Factors Contributing to Median Encroachments and Cross-Median Crashes (2014)

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.

72 Introduction This appendix describes methods to reduce the frequency and severity of median-related crashes on divided highways. Guidelines for reducing median-related crashes can be classified into three general approaches: 1. Design guidance to reduce consequences of median encroachments, 2. Design guidance to reduce likelihood of median encroach- ments, and 3. Countermeasures to reduce likelihood of median encroachments. Highway agency efforts to improve median safety on divided highways have focused primarily on the first approach pre- sented above—designing or retrofitting medians of divided highways to reduce the consequences of median encroach- ments. The AASHTO Roadside Design Guide and other resources provide guidance on implementing this approach. Although this research encourages the first approach, it also directs highway agency attention to the second and third approaches, which are intended to reduce the likelihood that motorists will run off the road into the median in the first place. An effective program to reduce median-related crashes should consider a mix of these three approaches. Table D-1 summarizes the design guidance and counter- measures that should be considered in median safety programs. This table includes both general objectives for implement- ing each of the three approaches to reducing median-related crashes and specific design guidance or countermeasures that should be considered for implementation. The following sections review specific design guidelines and countermeasures and include, where available, estimates of their safety effectiveness. The safety effectiveness of design guidelines and countermeasures is based on three general sources: the Roadside Safety Analysis Program (RSAP) model (Mak and Sicking, 2003); the new safety prediction procedure for freeways developed in NCHRP Project 17-45 (Bonneson et al., 2012) for implementation in the AASHTO Highway Safety Manual (HSM) (AASHTO, 2010); and other individ- ual sources in the literature. Both RSAP and the HSM include safety prediction models and cost-effectiveness/benefit-cost procedures for assessing design treatments and counter- measures. Unfortunately, the RSAP and HSM safety predic- tion models are based on inconsistent assumptions, since RSAP is based on estimated median encroachment frequen- cies and the HSM is based on reported crash frequencies. Research is under way in NCHRP Project 17-54 to try to make the RSAP and HSM more compatible, but no results are yet available. Design Guidance to Reduce Consequences of Median Encroachments The consequences of median encroachments can be reduced through the following design treatments: • Remove, relocate, or use breakaway design for obstacles in median; • Provide barrier to shield objects in median; • Provide wider medians; • Provide continuous median barrier; • Flatten median slopes; • Provide U-shaped (rather than V-shaped) median cross sections; and • Provide barrier to shield steep slopes in medians. All of these design improvements have the potential to reduce crash severity for motorists who run off the road into the median. Current design guidance for these treatments is presented in the AASHTO Green Book (AASHTO, 2011a) and the AASHTO Roadside Design Guide (AASHTO, 2011b). The A p p e n d i x d Guidelines for Reducing the Frequency and Severity of Median-Related Crashes on Divided Highways

73 following discussion reviews each design treatment and doc- uments known effectiveness measures for those treatments. Several of these treatments, by their nature, have high costs and may be most appropriate for consideration in new con- struction and reconstruction projects, rather than for appli- cation to existing roadways. Remove, Relocate, or Use Breakaway Design for Obstacles in Median A key strategy for improving median safety is to remove or relocate fixed objects located within the median in positions were they may be struck by vehicles that leave the roadway and enter the median. Such objects may include trees, util- ity poles, luminaire supports, signs and sign bridge supports, culverts and their appurtenances, bridge abutments and bridge supports, and other miscellaneous fixed objects. For any median fixed object that is reviewed, all three potential treatments—removing the object completely, relocating it to a less exposed position (farther from the roadway or less likely to be struck), or converting the object to a breakaway design—should be considered. Another potential alternative is providing guardrail to shield objects in the median (see next section on shielding). The AASHTO Roadside Design Guide (AASHTO, 2011b) provides detailed guidance on treatment of roadside obstacles. The RSAP model provides a benefit-cost analysis tool to assess roadside design issues, including the location and type of roadside fixed objects (Mak and Sicking, 2003). RSAP is intended for analysis of the roadside on the outside of the Objective Design guidance/countermeasure Design guidance to reduce consequences of median encroachments Minimize potential for collision with fixed objects Relocate or remove fixed objects in median Reduce consequences of collision with fixed objects Provide barrier to shield objects in median Reduce likelihood of cross-median collisions Provide wider medians Provide continuous median barrier Reduce likelihood of vehicle overturning Flatten median slopes Provide U-shaped (rather than V-shaped) median cross section Provide barrier to shield steep slopes Design guidance to reduce likelihood of median encroachments Improve design of geometric elements Provide wider median shoulder Minimize sharp curves with radii less than 3,000 ft Minimize steep grades of 4% or more Improve design of mainline ramp terminals Increase separation between on- and off-ramps Minimize left-hand exits Improve design of merge and diverge areas by lengthening speed-change lanes Simplify design of weaving areas Increase decision sight-distance to on-ramps Countermeasures to reduce likelihood of median encroachments Reduce driver inattention Provide edgeline or shoulder rumble strips Decrease side friction demand Improve/restore superelevation at horizontal curves Increase pavement friction Provide high-friction pavement surfaces Improve drainage Improve road surface or cross-slope for better drainage Reduce high driver workload Improve visibility and provide better advance warning for on- ramps Improve visibility and provide better advance warning for curves and grades Improve delineation Encourage drivers to reduce speeds Provide transverse pavement markings Minimize weather-related crashes Provide weather-activated speed signs Provide static signs warning of weather conditions (e.g., bridge freezes before road surface) Apply sand or other materials to improve road surface friction Apply chemical de-icing or anti-icing as a location-specific treatment Improve winter maintenance response times Install snow fences Raise the state of preparedness for winter maintenance Table D-1. Design guidelines and crash countermeasures to reduce median-related crashes.

74 New HSM procedures for freeways developed in NCHRP Project 17-45 (Bonneson et al., 2012) include the following crash modification functions (CMFs) for median width: CMF 1.0 –P exp a W – 2W – 48 P exp a 2W – 48 (D-2) ib m is ib icb ( ) [ ] [ ] ( ) ( ) = + where Pib = proportion of segment length with a barrier present in the median Wm = median width (measured between the inside edges of the traveled way) (ft) Wicb = distance from edge of inside shoulder to barrier face Wis = width of inside shoulder (ft) a = regression coefficient: –0.00302 for multiple-vehicle FI crashes –0.00291 for multiple-vehicle PDO crashes 0.00102 for single-vehicle FI crashes –0.00289 for single-vehicle PDO crashes Provide Continuous Median Barrier Highway agencies often provide continuous median barrier to reduce the risk of vehicles running across the median— CMC and NCMC crashes—to near zero. The AASHTO Road- side Design Guide (AASHTO, 2011b) provides guidance on situations in which a median barrier should be considered on divided highways as a function of median width and traffic volume (see Figure D-1). The figure shows that median bar- riers are considered warranted in specific design situations where the traffic volume on the divided highway exceeds 30,000 vehicles per day. Three types of median barrier are in general use: concrete barrier, metal guardrail, and cable barrier. Concrete barrier is generally placed at the center of a median. Metal guard- rail is generally placed in the center of the median only in relatively narrow medians; in wider medians, metal guardrail may be placed near the edge of the inside shoulder so that any vehicles that strike the guardrail will do so at a relatively shallow angle. Cable barrier has typically been placed in the center of the median, but some highway agencies have begun placing cable barrier just outside the shoulder on one side of the median or on the median slope toward one side of the median. NCHRP Project 22-22 has been charged with devel- oping placement guidelines for median barriers, but is not yet complete. Research by Graham et al. (2011) found that cable barri- ers may be cost-effective in some divided highway medians with average daily traffic volumes as low as 10,000 vehicles per day. roadway, but analysis of the equivalent fixed objects provides the best available method to analyze objects in the median. Provide Barrier to Shield Objects in Median Another alternative for consideration where objects in the median of a divided highway cannot be removed, relocated, or converted to breakaway design is to provide a guardrail or other barrier to prevent vehicles that run off the road into the median from striking the objects. This approach is generally less desirable than removing, relocating, or converting the object to breakaway design because guardrail is itself a road- side obstacle that can kill or injure drivers and passengers in vehicles that strike the guardrail. Thus, guardrail is generally provided for roadside obstacles only when removing, relocat- ing, or converting the object to breakaway design is impractical and only when it is verified that the risk to vehicle occupants from striking the guardrail is less that the risk of striking the object. Tradeoff analyses of the relative risks of striking guard- rails or obstacles are generally performed with the RSAP model (Mak and Sicking, 2003; RoadSafe LLC, 2012a; RoadSafe LLC, 2012b). Concrete barriers and cable barriers can be considered as an alternative to metal guardrail in these situations. Provide Wider Medians Wider medians generally reduce the severity of median encroachments by making it less likely that vehicles that run off the road and enter the median will cross the entire median, enter the opposing roadway and become involved in a cross- median collision (CMC) crash or a non-collision cross-median crash (NCMC). A wider median should provide more oppor- tunity for drivers of vehicles entering the median to be able to recover control of the vehicle or come safely to rest. Design guidelines for median width are presented in the AASHTO Green Book (AASHTO, 2011a). The design guidance in the AASHTO Roadside Design Guide (AASHTO, 2011b) focuses specifically on design criteria for median barriers (which are, in part, influenced by median width) and design criteria for fixed objects in the median. Pennsylvania research has quantified the influence of median width on CMC crashes as follows (Mason et al., 2001): 0.2 (D-1)–18.203 1.770 –0.0165N e L AADT eCMC MW= × × × × where NCMC = number of CMC crashes per year for one direction of travel L = segment length (mi) ADT = average daily traffic (veh/day) MW = median width (ft)

75 (AASHTO, 2011a) recommends the use of 1V:6H slopes in design of medians. Recent research in NCHRP Project 22-21 (Graham et al., 2011) has recommended that flatter 1V:8H slopes be considered instead. Provide U-Shaped (Rather than V-Shaped) Median Cross-Sections The AASHTO Green Book (AASHTO, 2011a) does not pro- vide specific guidance on the overall shape of median cross sections. However, many highway agencies build divided highway medians with V-shaped cross sections, as continuous slopes from the edge of each inside shoulder meet at the cen- ter of the median. Recent research in NCHRP Project 22-21 (Graham et al., 2011) recommended that U-shaped median cross sections be considered in preference to V-shaped medi- ans. The flat area in the center of the median appears to create a preferable design. New HSM procedures for freeways developed in NCHRP Project 17-45 (Bonneson et al., 2012), include the following CMFs for placement of continuous median barrier: ( )= +   CMF 1.0 –P P exp a W (D-3)ib ib icb where a = regression coefficient (0.240 for FI crashes and 0.169 for PDO crashes) Flatten Median Slopes Providing flatter median slopes can reduce the likelihood that drivers of vehicles that run off the road into the median will be able to avoid overturning and will be able to regain control of their vehicles or come to a stop before crossing the median or striking a fixed object. The AASHTO Green Book Figure D-1. Guidelines for median barriers on high-speed, fully controlled access roadways (AASHTO, 2011b).

76 Provide Wider Median Shoulder The median or inside shoulder of a divided highway is intended as a primary recovery area for vehicles that begin to leave the roadway and, in some cases, as a primary stopping and recovery area. This distinction is made because many highway agencies prefer to provide relatively narrow (4-foot) median shoulders to discourage motorists from stopping on the median shoulder; a wider shoulder is provided on the right side of the roadway and often highway agencies pre- fer that motorists that need to stop use the right or outside shoulder for that purpose. The new HSM freeway procedure developed in NCHRP Project 17-45 includes the following CMF for inside or median shoulder width: [ ]( )= −CMF exp a W 6 (D-4)is where a = regression coefficient (-0.0172 for FI crashes and -0.0153 for PDO crashes) Table D-2 shows the crash modification factor (CMF) for converting from one inside shoulder width to another, based on Equation D-4. These CMFs apply to all crash types, not specifically to median-related crashes. There are no equiva- lent CMFs for divided highways that are not freeways. It should be noted that the CMFs in Table D-2 do not show any reason to discourage wider inside shoulders—the wider the inside shoulder, the lower the resulting crash rate. It should be recognized, however, that crash prediction models are not necessarily good tools to reflect such design guidance. Nothing in the new HSM material should be taken as imply- ing that highway agency preferences for using narrow insider Provide Barrier to Shield Steep Median Slopes Just as guardrail and other barriers can be used to shield fixed objects in the median from being struck by vehicles that run off the road, guardrail or other barriers also can be used to ensure that vehicles that run off the road do not encounter steep slopes on which they might overturn. As in the case of barrier to shield fixed objects, the most applicable tools of barrier to shield steep slopes in median areas are the RSAP model (Mak and Sicking, 2003; RoadSafe LLC, 2012a; RoadSafe LLC, 2012b) and the new freeway models for the HSM (AASHTO, 2010) developed in NCHRP Project 17-45 (Bonneson et al., 2012). Design Guidance to Reduce Likelihood of Median Encroachments The likelihood of median encroachments can be reduced through the following design treatments: • Provide wider median shoulder, • Minimize use of sharp curves with radii less than 3,000 feet, • Minimize use of steep grades of 4 percent or more, • Increase separation between on- and off-ramps, • Minimize left-hand exits, • Improve design of merge and diverge areas by lengthening speed-change lanes, • Simplify design of weaving areas, and • Increase decision sight-distance to on-ramps. The importance of these design treatments to the likelihood of median-related crashes is not well addressed in current design guidance and, therefore, is addressed here. Inside Shoulder Width (ft) (Before) Inside Shoulder Width (ft) (After) 2 4 6 8 10 12 Fatal-and-Injury Crashes (FI) 2 1.00 0.97 0.93 0.90 0.87 0.84 4 1.03 1.00 0.97 0.93 0.90 0.87 6 1.07 1.03 1.00 0.97 0.93 0.90 8 1.11 1.07 1.03 1.00 0.97 0.93 10 1.15 1.11 1.07 1.03 1.00 0.97 12 1.19 1.15 1.11 1.07 1.03 1.00 Property-Damage-Only Crashes (PDO) 2 1.00 0.97 0.94 0.91 0.88 0.86 4 1.03 1.00 0.97 0.94 0.91 0.88 6 1.06 1.03 1.00 0.97 0.94 0.91 8 1.10 1.06 1.03 1.00 0.97 0.94 10 1.13 1.10 1.06 1.03 1.00 0.97 12 1.17 1.13 1.10 1.06 1.03 1.00 Table D-2. CMFs for changing inside shoulder width on freeways (computed from results of Bonneson et al., 2012).

77 Increase Separation Between On- and Off-Ramps Research by Harwood et al. (2014) found that short sepa- ration distances between ramps are a contributing factor to median-related crashes. Divided highway sections adjacent to both on- and off-ramps were found to have higher crash rates than divided highway sections not located adjacent to ramps. Interdisciplinary field reviews concluded that short separa- tions between adjacent ramps were a contributing factor to median-related crashes. Most interchanges consist of an off-ramp followed by an on-ramp. Other ramp combinations, with ramps spaced less than 0.3 miles apart (gore-to-gore spacing), within the same interchange or adjacent interchanges, appear to be contribut- ing factors to median-related crashes, including: • On-ramp followed by off-ramp, • On-ramp followed by on-ramp, and • Off-ramp followed by off-ramp. It is clearly impractical to reconstruct existing interchanges solely to reduce median-related crashes, so this design guide- line applies to new divided highways that may be constructed in the future. The role of on- and off-ramps in median-related crashes also suggests that, where median improvements to reduce crash severity (like those discussed above) are being considered, it may be appropriate to apply those treatments in the vicinity of ramps—especially closely spaced ramps—even when they are not applied along the full length of the roadway. Minimize Left-Hand Exits Research by Harwood et al. (2014) suggests that left-hand exits are a contributing factor to median-related crashes. It is clearly impractical to reconstruct existing interchanges solely to reduce median-related crashes, so this design guideline applies to new divided highways that may be constructed in the future. Equations D-5 and D-6 show CMFs indicating that left- hand ramps experience substantially more crashes than do right-hand ramps. These CMFs, of course, apply to all crash types and not only to median-related crashes. The role of left-hand exits in median-related crashes also suggests that, where median improvements to reduce crash severity (like those discussed above) are being considered, it may be appropriate to apply those treatments in the vicin- ity of left-hand exits—especially closely spaced ramps— even when they are not applied along the full length of the roadway. shoulders and wider outside shoulders necessarily needs to be changed. Minimize Use of Sharp Curves with Radii Less Than 3,000 Feet Research by Harwood et al. (2014) identified sharp curves with radii less than 3,000 feet as a key contributing factor to median-related crashes. This finding was suggested in inter- disciplinary field reviews of crash history and roadway char- acteristics at selected divided highway sites and confirmed in crash data analyses for rural freeways in Washington. To minimize median-related crashes, it is recommended that curve radii less than 3,000 feet be avoided on divided high- ways, whenever practical. It is clearly impractical to change the curve radii for existing roadways and it may be impractical to avoid designing sharp curves in mountainous terrain, so this design guideline applies to new divided highways that may be constructed in the future. The role of sharp curves in median-related crashes also suggests that where median improvements to reduce crash severity (like those discussed above) are being considered, it may be appropriate to apply those treatments to curves with radii less than 3,000 feet, even when they are not applied along the full length of the roadway. Minimize Use of Steep Grades of 4 Percent or More Research by Harwood et al. (2014) identified steep grades of 4 percent or more as a key contributing factor to median-related crashes. This finding was suggested in interdisciplinary field reviews of crash history and road- way characteristics at selected divided highway sites and confirmed in crash data analyses for rural freeways in Washington. To minimize median-related crashes, it is recommended that steep grades of 4 percent or more be avoided on divided highways, whenever practical. It is clearly impractical to change the curve radii for existing roadways, and it may be impractical to avoid designing steep grades in mountainous terrain, so this design guideline applies to new divided high- ways that may be constructed in the future. It should be noted that the design guidelines for the Interstate Highway System generally limit grades to 3 percent, except in mountainous terrain. The role of steep grades in median-related crashes also suggests that, where median improvements to reduce crash severity (like those discussed above) are being considered, it may be appropriate to apply those treatments to steep grades of 4 percent or more, even when they are not applied along the full length of the roadway.

78 c = regression coefficient (0.001 for all crashes) d = regression coefficient (0.198 for FI crashes and 0.0 for PDO crashes) Equations D-5 and D-6 illustrate that longer speed-change lanes result in lower crash frequencies. These CMFs apply to all crash types, not specifically to median-related crashes. Simplify Design of Weaving Areas Some short weaving areas require drivers to make two lane- change maneuvers to complete a merge or diverge maneuver. Driver workload can be substantially reduced by modifying the design or the marking plan so that only one lane-change maneuver is required. Driver lane-change distance requirements, including the time to search for and recognize gaps, were studied by McGee et al. (1978). They found that in low-volume conditions a single lane change requires 8 seconds, and, in high volume conditions, 9.8 seconds are required. These data were based on observed single lane change times from a freeway study combined with very limited data on gap search and recogni- tion time from 12 young male drivers. There are better data available on times for two lane changes, based on time and distance recorded for 20 subjects driving an instrumented vehicle on a multi-lane highway in light (725 vehicles/hour or less), medium (726 to 1,225 vehicles/ hour), and heavy (> 1,225 vehicles/hour) traffic. The posted speed limit was 55 mph. Distance was calculated according to the speed traveled and the time taken from signaling to turn from the left-most lane until all four wheels had crossed into the right-most lane. On high-speed roadways (55 mph and above), the average time to complete two lane changes was 15 seconds in medium traffic and 17 seconds in heavy traffic (McNees, 1982). At a design speed of 120 km/h (75 mph), time for one and two lane changes corresponds to travel dis- tances of 1,080 feet and 1,870 feet, respectively. The reduced workload by means of reducing a merge from two to one lane change is associated with a crash risk reduction of 32 percent for crashes in the merging lane (AASHTO, 2010). The design of weaving areas also can be improved by pro- viding collector-distributor roads so that weaving maneuvers do not occur adjacent to the mainline roadway. Redesign of interchanges in this way is, of course, very expensive. Increase Decision Sight-Distance to On-Ramps Research by Harwood et al. (2014) suggests that limited- sight-distance for traffic approaching on-ramps at some sites is a contributing factor to median-related crashes. Improve Design of Merge and Diverge Areas by Lengthening Speed-Change Lanes Research by Harwood et al. (2014) highlighted on- and off- ramps as key contributing factors in median-related crashes. A key element that affects the operation of on- and off-ramps is the design of the merge and diverge areas and their associ- ated speed-change lanes (acceleration and deceleration lanes). A study of driver physiological workload suggested that the physiological stress associated with merging onto a freeway was 2.2 times as high as in a “basic driving section,” peaking at 260 feet beyond the gore for a ramp posted at 50 km/h (31 mph) and a freeway posted at 80 km/h (50 mph), and subsiding 4 seconds after merging (Chang et al., 2001). An acceleration lane that does not allow drivers a comfortable amount of time to reach freeway speed is likely to increase this stress. HSM Part D indicates that lengthening the speed change lane by 100 feet from a base condition of 690 feet is associated with a 7 percent reduction in crashes (AASHTO, 2010). How- ever, the confidence interval is large and includes the possibil- ity of a crash increase. The new HSM freeway procedure developed in NCHRP Project 17-45 includes the following CMF for ramp exits (deceleration lanes): = + CMF exp a I b L (D-5)left ex where Ileft = side of road for ramp (1 for left-side ramp; 0 for right-side ramp) Lex = length of deceleration lane (ft) (range: 160 to 1,600 ft) a = regression coefficient (0.594 for FI crashes and 0.824 for PDO crashes) b = regression coefficient (0.0116 for FI crashes and 0.0 for PDO crashes) The comparable equation for ramp entrances (accelera- tion lanes) is [ ]= + + CMF exp a I b L d ln c AADT (D-6)left en r where Len = length of acceleration lane (ft) (range: 370 to 1,600 ft) AADTr = average daily traffic volume for ramp crashes (veh/day) a = regression coefficient (0.594 for FI crashes and 0.824 for PDO crashes) b = regression coefficient (0.0318 for FI crashes and 0.0252 for PDO crashes)

79 • Improve delineation; • Provide transverse pavement markings; • Provide weather-activated speed signs; • Provide static signs warning of weather conditions (e.g., bridge freezes before road surface); • Apply sand or other materials to improve road surface friction; • Apply chemical de-icing or anti-icing as a location-specific treatment; • Improve winter maintenance response times; • Install snow fences; and • Raise the state of preparedness for winter maintenance. The importance of these countermeasures is supported by the results of research by Harwood et al. (2014). That research found wet and snow-covered pavements to be a key contributing factor in median-related crashes. Therefore, any countermeasures that reduce the likely duration of exposure of motorists to adverse pavement conditions or the conse- quences of that exposure should help reduce median-related crashes. With the exception of the first countermeasure dis- cussed, shoulder rumble strips, the role of these countermea- sures in reducing the likelihood of median-related crashes is not well addressed in current design guidance and, therefore, is addressed here. Provide Edgeline or Shoulder Rumble Strips The most widely known countermeasure for reducing the likelihood of median encroachments is providing edgeline and shoulder rumble strips. Fatigue and distraction are commonly experienced by drivers and easily can lead to momentary in- attention to the road path and, thus, to a median encroachment. Edgeline and shoulder rumble strips alert drivers that they are about to leave the paved roadway and help drivers correct their maneuvers before encroaching on the median. Shoulder edge rumble strips have been shown to decrease single-vehicle run-off-road crashes by 18 percent on rural and urban freeways combined and by 21 percent on rural freeways (Griffith, 1999). Torbic et al. (2009) provided the following estimates on the safety effectiveness of shoulder rumble strips on freeways: Urban/Rural Freeways Rolled shoulder rumble strips – 18 percent reduction in single-vehicle run-off-the-road (SVROR) crashes (Standard Error [SE] = 7) and – 13 percent reduction in SVROR fatal and injury (FI) crashes (SE = 12). Rural Freeways Shoulder rumble strips – 11 percent reduction in SVROR crashes (SE = 6) and – 16 percent reduction in SVROR FI crashes (SE = 8). Decision sight-distance is defined as the sight distance that should allow drivers to detect an unexpected or difficult- to-perceive information source or condition, recognize the condition or its potential threat, select an appropriate speed and path, and initiate and complete the maneuver safely and efficiently (Alexander and Lunenfeld, 1975). Examples of traffic control devices and road geometric elements that are high priority with respect to the need to apply or consider decision sight-distance so that drivers can change lanes com- fortably include the following among others: • A guide sign; • Lane markings indicating a change in cross-section; and • A change in cross-section (two-lane to four-lane, four- lane to two-lane, passing lane, climbing lane, lane drop, optional lane split, deceleration lane, channelization) (Lerner et al., 2004). On a freeway, drivers are assisted in identifying the exit or entrance if the gore area has good sight-distance. It has been shown that a significant proportion of drivers wait until they can see the road layout before changing lanes (McGee et al., 1978). The best data on the amount of sight distance that drivers require is probably the data on decision sight-distance col- lected by McGee et al. (1978). Drivers were observed as they responded to unusual highway features such as lane drops at an exit or a lane split. Despite warning devices of various kinds, in approximately half the approaches, subjects did not start responding until they actually saw the feature in ques- tion. Detection and recognition times varied from a mini- mum of 1.5 to 3.0 seconds depending on the complexity of the change in the roadway. Decision and response initiation, which includes time to search for a gap prior to changing lanes, took 4.2 to 7.0 seconds and lane change required 3 to 4.5 sec- onds. On the basis of this work, a decision sight-distance of 11 to 14 seconds at the operating speed is recommended. Countermeasures to Reduce Likelihood of Median Encroachments The likelihood of median encroachments can be reduced through the following traffic control and weather-related countermeasures: • Provide edgeline or shoulder rumble strips; • Improve/restore superelevation at horizontal curves; • Provide high-friction pavement surfaces; • Improve road surface or cross-slope for better drainage; • Improve visibility and provide better advance warning for on-ramps; • Improve visibility and provide better advance warning for curves and grades;

80 more visible. No formal effectiveness measures for this treat- ment are available. Improve Visibility and Provide Better Advance Warning for Curves and Grades It was noted in the previous section that sharp curves (radius less than 3,000 feet) and steep grades (4 percent or more) are associated with increased levels of median-related crashes. Geometric realignment to flatten the curves or reduce the grades is likely to be impractical for existing sites, but advance signing may assist drivers to anticipate the curves and grades. No formal effectiveness measures for this treat- ment are available. Improve Delineation Improvements to delineation improve driver path preview in poor weather and at night. This in turn increases driver confidence, and reduces driver workload. In some geometric conditions, improved delineation reduces crash risk. In other conditions, improved delineation can lead to higher speeds and increased crashes. A study in Finland examined the impact of post-mounted delineators on roads of various standards, those with higher standards posted at 100 km/h (62 mph) and those with lower standards posted at 80 km/h (50 mph) (Kallberg, 1993). Twenty pairs of road sections were selected, and one of each pair was randomly assigned to have post-mounted delinea- tors, thus avoiding a potential regression-to-the-mean effect. Improved guidance with post-mounted delineators had min- imal effects on nighttime speed or on crashes on 100 km/h (62 mph) roads. On the 80 km/h (50 mph) roads; however, the speeds increased at night, with a highly significant 40 to 60 percent increase in nighttime injury crashes. An NCHRP evaluation of raised pavement markers on undivided highways and freeways applied Empirical Bayes (EB) statistical techniques to data from several hundred miles of roadway (Bahar et al., 2004). The results strongly confirmed Kallberg’s findings that improving path delinea- tion on lower standard roadways (i.e., sharper curves) leads to more crashes. Raised pavement markers placed on sharp curves—that is, with greater than 3.5 degrees of curvature— on low-volume roadways with an average annual daily traffic (AADT) of less than 5,000 vehicles per day were associated with a 43 percent increase in crashes when compared to the number of crashes when the pavement markers were not present. The best effect is on freeways with traffic volumes of more than 60,000 veh/day—the number of crashes on roads equipped with raised pavement markers is 67 percent of the number of crashes on roads not so equipped. Improve/Restore Superelevation at Horizontal Curves Appropriate superelevation is important in helping motor- ists maintain control of their vehicles on horizontal curves. Given the importance of wet pavement conditions as a contrib- uting factor in median-related crashes, as found by Harwood et al. (2014), providing or restoring design superelevation on horizontal curves so that pavements on curves drain properly, may be particularly important. There are no crash-reduction effectiveness estimates for restoring superelevation of divided highway curves. Provide High-Friction Pavement Surfaces High-friction pavement is best implemented in areas where drivers are expected to be braking (e.g., in weaving areas, in areas where queuing regularly occurs, on downgrades, and at the entrance to sharp curves). Elvik and Vaa (2004) state that when the friction coefficient is initially higher than 0.7, improvements in friction have no effect on crashes. When the friction coefficient is less than around 0.7, a reduction of the total number of crashes on bare roads of the order of magnitude of 5 to 10 percent can be achieved. This reduc- tion is entirely attributable to fewer crashes on wet and snow- covered road surfaces. FHWA has encouraged the provision of high-friction pavement surfaces at high-friction-demand areas, such as horizontal curves. To minimize median-related crashes, use of high-friction pavement surfaces at all areas where median- related crashes are most common—on-ramps, off-ramps, sharp curves (radius less than 3,000 feet), and steep grades greater than 4 percent—should be considered. Improve Road Surface or Cross-Slope for Better Drainage Just as superelevation was noted as important in maintain- ing proper drainage to minimize crashes on wet and snow- covered roads, maintenance of the road surface and the normal pavement cross slope (typically 1.5 to 3 percent) also is important. There are no formal crash reduction measures for road surface condition or normal pavement cross slope. Improve Visibility and Provide Better Advance Warning for On-Ramps It was noted in the previous section that limited sight-distance to on-ramps for mainline drivers can lead to median-related crashes. Geometric realignment to correct the limitedsight- distance may be impractical for existing sites, but pavement markings and signing may be improved to make the on-ramp

81 speed comes from the streaming of information or “optical flow” in peripheral vision. After driving long distances, driv- ers become “speed-adapted” and have difficulty in reducing speed, even when they are aware of the adaptation and try to counter it. Thus, stimulating peripheral vision with close-by stimuli should lead to a modest decrease in speed. Progres- sively closer transverse lane markings have been successfully used to accomplish this. Progressively closer transverse bars and converging chevron patterns have been applied at entries to roundabouts, approaches to intersections, the end of free- way ramps, and other hazardous locations. Results found have been an immediate reduction in speed with reduced effectiveness over time and a small reduction in overall crash risk (Griffin and Reinhardt, 1996). A more recent study has shown greater effects on roads expected to have the great- est proportion of unfamiliar drivers. After the implementa- tion of transverse bars, Katz et al. (2006) found the following reduction in average speeds after 4 months: • 2.4 mph for off-ramps, • 1.2 mph for sharp curves on highways, and • 0.2 mph for sharp curves on local roads. Godley et al. (1999) found that constant spacing was as effective as progressively closer spacing and peripheral trans- verse lines were almost as effective as full transverse lines. Constant spaced peripheral markings are preferable, as they are easier to implement and maintain. Provide Weather-Activated Speed Signs Weather-activated signs may be most appropriate on downgrades and curve approaches where inappropriately high speed is most likely to lead to loss of control. A study of activated speed signs at an arterial road lane drop with speed limit reduction found a significant drop in mean speed and a 40 percent drop in the proportion of vehicles travel- ling 9 mph over the speed limit (Maroney and Dewar, 1987). A study in a construction zone found a 4-mph drop in mean speed and 4 percent drop in the proportion of vehicles travel- ing 16 km/h over the speed limit (McCoy et al., 1995). A study by Van Houten and Nau (1981) found a 46 percent reduction in crashes of all types, although the HSM (AASHTO, 2010) (pp. 13–31) indicates that there is considerable variability in this reduction estimate. A Finnish study investigated the effects of weather con- trolled speed limits and signs for slippery road conditions on driver behavior. In winter, with a lowering of the speed limit from 100 km/h to 80 km/h (62 to 50 mph), the signs were associated with a 2.1 mph reduction in speed, on top of the 3.9 mph reduction related to the road and weather conditions. A somewhat lesser effect was found when poor road conditions In 2005 and 2006, the Missouri Department of Transporta- tion (MoDOT) undertook a major program to improve both the rideability and the visibility of over 2,300 miles of major roadways in Missouri, including most of the Interstate High- way System in Missouri, as well as freeways and expressways; some multilane and two-lane undivided roads were included. The striping and delineation improvements included the following: • Wider and higher visibility lane lines; • Wider edgelines with rumble strips; • Centerline rumble strips (on undivided highways only); • Barrier-mounted delineators (on concrete barriers, guard- rails, and cable barriers); and • Emergency reference marker signs (on interstate highways only). A before/after evaluation was performed using the EB method (Potts et al., 2011). The striping and delineation pro- gram was found to have resulted in an overall reduction of 16 percent in fatal-and-disabling-injury crashes and 11 per- cent in fatal-and-all-injury crashes. The evaluation results for total crashes (all crash severity levels combined) show a statis- tically significant 4 percent reduction. The program appears to be particularly effective in reducing multiple-vehicle crashes on improved roadways. By contrast, single-vehicle crashes appear to have increased. This increase was considered likely to have resulted from a statewide trend of increases in lane- departure crashes rather than from an effect of the striping and delineation improvements. The program was associated with consistent crash reductions for all road types in daytime, but there was a statistically significant 24 percent increase in nighttime fatal-and-disabling-injury crashes on urban freeways. Human factors studies are needed to explain the inter- action of road geometry, delineation, confidence, visibil- ity, road condition, and speed. But Kallberg’s findings that speed increased at night when post-mounted delineators were installed, as well as older research by Allen showing that higher contrast lane markings led to higher speeds both in simulators and on real roads, indicate a likely explanation for findings that, in some circumstances, speed increases when the road is better delineated (Allen et al., 1977). On roads with little room for error, and potentially on wet and snow-covered road surfaces, even small increases in speed can greatly increase the risk of a crash. Provide Transverse Pavement Markings Transverse pavement markings are best implemented in areas where drivers may be travelling faster than desirable (e.g., on the entry to a tight off-ramp). A driver’s main cue for

82 were difficult to detect. In both cases speed variance decreased; however, headway was not affected (Rämä, 1999). Provide Static Signs Warning of Weather Conditions A static sign (e.g., bridge freezes before road surface) may change driver behavior at the location where it is posted. It also serves to educate drivers about the problem. This countermeasure is listed in HSM Part D, but the HSM indi- cates that the crash-reduction effects of this treatment are unknown. Apply Sand or Other Materials to Improve Road Surface Friction This countermeasure is listed in HSM Part D, but the HSM indicates that the crash-reduction effects of this treatment are unknown. Apply Chemical De-Icing or Anti-Icing as a Location-Specific Treatment Chemical de-icing (e.g., application of salt) prevents snow from sticking to the road surface and is only effective above 21° F. Preventative de-icing, also known as anti-icing, involves applying salt or other chemicals to the road before the storm begins. Based on Elvik and Vaa (2004), the HSM states that anti-icing appears to reduce injury crashes. Improve Winter Maintenance Response Times Based on several international studies cited by the HSM, the crash-reduction effect of raising winter maintenance response times standards (based on traffic volume and road function) by one class level is a potential reduction of 11 per- cent in injury crashes and 27 percent in all crash types. Install Snow Fences Based on Elvik and Vaa (2004), the HSM states that install- ing snow fences on mountainous highways appears to reduce all types of crashes of all severities. However, the magnitude of the crash-reduction effect is uncertain (AASHTO, 2010). Raise the State of Preparedness for Winter Maintenance Based on Elvik and Vaa (2004), the HSM states that limited research suggests that such measures as putting maintenance crews on standby or having inspection vehicles drive around the road system may reduce the number of crashes in some cases but not others. The research suggests the measure may be more effective in the early morning hours. References AASHTO, A Policy on Geometric Design for Highways and Streets, Washington, D.C., 2011a. AASHTO, Highway Safety Manual, 2010. AASHTO, Roadside Design Guide, Washington, D.C., 2011b. Alexander, G., and H. Lunenfeld, Positive Guidance in Traffic Control, FHWA, 1975. Allen, R. W., J. F. O’Hanlon, and D. T. McRuer, Drivers’ Visibility Requirements for Roadway Delineation, Volumes I and II: Effects of Contrasts and Centrifugation on Driver Performance and Behavior, FHWA-RD-77-165 and -166, FHWA, 1977. Bahar, G., et al., NCHRP Report 518: Safety Evaluation of Permanent Raised Pavement Markers, Transportation Research Board, Nation- al Research Council, Washington, D.C., 2004. Bonneson, J. A., et al., “Chapter 19: Freeways,” draft HSM chapter pre- senting safety prediction method for freeway segments, Final Report of NCHRP Project 17-45, Texas Transportation Institute, 2012. Chang, M., et al., “Evaluation of Driver’s Psychophysiological Load at Freeway Merging Area,” presented at the 80th annual meeting of the Transportation Research Board, 2001. Elvik, R., and T. Vaa, The Handbook of Road Safety Measures, 2004. Godley, S., et al., Perceptual Countermeasures: Experimental Research, Report No. CR182, Australian Transport Safety Bureau, Common- wealth Department of Transport and Regional Services, Canberra, Australia, 1999. Graham, J. L., et al., Median Cross-Section Design for Rural Divided Highways, Final Report, NCHRP Project 22-21, MRIGlobal, 2011. Griffin, L. I., and R. N. Reinhardt, A Review of Two Innovative Pavement Patterns That Have Been Developed to Reduce Traffic Speeds and Crashes, AAA Foundation for Traffic Safety, Washington, D.C., 1996. Griffith, M. S., “Safety Evaluation of Rolled-In Continuous Shoulder Rumble Strips Installed on Freeways,” Transportation Research Record 1665, Transportation Research Board, National Research Council, Washington, D.C., 1999. Harwood, D. W., et al., NCHRP Report 790: Factors Contributing to Median Encroachments and Cross-Median Crashes, Transportation Research Board, National Research Council, Washington, D.C., 2014. Kahlberg, V. P., “Reflector Post—Signs of Danger?” Transportation Research Record 1403, Transportation Research Board, National Research Council, Washington, D.C., 1993. Katz, B. J., D. E. Duke, and H. A. Rakha, “Design and Evaluation of Peripheral Transverse Bars to Reduce Vehicle Speed,” presented at the 85th Annual Meeting of the Transportation Research Board, 2006. Lerner, N., et al., NCHRP Web-Only Document 70: Comprehensive Human Factors Guidelines for Road Systems, Transportation Research Board, National Research Council, Washington, D.C., 2004. Mak, King K. and D. L. Sicking, NCHRP Report 492: Roadside Safety Analy- sis Program (RSAP)—Engineer’s Manual, Transportation Research Board, National Research Council, Washington, D.C., 2003. Maroney, S., and R. E. Dewar, “Alternatives to Enforcement in Modi- fying Speeding Behavior of Drivers,” Transportation Research Record 1111, Transportation Research Board, National Research Council, Washington, D.C., 1987.

83 Mason, J. M., et al., Median Safety Study (Interstates and Expressways), FHWA-PA-2001-009-97-04, Pennsylvania Transportation Insti- tute, June 2001. McCoy, P. T., J. A. Bonneson, and J. A. Kollbaum, “Speed Reduction Effects of Speed Monitoring Displays with Radar in Work Zones on Interstate Highways,” Transportation Research Record 1509, Trans- portation Research Board, National Research Council, Washington, D.C., 1995. McGee, H. W., et al., Decision Sight Distance for Highway Design and Traffic Control Requirements, FHWA-RD-78-78, FHWA, 1978. McNees, R. W., “In Situ Study Determining Lane-Maneuvering Distance for Three- and Four-Lane Freeways for Various Traffic Volume Conditions,” Transportation Research Record 869, Transportation Research Board, National Research Council, Washington, D.C., 1982. Potts, I. B., et al., Benefit/Cost Evaluation of MoDOT’s Total Striping and Delineation Program: Phase II, Report No. RI06043, Missouri DOT, 2011. Rämä, P., “Effects of Weather-Controlled Variable Speed Limits and Warning Signs on Driver Behavior,” Transportation Research Record 1689, Transportation Research Board, National Research Council, Washington, D.C., 1999. RoadSafe LLC, Roadside Safety Analysis Program (RSAP) Update, Appendix A: User’s Manual, Version 3.0.0, Final Report of NCHRP Project 22-27, http://rsap.roadsafellc.com, 2012a. RoadSafe LLC, Roadside Safety Analysis Program (RSAP) Update, Appen- dix B: Engineer’s Manual, Version 3.0.0, Final Report of NCHRP Project 22-27, http://rsap.roadsafellc.com, 2012b. Torbic, D. J., et al., NCHRP Report 641: Guidance for the Design and Application of Shoulder and Centerline Rumble Strips, Transpor- tation Research Board, National Research Council, Washington, D.C., March 2009. Van Houten, R., and P. Nau, “A Comparison of the Effects of Posted Feedback and Increased Police Surveillance on Highway Speeding,” Journal of Applied Behavior Analysis, 14(261), 1981.

Abbreviations and acronyms used without definitions in TRB publications: A4A Airlines for America AAAE American Association of Airport Executives AASHO American Association of State Highway Officials AASHTO American Association of State Highway and Transportation Officials ACI–NA Airports Council International–North America ACRP Airport Cooperative Research Program ADA Americans with Disabilities Act APTA American Public Transportation Association ASCE American Society of Civil Engineers ASME American Society of Mechanical Engineers ASTM American Society for Testing and Materials ATA American Trucking Associations CTAA Community Transportation Association of America CTBSSP Commercial Truck and Bus Safety Synthesis Program DHS Department of Homeland Security DOE Department of Energy EPA Environmental Protection Agency FAA Federal Aviation Administration FHWA Federal Highway Administration FMCSA Federal Motor Carrier Safety Administration FRA Federal Railroad Administration FTA Federal Transit Administration HMCRP Hazardous Materials Cooperative Research Program IEEE Institute of Electrical and Electronics Engineers ISTEA Intermodal Surface Transportation Efficiency Act of 1991 ITE Institute of Transportation Engineers MAP-21 Moving Ahead for Progress in the 21st Century Act (2012) NASA National Aeronautics and Space Administration NASAO National Association of State Aviation Officials NCFRP National Cooperative Freight Research Program NCHRP National Cooperative Highway Research Program NHTSA National Highway Traffic Safety Administration NTSB National Transportation Safety Board PHMSA Pipeline and Hazardous Materials Safety Administration RITA Research and Innovative Technology Administration SAE Society of Automotive Engineers SAFETEA-LU Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users (2005) TCRP Transit Cooperative Research Program TEA-21 Transportation Equity Act for the 21st Century (1998) TRB Transportation Research Board TSA Transportation Security Administration U.S.DOT United States Department of Transportation