for use at hazardous curves where adequate geometric design features cannot be provided. Because pavement markings are subject to constant wear from vehicle contact and the natural environment, they require periodic inspection and maintenance. ACKNOWLEDGMENT This work was supported by the Insurance Institute for Highway Safety (IIHS). The authors acknowledge the Northern Virginia District of the Virginia Department of Transportation for participating in this research, Pavemark TM and 3M TM for providing marking materials for this research and Michael Greene for statistical support. References 1. National Highway Traffic Safety Administration. Traffic Safety Facts Report DOT- HS U.S. Department of Transportation (USDOT), Washington, D.C., USA, Insurance Institute for Highway Safety. Fatality Facts. IIHS, Arlington, Va., USA, Troxel, L.A., M.H. Ray and J.F. Carney. Accident Data Analysis of Side-impact, Fixed-Object Collisions. Report RD Federal Highway Administration (FHWA), USDOT, Zegeer, C.V., R. Stewart, D. Reinfurt, F. Council, T. Neuman, E. Hamilton, T. Miller and W. Hunter. Cost Effective Geometric Improvements for Safety Upgrading on Horizontal Curves. Report RD FHWA, USDOT, Viner, J.G. The Roadside Safety Problem. Transportation Research Circular 435. Transportation Research Board (TRB), National Research Council (NRC), Washington, D.C., USA, Wright, P.H., and L.S. Robertson. Priorities for Roadside Hazard Modification: A Study of 300 Fatal Roadside Object Crashes. IIHS, Washington, D.C., USA, Puvanachandran, V.M. Effect of Road Curvature on Accident Frequency: Determining Design Speeds to Improve Local Curves. Road and Transport Research, 4 (1995): Agent, K.R. Transverse Pavement Markings for Speed Control and Accident Reduction. Kentucky Department of Transportation, Lexington, Ky., USA, Hungerford, J.C., and T.H. Rockwell. Modification of Driver Behavior by Use of Novel Roadway Delineation Systems. Report DOT HS USDOT, Rockwell, T.H., J. Malecki and D. Shinar. Improving Driver Performance on Rural Curves through Perceptual Changes Phase III. Ohio Department of Transportation, Columbus, Ohio, USA, Helliar-Symons, R.D. Yellow-Bar Experimental Carriageway Markings Accident Study. LR Transport Research Laboratory, Crowthorne, Berkshire, USDOT. Manual on Uniform Traffic Control Devices. Report SA FHWA, USDOT, Joksch, H.C. Velocity Change and Fatality Risk in a Crash A Rule of Thumb. Accident Analysis and Prevention, 25 (1993): Zador, P.L., H.S. Stein, P.H. Wright and J.W. Hall. Effects of Chevrons, Post-Mounted Delineators, and Raised Pavement Markers on Driver Behavior at Roadway Curves. Transportation Research Record TRB, NRC, Washington, D.C., USA, Lyles, R.W. An Evaluation of Warning and Regulatory Signs for Curves on Rural Roads. Report RO FHWA, USDOT, Agent, K.R., and T. Creasey. Delineation of Horizontal Curves Interim Report. Report No. UKTRP864. Kentucky University, Lexington, Ky., USA, Stokes, R.W., M.J. Rys and E.R. Russell. Motorist Understanding of Selected Warning Signs. ITE Journal, 66 (1996): Oei, H. Automatic Speed Management in the Netherlands. Transportation Research Record TRB, NRC, Washington, D.C., USA, Elvik, R. Effects on Accidents of Automatic Speed Enforcement in Norway. Transportation Research Record TRB, NRC, Washington, D.C., USA, Harwood, D. Synthesis of Highway Practice 191 Use of Rumble Strips to Enhance Safety. TRB, Washington, D.C., USA, Alexander, G.J., and H. Lunenfeld. Driver Expectancy in Highway Design and Traffic Operations. Report TO FHWA, USDOT, Zegeer, C.V., J.M. Twomey, M.L. Heckman and J.C. Hayward. Safety Effectiveness of Highway Design Features Volume II: Alignment. Report RD FHWA, USDOT, ITE JOURNAL / SEPTEMBER 1998

6 23. McGee, H.W., W.E. Hughes and K. Daily. Effect of Highway Standards on Safety. National Cooperative Highway Research Program 374. TRB, NRC, Washington, D.C., USA, RICHARD A. RETTING is a Senior Transportation Engineer with the Insurance Institute for Highway Safety. He is a Member of ITE and serves as Chair of the ITE Transportation Safety Council. CHARLES M. FARMER is a Senior Statistician with the Insurance Institute for Highway Safety. 36 ITE JOURNAL / SEPTEMBER 1998