Application of Pedestrian Crossing Treatments for Streets and Highways (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.

35 CHAPTER FOUR RECOMMENDED APPLICATIONS, EFFECTIVENESS, AND CURRENT USE OF PEDESTRIAN CROSSING TREATMENTS This chapter summarizes practices regarding applications of 25 specific pedestrian crossing treatments. The next sections describe the treatments and their intended uses as described by the current recommended practice documents summa- rized in chapter three, provide a summary of effectiveness from the literature synthesis, and characterize uses of the treatments reported by states and select local jurisdictions. There are numerous publications, guides, and web resources that are useful in selecting the most appropriate types of pedestrian crossing treatment(s) for a given location. Some of these, including the PEDSAFE Countermeasure Selection Tool, were described in chapter three. PEDSAFE provides guidance for selecting appropriate measures and using the most up-to-date safety research and expert knowl- edge available, and provides numerous case studies dem- onstrating how jurisdictions are applying treatments to specific types of locations. PEDSAFE describes more than 60 types of pedestrian treatments, most of which are appro- priate at pedestrian crossing locations. The Safe Routes to School Guide, also described in chapter three, summa- rizes engineering countermeasures that may be appropriate for crossings near schools or along school walking routes. Comprehensive measures such as adult crossing guards, enforcement, Walking School Buses, and other programs to supplement engineering improvements in school zones and along school walking routes are also described in resources on the SRTS website (http:/www.saferoutesinfo.org/). As mentioned in chapters one and two, there are a vari- ety of issues and other policy frameworks that may also affect decision making. For example, a number of cities have recently embraced a TZD or Vision Zero framework in which zero traffic deaths are acceptable. In a Vision Zero frame- work, wherever pedestrians are expected, they should be provided safe and convenient access to all destinations, with designs and traffic controls in place to minimize the chances of serious harm in the event that road users make mistakes. Measures that unnecessarily delay pedestrians in favor of motorized mobility may also be considered generally inap- propriate, and may lead to safety problems because pedes- trians may choose to cross without a safe gap or at less-safe locations. In addition to countermeasures safety resources, a number of jurisdictions also use design guidance such as NACTO’s Urban Street Design Guide, or locally produced design and countermeasures guidance to aid contextually appropriate decision making and development of a self- explaining and self-enforcing network of streets. Traffic speed and driver expectation—two measures that are often unavailable in existing studies—are other fac- tors that may play roles in treatment selection. Midblock and uncontrolled approaches to intersections tend to have higher severity outcomes when crashes occur—these more severe outcomes may relate to both of those issues, driv- ing speed and a lack of expectation of pedestrians at such locations—as well as lighting and other factors. Speed is often indicated as a consideration for selecting appropriate treatments, particularly at uncontrolled locations [as in the FHWA crosswalk study by Zegeer et al. 2005, TCRP Report 112/NCHRP Report 562 Improving Pedestrian Safety at Unsignalized Crossings (Fitzpatrick et al. 2006a), and PED- SAFE (Zegeer et al. 2013)]. Fitzpatrick et al. (2006a) also reported that speed is a primary concern of pedestrians in perceiving whether a location is safe to cross. In a Vision Zero or Toward Zero Deaths approach, speed is a major con- sideration for how much separation to provide users (through either designs or traffic controls). Studying the safety effects of treatments to improve con- ditions for pedestrians is especially challenging because of the low number of pedestrian crashes at most locations, and also a widespread lack of data on pedestrian volumes or counts. This condition affects both the application of treat- ments and the study of treatment effects because the amount of exposure (measures such as the numbers of crossing pedestrians and auto traffic volume) typically have a large effect on pedestrian safety. Because crashes also tend to migrate from one “hot spot” to another, partly by chance, many jurisdictions are also looking for more proactive ways to assess risks and provide safety improvements that may lead to more widespread benefits (systemic approach). In the following summaries of safety effects, bear in mind that it is challenging to incorporate enough treated and control locations in either before–after or cross-sec- tional studies to assess the safety impacts of the treatments of interest, and to isolate what factors or site conditions may affect the outcomes in a positive or negative direction. The many variations in conditions from site to site or before and after treatments often overwhelm the pedestrian–motor vehicle crash numbers, and especially more severe crashes,

36 so that generalizable conclusions about safety effects can be difficult to reach. Therefore, other measures of effec- tiveness (MOEs) are often utilized in lieu of crash- and injury-based measures. MOEs often used include motorist yielding, pedestrian–motor vehicle conflicts, and measures of mobility effects (pedestrian delay, gap analysis, etc.). Motorist speed could also be used more often as a mea- sure of safety impact. Operating speed is a measure that has well-established relationships with safety, including several studies that developed risk curves for pedestrian fatalities related to preimpact or impact speeds (Leaf and Preusser 1999; Rosén and Sander 2009; Tefft 2011). All have found the same general relationship, with a sharp rise in risk at increasing speeds. The Rosén and Sander study estimated fatality risk for impact speeds of 50 km/h to be twice as high as the risk at 40 km/h and more than five times higher than the risk at 30 km/h. Thus, the results of studies of behav- ioral/operating measures are quite useful for helping select appropriate treatments when sufficient crashes are unavail- able or require too long or too many sites to evaluate. These measures may be observed in before-to-after assessments or cross-sectional studies of treatment and comparison locations. Before–after behavioral assessments may be somewhat more rigorous for identifying effects of treatments. Ideally, these assessments are performed under similar conditions and times. Comparison locations can be used to assess whether behaviors or crashes may be chang- ing over time for other reasons such as changes in the level of enforcement, traffic or pedestrian volumes, distraction, communications or educational campaigns, or other factors. However, in practice, many of these potential confounders are often not considered or data are unavailable to control for them in real-world research designs. If before period data are unavailable, cross-sectional studies, either behavioral, or crash based, may be used. These types of studies also have limitations, because it is challenging to identify comparison sites that are compa- rable to the treated location(s), except for the treatment of interest. However, studies that at least attempt to control for potentially important factors by using similar comparison locations are better than those that do not. Also, in order to detect factors that may affect the degree of success, it is important to have a number of sites and comparison sites that do vary with respect to those traits. Consistent findings from multiple studies also provide more support for the direction and degree of effects. This chapter summarizes the safety effectiveness evi- dence available for each countermeasure or treatment. Tables showing study information, site characteristics, out- come measures, and references are provided in Appendix B. As other details of studies can help in understanding how and where treatments have been applied and evaluated, the reader is also referred to the comprehensive literature review by Mead et al. (2014), which summarizes most of the studies included (through at least 2013). “When available, we have reported crash effect estimates (CMFs). However, users are cautioned that crash effect estimates should be developed or at least calibrated using local data, that actual effects may vary, and specific site conditions should always be investigated before any treatment is implemented.” ROADWAY DESIGN FEATURES Eleven different design elements were investigated through the literature as well as the survey of practices. Following are results from the literature synthesis. Narrow Lane Width Description and Purpose of Treatment Narrowing lane widths, or a lane diet, can provide multiple benefits, including reducing crossing distance and time that pedestrians are exposed to motor vehicle traffic (Figure 5). Depending on how they are implemented (such as through rechannelization or extending the curb line), narrower lanes may also provide space for other uses such as wider side- walks, enhanced buffers, or space for bike lanes, which can also provide a buffer for pedestrians from adjacent traffic (Zegeer et al. 2013, PEDSAFE Lane Narrowing countermea- sure). By shortening crossing distance, the delay to motor- ists may also be reduced. If lanes are narrowed to add buffers for sidewalks, bike lanes, or shoulders, the added space can also provide an additional buffer between motorized traffic and pedestrians waiting to cross the street or walking along the roadway. Also see the section on Curb Extension/Bulb- Out countermeasure below, on another way to narrow lanes at crossing locations. FIGURE 5 Lane diet on Harvard Avenue in Boston, Massachusetts. Source: PEDSAFE/Bill Schultheiss, Toole Design Group.

37 NACTO’s Urban Street Design Guide indicates that lane widths of 10 ft are appropriate in urban areas and have a posi- tive impact on a street’s safety without affecting traffic opera- tions. NACTO also mentions that some cities may want to consider 11-ft lanes (versus 10) on routes used by trucks and buses, and that turn radii at intersections also be considered. Summary of Effects Although no studies were identified that specifically evaluated effects on pedestrian safety of narrowing lane widths, literature suggests that there are negligible operational (capacity) or safety differences for motorists between lane widths from 10 to 12 ft (Petritsch n.d.). The HSM is silent on the effects of lane width in safety performance. Narrower lanes have been associated with lower speeds on suburban/urban streets (Fitzpatrick et al. 2003), but results have been mixed when lanes were narrowed using different techniques and at different types of locations. The evidence thus far is also mixed regarding crash effects of narrowing lanes. The variable results may relate, in part, to the purpose of the lane-narrowing project. For example, Harwood found that when the purpose of lane narrowing was to increase the total number of through lanes on an arterial street, inter- section crashes increased, but there was no effect on segment crashes. When narrower lanes were created to provide space for installation of a center two-way left-turn lane, total crashes were generally reduced by 24% to 53% (Harwood 1990). The effects of lane diets on speeds and crashes may depend on inter- actions with shoulder width and type, and number and width of other adjacent lanes as well as other environmental cues and design speed factors (Boodlal et al. 2015). Use of Narrower Lanes In terms of using narrower lane widths, 67% of the states and 78% of the select local jurisdictions responding to the survey indicated that they use this measure to improve safety conditions for pedestrians at crossing locations. The range of widths considered “narrow” by different jurisdictions varied from 9 ft to less than 12 ft. Local jurisdictions responding to the survey for this study most often reported using widths of 10 ft or less (nearly 80% of those answering the question). Ann Arbor, Michigan, uses 9 ft for turn lanes and 10 ft for through lanes. Cam- bridge, Massachusetts, uses 10 ft for major roads and 9 ft for minor. A few other cities, including Eugene, Oregon, and Charlotte, North Carolina, use 9-ft lanes on occasion. The most common width considered “narrow” among states was 11 ft or less; but at least seven states would use 10 ft or less in some situations. North Carolina DOT indicated that it depends on the context and that generally a lane at or under 10 ft wide would be considered “narrow” for most nonresidential conditions. Three states indicated that any- thing at or less than 12 ft is considered narrow. Road Diets Description and Purpose of Treatment A road diet is a reallocation of road space cross-section through reduction in the number of motorized traffic lanes (also known as road conversions) (Figure 6). Road diets are of interest to communities that may be seeking to reallocate space for other uses, smooth traffic flows and reduce traf- fic speeds, improve access management, reduce crashes, improve safety and accessibility for pedestrians or bicyclists, improve parking utilization, improve economic function of the street, or gain space for other uses more compatible with the purposes of the street (Thomas 2013). FIGURE 6 Before (left) and after (right) road diet on Edgewater Road, Orlando, Florida. Source: PEDSAFE. The most common road diet configuration involves con- verting a four-lane, undivided two-way road to three lanes, with one travel lane in each direction and a center two-way left-turn lane (TWLTL). It is important that a traffic analysis be conducted to assess the feasibility of a road diet, which is generally considered feasible from lower volumes up to about 25,000 ADT in some cases [Zegeer et al. 2013, PEDSAFE, Lane Reduction (Road Diet) countermeasure description]. FHWA released a Road Diet Informational Guide (Knapp et al. 2014) that summarizes the history of road diet conversions, safety effects, and other uses and benefits. According to the guide, road diets or reallocation of the roadway cross section may be used to create pedestrian median islands, bike lanes, transit stops, or parking. Pedestrians also have fewer motorized lanes to cross, and medians and median islands may further be used to shorten crossing distances. If bike lanes are added, these create an addition buffer between motorized traffic and pedestrians on walkways (Knapp et al. 2014). Summary of Effects A study of road diets at 460 sites in New York City found that road diets have contributed to a trend of fewer pedes- trian crashes between intersections, and decreases in total and injurious crashes at segments and intersections abutting the road diet segments. The range of expected effects on total crashes is from 19% to 47% reduction. The expected effects are in the lower range in large urban areas, and in the higher range for main roads passing through smaller towns

38 and suburban areas. An average 29% reduction (CMF of 0.71) for total crashes across all sites was estimated (Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes 2013, citing Harkey et al. 2008). Road diets have also been found to reduce motor vehicle operating speeds, which are also associated with expected reductions in injury and fatal crashes. Use of Road Diets Similar percentages of states (78%) and the select local juris- dictions in the survey for this study (72%) have used road diets to improve crossing conditions for pedestrians. Most of the jurisdictions using the treatment mentioned applying a road diet where volumes allow or there is excess capacity, or to allocate space for bike lanes or improve conditions for pedestrians. Road diet conversions are also used if the current cross section is not operating well, the local community has pushed for the change, or there is a change in function from highway to local collector. High speeds were mentioned as one of the issues of concern. North Carolina mentioned taking care to address transitions on either end of a road-dieted sec- tion with regard to driver expectation, lane continuity, posi- tive guidance, proper advance signs, and markings. A number of state and local jurisdictions mentioned imple- menting bike lanes and median islands with road diets. Fewer mentioned parking. New York City implements medians, median islands, and curb extensions with road diet conversions. Eugene, Oregon, implements bike lanes and median islands. Others use two-way, center two lanes. Three jurisdictions men- tioned they will implement RRFBs in conjunction with road diets (Nevada; Portland, Oregon; and Davis, California). Davis also adds a center median and marked crosswalks. A few juris- dictions mentioned installing sidewalks and lighting. Raised Median and Pedestrian Median Islands Description and Purpose of Treatment Medians with pedestrian crossings, and pedestrian median islands—sometimes referred to as center islands, refuge islands, or pedestrian islands—are raised areas intended to help protect pedestrians who are crossing the road at either intersections or midblock locations (Figure 7). The presence of a median island in the middle of a street or intersection allows pedestrians to focus on one direction of traffic at a time as they cross and gives them a place to wait for an adequate gap between vehicles before finishing the second leg of the crossing (Figure 8). Islands are appropriate for use at both uncontrolled and signalized crosswalk locations. According to PEDSAFE, where the road is wide enough and on-street parking exists, center islands can be combined with curb extensions to further reduce the crossing distance (Zegeer et al. 2013, PEDSAFE, Crossing Islands). Summary of Effects The available evidence suggests that medians and median crossing islands provide a protective effect for pedestrians (as well as motorists) with regard to crashes by reducing pedestrian exposure to traffic. A CMF of 0.75 (25% reduc- tion) for pedestrian crashes was estimated from a Florida study (Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes 2013, citing Gan et al. 2005). CMF estimates of 0.54 (46% reduction) for raised median with a marked crosswalk at unsignalized inter- section, and 0.61 (39% reduction) for raised median with unmarked crosswalk at unsignalized intersection were estimated from the FHWA crosswalk study (Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes 2013, citing Zegeer et al. 2002). FIGURE 7 Landscaped, raised median with pedestrian crossing. Source: Zeeger et al. (2013), citing PBIC (n.d) FIGURE 8 Midblock pedestrian crossing island with curb extensions, high-visibility crosswalks, advanced stop lines, lighting, and angled cut-through. Note that a cut-through must include detectable warnings. Source: Oregon Bicycle and Pedestrian Design Guide: Oregon Highway Design Manual (2011).

39 Studies for NCHRP Report 562 found that motorist yield- ing rates varied widely across sites with median islands and marked crosswalks. Some of the explanatory variables included the number of through lanes and posted speed limit (Fitzpatrick et al. 2006a). Other treatments may be needed to improve yielding at uncontrolled locations or improve pedestrian access and safety at multilane uncontrolled road crossing locations, particularly with higher volumes or speeds or both (see the matrix in chapter three). Treatments recommended by Zegeer et al. (2005) and Fitzpatrick et al. (2006a) include advance stop/yield lines and signs, beacon signals with a red (stop) indication (such as PHBs), and signs and flashing beacons, although the latter have tended to have a wider range of effectiveness. Use of Raised Medians Raised medians were used by 89% of states responding and 67% of local jurisdictions. The most common reasons or sit- uations mentioned by state or local agencies (44 responses in total) for using raised medians to improve pedestrian crossing safety were to manage access/reduce conflict points. Raised medians are frequently used on multilane, wide, or arterial roadways, or on higher-speed and higher-volume roads, and may be used where there is a need for a midblock crossing. Raised medians are also used to provide a pedestrian refuge, and to reduce conflicts and crashes at intersections. Two state agencies mentioned that they sometimes used medians in conjunction with fencing as a means to discour- age pedestrians from crossing or “jaywalking,” but there was no further description of the type of roads or land uses where this type of treatment was applied. Some other issues mentioned include whether there was sufficient width, and Scottsdale, Arizona, mentioned that a raised median is a design criterion for all arterials, whereas Portland uses raised medians on any four-lane road with a speed limit higher than 25 mph. The most common other treatments used in addition to raised medians were the following: • Marked crosswalks (and signs) are used most often in conjunction with other treatments, but sometimes alone or with signs only. • RRFBs or overhead or other flashing devices are used by Oregon, Utah, and Minnesota, which will also incorporate striping, curb extensions, and pedestrian countdown timers at signals. • Tucson, Arizona, will consider a PHB (HAWK) or PELICAN (PEdestrian LIght Control ActivatioN system, a type of two-stage pedestrian crossing) for multilane roads with a median. Delaware will also consider mid- block crosswalks enhanced by a PHB, whereas Florida may implement a pedestrian signal and markings. • Washington uses street trees and context-sensitive design in conjunction with raised medians. • Scottsdale, Arizona, will implement a pedestrian refuge and raised crosswalk in conjunction with some medians. Three states mentioned that median fencing or barriers are sometimes installed in conjunction with raised medians but were not more specific about the context or type of road. Use of Pedestrian Median Islands Pedestrian median islands are used by 97% of states and 94% of the select group of local jurisdictions to provide enhance- ments for pedestrian crossings. There appear to be two basic approaches to providing median refuges. Many jurisdictions consider the characteristics of the street that may cause pedestri- ans to need a refuge—such as high speed, high volume, width, and multiple lanes—in their decisions. Most of the answers were along these lines, with one jurisdiction (Portland, Oregon) indicating that it might apply a pedestrian refuge on “any four- lane road with a speed limit higher than 25.” Another indicated it might on “most three- and five-lane roadways.” A few juris- dictions mentioned that they use such refuges at unmarked and some marked crosswalks on urban arterials where signalized intersections are infrequent. A couple of other jurisdictions mentioned they would use refuges at urban arterial signalized intersections, at nonintersection path crossings, or where there is a high volume of pedestrian–vehicle conflicts. Other jurisdictions take a somewhat different approach, at least as reflected in their survey responses. These juris- dictions described a more opportunistic approach and men- tioned criteria, such as “already-existing median, sufficient width or space, TWLTL designs, when four- to three-lane road diets are implemented,” or “crossing with storage/ref- uge area—ample sight distance, good spacing from adjacent intersections and conflicts” that would lend themselves to implementation of pedestrian refuge islands. Others men- tioned both pragmatic concerns and needs. Raised Crosswalk/Speed Table Description and Purpose of Treatment According to PEDSAFE, raised pedestrian crossings are a type of vertical traffic control measure that can reduce vehicle speeds, reduce the need for curb ramps (though truncated domes should still be included), and enhance the pedestrian crossing environment (Figure 9). A raised pedestrian crossing is similar to a raised intersection, but it is only the width of a crosswalk, usually 10 to 15 ft. Construction involves providing ramps on each vehicle approach, which elevates the entire crosswalk to the level of the sidewalk. The crosswalks can be built with a variety of materials, including asphalt, concrete, stamped con- crete, or brick pavers (Zegeer et al. 2013, PEDSAFE, Speed Tables, Raised Pedestrian Crossings countermeasures).

40 A speed table is a long, broad speed hump or a flat-topped speed hump typically used to reduce vehicle speeds on resi- dential roads. Speed tables can be used in combination with curb extensions where parking exists. Speed tables are paved (usually asphalt) and approximately 3 to 4 in. high at their center, and extend the full width of the street with height tapering near the drain gutter to allow unimpeded bicycle travel. Speed tables should not be confused with the speed “bump” that is often found in mall parking lots. There are several designs for speed tables. The traditional 12-ft table has a design speed of 15 to 20 mph, the 14-ft table has a design speed of a few mph higher, and the 22-ft table has a design speed of 25 to 30 mph. The longer tables are much gentler for larger vehicles (Zegeer et al. 2013). FIGURE 9 Raised crosswalk in Salt Lake City, Utah. Source: http://www.pedbikeimages.org, Dan Burden. Speed tables are used to slow vehicular traffic at two-way path crossings in Canberra and Yarra, Australia, and Malmö, Sweden, and help enforce path user right-of-way priority (Thomas et al. 2015a). Image source: Thomas et al. 2015a, Alistair McDonald. Speed humps in Helsinki, Finland, were implemented in advance of similar, two-way path crossings, and an evaluation found that drivers were slowed on the approach and scanned in both directions more often before driving across the path after the speed humps were installed (Thomas et al. 2015a; citing Summala et al. 1996). Existing guidance suggests that raised pedestrian cross- walks or speed tables (most often flat-topped, but may also be parabolic) may be applied at midblock locations on low-speed local streets, collector roads, and other loca- tions such as airport drop-off and pick-up zones, campus settings, and shopping centers. Typically, they function as an extension of the sidewalk (per AASHTO guidance), but may have gaps to allow bicycles and motorcycles to pass through (AASHTO Guide 2014; Fitzpatrick et al. 2006a). Because raised crosswalks create a vertical displacement of the vehicle as it drives over the devices, lower speeds are expected if they are designed for speeds slower than initial operating speeds. Because vertical displacement can also cause discomfort and noise issues (especially with larger vehicles) as well as potential drainage issues, PEDSAFE recommends that these devices go through a public process that involves emergency response agencies. However, raised devices such as speed tables and speed humps tend to be consistently effective at lowering speeds (FHWA 2014) and at lowered total and injury crashes of all types, as demonstrated in a study from the United King- dom (Mountain et al. 2005). They are also thought to call increased attention to pedestrian crossings. Summary of Effects In research conducted for the HSM, CMF estimates of 0.70 (30% reduction) for all crashes and 0.64 (36% reduction) for fatal injury crashes were developed, although the CMFs were not ultimately included in the HSM (Toolbox of Coun- termeasures and Their Potential Effectiveness for Pedes- trian Crashes 2013, citing Bahar et al. 2007). The effects of raised crossings on motorist yielding and on pedestrian crashes are not well-documented. However, lower speeds, and potentially improved search behaviors (see sidebar), should reduce severe crashes according to HSM estimates of effects of lower speeds. Speed reductions for speed tables have been documented in five studies cited by FHWA (Engi- neering Speed Management Countermeasures 2014). Use of Raised Crosswalk/Speed Table Only 17% of states indicated they use raised crosswalks or speed tables. In contrast, 72% of the local jurisdictions responding indicate that they do. Cases where raised crosswalks have been implemented include when a street is “functionally” classified as a local street, most frequently mentioned in association with residen- tial areas or new residential developments. Speed tables have also been used near schools, parks, or other pedestrian traffic generators. The devices are used to slow traffic for midblock crossings, and also used at signalized locations with a “free” or yield-controlled right turn. Speed tables are also considered on a case-by-case basis, although the treatment is not nor- mally used on state roads according to some state agencies.

41 The only treatments often mentioned in survey responses as being used in conjunction with raised crosswalks/speed tables were signs and pavement markings. Curb Extension/Bulb-Out Description and Purpose of Treatment Curb extensions are installations at intersections or mid- block locations that extend the sidewalk or curb line into the street or parking lane, thus reducing the street width and crossing time. It is essential that they not extend into bicycle travel lanes or shoulders (Zegeer et al. 2013, PEDSAFE, Curb Extensions) (Figure 10). FIGURE 10 Integrating curb extensions and on-street parking into the sidewalk corridor decreases pedestrian exposure and enhances the walking experience. Source: PEDSAFE, Michele Weisbart. Curb extensions have the potential to enhance pedestrian safety in several ways: by making pedestrians and motorists more visible to each other; when used at intersections, by keeping motor vehicles from parking too close to the inter- section and blocking sight lines; by reducing crossing dis- tance (which may also allow less signal time to be devoted to the pedestrian phase at signalized locations); and by nar- rowing radii at intersections, which may slow turning traffic. Curb extensions also tend to allow for better placement of curb ramps and prevent ramps from being blocked by vehi- cles that park at the corner. They are useful in conjunction with parking restriction policies, such as daylighting. Current guides suggest that curb extensions, also known as bulb-outs and neck-downs, are appropriate only where there is on-street parking, and may not be desirable where the street is too narrow or the extension would interfere with a bicycle lane (Guide for the Planning, Design, and Opera- tion of Pedestrian Facilities 2004; Knapp et al. 2014; Zeeger et al. 2013). Summary of Effects Although curb extensions are expected to have safety ben- efits based on exposure and design principles, behavioral or crash effects of curb extensions have not been widely stud- ied, as yet. Curb extensions may be effective at improving motorist yielding as part of a larger package of treatments or designs (including crosswalks with advance stop/yield lines and median islands). As mentioned, curb extensions are expected to slow turning speeds by narrowing the effective turn radius, and to improve sight lines between pedestrians and motorists (but this depends on implementation; for example, tall shrubs are not to be used as part of landscaped areas). They may also help lower overall crash severity based on a New York City study that attempted to measure crash effects at six locations (King 1999). Use of Curb Extensions/Bulb-Outs Of the states and local jurisdictions surveyed, 94% use curb extensions/bulb-outs in certain situations. Jurisdictions most often mentioned using this treatment in downtown and urban settings, or along main roads/trunk lines that pass through towns and cities. Other criteria mentioned were high-pedestrian locations, and the need to shorten crossing distances/reduce pedestrian exposure. Sev- eral jurisdictions mentioned on-street parking as a require- ment, which is consistent with the current recommended practice guides. One jurisdiction mentioned the need to con- sider drainage issues. Treatments mentioned most often as used in combination with curb extensions or bulb-outs include the following: • Marked crosswalks were mentioned 10 times in total, seven of those times in combination with additional other treatments including pedestrian signals, signs, RRFBs, LPIs, and lighting. Oregon mentioned a com- prehensive list of treatments used sometimes, with curb extensions, including marked crosswalks, LPIs, pedes- trian countdown signal heads, RRFBs, and lighting. • High-visibility crosswalk markings were mentioned by three jurisdictions, one of those times in conjunction with pedestrian crossing warning signs. • ADA curb ramps were also mentioned by several jurisdictions. Treatments mentioned only once include landscaping, back-in parking, and water harvesting. Reduce Corner Radius Description and Purpose of Treatment Large curb radii facilitate high-speed turns for vehicles (Figure 11). Curb radius reduction forces sharper turns, thus reduced turning speeds. Curb radius reduction also creates larger waiting areas for pedestrians wishing to cross. Curb radii sizes are normally based on design vehi- cle types, often for large emergency vehicles. By reducing radius sizes, emergency vehicles may find navigating the turn more difficult.

42 FIGURE 11 The effective corner radius controls turning speeds and the ability of large vehicles to turn. Source: PEDSAFE, Michele Weisbart. However, the distinction between the actual and effective curb radius is important when designing the intersection. The actual curb radius is the radius of the curb itself whereas the effective radius is determined by the path of a turning vehicle, which is increased by the presence of on-street park- ing, bicycle lanes, or striping advance stop lines on the desti- nation street of multilane roadways. An actual curb radius of 5 to 10 ft, where possible, is recommended. For the effective curb radius, 15 to 20 ft on urban streets with high numbers of pedestrians is recommended. An arterial with buses and trucks may have an effective curb radius of 25 to 30 ft. Curb radius reductions are important for skewed intersections, where it is necessary to create the smallest possible radius on the obtuse-angle corner to avoid high-speed turns. Important issues for curb radius reduction are as follows: • Parking or bicycle lanes at an intersection can increase the effective radius. • If curb radius is small in an area with high traffic vol- umes, vehicles may drive over the curb. • Consider emergency vehicle access (Zegeer et al. 2013, PEDSAFE, Curb Radius Reduction). Summary of Effects No research studies were identified that have evaluated the safety or behavioral effects for pedestrians of reducing cor- ner radii, which as mentioned, are intended to slow motor- ist turning speeds and increase chances for motorists to perceive and yield to crossing pedestrians. Slower speeds are also expected to reduce the severity of collisions that do occur. There is an equation to estimate a CMF for all crashes based on the skew angle, which can also lead to wide radii at some corners. The equation can be used to estimate effects compared with an intersection without any skew (all right angles) (see Thomas et al. 2015b or the CMF Clearinghouse). Use of Reduced Corner Radius A lower percentage of states, 64%, and local jurisdictions, 78%, use the approach of reducing the corner radius at intersections in comparison with the proportions using curb extensions. Of those, at least five jurisdictions—including Ann Arbor, Michigan; Santa Barbara, California; and Port- land, Oregon—seem to try to keep or modify turning radii to the minimum allowable based on the design vehicle or road type, or in the case of Cambridge, Massachusetts, as standard design for city streets. Seattle, Washington, also reduces curb radius at intersections to slow motorist speeds (see a case example from BikeSafe, Curb Radii/Curb Revi- sions, Sundstrom et al. 2014). Corridorwide Speed Calming Description and Purpose of Treatment Areawide traffic calming is a comprehensive strategy compris- ing a variety of countermeasures with the objectives of reduc- ing traffic speed or volume or both; reducing conflict between local traffic and through traffic; making it easier for pedestri- ans to cross roads and avoid conflicts with motor vehicles; and enhancing the overall environment by reducing noise from traffic (Mead et al. 2014). In urban areas, crashes can be scat- tered widely, making traditional treatments focused on specific sites less effective. Often, similar crash or risk factors (such as traffic speed, number of lanes, or transit presence) occur at dif- ferent locations along a corridor. A corridorwide approach to speed calming may therefore be appropriate. Ann Arbor, Michigan, assesses whether most roads planned for reconstruction are suitable for road diets, curb extensions, curb radius reductions, and narrower lanes. In addition, the city uses corridorwide speed calming, which may include lane narrowing, chicanes, median islands, gateway signs, and roundabouts or traffic circles at intersections. Also, see the case study from Cambridge, Massachusetts, Traffic Calming Project Development and Evaluation, in chapter five, on use of reconstruction projects to implement pedestrian safety improvements. Summary of Effects No formal studies were found that assessed the impacts of corridorwide speed calming on pedestrians as intended by this treatment. However, studies on areawide traffic-calming measures of effectiveness are the closest approximation to cor- ridorwide speed calming available. Areawide traffic calming, using a combination of multiple types of countermeasures, has resulted in an estimated percent reduction in fatalities and

43 injury collisions for all road users (Elvik and Vaa 2004). Treat- ments commonly applied in traffic-calming schemes, such as street closures, cul-de-sacs, diverters, traffic circles, shared street design, chicanes, flares/chokers, speed humps, speed limit signs and speed zones, enforcement programs, walk- ways, parking controls, and other signage, have also yielded reductions in vehicle speeds. Results of the meta-analysis indicated that these types of schemes reduced the number of total injury collisions for all road users by about 11% across all urban road types according to the CMF Clearinghouse review of this study (CMF ID 588 Details, n.d.). Effects were greatest on residential/local streets (18%; CMF ID 587 Details, n.d.) and lower on main roads (6%; CMF ID 586 Details, n.d). See more discussion of traffic-calming effects in the sec- tions on Speed Tables, Pedestrian Median Islands, and Road Diets as well as review tables in Appendix B. Use of Corridorwide Speed Calming Jurisdictions described a variety of considerations for cor- ridorwide speed calming, including the following: • Sometimes speed calming measures are implemented during corridor reconstruction projects; one juris- diction always performs an assessment to determine whether this is possible • As part of “greening” (planting) corridor projects • Used for urban arterials or locations that are suburban- izing and where there are high-speed differentials. Other jurisdictions viewed corridorwide speed calm- ing as applying to neighborhood streets (when requested), downtown/urban center streets, or main streets. Jurisdictions apply a wide variety of techniques and treat- ment combinations to calm corridors, which may include speed humps, tables, and raised intersections as well as narrower lanes and street trees. California uses road diets and St. Petersburg, Florida, mentioned using traffic signal progression. Jurisdic- tions also variously use roundabouts and mini circles, striping, high-visibility crosswalks with in-roadway yield signs, advance stop bars, chicanes, and curb extensions to calm streets. Oregon indicates it uses curb extensions, raised medians, street trees, buildings close to the sidewalk (no setback), and sidewalks to calm corridors. Lowering posted speed limits was mentioned by only one jurisdiction, as was enhanced enforcement and use of radar feedback about speeds. Pedestrian Overpass/Bridge and Underpass/Tunnel Description and Purpose of Treatment Pedestrian overpasses or bridges, and underpasses or tun- nels, are used when it is necessary to completely sepa- rate pedestrians from vehicular traffic. This is most often required with freeways, railways and natural barriers, or busy roads that may hinder the use of traditional pedestrian crossing facilities. Overpasses and underpasses allow for uninterrupted flow of pedestrian movement by providing crossings where no other pedestrian facility is available or possible (Figure 12). They often connect off-road trails and paths across major barriers (Zegeer et al. 2013, PEDSAFE, Pedestrian Overpasses/Underpasses). FIGURE 12 Pedestrian overpass connects pedestrian areas across a natural barrier—Liberty Bridge over Reedy River and falls in downtown Greenville, South Carolina. Source: Libby Thomas/author. Summary of Effects Clearly, overpasses and underpasses of highways and other roadways can enhance pedestrian safety, if they are used by a majority of pedestrians. In some instances, there may be no feasible alternative to providing separated-grade facilities to provide pedes- trian movement across major highways and other types of barriers such as rivers. In other situations, it will be impor- tant to assess the feasibility, convenience to pedestrians, and other issues of providing accessible and user-friendly overpasses and underpasses versus providing comprehen- sive at-grade facilities, and these issues should be factored into the cost equations. Pedestrian overpasses and underpasses can be effective in reducing pedestrian crashes in certain locations. How- ever, grade-separated crossings are expensive structures and may not be used by pedestrians if not perceived to be safer and more convenient than crossing at street level. An important measure of effectiveness for overpasses and underpasses is how much they are used by pedestrians. Overpasses and underpasses can also present problems for disabled users, so it is important that proper attention to accessibility be taken into account.

44 Use of Pedestrian Overpass/Bridge Although a majority of agencies sometimes use or have used overpasses (83% of states and 61% of local jurisdic- tions) per the respondents to this survey, some are no longer using them or trying to remove them owing to a lack of ADA compliance, or use them rarely. Typical responses to when overpasses might be used is for regional trails, “when pos- sible,” over major highways or other barriers, and depending on the volume of pedestrians and conflicts. Another situation is when a developer wants to include an overpass in a project and there is political pressure for the state to maintain it. Lighting and positive guidance or fencing to discourage at-grade crossings were commonly mentioned as associ- ated treatments. Use of Pedestrian Underpass/Tunnel Fewer jurisdictions use underpasses or tunnels than over- passes/bridges. When they do, the reasons are usually simi- lar, and lighting is typically considered, as well as signing, and positive guidance or even fencing to discourage at-grade crossings. They may also be implemented along with flood control improvements, although one jurisdiction mentioned environmental impacts as also discouraging use. Enhanced Illumination at Crossings Description and Purpose of Treatment Road lighting sometimes has been centered on the needs of motorists and not of pedestrians. Illumination at pedestrian crosswalks must be considered to enhance visibility of pedestri- ans who are crossing during nighttime and other low-visibility conditions. In commercial areas with heavy nighttime pedes- trian activity, streetlights and building lights can enhance the visibility in the area. Streetlights along both sides of arterials at consistent intervals enhances overall ambience. Pedestrian crossing areas should be augmented with additional brighter lighting both on the approaches to and at the crosswalks (Zegeer et al. 2013, PEDSAFE, Lighting and Illumination) (Figure 13). Summary of Effects Studies have shown that pedestrian safety and motorist and driver behaviors such as pedestrians’ use of crosswalks and driver yielding are improved by the use of enhanced illumi- nation at crossing locations. Lighting is estimated to reduce injury crashes at intersections by 27% (CMF 0.73), and reduce all severity of crashes at intersections by 21% (CMF of 0.79). For segments, lighting improvements are estimated to reduce injury crashes by 23% (CMF 0.77) and all crashes by 20% (CMF 0.80) (Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes 2013, citing Harkey et al. 2008). FIGURE 13 Enhanced lighting at raised crosswalk in West Lafayette, Indiana. Source: http://www.pedbikeimages.org, Dan Burden. Use of Enhanced Illumination at Crossings Seventy-two percent of states and 78% of the local agencies also use enhanced illumination at crossings on some occa- sions. A number of jurisdictions replied that staff always checks to make sure adequate lighting exists at locations under review for traffic changes. Several mentioned using lights in combination with RRFBs or PHBs, intersections with major trail crossings, or locations with high nighttime crashes. One jurisdiction mentioned that lighting may be underused, and potentially needs more research. However, one state indicated that lighting is most often provided by the municipal partner agency for nonfreeway roads because of maintenance, energy costs, and dark skies initiatives at the local level. Other Types of Design Features Used As far as other or innovative types of design features being used to improve safety for pedestrians, most of those men- tioned were also included in the lists of designs or traffic- calming devices. However, roundabouts; dedicated, separate multiuse paths; and removal of free right turns were uniquely mentioned. There was no other frequently mentioned design that appears to be receiving widespread implementation. Roundabouts were not a focus of this synthesis. Summary of Effectiveness of Pedestrian Design Features Table 7 provides an overview of the state of knowledge of key safety effects, including estimates of expected crash effects when available, and behavioral measures of pedes- trian design features. Summary of Applications of Roadway Design Features Table 8 summarizes the responses for which design treat- ments are used by state and local jurisdictions. Note that the question was phrased as yes or no to sometimes used, and

45 TABLE 7 OVERVIEW OF STUDIES AND FINDINGS FOR EFFECTS OF DESIGN TREATMENTS ON PEDESTRIAN SAFETY Design or Treatment Area Types Studied Key Measured Effects Raised median/ pedestrian crossing Mostly urban, suburban multi- lane arterials; controlled and uncontrolled crossings • CMF estimate of 0.54 pedestrian crashes at uncontrolled locations with marked crosswalk • CMF estimate of 0.61 pedestrian crashes at uncontrolled locations (unmarked crosswalk) • Raised medians may sometimes contribute to increased speeds. • Other treatments may also be needed on multilane, higher-volume, higher- speed roads. Pedestrian median crossing island Major and minor arterials; res- idential streets; urban/subur- ban; midblock and intersec- tion; controlled and uncontrolled • See CMFs for medians. • Motorist yielding may increase, especially if used with other treatments such as curb extensions and advance stop lines and signs, or stop or warning beacon devices. • May help reduce speeding when implemented with multiple measures • Mixed results for pedestrian delay • Potential for motor vehicles to strike the islands Raised crosswalk or speed table Mostly two-lane streets and residential collectors Raised intersections have been used in residential, central business district, and other commercial zones. CMF study for the HSM—esti- mates not incorporated • Lower speeds • Improved motorist yielding at some locations • CMF estimate of 0.70 for all crashes • CMF estimate of 0.64 for all fatal injury crashes Curb extension/bulb-out New York City (urban) study of crash effects Various in behavioral assess- ments including urban residen- tial and arterial streets • Trend in reduced crash severity • Potential motorist yielding improvements combined with advance stop bars and median island • Unclear effects on motorist yielding, motorist speed, pedestrian delay when used alone • See reduce corner radius—expected effects are similar. Reduce corner radius No peer-reviewed studies identified measuring safety effects for pedestrians • Expert-based guides suggest that narrower radii can improve pedestrian safety by – Requiring motorists to reduce vehicle speed to make sharper turns; – Shortening pedestrian crossing distances, which reduces pedestrian exposure, which may also shorten time needed for the pedestrian crossing; – Providing larger pedestrian waiting areas at corners; and – Improving sight lines. • Pedestrian service may also be improved by allowing for greater flexibility of curb ramp placement in line with walkways. Road diet Urban streets (New York City) Urban/suburban in large metro areas, and rural arterials pass- ing through smaller towns/cit- ies—effects on all crash types and severities • Downward trend in midblock pedestrian crashes (one study) • Estimated reductions in total crashes: – CMF of 0.81 for arterials in larger urban areas; and – CMF of 0.53 for main highways passing through smaller towns. • Gains space for other uses, including pedestrians, but often bike lanes or parking • Calms streets/reduces speeds, but effects may vary • Good practice guides caution for the need for site-specific assessment and describe many of the relevant considerations. Narrow lane width • Various—but none examining impacts on pedestrian crashes • There seem to be no negative safety or operational impacts for lanes as narrow as 10 ft in urban areas, but cautions are urged with regard to large vehicle volumes (Petritsch n.d.). • No peer-reviewed studies identified assessing effects on pedestrian crashes • Study results are mixed with respect to effects on all crashes and may depend on how lanes are narrowed, why they are narrowed (e.g., if to add more motor vehicle lanes, speeds may increase), and adjacent lane and shoulder presence and width. • Speed effects may depend on road configuration, including adjacent lanes, shoulders, and other design and environmental characteristics (as well as enforcement).

46 does not indicate the frequency or intensity of use of differ- ent treatments. TABLE 8 SUMMARY OF STATES AND OF LOCAL JURISDICTIONS USING EACH OF THE DESIGN TREATMENTS TO IMPROVE THE SAFETY OF PEDESTRIAN CROSSINGS Design Features States Using Treatment Local Jurisdictions Using Treatment Raised median 88.9% 66.7% Pedestrian refuge/median crossing island 97.2% 94.4% Raised crosswalk or speed table 16.7% 72.2% Curb extension/bulb-out 94.4% 94.4% Reduce corner radius 63.9% 77.8% Road diet 77.8% 72.2% Narrow lane width 66.7% 77.8% Corridorwide speed calming 33.3% 66.7% Pedestrian overpass/bridge 83.3% 61.1% Pedestrian underpass/tunnel 61.1% 44.4% Enhanced illumination at pedestrian crossings 72.2% 77.8% Total responding 36 18 The design treatments most often used (at least some- times) by states and the local jurisdictions responding were the following: • Pedestrian refuges/median crossing islands are used by 97% of states and 94% of select local jurisdictions. • Curb extensions/bulb-outs are used by 94% of states and 94% of select local jurisdictions. • Raised medians are used by 89% of states and 67% of local jurisdictions. However, a few agencies quali- fied that their primary reason for using medians was for motor vehicle reasons, or that this was the typical design for all arterial streets. • Pedestrian overpasses/bridges are used by 83% of states and 61% of local jurisdictions. • Road diets are used by 78% of states and 72% of local jurisdictions in the survey. Other treatments were also fairly widely used, including the following: • Enhanced illumination at pedestrian crossings are used by 72% of states and 78% of local jurisdictions. • Reducing the corner radius is used by 64% of states and 78% of local jurisdictions. • Pedestrian underpasses/tunnels are used by 61% of states and 44% of local jurisdictions. The two types of speed- or traffic-calming measures asked about were used much less often by states than by the local jurisdictions, which may pertain to some extent to the types of roads managed: • Corridorwide speed calming is used by 33% of states and 67% of local jurisdictions. • Raised crosswalks or speed tables are used by 17% of states and 72% of local jurisdictions. Design or Treatment Area Types Studied Key Measured Effects Pedestrian overpass/bridge One crash-based study from Japan Assessment of use by conve- nience measure study • Reductions in crashes (if used) – CMF estimate of 0.10 pedestrian fatal/injury crashes – CMF estimate of 0.14 pedestrian total crashes – CMF 0.87, pedestrian total crashes when replaces unsignalized intersection • Best-practice guides (and research) suggest that convenience or time it takes to cross using bridge/tunnel versus at grade will affect how much pedestrians use an overpass and underpass. • Accessibility is also a key concern. Pedestrian underpass/tunnel See pedestrian overpass/bridge See pedestrian overpass/bridge Corridorwide speed calming No studies identified assessing corridorwide safety effects on pedestrians Meta-analysis of multiple studies of areawide effects on all crashes • Total crash reductions from 10% (main roads) to 25% (residential/local streets)—from meta-analysis of areawide traffic calming • The HSM provides estimates of effects of lowering speeds for all injury and fatal crashes (not just pedestrians) that could be used to generate estimates of effectiveness for speed calming measures. Enhanced illumination at crossing Urban, midblock; high night- time collisions and motorists failing to yield • CMF estimate of 0.58 for pedestrian nighttime crashes • CMF estimate of 0.73 intersection injury crashes • Improved yielding by drivers • Increase in pedestrians using crosswalk • In addition to safety and security considerations, light pollution and energy costs must be considered. Note: See text for each countermeasure description and FHWA’s Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes (2013) for most CMFs and references; see Appendix B tables for study details including behavioral effects. Table continued from p. 45

47 TRAFFIC CONTROL DEVICES If road designs are such that traffic speeds and conflict potential between motor vehicles and pedestrians are high, it becomes more important to select traffic control devices to safely manage the interactions of pedestrians and motorists. Treatment descriptions, safety effects, and use of various traffic control devices (TCDs) are described in the following sections, after some information regarding device activation and yielding laws that may affect use and success of differ- ent treatments. However, as previously discussed, there are many other factors, including enforcement and driving cul- ture, that can affect success of treatments in different locales and area types. Active Versus Passive Activation Pushbuttons are generally more appropriate at locations or time periods with low or intermittent pedestrian activ- ity (PBIC, SRTS Guide, Traffic Signals, n.d.). If used, they should be in clear view and meet accessibility requirements (Figure 14). With devices that are pedestrian activated, whether the pedestrian does in fact activate the device may have a great bearing on whether the motorists yield. Fitz- patrick et al. (2015) found that drivers were more than three times as likely to yield when an RRFB was activated than when it was not. FIGURE 14 Pushbutton activation with direction and use information—West Grand, Michigan. Source: http://www. pedbikeimages.org, Dan Burden. As pedestrian detection technologies improve, auto- mated detection has the potential to remove much of the uncertainty around whether the pedestrian will activate a pedestrian signal (Figure 14). In addition, automated detec- tion technologies can be used to extend the walk phase for slower pedestrians, further helping prevent conflicts. Con- cerns with reliability and maintenance issues may have limited U.S. jurisdictions’ use of these technologies in the past. However, some international jurisdictions, including Britain, and more recently, Sydney, Australia, have exten- sive experience with using these technologies in Pedestrian User-Friendly Intelligent (PUFFIN) signals (Fischer et al. 2009; Thomas et al. 2015a). In the United States, Tucson has used PUFFIN technology in conjunction with PHBs. See PEDSAFE 2013, Automated Pedestrian Detection (Zegeer et al. 2013) for more information. Pedestrians may also be more willing to pushbuttons at signal locations with PCSs in place. When a treatment gives more information to pedestrians, such as through the pedes- trian countdown timers, there may be more incentive to push the buttons. Also, when a pushbutton activation lets pedes- trians know that the signal call has been activated through a sound or a light, activation may be increased (Redmon 2005). Stop Versus Yield Laws State laws vary with regard to requirements for motorists to stop versus yield for pedestrians at or in uncontrolled cross- walks. Nine states and the District of Columbia have laws requiring motorists to stop when approaching a pedestrian in an uncontrolled crosswalk per a report by the National Conference of State Legislatures (2015). Most states require motorists to yield to pedestrians only in uncontrolled cross- walks, and states vary in the strictness of yielding laws as well (such as whether motorists are required to yield at unmarked crosswalks, or whether the pedestrian must be in the same half of the roadway, in the next lane, etc.). Research into the potential safety impacts of these different laws was not found. However, the differences in laws sometimes affect applica- tion of treatments. There is also reason to think that these laws may affect motorist and pedestrian understanding and interactions, including motorists giving way to pedestrians at crosswalks and consequently the amount of pedestrian delay at uncontrolled crosswalks (McGrane and Mitman 2013). High-Visibility Crosswalks Description and Purpose of Treatment The 2009 MUTCD states that “marked crosswalks provide guid- ance for pedestrians who are crossing roadways by defining and delineating paths on approaches to and within signalized inter- sections, and on approaches to other intersections where traffic stops” (FHWA 2009) (Figure 15). High visibility describes the crosswalk marking type that is usually ladder, continental, bar pairs, and triple four markings (Mead et al. 2014). Summary of Effects Studies have estimated high visibility crosswalks reduce pedestrian crashes by 48 percent in general (CMF of 0.52),

48 and by 37 percent (CMF of 0.63) when applied in school cross- ing zones (Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes 2013 citing Chen et al. 2012, and Feldman, Manzi and Mitman 2010, respectively) but local effects may vary. High-visibility markings have also proved to be more readily detected by drivers and increase yielding, which may be especially important at uncontrolled locations (McGrane and Mitman 2013). FIGURE 15 Bar-pair-style high-visibility crosswalk markings in Bainbridge Island, Washington, can help reduce wear from tire tracks. Source: http://www.pedbikeimages.org, Carl Sundstrom. Use of High-Visibility Crosswalks Nearly all jurisdictions use high-visibility crosswalks in some circumstances. One-fifth of jurisdictions responding to the question about where they have been applied indicated that they now use high-visibility crosswalks at all marked crosswalk locations. Others mentioned using these at some locations, such as areas with high pedestrian activity or downtown develop- ment districts, midblock or uncontrolled locations, high-speed or wide roads, high-priority intersections, or school zones/ areas. One jurisdiction mentioned that it is currently updating its guidance but continues to prefer using them at select loca- tions because of long-term maintenance costs. Advanced stop/yield bars and signs were mentioned sev- eral times by agencies concentrating high-visibility cross- walks at midblock or uncontrolled locations. Other types of signs, signals, and warnings were mentioned, including RRFBs, overhead flashers and oversized signs, and R1-6a (in-street, stop for pedestrian) signs. One state also requires reflective pavement markers at midblock crosswalks. Advance Stop/Yield Bars and Signs Description and Purpose of Treatment Advance stop or yield markings, used in conjunction with signs, are a type of pavement marking placed before a cross- walk on multilane roads to increase the distance at which drivers stop or yield to allow pedestrians to cross (Figure 16). The advance stop or yield line would be supplemented with “Stop Here for Pedestrians” or Yield Here to Pedestri- ans” signs (R1-5 or R1-5a) to alert drivers where to stop to let a pedestrian cross. FIGURE 16 Advance yield bars used with “Yield Here to Pedestrians” (R1-5) sign in Milwaukee, Wisconsin. Note additional pedestrian warning sign and arrow at crosswalk. Source: PBIC (n.d.). An advance stop or yield line placed 20 to 50 ft ahead of the crosswalk can reduce the likelihood of a multiple-threat crash at unsignalized midblock crossings. The line or mark- ings encourage drivers to stop far enough in advance of the crosswalk that a pedestrian can see if a second motor vehicle is not stopping and be able to take evasive action, and can also increase the second driver’s chances of detecting the pedes- trian in time to stop. A setback of 30 ft for the line has been found to be a good distance for most purposes. Parking is also to be restricted between the stop or yield line and the cross- walk to allow for better visibility. See PEDSAFE, Advance Yield/Stop Lines (Zegeer et al. 2013) for more information. Summary of Effects A number of studies by Van Houten and colleagues have found reductions in conflicts with advance stop bars or yield markings and signs of various configurations and intensi- ties of applications. Although the increases in percentages of motorists who yield or stop tends to be small, the distance at which motorists stop or yield has increased, which improves sight lines between motorists in other lanes and pedestrians and helps reduce the multiple-threat crash type (Malenfant

49 and Van Houten 1990; Van Houten 1998; Van Houten and Malenfant 1992; Van Houten, Malenfant, and McCusker 2001; Van Houten et al. 2002). The addition of enforcement and educational measures may also improve the percentages of motorists who yield, as found in a study from three Cana- dian cities (Malenfant and Van Houten 1990). Other mea- sures could also be tried to increase yielding. Crash-based evidence is not yet available, but NCHRP Project 17-56 aims to develop CMFs for advance stop/yield lines and three other treatments, including median islands, PHBs, and RRFBs. Use of Advance Stop/Yield Bars and Signs Nearly 90% of both state and local agencies responding indi- cated that they use advance stop/yield bars and signs. This low-cost treatment combination can be viewed as a supportive device in that it has substantially increased stop/ yield distances from uncontrolled crosswalks and reduced motor vehicle–pedestrian conflicts. The device on its own is not expected to gain much additional yielding on the part of motorists. Two agencies mentioned they use RRFBs and other treatments such as median islands, curb extensions, and lighting along with advance stop/yield bars and signs to improve conditions for pedestrians. One jurisdiction men- tioned using advance stop/yield lines along with PHBs. Most of the remaining 19 jurisdictions that provided information mentioned using marked or high-visibility crosswalks, or signs and other kinds of flashers or warning signs, in con- junction with advance stop lines. In-Roadway “Yield to Pedestrians” Signs Description and Purpose of Treatment In the roadway, “Yield to Pedestrian” signs are placed in the middle of an uncontrolled crosswalk, often but not always on crossing islands, and direct drivers to yield to crossing pedestrians (Figure 17). Summary of Effects The use of in-roadway “Yield to Pedestrians” signs has been found to increase motorist yielding at crosswalks substan- tially, whether signs are placed directly at the crosswalk or farther away. However, trends suggest that motorist yielding may be increased with signs placed directly at the crosswalk, by using multiple signs at the crosswalk, or at the crosswalk plus locations farther away. There is little information on effects on conflicts, but one study found no consistent effect on motorist and pedestrian conflicts. There is also limited evidence about whether more pedestrians use the crosswalks treated with these signs, or whether delay is reduced. A cou- ple of studies have found evidence of significant traffic speed reductions, although one of these studies evaluated signs used in conjunction with removable islands. The use of signs along with PHBs also resulted in a higher yielding rate com- pared with PHBs alone or signs alone. Crash-based evidence of safety effects of these devices is also lacking. A further consideration noted by some studies is that the devices are damaged or removed in some cases. FIGURE 17 In-roadway “Yield to Pedestrian” signs (R1-6 signs) used with high-visibility crosswalk, median island, and detectable ramps. Source: http://www.pedbikeimages.org, Lyubov Zuyeva. Use of In-Roadway “Yield to Pedestrians” Signs About 70% of states and local jurisdictions use in-roadway “Yield to Pedestrians” signs. Pedestrian Warning Signs Description and Purpose of Treatment Advance pedestrian warning signs are typically used where pedestrian crossing are not expected. According to PED- SAFE, fluorescent yellow/green color is approved for pedes- trian, bicycle, and school warning signs (FHWA 2009, Section 2A.11; Zegeer et al. 2013, PEDSAFE Signing) (Figure 18). Summary of Effects Signs in conjunction with other treatments or improved con- spicuity through enhanced lighting treatments or both may improve driver yielding. Some treatments and studies have found improvements in pedestrian–vehicle conflicts, but results vary according to treatments and studies. In addition, signs alone may not bring yielding rates to very high levels. Crash- based evidence is lacking. Definite conclusions about use and optimal types and placement of signs are difficult to reach. Use of Pedestrian Warning Signs All 40 states surveyed and 88% of local jurisdictions apply pedestrian warning signs on occasion. Several mentioned

50 that they use these at midblock or uncontrolled crossings; several indicated at all such crossings. Other locations men- tioned were near bus stops; where there are high pedestrian volumes; near schools; or where sight distance issues, docu- mented crashes, or conflicts exist. One jurisdiction indicated that it used custom-made Yield to Pedestrians signs before other treatments were available. Finally, several jurisdic- tions mentioned they use pedestrian warning signs in all settings, or at all marked crosswalks, and sometimes other locations (all-way stops and traffic signals). FIGURE 18 Pedestrian crossing warning (W11-2 in foreground) with crossing sign and arrow pointing at crosswalk in background, Chapel Hill, North Carolina. Source: http:// www.pedbikeimages.org, Dan Burden. In-Pavement Flashing Lights (Associated with Crosswalks) Description and Purpose of Treatment In-pavement lights or flashing crosswalk warning systems are sometimes used to alert motorists to the presence of pedes- trians crossing at uncontrolled locations. Both sides of the crosswalk are lined with encased raised pavement markers, which sometimes contain LED strobe lighting. The lights dis- play outward in the directions of oncoming traffic and may be activated by passive detection of pedestrians waiting in the crossing area, or activated by pushbutton (Figure 19). Summary of Effects Effects on motorist speeds and yielding have varied widely across studies. Short-term improvements in motorist yielding have been reported, and were generally greater at nighttime for most locations. Yielding improvements tended to reduce over time. Although in some cases the treatment has been shown to also reduce vehicle approach speeds, a major drawback to this device is maintenance and the tendency for the installations to become damaged or malfunction (Thomas 2006). FIGURE 19 In-pavement flashing lights at crosswalk. Source: U.S. Air Force photo, Jason Minto; Dover Air Force Base website (http://www.dover.af.mil/News/Article-Display/ Article/230536/dover-installs-lighted-crosswalks). Use of In-Pavement Flashing Lights Several state and local jurisdictions mentioned having used these in the past, but that they are removing them as a result of unreliability, maintenance issues, and ineffectiveness com- pared with other TCDs. Concerns include the inability of fol- lowing drivers to see the devices compared with overhead or side-mounted devices, and lack of visibility in daytime. One state mentioned it now preferentially uses PHBs or RRFBs. Overhead or Roadside-Mounted Flashing Beacons Description and Purpose of Treatment Overhead and roadside-mounted pedestrian-activated yel- low beacons are used to alert motorists that pedestrians are crossing the roadway (Figure 20). FIGURE 20 Overhead flashing beacons used with crosswalk warning signs (W11-2A) and median island in Fargo, North Dakota. Source: http://www.pedbikeimages.org, Dan Burden. Summary of Effects Research has shown that overhead pedestrian signs with flashing beacons may encourage motorists to yield for pedestrians more often than when there are no signs (Mead et al. 2014, citing Nitzburg and Knoblauch 2001, Huang et

51 al. 2000, and Van Houten et al. 1999). These positive effects, may, however, vary or be modest because (1) yellow warning beacons are not exclusive to pedestrian crossings, so driv- ers may not necessarily expect a pedestrian when they see a flashing beacon; and (2) motorists learn that many pedestri- ans are able to cross the road more quickly than the timing on the beacon allows and therefore may think the person has already finished crossing the road if a yielding or stopped car blocks the pedestrian from sight. In field studies for TCRP Report 112/NCHRP Report 562, Fitzpatrick et al. (2006a) measured motorist yielding rates to “natural” population pedestrians from 36% to 62% at three pushbutton-activated sites, and ranging from 61% to 73% for passive/automatically activated overhead flashing beacons. Overhead flashing beacons have sometimes, but not always, achieved increases in motorist yielding or other behavioral improvements. The varied types of devices stud- ied and application with other warning signs and treatments and location types provides a limited basis for conclusions about safety and mobility effects for pedestrians. Use of Overhead or Roadside-Mounted Flashing Beacons About 80% of states, but only 41% of local jurisdictions, indicated that they currently sometimes use these types of beacons. Some jurisdictions use overhead or roadside- mounted flashing beacons to call attention to crossings on busy streets or high-speed or wide/arterial roadways with pedestrians crossing, school crossings, where a median island cannot be installed, or other areas with safety issues such as poor sight distance or conflicts. However, it appears these may be more of a “legacy” device. A number of juris- dictions indicated that they are no longer using overhead or roadside-mounted flashing beacons, but are now using RRFBs or PHBs in locations where they might have used the beacons in the past. One jurisdiction has, however, used them in advance of RRFBs. When asked about use of these devices, San Francisco practitioners responded, “Costs the same as installing a traffic signal but without the same safety benefits.” Massachusetts indicated that these are no longer installed at new locations, and that eligible locations now tend to get RRFBs. Rectangular Rapid Flash Beacon Description and Purpose of Treatment The rectangular rapid flash beacon device includes amber LED flashing lights installed to enhance pedestrian cross- ing warning signs at uncontrolled crosswalks (Zegeer et al. 2013, PEDSAFE, Rectangular Rapid Flash Beacon). The crosswalks should not be controlled by signals, stop signs, or yield control signs. RRFBs are placed on both sides of an uncontrolled crosswalk, below a pedestrian crossing sign, and above an arrow pointing at the crosswalk (Figure 21). The beacons differ from standard flashing beacons by using a rapid flash frequency (approximately 190 times per minute), brighter light intensity, and ability to aim the LED lighting. RRFBs can be automated or pedestrian actuated, and feature an irregular, eye-catching flash pattern to call attention to the presence of pedestrians. The RRFB was given interim approval as a crossing sign enhancement by the FHWA in 2008 (FHWA 2009). PEDSAFE indicates that RRFBs are good for two-lane streets but may be less suited for use on multilane roads because of the multiple-threat crash potential (Zegeer et al. 2013). FIGURE 21 RRFB installation with median islands, advance yield bars, and high-visibility crosswalk in Chicago, Illinois. Source: Flickr.com, Steven Vance—Chicago. Summary of Effects Research has frequently demonstrated that installing RRFBs on roadside pedestrian crossing signs can significantly increase motorist yielding behavior at uncontrolled crossings from very low levels before installation. In general, yielding has risen to anywhere from 35% of drivers up to nearly 100% nighttime yielding when two beacons per approach direction were installed. Whether pedestrians activate the device, if pushbutton activation is required, also appears to influence outcomes (Fitzpatrick et al. 2014). A 2010 study for FHWA evaluated yielding when the devices were implemented on 22 multilane roads in St. Petersburg, Florida; Washington, D.C.; and Mudelein, Illinois, and documented both rapid and sus- tained increases in yielding from an average of 4% of drivers before treatment to an average of 84% across all sites at 2 years postinstallation (Shurbutt and Van Houten 2010). How- ever, questions remain about the potential for multiple-threat collisions, which can occur when one driver yields and an approaching driver in another lane does not notice the pedes- trian, who may be obscured by the yielding vehicle. Advance stop/yield bars can help reduce the potential by encouraging yielding farther back from the crosswalk. Low levels of activation have also been reported (Hua et al. 2009). Foster et al. (2014) reported, however, that 94%

52 and 83% of pedestrians at two sites in Portland, Oregon, acti- vated the devices. Yielding may also increase over time as drivers become more familiar with the devices. Pedestrians may also be learning to use the devices more often, or there may be variation in results across geographic locations. Dif- ferences in enforcement and driving culture are also likely influences, although unequivocal evidence to support these conjectures is lacking. There remains some uncertainty about the other factors that may affect success of this treatment. Other studies (not of RRFBs) have shown driver yielding is strongly affected by actual driving speed on approaches to uncontrolled cross- walks (Gårder 2004; Bertulis and Dulaski 2014). A 2014 Texas study indicated that yielding rates varied by city, and that wider roads may be associated with lower yielding rates at RRFB sites. However, this study also found that yielding appeared to be higher at locations with higher speed limits, within the range of those studied (30 to 45 mph) (Fitzpatrick et al. 2014). However, speed limits were overlapped some- what with cities in the study, and actual driving speeds were not documented in Fitzpatrick et al. (2014). In addition, all of the Texas sites were also marked as school crossings, which may have contributed to higher yielding rates. It is not known at present if there is some threshold level of use of RRFBs or of PHBs (see the section on PHBs) that may result in an increasing or diminishing effect for one of both of these types of devices. Fitzpatrick et al. (2014) sug- gests that cities with more of either type of device have better outcomes. However, it is important that the use of warning- type beacons be carefully considered in the context of both the particular location and the network as a whole, given that experts caution that overuse of signs and beacons may result in drivers learning to disregard such devices. An active (at the time of this synthesis) NCHRP Proj- ect 17-56, “Development of Crash Modification Factors for Uncontrolled Pedestrian Crossing Treatments,” aims to vali- date safety benefits and develop crash modification factors for RRFBs and other pedestrian crossing treatments, includ- ing PHBs, advance stop/yield lines, and median islands. St. Petersburg, Florida, reports the following experience: “Three PHBs were installed along a regional trail and we are only achieving 78% motorist yielding compliance to a solid red beacon. In contrast, RRFBs at 54 locations are achieving 90% + compliance.” Use of Rectangular Rapid Flash Beacon Three-fourths of the states and 82% of the local jurisdictions have tried RRFBs, and some have quite a lot of experience with them. There were diverse practices in terms of the criteria or location types where RRFBs are being applied. It is clear that a number of jurisdictions are applying RRFBs on mul- tilane roads. Some jurisdictions indicated they use the treat- ment at high-volume or high-speed intersections or at special situations (school crossings), whereas others specified cross- ings that require enhanced safety but do not have high speeds (which these jurisdictions indicated as 40 mph and up) and have adequate sight distance. Other jurisdictions, such as, for example, Ann Arbor, Michigan, appear to have a more stan- dard policy, and may use RRFBs at all multilane unsignalized crosswalks, depending on available resources. St. Petersburg has installed 55 of the devices thus far, with 55 more sched- uled, and is pleased with their success (see sidebar). Other jurisdictions mentioned pedestrian volumes as a criterion, whereas Santa Barbara mentioned that it uses the 2005 Zegeer et al. report, Safety Effects of Marked Versus Unmarked Crosswalks, to identify locations that may require enhancements such as RRFBs. A couple of jurisdictions men- tioned use of these devices in association with roundabouts. Studies from Texas indicate that RRFBs are being used some- times on two-lane, two-way streets; one-way streets with four and five lanes; and two-way streets with up to six lanes. All of the studied locations are also in school zones, which the authors thought could have contributed to the high yielding rates in this study (Fitzpatrick et al. 2014). Advance yield bars and “Yield Here” signs were not described by the authors, but were present in images of the study locations in the article. Treatments mentioned as being used with RRFBs include crosswalks or high-visibility and ADA-compliant crosswalks. Signs and advance warning signs were also mentioned by a few agencies. California uses advanced yield lines and signs (recommended at uncontrolled crossings to reduce multiple- threat crash risk), pedestrian crossing warning signs, and high- visibility markings in conjunction with RRFBs. Arizona; Kentucky; and Portland and Eugene, Oregon, also mentioned advance stop/yield lines. The state of Oregon sometimes uses median islands, curb extensions, and lighting; lighting was also mentioned by other jurisdictions. Michigan typically uses RRFBs and crossing guards at school crossing locations that would benefit from greater conspicuity. Virginia mentioned crosswalks and advance signing, as well as public awareness and stakeholder (e.g., law enforcement) awareness. Pedestrian Hybrid Beacon (Formerly Called HAWK Signal) Description and Purpose of Treatment Pedestrian hybrid beacons, also known as HAWK beacons (short for High-Intensity Activated Cross Walk Beacon), are traffic control devices developed by Tucson, Arizona, traffic engineer Richard Nassi in the late 1990s as a means of provid- ing safe pedestrian crossings where minor streets intersected with major arterials (Mead et al. 2014; Fitzpatrick and Park 2010a, citing Nassi and Barton 2008) (Figure 22). The devices

53 have also been used at midblock crosswalks, and although there are pedestrian volume warrants outlined by the MUTCD, they are significantly lower than for traffic signals [Zegeer et al. 2013, PEDSAFE, Pedestrian Hybrid Beacon (PHB). FIGURE 22 Pedestrian hybrid beacon signal, Tucson, Arizona. Source: http://www.pedbikeimages.org, Mike Cynecki. A research effort carried out to assess alternative traffic control devices for Oregon resulted in a decision matrix between PHBs and RRFBs, or neither (Hunter-Zaworski and Mueller 2012, p. 78, report available at http://www.oregon.gov/odot/td/ tp_res/docs/reports/2012/spr721pedreport.pdf). At speeds greater than 40 mph, the decision guidance defaults to PHBs. Other criteria that are considered when speeds are at or below 40 include the existence or possibility of a median, visibility, whether there are more than three lanes or high traffic volume or both, and if the crossing is located in a high-risk surrounding environment (e.g., environments that include visual clutter and factors such as transit stops, retail and housing, schools, recreation and senior centers, or other vulnerable pedestrians). The PHB signal beacons are located both on the roadside and on mast arms over the major approaches to a crossing location. The head contains two red lenses above a single yellow lens. The lights are normally dark, but when acti- vated, the beacon first displays a few seconds of flashing yel- low, followed by a steady yellow interval, and then “displays a steady red indication to drivers and a WALK indication to pedestrians, thereby allowing them to cross the major road while traffic is stopped” (FHWA 2014). During the flashing pedestrian clearance interval, the red indications of the PHB changes to an alternating flash pat- tern to allow drivers to proceed after they have stopped, if the pedestrian has cleared the roadway, thereby contribut- ing less to vehicle delay than traditional stop-and-go signals. The first PHB was installed in Tucson in 2000. The PHB was considered an experimental treatment until 2009, when it was included for the first time in the MUTCD (FHWA 2009). Pedestrian volume warrants are, however, considerably lower for implementing a PHB than for a traffic signal (Zegeer et al. 2013). See Tucson’s 2014 ranking criteria for prioritizing crosswalks for treatment with PHBs in chapter five. Summary of Effects Motorist compliance or yielding at crosswalks is improved with PHBs compared with no control and compared with warning- type devices. Yielding has generally been consistently high, 90% and greater, at study locations with PHBs installed. A pedestrian crash CMF estimate of 0.31 (69% reduction), and a total crash CMF of 0.71 (29% reduction) were also derived from a 2010 FHWA study at 21 sites in Tucson (Fitzpatrick and Park 2010a, 2010b; Fitzpatrick et al. 2011). Fewer pedestrian conflicts have also been observed after PHB installation. In a study conducted by the Texas Transportation Institute, wider crossings were associated with higher yielding rates at sites in Texas, and posted speed did not have an effect on yielding, suggesting that this countermeasure is effective for wider and higher-speed locations. However, results from one study sug- gest that drivers might use the signal indications—especially when the beacon is in dark mode—to perceive that pedestri- ans are not present and increase speed. There is also presum- ably some risk of multiple-threat collision when the flashing red phase is active or if, as some states indicated in survey responses, drivers do not, in fact, stop. See results from more studies in Appendix B. Use of Pedestrian Hybrid Beacons Three-fourths of states and a smaller percentage of local jurisdictions (65%) indicate that they sometimes use PHBs. PHBs are being applied, according to responses, at midblock locations; at other (uncontrolled) approaches with known safety concerns or high crashes; or on roadways with mul- tiple lanes, high speeds, and high user demand. Because these devices combine elements of stop and go with “warn- ing” devices (with warrants outlined in the MUTCD), many jurisdictions’ respondents mentioned the importance of high volumes of traffic or high demand by pedestrians, bicyclists, equestrians, or warrants as being a decision factor. One PHB has been installed in San Francisco, a Vision Zero city, on a street managed by CALTRANS (state agency). However, “SFMTA [San Francisco Municipal Transportation Agency] feels that for the investment of funds, a full traffic signal is a better alternative at urban locations. Our engineers feel that Hawk signals [PHBs] prioritize auto service levels over pedestrian safety. It may be appropriate in a suburban context, but not in an urban context where there are higher volumes of people crossing, and reducing vehicle delay is not a priority.” At this point, SFMTA has no plans to implement PHBs on streets managed by the city.

54 Also, as mentioned by a few agencies, PHBs are considered when warrants for regular traffic signals are not met. Two juris- dictions, Washington State and Oakland County, Michigan, have used PHBs at pedestrian crossings at roundabouts. Wash- ington also mentioned using PHBs at limited-visibility locations. Several jurisdictions are just beginning to look at using PHBs, or have a few in planning stages, and one state indi- cated that the device is counter to state law and is not used on the state system. When used, treatments sometimes used with PHBs include the following: • California mentioned advanced stop bars, pedestrian countdown signal with accessible pedestrian signals (APS), and crosswalk stop on red (R10-23) sign. • Virginia mentioned crosswalks and advanced signing, as well as public awareness and stakeholder (e.g., law enforcement) awareness. • North Carolina mentioned public outreach and education. • Oregon may use median islands, curb extensions, and lighting. • Wisconsin has a draft policy in progress for PHBs and RRFBs. Use of RRFBs and HAWK installations at multilane roundabouts were studied for the effects on safety and accessibility for blind persons at two multilane roundabouts in Oakland County, Michigan. The complete study is available on the Road Commission for Oakland County website at http://www.rcocweb.org/Lists/Publications/ Attachments/126/HAWK%20Final%20Report%202011.pdf. One agency mentions a point that public education and outreach may be important for newer (even if MUTCD- approved) types of devices, especially if they have not been used in an area. A Texas study cited earlier (in the RRFB section) also suggests that effects tend to be more positive in cities or areas with multiple installations of PHB devices compared with locations with one to few, suggesting a learn- ing curve that education and outreach may help accelerate. Other jurisdictions that have used the devices report favor- able results: Scottsdale, Arizona, reports, “Our HAWKs are operating successfully and our RRFBs have been well accepted.” Soon after meeting the minimum warrant of 20 pedestrians per hour and installing a PHB, the volume of pedestrians at the crossing increased to 360 per hour. Scottsdale, Arizona, indicates: “The HAWK crossing Scottsdale Road for the new Scottsdale Quarter on the border of Phoenix and Scottsdale just met the pedestrian warrant of 20 pedestrians per hour. Only a month later there were 360 pedestrians in an hour and it went as high as 660 pedestrians per hour. This two- stage crossing treatment connects two major commercial centers in both cities,” and the city deems the installation very successful. Install Traffic Signal Without Pedestrian Countdown Signals Description and Purpose of Treatment Traffic signals are intended to create a gap for pedestrians to cross the road where they would otherwise experience safety issues, significant delay or other difficulties in crossing. Tra- ditional signals with WALK/DON’T WALK pedestrian indications may remain for their useful lives, but pedestrian countdown signals should be installed for all new treatments. The MUTCD provides warrants for the installation of traffic signals based on pedestrian and motor vehicle volumes, but also describes special considerations. These include the lack of a signal possibly inhibiting “young, elderly, and/or per- sons with physical or visual disabilities” from crossing and/or “nearby facilities and activity centers” intended to serve these populations possibly generating demand so that counts may not accurately reflect pedestrian demand. Thus, engineering judg- ment is expected to consider pedestrian needs comprehensively (Zegeer et al. 2013, PEDSAFE, Traffic Signals). PEDSAFE also indicates that where pedestrians are regular and frequent, a pedestrian phase should be actuated automatically. Summary of Effects It is not clear how the use of a traditional walk/don’t walk signal affects pedestrian crash risk for intersections without pedes- trian signals, because the need for pedestrian signals may be much greater at some sites than others. More research is needed to better understand the effects of pedestrian signals, but it is clear that pedestrians may be confused at intersections with wide street crossings where no pedestrian signals (or inad- equate crossing time or conflicts with turning vehicles) exist. Certain types of traffic signal phasing, such as an exclu- sive phase for pedestrians or restricted left-turn phasing, have been found to reduce pedestrian–motor vehicle collisions, as compared with traditional concurrent signal phasing (Tool- box of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes 2013, citing Harkey et al. 2008 and Chen et al. 2012). More information on other phasing or signal control strategies are described under the leading pedestrian interval countermeasure and right-turn-on-red restrictions. Use of Installation of Traffic Signal Without PCS Although several jurisdictions continue to have locations with signals that have not been updated yet, most jurisdic- tions are routinely replacing these, or have mostly done so with signals with pedestrian countdown signals (PCSs) in keeping with MUTCD guidelines. Only one jurisdiction mentioned that it sometimes installs signals without pedes- trian countdown timers in locations with no pedestrian accommodation or connectivity and no indication of pedes- trian demand, mostly in rural areas.

55 Install Traffic Signal with Pedestrian Countdown Signal Description and Purpose of Treatment Countdown signals are indications designed to begin count- ing down at the beginning of the clearance interval (flashing “WALK”/“DON’T WALK”) and can be on fixed-time or push- button operation (Figure 23). They indicate to the pedestrian how much time is left in the crossing phase. According to the MUTCD, countdown pedestrian indications are required for all newly installed traffic signals where pedestrian signals are installed. See PEDSAFE Traffic Signals, Pedestrian Signals, Pedestrian Signal Timing, and other signal-related countermea- sures in PEDSAFE for more information (Zegeer et al. 2013). FIGURE 23 PCS with countdown indicators. (Top) Countdown timer below walk/don’t walk indicators, Washington, D.C. (Bottom) Side-by-side walk and countdown indicators, New York City. Source: Flickr. com, Eric Fischer (lower image modified by cropping). Summary of Effects A CMF of 0.75 for pedestrian crashes was developed from one study of the effects of replacing traditional walk/don’t walk pedestrian signals with pedestrian countdown signal heads (Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes 2013, citing Markowitz et al. 2006). However, other studies have found more mixed results from widespread installation. Another positive finding: motorist behaviors have not been shown, thus far, to be significantly negatively affected by information from the countdown timers. Pedestrians may or may not comply more often with the walk phase of the signal, and results are somewhat mixed as to whether they tend to be stranded less often in the crosswalk at a signal change. Regarding crashes, as with traditional signals, safety benefits may be affected by the particular phasing and tim- ing strategies used, such as allowing more pedestrian walk time (CMF = 0.49), or providing exclusive phases (0.65) (Toolbox of Countermeasures and Their Potential Effective- ness for Pedestrian Crashes 2013, citing Chen et al. 2012. These studies are also summarized in Appendix B of this study, under Install Traffic Signal with Pedestrian Count- down Signal. Also see the CMF Clearinghouse for more CMFs on signal phasing strategies. Other signal timing and traffic control strategies are covered under the leading pedestrian interval and no-turn-on-red restrictions counter- measures sections. Use of Installation of Traffic Signal with PCS Again, almost all the jurisdictions responding are using PCSs for replacements or at new signalized intersections with pedestrian indications or anywhere pedestrians are present. A couple of jurisdictions qualified that they use PCSs at locations near schools and other pedestrian gen- erators, or at locations with crash histories or bike crossing improvement projects. Leading Pedestrian Interval Description and Purpose of Treatment A leading pedestrian interval at a signalized intersection gives pedestrians anywhere from a 3- to 7-s head start before traffic is released by the parallel-path corresponding green signal. This treatment is basically intended to help enforce the pedestrian right of way and encourages turning motorists to yield to parallel-path crossing pedestrians by delaying the release of the motorists until pedestrians have established presence in the crosswalk (Zegeer et al. 2013, PEDSAFE Leading Pedestrian Interval).

56 Summary of Effects LPIs, which have been studied predominantly at high- pedestrian sites, are another countermeasure that has some crash-based evidence of effectiveness. A CMF of 0.95 (5% reduction) for pedestrian crashes was estimated for LPIs (FHWA, citing ITE Briefing Sheet 8, 2004), but a more recent study estimates reductions of 59% in pedestrian– vehicle crashes at 10 treated intersections (significant at the 95% confidence level) (Fayish and Gross 2010). Behavioral studies provide additional support for crash reductions, showing a reduction in conflicts with turning vehicles, par- ticularly left-turning vehicles. RTOR restrictions may also be needed to further reduce conflicts and crashes with right- turning vehicles and improve ability of pedestrians to access the crosswalk when waiting on the curb in some situations such as intersections with high volume of right-turning vehi- cles (Hubbard et al. 2008). Use of Leading Pedestrian Intervals About 61% of states and 77% of local jurisdictions surveyed reported using LPIs. These are often used on a case-by-case basis, either where complaints come in or collision history or conflicts suggest their use. Maryland indicates, “Generally used where motorists are more aggressive and pedestrian vol- umes are not high enough to command that motorists yield. More specific guidance is currently being developed.” Other situations mentioned were for crosswalks along approaches with heavy turning traffic; locations with dual turning move- ments, high pedestrian volumes, and high elderly population; or school crosswalks or other identified safety improvement locations. A couple of jurisdictions appear to be routinely employing LPIs or using them at locations with concurrent signal phasing and significant pedestrians or turning con- flicts. Maryland indicates it is currently developing more spe- cific guidance for implementation of LPIs. Use of Pedestrian-Only Walk Phases Pedestrian-only walk phases (0.65 CMF; Toolbox of Coun- termeasures and Their Potential Effectiveness for Pedes- trian Crashes 2013, citing Chen et al. 2012) typically have more limited use, and were primarily based on volumes of pedestrians. One state mentioned that these had been tra- ditionally used at all signals, but that state is moving away from them to improve efficiency for all users. Right Turn on Red Restrictions Description and Purpose of Treatment During the mid-1970s, a number of states in the eastern por- tion of the United States adopted the “permissive” type of RTOR that was already common in some western states. This approach allowed RTOR at all locations that were not otherwise marked by a prohibitory sign. Before making a RTOR, motorists are required to stop and yield to pedestri- ans, bicyclists, and oncoming vehicles. No-turn-on-red restrictions are meant to keep motorists from turning right during red lights across the path of con- flicting pedestrians, and are usually indicated through signs (Figure 24). Some communities are using active signs that can be adjusted for peak/nonpeak periods. Although laws require motorists to come to a full stop and yield to cross- street traffic and pedestrians before turning right on red, many motorists do not fully comply with the regulations, especially at intersections with wide turning radii. Motor- ists are often intent on looking for traffic approaching from their left and may not look for pedestrians approaching on their right. In addition, motorists often pull across the cross- walk to wait for a gap in traffic, blocking pedestrian cross- ing movements. In some instances, motorists simply do not come to a full stop (Zegeer et al. 2013, PEDSAFE Right- Turn-on-Red Restrictions). PEDSAFE also indicates that restrictions are to be considered where exclusive pedestrian phases or high pedestrian volumes are present. FIGURE 24 No-turn-on-red time-based restriction in the District of Columbia. Source: Flickr.com, M.V. Jantzen. Summary of Effects Although collisions and injuries to pedestrians resulting from right turns on red usually make up a small percentage of pedestrian and bicycle crashes, both pedestrian and bicy- cle crash rates were observed to increase in several states following implementation of RTOR allowance. Pedestrian conflicts with right-turning vehicles that violate right-turn- on-red laws can be significant, given the high percentage of motorists that do not fully stop and the observed behavior of motorists to look only left before pulling out. Active signs may be more effective than static ones at inducing motorists to stop before turning right on red. In addition, time-related

57 restrictions may be more effective than ones that call on the motorist to notice pedestrian presence (no RTOR when pedestrians present), but more evidence is needed regard- ing safety and mobility outcomes for pedestrians. NTOR has been estimated to reduce all crashes by about 3% (CMF = 0.97), but estimates of effects for pedestrians are not yet available (Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes 2013, citing Harkey et al. 2008). Several agencies, including Ithaca, New York, and Charlotte, North Carolina, reported using dynamic no-turn-on-red signs. Charlotte has tried these in conjunction with a red arrow for right-turning lanes at busy intersections. They are activated by pedestrian pushbutton and hold the right-turning traffic while pedestrians cross. During peak hours when right-turning traffic cannot be held through the entire walk phase, they switch to a “Yield to Pedestrians” message. Oregon uses curb extensions, lighting, and LPIs along with NTOR. Use of No-Turn-on-Red Restrictions A high percentage of states (86%) and local jurisdictions (94%) use no turn on red (NTOR) on occasion. New York City has a citywide regulation restricting right turns on red. Most agencies that provided answers mentioned sight dis- tance or vehicular sight distance as a main consideration for implementing NTOR restrictions. Others indicated they use or only use NTOR where there are high numbers of pedestri- ans or pedestrians/bicyclists (including where there are bike boxes or documented conflicts. A few agencies mentioned using NTOR at school crossings or other areas of special concern such as crossings that are also near a light rail light, at skewed intersections, or near freeway off ramps. See a PEDSAFE case study on an inexpensive treatment used in Hoboken, New Jersey, for “daylighting” intersections—that is, reducing cars illegally parked too close to the intersection— available at http://pedbikesafe.org/PEDSAFE/casestudies_ detail.cfm?CM_NUM=9&CS_NUM=74. Parking Restrictions Description and Purpose of Treatment Vehicles that are parked too close to pedestrian crossings obscure the visibility of both drivers and pedestrians. This increases the risk of dart-dash-type pedestrian collisions at both intersections and midblock crosswalks. Eliminating park- ing from the roadway near a crossing can improve pedestrian and driver sight lines and make it easier for pedestrians to cross (Figure 25). It is important to install physical barriers to negate the possibility of illegal parking. The removal of parking spaces can also free up the roadway for other uses, such as curb bulb- outs/extensions, landscaping, or bicycle parking (Zegeer et al. 2013, PEDSAFE, Parking Restrictions). Typically, jurisdic- tions should restrict vehicles from parking within at least 20 ft of an intersection. In its Urban Street Design Guide, NACTO (n.d.) suggests 20–25 ft restricted areas. Summary of Effects One study estimated restricting parking near intersections to off street to yield a CMF of 0.7 in pedestrian crashes (Tool- box of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes 2013, citing Gan et al. 2005). About 75% of states and 88% of the local jurisdictions use parking restrictions in some instances. Oregon may use curb exten- sions, lighting, LPIs, medians, and RRFBs in conjunction with parking restrictions. Some jurisdictions mentioned using restrictions near both midblock and intersection cross- ings. One jurisdiction indicated that it would like additional guidance on using parking restrictions. FIGURE 25 Parking restrictions improve sight lines between motorists and pedestrians near crosswalks. Source: PEDSAFE, Peter Lagerwey. Summary of Effectiveness of Traffic Control Devices and Treatments Table 9 provides an overview of pedestrian safety effects from evaluations of traffic control devices. When available, estimates of effects on pedestrian crashes are provided; some- times estimates of total crash effects are included, especially when these are high-quality estimates. Often the known safety effects are only from observed behaviors measures

58 TABLE 9 OVERVIEW OF STUDIES AND FINDINGS FOR TRAFFIC CONTROL DEVICE TREATMENTS Design or Treatment Area Types Studied Key Measured Effects High-visibility crosswalks General urban crossings, urban school zone crossings; behavioral effects also studied at university campus midblock locations • CMF 0.52 pedestrian crashes • CMF 0.73 pedestrian crashes in school zones • High-visibility markings have also proved to be more readily detected by drivers and to increase yielding. Advance stop/yield bars plus signs Mostly multilane uncontrolled crossings in urban areas. One study included rural pedestrian areas where pedestrian crashes had occurred. • Motorist yielding/stopping distance increases at uncontrolled locations. • Pedestrian–motor vehicle conflicts decrease substantially at uncontrolled locations. • Generally only small increases in motorist yielding, unless combined with enforcement and educational measures. Other engineering measures may also increase yielding in association with the treatment, but the evidence base is thin. • Driver yielding, vehicle stop position, or pedestrian–vehicle conflicts may not be affected at signalized locations. • Crash-based evidence is unavailable at present. In-roadway “Yield to Pedestri- ans” sign Uncontrolled midblock and inter- sections; multilane; unknown number of lanes; state highways; one-way and two-way streets • Substantial increase in motorist yielding at all types of crosswalk locations studied compared with no signs • Trends suggest motorist yielding may be highest with signs placed directly at the crosswalk, compared with farther away, or by using multiple signs at the crosswalk with or without signs at locations farther away. • There is little information on effects on conflicts. • There is also limited evidence about pedestrian mobility/delay effects or changes in pedestrian behaviors. • Traffic speed reductions have been observed in two studies, although one of these combined signs with removable refuge islands. • The use of signs along with PHBs also resulted in a higher yielding rate compared with PHBs alone or signs alone. • Crash-based evidence is unavailable at present. • Prevalence of damage or device removal may affect the benefits. Pedestrian warning signs Varies • Signs, especially in conjunction with other treatments or improved conspicuity through enhanced lighting treatments may improve driver yielding. • Some treatments and studies have found improvements in pedestrian–vehicle conflicts, but results vary according to treatments and studies. • Signs alone may not bring yielding rates to very high levels. • Definite conclusions about signs and types and placement of signs for maximal effectiveness for different types of locations are difficult to reach. • Crash-based evidence is unavailable at present. In-pavement flashing lights (associated with crosswalks) Mix of urban and suburban loca- tions, including midblock, T-inter- sections; road types from local/ campus to undivided and divided arterials • Motorist yielding generally improved, but often to much less than 50%. • Speed results varied. • Yielding rates have tended to degrade over time. • Nighttime yielding may have improved by a larger percentage than daytime, but initial nighttime rates were usually very low. • Treatment not sufficient to channelize pedestrians to the crossing in some situations • Varied study locations and conditions, but no studies have compared effects in different environments • Crash-based evidence is unavailable. Overhead or roadside-mounted flashing beacon Urban and suburban uncontrolled crosswalks • Some improvements in motorist yielding may be observed; possibly higher in conjunction with other treatments. • Crash-based evidence is unavailable. Rectangular Rapid Flashing Beacon Urban and suburban uncontrolled crossings, including one trail crossing • Significant increases in motorist yielding behavior at uncontrolled crosswalks up to high rates (90% and above) in many locations, including at multilane sites, but yielding rates as low as 34% in some cities • Variations in outcomes may relate to time since installation, city of implementation, number of treatments in a city, width of crossing (and possibly number of lanes), traffic speed, and whether the beacon is activated • Higher speed limits (within the range of those studied, 30–45 mph from Texas sites) were associated with slightly higher rates of yielding at 22 sites in three cities; however all sites also had “School Zone” signs. • Crash-based evidence is unavailable at present. • Concerns about multilane use, potential multiple-threat crash situation

59 Design or Treatment Area Types Studied Key Measured Effects Pedestrian hybrid beacon (for- merly called HAWK signal) Urban/suburban midblock cross- ings and uncontrolled intersections • 0.31 CMF pedestrian crashes • 0.71 CMF all crashes • Increase to high rate of motorist yielding/compliance at crosswalks— generally 90% and higher • Wider crossing distance was positively associated with driver yielding in a Texas study of 32 sites in four cities. Install traffic signal without pedestrian countdown signal Varied intersections in urban areas • Exclusive phasing or restricted turn phases provide safety benefits to pedestrians. Install traffic signal with pedes- trian countdown signal (PCS) Varied intersections in urban areas • CMF of 0.75 for pedestrian crashes • Other crash effects also reported⎯mixed results on crash effects of widespread implementation of PCS (replacing traffic signals without PCS) – Nonsignificant, slight increase in pedestrian collisions when PCS signals installed at formerly nonsignalized locations • CMF of 0.65 for pedestrian crashes, with exclusive pedestrian phase • Giving pedestrians more time in pedestrian walk phase associated with safety (and mobility) benefits • Drivers do not appear to increase their speed on a late green (per information on pedestrian countdown). • Pedestrian compliance starting on “WALK” indication—mixed results • Pedestrians remaining in crosswalk at signal change to green for opposing traffic—mixed results Leading pedestrian interval Varied signalized intersections of urban, suburban, and small town thoroughfares • 59% statistically significant reduction in pedestrian–vehicle crashes at 10 treated intersections (CMF of 0.41) • Reduced conflicts, especially with left-turning motorists • RTOR restrictions may also be needed to reduce conflicts and crashes with right-turning vehicles. “No turn on red” restrictions Varied signalized intersections in urban and suburban locations • 0.97 CMF all crashes; no crash evidence for pedestrian crashes exclusively • Active signs may be more effective than static ones at inducing motorists to stop before turning right on red. • Time-related restrictions may be more effective than ones that call on the motorist to notice pedestrian presence (no RTOR when pedestrians present). Parking restrictions One study • One study estimated restricting parking near intersections to off street to yield a CMF of 0.7 in pedestrian crashes. Note: See text for description of each countermeasure and Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes (FHWA 2013) for most CMFs and references; see Appendix B tables for study details including behavioral effects. TABLE 10 SUMMARY OF STATE AND LOCAL JURISDICTION USE OF TRAFFIC CONTROL DEVICES TO IMPROVE PEDESTRIAN SAFETY AT CROSSINGS Traffic Control Devices States Using Treatment Local Jurisdictions Using Treatment Pedestrian warning signs 100% 88.2% Traffic signal with pedestrian countdown signal 97.3% 100% High-visibility crosswalks (e.g., continental, zebra, bar pairs) 97.3% 100% Advance stop/yield bars and signs 89.2% 88.2% “No turn on red” restrictions 86.5% 94.1% Overhead or roadside-mounted flashing beacon 78.4% 41.2% Rectangular rapid flashing beacon 75.7% 82.4% (Pedestrian) hybrid beacon (formerly called HAWK signal) 75.7% 64.7% Parking restrictions 75% 88.2% In-roadway “Yield to Pedestrians” signs 73.0% 70.6% Leading pedestrian interval 62.2% 76.5% Pedestrian-only walk phase or pedestrian scramble 43.2% 52.9% Install traffic signal without pedestrian countdown signal 18.9% 11.8% In-pavement flashing lights (associated with crosswalks) 16.2% 0 Total responding 37 17

60 such as motorist yielding, speed, or pedestrian delay or com- pliance with crosswalks, because as mentioned previously, crash-based estimates of pedestrian safety effectiveness require many study locations and robust analysis methods. See Appendix B for tables with summaries of each study reviewed for each treatment, including crash-based and behavioral effects of the treatments studied. Also see the Toolbox of Countermeasures and Their Potential Effectiveness for Pedestrian Crashes (2013), which sum- marizes most CMFs included, and provides links to the original studies. SUMMARY OF APPLICATION OF TRAFFIC CONTROL DEVICES Table 10 summarizes the responses for which traffic con- trol devices are used by jurisdictions, by type of jurisdic- tion. Recall that the local jurisdictions surveyed were a select sample of cities thought to have good practices in place. Washington, D.C., responses are included with the states. Most traffic control devices were used, at least sometimes, to enhance pedestrian crossings. The treatments most often used were the following: • Pedestrian warning signs are used by all the states responding and 88% of the local jurisdictions. • Traffic signals with pedestrian countdown timers (as currently recommended in the MUTCD) are used by nearly all states (97%) and local jurisdictions (100%). • High-visibility crosswalk markings are used by 97% of the states and 100% of the local jurisdictions. • High-visibility crosswalks are also used by nearly all states that responded (97%), and by all of the local jurisdictions that responded. The local jurisdictions responding tended to make less use of overhead beacons (41% of those responding) than states (78%), with some cities indicating that they cost nearly the same as a traffic signal, but are much less effective, or that RRFBs are now being used where overhead beacons might formerly have been used. Among the treatments least often used were traffic sig- nals without pedestrian countdown signals (19% of states and 12% of local jurisdictions responding). Another treat- ment seldom being used at present, according to these juris- dictions, is in-roadway flashing crosswalk lights (16% of states and none of the set of local jurisdictions).