Practice of Rumble Strips and Rumble Stripes (2016)

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.

21 interior sound levels, and may not produce adequate feedback to alert inattentive drivers (Terhaar 2015). Caltrans’s design also had a better tonal quality than MnDOT’s current design. Minnesota’s design produces a single, strong tonal peak at 125 hertz, which stands out against ambient noise because few sounds in the natural envi- ronment produce similar tones. California’s design produces two smaller peaks at 100 hertz and 200 hertz, so the sound is less abrupt. The noise of a rumble strip is considered detectable if it produces a sound level at a listener’s location greater than the ambient noise at any frequency. In passenger vehicles driving at 60 mph, the Caltrans design has been modeled as just detectable at 3,000 feet, whereas the Minnesota design would be detectable at well beyond 3,000 feet. This research showed, however, that Caltrans design only produces its greatest volume when a tire is fully on the rum- ble strip. MnDOT’s design provided feedback to the driver immediately after the tire made contact with the rumble strip. The Ohio DOT deals with noise issues in residential areas by adjusting the distance from the edge of where the rumble will be installed. The policy reads: In residential areas, noise generated by rumble strips could be objectionable. Rumble strips installed in these areas may be placed further from the edge of the traveled lane to reduce the frequency of contact while still providing some degree of warning to drifting drivers. The distance from the edge of the traveled lane to the rumble strip pattern should not exceed 2.0 feet on the outside shoulder. Also, the use of either rolled or formed rumbles is preferable to the use of the milled rumble in these areas. The Montana DOT deals with the noise issue by modifying the shoulder rumble strip dimensions. Its policy reads: The rumble strip dimensions shown in the MDT Detailed Draw- ings should be used for most installations. However, situations will occur when modifications to these dimensions must be evaluated. The following factors should be considered in the decision to modify the rumble strip dimensions: Depth of rumble strip—A 5⁄8 inch rumble strip depth is typi- cally used. The depth of a rumble strip can be reduced to minimum of 3⁄8 inch to provide a “quieter” pattern near residential areas. The 3⁄8 inch depth will not provide adequate noise/vibration after a chip seal has been placed, so the rumble strip would have to be This chapter provides an in-depth analysis of four key issues— noise, bicycle, safety benefits, and pavement deterioration— identified from the synthesis results; and discusses how some state DOTs are addressing those issues. NOISE ISSUES The survey found noise to be one of the most important issues identified by the state DOTs when it comes to rumble strip and stripes. Twenty-four agencies (almost 60%) ranked noise issues as the highest importance or one level below (5 or 4). When asked if the agency had developed a policy to address noise issues, 27 agencies (66%) responded yes. The informa- tion that follows is a sample of how some state DOTs addressed the noise issues resulting from rumble strip and stripes starting with alternative designs (Caltrans and MnDOT) and also the more traditional ways. In 2012, Caltrans conducted the most extensive background search by a state agency to date on noise issues by conduct- ing a comprehensive literature review (at national and inter- national levels), interviewing state DOTs, and conducted an evaluation of two alternative rumble designs to investigate their noise impacts. An excerpt from its report follows: External noise caused by traffic crossing rumble strips, while hav- ing beneficial safety effects for drivers, is a cause of concern for Caltrans and other departments of transportation (DOTs) because it generates complaints from homeowners and may affect pro- tected wildlife species. Caltrans is testing two potentially quieter forms of rumble strip: the sinusoidal rumble strip and the 5⁄16-in. milled rumble strip with thermoplastic stripe (CTC 2012). The study found that sinusoidal rumble strips provide suffi- cient audible and tactile warning to drivers and at the same time reduce noise outside the vehicle. In 2014, MnDOT initiated a research study to evaluate its existing rumble design (3⁄8-inch to 1⁄2-inch depth, 16 inches wide with 12-inch spacing) in comparison to the Caltrans (1⁄32- to 5⁄8-inch depth, eight inches wide with 14-inch spacing) and the Pennsylvania contractor pilot project (1⁄8- to 1⁄2-inch deep, eight inches wide with 24-inch spacing) sinusoidal rumble designs (Figure 10). The study showed that the California’s rumble strip design had the most efficient exterior-to-interior sound ratio: It produced as much noise inside the vehicle as the Minnesota design, but less sound outside of the vehicle. PennDOT’s contractor design produced lower exterior and chapter four CASE EXAMPLES

22 re-milled after every chip seal. Depths shallower than 3⁄8 inches do not provide enough noise/vibration to alert the errant driver. Rumble strips have often been terminated through residential areas due to the nuisance noise from incidental contact. MDT’s position is that rumble strips should be placed on highways in residential areas where the distance to the residences is adequate to minimize the adverse effect of rumble strip noise. Two options are available to decrease rumble strip noise through these areas: • The first option is to increase the offset from the edge of the travel lane. A greater offset can be beneficial where there is substantial truck traffic, because trucks tend to crowd the shoulder resulting in more “nuisance” contact. However, the greater offset will somewhat reduce the effectiveness of the rumble strip. • The second option is to use a quieter (less aggressive) rumble strip. The depth of the rumble strip can be reduced to 3⁄8 inch to decrease noise. As noted above, the 3⁄8 inch rumble strip will not provide adequate noise after a chip seal has been placed over the strip, so they will have to be re-milled after every chip seal application. • If the decision is made to eliminate rumble strips, research has indicated that terminating rumble strips 650 feet (200 m) from residences resulted in tolerable noise levels. The policy of the Saskatchewan Ministry of Highways and Infrastructure in Canada states, “Installation of shoul- der rumble strips through urban areas is not recommended because of the noise.” BICYCLE ISSUES The survey found bicycle issues is second only to noise among issues identified by the state DOTs when it comes to rumble strip and stripes. Twenty-one (21) agencies, just over half, ranked bicycle concerns at the highest importance or a level below (5 or 4). When asked if the DOT developed a policy to address noise issues, 34 agencies (83%) responded yes. Fol- lowing is a sample of agency approaches: The South Carolina DOT (SCDOT) addressed bicycle con- cerns by establishing a policy that adjusts their current rumble strip design standards. The rumble strip design details and location criteria were developed after receiving input from the South Carolina cycling community, FHWA and other state DOTs. The SCDOT has implemented the following accom- modations to address the presence of cyclists: • Reducing maximum depth of milled groove, • Providing an option of various width rumble strips based on width of paved shoulder, • Providing an option of a skip pattern for the milled-in rumble strips, • Establishing minimum ADT threshold for rumble strip application, • Establishing a minimum roadway width for rumble strips, and • Where RS are placed on bike lanes, a minimum width of 3 feet and 6 inches will remain undisturbed. To ensure that bicycle concerns are addressed, the Mary- land State Highway Administration rumble strip and stripe guidelines regarding bicycles indicates that All future revisions to rumble strip design document that may impact bicyclists shall require the notification to both the [State Highway Administration] Bicycle and Pedestrian Coordinator within the Office of Planning and Preliminary Engineering and the [Maryland] DOT Director of Bicycle and Pedestrian Access within the Office of Planning and Capital Programming to gain their input on proposed changes. FIGURE 10 Comparison of rumble strip cross sections (Minnesota, California, and Pennsylvania) (Terhaar 2015).

23 This basically ensures that bicycle concerns will always be considered as part of the design process. The Arkansas DOT indicated that a common complaint of bicyclists is that a continuous shoulder rumble strip along a narrow shoulder sometimes requires bicyclists to move into the travel lane. To address this concern, it is important that shoulder rumple strips with a gap pattern be installed on highways that do not have full access control and have at least four feet of shoulder beyond the rumble strip. The typi- cal longitudinal pattern will consist of 48 feet of rumble and 12 feet without rumble. When driveways or intersections are present, the use of a gap pattern will be adjusted at the discre- tion of the Engineer so that the driveway or intersection may be utilized as a gap. In addition, the ³⁄8-in. depth of grooves of rumble strips on rural, undivided highways is considered to be less disruptive to bicyclists traveling on these routes. The Ohio DOT takes a different perspective on dealing with bicycle concerns. Its rumble policy indicates that rumble strips generally are not to be used on the shoulders of road- ways designated as bicycle routes or having substantial vol- umes of bicycle traffic, unless the shoulder is wide enough to accommodate the rumble strips and still provide a minimum clear path of four feet from the rumble strip to the outside edge of the paved shoulder or five feet to adjacent guardrail, curb, or other obstacle. In areas designated as bicycle routes or having substan- tial volumes of bicycle traffic, the rumble strip pattern would not be continuous but consist of an alternating pattern of gaps and strips, each 10 feet in length. Also, gaps are to be provided in the rumble strip pattern ahead of intersections, crosswalks, driveway openings, and at other locations where bicyclists are likely to cross the shoulder. The Kansas DOT specifies that a minimum three-foot paved area outside of the shoulder rumble strip should be provided for bicyclists on highway routes on the American Discovery Trail Route, Trans America Route and other sug- gested cross-state bicycle routes per the latest edition of the Kansas Bicycle Guide. Also, a recent research study com- pleted in 2007 (KSU-00-4: Comparison of Football Shaped Rumble Strips Versus Rectangular Rumble Strips), found that bicyclists prefer the “football” shape shoulder rumble strip when traveling on highway shoulders. The football shape, according to this study, is easier to traverse by a cyclist, but still provides the same noise levels for motorists. Montana DOT also modifies its rumble strip design to address bicycle concerns. The policy states that where bicy- cle usage is a consideration, the following modifications to rumble strip installation should be evaluated: • Where the shoulder width is between 1 and 4 ft (1 ft < shoulder width < 4 ft) reduced lateral width rumble strips should be installed adjacent to the outside edge of the pavement. • Where shoulder widths are greater than 4 feet, the use of a 2-ft offset from the edge of travel lane should be evaluated. This offset will provide an area for bicyclists between the edge of travel lane and the rumble strip. • Where center line rumble strips will be used on narrow roads, it should be noted that drivers tend to shy away from center line rumble strips which could adversely affect vehicle/bicycle interaction. The use of a modified lateral width rumble strip should be evaluated. In all cases the benefits to bicyclists must be weighed against the potential for roadway departure incidents, because greater offsets reduce the effectiveness of rumble strips. In Arizona, the DOT developed a shoulder width policy to address bicycle concerns. The policy states that: If appreciable bicycle traffic exists or is anticipated then a mini- mum effective clear shoulder width of three-feet and five-inches (3′-5″) should be provided from the outside edge of the rumble strip groove to the front face of the barrier or guardrail. If this clear area cannot be maintained then a change of configuration and/or deletion of the rumble strip should be considered. SAFETY BENEFITS OF RUMBLES Over the years, multiple state DOTs have conducted studies to examine the safety benefits of installing rumbles on the edge line, center line, or both. NCHRP Report 641 documented 11 state studies and conducted a national study to deter- mine the safety benefits. Research has shown that installing center line rumble strips can reduce severe crashes as much as 45% on rural two-lane roads and by 64% on urban two- lane roads; and that shoulder rumbles can reduce crashes as much as 36% on rural two-lane roads and by 17% on rural freeways. Of the 41 state DOTs responding to the survey, 10 indi- cated that CMFs effectiveness values have been determined for rumble strips and stripes; 11 discussed cost effective- ness based on crash reductions (fatal and injury crashes) compared with the cost of installing rumbles (six agencies responded to both options). PennDOT developed effectiveness graphs to show the impact of installing rumbles (edge and center line) on crashes, and published them in its state highway safety report. Fig- ure 11 shows the effectiveness of the edge line rumble strip in reducing run-off the road fatalities, and Figure 12 shows the effectiveness of the center line rumble in reducing head- on fatalities. Figure 11 shows the relationship between the number of run-off-the-road fatalities (based on a five-year running aver- age) and the total number of miles that have edge line rumbles

24 FIGURE 11 Pennsylvania DOT edge line rumble effectiveness. FIGURE 12 Pennsylvania DOT center line rumble effectiveness.

25 installed. The data shows that as the number of miles increased, the average number of run-off-the-road fatalities decreased from a high of 745 in 2008, with more than 1,100 miles of edge line rumbles, to a low of 612 in 2014 with more than 4,300 miles of edge line rumbles. That is a reduction of 133 run-of the road fatalities in 6 years, representing an approx- imately 18% reduction. Figure 12 shows the relationship between the number of head-on fatalities (based on a 5-year running average) and the total number of miles with center line rumbles installed. The data show that as the number of miles with center line rumbles increased, the average number of head-on crashes decreased from a high of 168 in 2008, with more than 2,500 miles of center line rumbles, to a low of 148 with more than 5,100 miles of center line rumble. This is a reduc- tion of 20 head-on fatalities over 6 years representing about a 14% reduction. As part of the MnDOT “Towards Zero Deaths” initiative, the DOT developed a brochure to show the safety impact of rumble strips “Saving Lives by Keeping Drivers Focused: Noise from Centerline Rumble Strips” (MnDOT 2015). The brochure presents information from one two-lane rural high- way from 1986 to 2011. The data show that 25 people were killed in head-on crashes (an average of one per year). The brochure then addresses how rumble strips reduce those types of crashes. Figure 13 shows a partial image of the brochure. A study by the Michigan DOT (Datta 2015) addressed the impact of center line rumble strips on safety. The analysis revealed a total of 2,488 “before” and 1,306 “after” target crashes (incidents involving at least one vehicle crossing or encroaching on to the center line, resulting in a crash) with regard to the center line rumble strip installation period. The crash analysis indicated statistically significant reductions after rumble installations in all target crashes, including FIGURE 13 Minnesota DOT brochure on how rumble strips save lives (MnDOT 2015).

26 head-on, sideswipe opposite, and single vehicle run-off-the- road incidents. The study of crash severity resulted in the reduction in fatalities and all categories of injury crashes (Table 19). These reductions were statistically significant and ranged from 43% to 55%. A study of the safety impact of various traffic volume (average annual daily traffic) groups and crash factors also indicated statistically significant crash reduc- tions after center line rumble strip installation. Passing-related target crashes were reduced by 47%, those occurring on wet pavement by 54%, and others that involved impaired drivers as a result of alcohol or drug use were reduced by 35%. PAVEMENT DETERIORATION There appears to be little consensus concerning the issue of pavement deterioration resulting from rumble installation. When the DOTs ranked the issues in terms of importance, pavement deterioration (focused more on center line joint deterioration) was third, after noise and bicycle issues. How- ever, a closer look at the survey results reveals that while 17 DOTs rated the issue as important (5 or 4), 18 ranked it as not important (1 or 2). In addition, 10 ranked pavement deterioration at 5; the same number ranked it at 1, the lowest level. Of the 17 agencies who identified pavement deteriora- tion as a high-level issue, 12 reported they apply sealants over the rumble to protect the pavement. As was indicated in the literature review, FHWA states on its rumble strip website (FHWA 2015a) that “Maintenance crews were initially concerned that heavy traffic would cause shoulder pavements with rumble strips to crumble faster, or that the freeze-thaw cycle of water collecting in the grooves would crack the pavement. These worries have proved to be unfounded where rumble strips were installed in pave- ments in fair to good condition. Rumble strips have little if any effect on the rate of deterioration of new pavements.” This provides anecdotal evidence that the pavement deterio- ration resulting from rumbles is not an issue, but does not provide specific information based on field research to back this statement. Two states, Kentucky and Michigan, have investigated the topic. In a study on rumble trips conducted for the Kentucky Transportation Cabinet (Kirk 2008), the effect of the rum- ble on pavement deterioration was of special concern. To address these concerns, a special meeting was held with maintenance personnel from Districts 6, 9, and 11, where limited applications of center line rumble strips exist on the Daniel Boone Parkway, Mountain Parkway, and AA High- way. Pavement deterioration along the center line joint was noted on the Mountain Parkway and Daniel Boone Park- way; however, it was noted that this was a retrofit application and pavement performance was poor before the rumble strip placement. Other applications on new pavement were not reported to have suffered any significant problems of pavement deterioration. Though the Kentucky study provided further anecdotal evidence that rumbles do not negatively impact pavement condition, the scope was very limited and not based on com- paring field pavement performance data. The Michigan DOT study (Datta et al. 2012) investi- gated the short-term pavement performance and the impact of the rumble on non-freeway segments. The effects of cen- ter line rumble strips were assessed by comparing the rate of crack propagation between road segments where rumble strips were installed, and similar control segments where rumble strips were not installed. The study considered other factors such as traffic, pavement age, and region in deter- mining their samples to conduct the comparison. The study showed that in each case, the increase in cracks during the two-year analysis period was marginally higher in the con- trol sections in comparison to rumble strip sections. The differences were not statistically significant, but these data TABLE 19 “BEFORE AND AFTER” SAFETY PERFORMANCE Source: Datta (2015).

27 suggest that rumble strips did not create adverse impacts on pavement performance in the short-term. Table 20 shows the comparison results. The Michigan DOT study on the impact of rumbles on pavement condition is the most comprehensive and field- data oriented. Even though it showed no impact, the study only considered short-term (2 years) effects, and pavement and material engineers may argue that it is the long-term performance of the pavement that is critical and that could be impacted by rumbles. This might be a topic for further research. TABLE 20 COMPARISON OF INCREASE IN CRACKS BETWEEN RUMBLE STRIP AND CONTROL SECTIONS Source: Datta (2012).