National Academies Press: OpenBook

Recent Roadway Geometric Design Research for Improved Safety and Operations (2012)

Chapter: Chapter Four - Cross-Section Elements

« Previous: Chapter Three - Elements of Design
Page 20
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 20
Page 21
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 21
Page 22
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 22
Page 23
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 23
Page 24
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 24
Page 25
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 25
Page 26
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 26
Page 27
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 27
Page 28
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 28
Page 29
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 29
Page 30
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 30
Page 31
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 31
Page 32
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 32
Page 33
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 33
Page 34
Suggested Citation:"Chapter Four - Cross-Section Elements." National Academies of Sciences, Engineering, and Medicine. 2012. Recent Roadway Geometric Design Research for Improved Safety and Operations. Washington, DC: The National Academies Press. doi: 10.17226/14661.
×
Page 34

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

21 Overview Researchers investigated operational and safety character- istics of passing lanes on two-lane highways, which grew in popularity during the decade in an attempt to maximize performance without widening to traditional four-lane high- ways. At the same time, the concept of the “road diet” was introduced to reduce traffic speeds and provide space for pedestrian or bicycle amenities without widening the right- of-way. The effects of the width of travel lanes and TWLTLs were examined. Shoulder treatments such as rumble strips received attention for their potential in reducing crashes, and researchers looked at new median and roadside treatments. AllOcAtiOn Of trAveled wAy width An NCHRP-sponsored scan tour looked at characteristics of 2+1 roads in several European countries to determine the potential applications of the design for use in the United States (Potts and Harwood 2003). A 2+1 road design has a continuous three-lane cross section with alternating pass- ing lanes. Although specifics of the designs in the respective countries varied somewhat, the authors made comparisons of some of the key design and operational criteria, which are summarized in Table 6. The NCHRP authors concluded that the benefits of 2+1 roads in Europe validated a recommendation for their use in the United States, to serve as an intermediate treatment between an alignment with periodic passing lanes and a full four-lane alignment. They also recommended that 2+1 roads were most suitable for level and rolling terrain, with installa- tions to be considered on roadways with traffic flow rates of no more than 1,200 veh/hr in a single direction. The authors discouraged the use of cable barrier as a separator, and they recommended that major intersections be located in the buf- fer or transition areas between opposing passing lanes, with the center lane used as a turning lane. Gattis et al. (2006) reported on a study of passing lane operations in Arkansas. The focus was on segments of con- tinuous three-lane cross sections with alternating passing lanes; for example, three-lane alternate passing or 2+1. They examined the effects of passing lane length on platooning, passing, speed, and passing lane crash rates. Five sets of field data were collected at four rural sites, and it was determined that platooning decreased and eventually stabilized after a vehicle entered the passing lane. They observed that passing activity was greatest at the beginning of the segments and the greatest benefits of decreased platooning and increased safety occurred within the first 0.9 mi of a passing lane seg- ment. Speed patterns were found to vary among sites, but average speed rose when a vehicle entered the passing lane section. Their study of crash rates included “five years of crash data from 19 sites; [they concluded that] even though the volumes for the passing lane segments were higher than the state average volume for rural two-lane roads, the pass- ing lane crash rates were generally lower than the statewide average crash rate for rural two-lane roads.” A study of similar roadways in Texas, called “Super 2” highways, found that passing lanes were beneficial at volumes approaching 15,000 vehicles per day, particularly on rolling terrain; the presence of passing lanes improved delay and per- cent time spent following (Brewer et al. 2011). Most passing occurred within the first mile of a passing lane, so additional length may be less useful than additional lanes in a Super 2 corridor, particularly at lower volumes. Empirical Bayes analysis of crash data showed that there was a statistically significant crash reduction of 35% for segment-only (i.e., nonintersection) injury crashes on the study corridors, as compared with the expected number of crashes without pass- ing lanes. In lieu of guidelines related to specific ADT values, researchers recommended including general principles for Super 2 design as part of their proposed revisions to the Texas Roadway Design Manual, such as avoiding intersections with state highways and high-volume county roads within passing lanes, consideration of terrain and right-of-way in determin- ing alignment and placement of passing lanes, avoiding the termination of passing lanes on uphill grades, and discourag- ing passing lane lengths longer than 4 mi. Volume 4 of NCHRP Report 500 (Neuman et al. 2003b) discusses the use of center TWLTL on four-lane and two- lane roads to reduce the likelihood of head-on and rear-end collisions. It can be accomplished either by the conversion of four-lane undivided arterials to three-lane roadways with a center left-turn lane or by the more conventional reconstruction of a two-lane road to include the TWLTL. Since the latter could be a costly con- version because it may require new right-of-way, the four-lane road conversion is considered more appropriate to the AASHTO emphasis on low-cost alternatives. However, where right-of-way chapter four crOss-sectiOn elements

22 cost is not a major consideration, the inclusion of TWLTLs on existing two-lane roads may be an even more effective treatment for head-on collisions since more of such collisions would likely occur on two-lane roads than on four-lane roads. The development of TWLTLs is usually for traffic operations rather than safety concerns. TWLTLs are usually implemented to improve access. When they are used in response to a safety con- cern, the use is traditionally to reduce driveway-related turning and rear-end collisions. However, because studies have also indi- cated a positive effect on head-on crashes, the strategy is included here. The principle behind the use of TWLTLs in this context is to provide a buffer between opposing directions of travel. The strat- egy is intended to reduce head-on crashes by keeping vehicles from encroaching into opposing traffic lanes through the use of the buffer. If available right-of-way, construction budget, and traffic vol- umes allow, incorporating TWLTLs into the design of new and reconstructed roads is more efficient than converting road- ways later. managed lanes Transportation agencies in many jurisdictions have exam- ined new ways to maximize the use of existing infrastruc- ture to improve capacity and reduce congestion, particularly on urban freeway corridors. One such method is the use of managed lanes, whereby one or more lanes in the corridor are reserved for use that is limited to specific types of vehi- cles, such as high-occupancy vehicles (HOV), buses, motor- cycles, or toll-paying drivers. Kuhn et al. (2005) conducted a multi-year study on a wide variety of planning, design, and operational issues related to managed lanes. A chapter in their Managed Lanes Handbook contains recommendations and guidelines for design elements of freeway managed lanes. In general, they recommended that the features of the man- aged lane be commensurate with the design vehicle that is selected to be appropriate for the facility. They recommended that the designer use the AASHTO Green Book templates in determining turning paths, lateral and vertical clearances, bus stops, and other elements associated with a project. In particu- lar, the design process might also account for the path of the vehicle overhang beyond the outside turning radius. According to Kuhn et al. (2005), “in most cases, the design speed of managed lanes will be the same as that used on the adjacent general-purpose lanes. However, there may be limited instances where the design speed of the man- aged lanes is lower than the adjacent general-purpose lanes, owing to the geometrics of the managed lanes facility or other limitations. The designated design speed of the facility should relate to the maximum speed the facility is expected to accommodate. Further, the design speed should accom- modate the vast majority of users (e.g., the anticipated 85th percentile speed).” The Managed Lanes Handbook also contains recommen- dations on horizontal and vertical clearance and curvature, gradient, SSD, cross slope, superelevation, and minimum turning radius. Their recommended guidelines are summa- rized in Table 7. hard shoulder running In the United States, the primary use of shoulders has been as a safety refuge area; however, in recent years there has been an increasing trend for transportation agencies to explore the use of shoulders as travel lanes during peak periods as a congestion management strategy. Also called “hard shoulder running,” the limited use of the shoulder as a travel lane has been primarily reserved for special users of the roadway system, most often transit vehicles. Overall, experience using shoulders for interim use has been posi- tive in the United States, and more agencies are considering the strategy to address growing congestion on their urban freeway networks. Several states have deployed temporary shoulder use for all vehicles on congested corridors with success. Kuhn (2010) describes several uses of hard shoul- der running in the United States, as noted in the following paragraphs. In San Diego, California, along I-805/SR 52, transit vehi- cles may use the freeway shoulder during congested periods, when general-purpose lane traffic slows to 30 mph or lower. TABLE 6 COMPARISON OF EUROPEAN 2+1 ROAD CHARACTERISTICS Germany Finland Sweden Critical Transition Length, m (ft) 180 (590) 500 (1,600) 300 (1,000) Non-Critical Transition Length, m (ft) 30–50 (100–160) 50 (160) 100 (330) Typical Passing Lane Length, km (mi) 1.0–1.4 (0.6–0.9) 1.5 (0.9) 1.0–2.0 (0.6–1.2) Separation between Opposing Traffic m (ft) 0.5 (1.6) 0.3 (1.0) 1.25–2.0 (4.1–6.6) Fatal+Injury Crash Rate, per 106 veh- km (106 veh-mi) 0.16 (0.26) 0.09 (0.14) 0.50 (0.80) Typical Volumes, veh/day 15,000– 25,000 14,000– 25,000 4,000– 20,000 Source: Potts and Harwood (2003). Note: Sweden’s 2+1 roads are separated by cable barrier, and their crash rates are specifically reported as crashes per million axle pair-km.

23 They may travel no more than 10 mph faster than traffic in general traffic lanes. The cross section of the shoulder is at least 10-ft wide throughout the deployment area. Pavement markings to indicate the operational strategy include text indi- cating “Transit Lane Authorized Buses Only” (Martin 2006). The Florida DOT, Miami–Dade Transit, and the Miami Dade Expressway Authority operate several shoulder use applications in the Miami region. Along the Florida Turn- pike, SR 826, and SR 836, buses are allowed to use the shoulder when the freeway is congested. Implemented in 2005, the program stipulates that buses may travel no faster than 35 mph on the shoulder when open to transit. The typi- cal cross section is a minimum 10-ft width, with a 12-ft width in high volume areas, and cross slopes of 2% to 6%. Pavement markings indicate “Watch for Buses on Shoulder” (Martin 2006). On GA 400 in Alpharetta, Georgia, buses are allowed to use the freeway shoulder in an effort to provide access between a local transit rail station and a park-and-ride lot. The Georgia Regional Transportation Authority and Georgia DOT operate the facility, which is functional whenever traf- fic slows to 35 mph or less. Buses can travel no more than 15 mph faster than general-purpose lane traffic and are required to reenter general traffic lanes before interchanges. A con- struction project was necessary to upgrade the travel surface by widening the shoulder by 2 ft and providing reinforcement of the shoulder pavement (Martin 2006). An extensive system around Minneapolis–St. Paul, Min- nesota, is also focused on buses. The operational strategy is considered interim, and some deployments have been removed since their inception. When operational, buses must yield to any vehicle entering, merging, or exiting through the shoulder, and buses must reenter the main lanes when the shoulder is obstructed. Typically, buses may use the shoul- der any time that traffic in the adjacent mainlines is moving at less than 35 mph. Buses may travel no more than 15 mph faster than mainline traffic with a 35 mph maximum allowed speed on the shoulder. The typical minimum shoulder width is 10 ft, with an 11.5-ft minimum at bridges, and a 12-ft mini- mum on new construction. Typically, buses travel through the entrance and exit ramps; where queues are long at ramps with metering, buses typically merge with traffic on the ramp and return to the shoulder after the ramp (Kuhn 2010). One segment of I-35W in Minneapolis has a unique combi- nation of strategies. Known as priced dynamic shoulder lanes (PDSL), the left shoulder is open during the peak periods; transit and carpools use the shoulder for free and MnPASS customers can use the shoulder for a fee. As shown in Figure 4, the left shoulder is open to traffic, with overhead sign gan- tries indicating its operational status. When the general- purpose lanes become congested, the shoulder is opened and the speed limit on the general-purpose lanes is reduced (Kuhn 2010). In dedicated shoulder-lane operations, either general- purpose or HOV-specific capacity has been added through the permanent conversion of shoulders. Most HOV appli- cations use the interior lane for HOV operations, whereas the exterior shoulder is used for general-purpose traffic so as to maintain the same number of general-purpose lanes that existed before implementation. A typical HOV appli- cation would convert a three-lane freeway with 12-ft lanes, 10-ft exterior shoulder, and 8-ft interior shoulder to 11-ft general-purpose lanes, 14-ft (including buffer striping) HOV lane, 5-ft exterior shoulder, and 2-ft interior shoulder (Kuhn 2010). Design Speed U.S. Customary Metric Desirable (70 mph) Reduced (50 mph) Desirable (110 km/h) Reduced (80 km/h) Alignment Stopping distance Horizontal curvature (radius) Maximum superelevation Rate of vertical curvature Crest, k Sag, k 730 ft 2,050–2,345 ft 0.04 ft/ft 247 181 425 ft 835–930 ft 0.06 ft/ft 84 96 220 m 560–635 m 0.04 m/m 74 55 130 m 250–280 m 0.06 m/m 26 30 Gradients Maximum (%) Minimum (%) 4.0 0.5 5.0 0.3 4.0 0.5 5.0 0.3 Clearance Vertical Lateral 16.5 ft 4 ft 14.5 ft 2 ft 5 m 1.2 m 4.4 m 0.6 m Lane Width Travel lanes 12 ft 11 ft 3.6 m 3.4 m Cross Slope Maximum Minimum 0.020 ft/ft 0.015 ft/ft 0.020 ft/ft 0.015 ft/ft 0.020 m/m 0.015 m/m 0.020 m/m 0.015 m/m Superelevation: Dependent on curve radii and design speed [0.10 ft/ft (0.10 m/m) maximum]. Source: Kuhn et al. (2005). TABLE 7 SUMMARy OF MANAGED LANES MAINLINE DESIGN CRITERIA

24 lAne width Potts et al. (2007b) investigated the relationship between lane width and safety for roadway segments and intersection approaches on urban and suburban arterials. Their research found no general indication that the use of lanes narrower than 12 ft (3.6 m) on urban and suburban arterials increased crash frequencies. Researchers stated that this finding suggested that geometric design policies can provide substantial flexibility for use of lane widths narrower than 12 ft (3.6 m). They added that inconsistent results suggested increased crash frequencies with narrower lanes in three specific design situations: • Lane widths of 10 ft (3.0 m) or less on four-lane undivided arterials. • Lane widths of 9 ft (2.7 m) or less on four-lane divided arterials. • Lane widths of 10 ft (3.0 m) or less on approaches to four-leg stop-controlled arterial intersections. The researchers recommended that “narrower lanes should be used cautiously in these three situations unless local experi- ence indicates otherwise.” Gross et al. (2009) studied a variety of crash data and roadway characteristics to determine the safety effectiveness of specific combinations of lane and shoulder width on rural, two-lane, undivided roads. In general, all else being equal, results were consistent with previous research efforts, show- ing crash reductions for wider paved widths, wider lanes, and wider shoulders. More specific to the research objec- tive, CMFs were provided for various lane–shoulder con- figurations. Individual state analyses did not indicate a clear preference for lane or shoulder width given a fixed paved width, but combined with findings from previous research, researchers described some potential trends: • For 26- to 32-ft total paved widths, 12-ft lanes provided the optimal safety benefit. The CMF ranged from 0.94 to 0.97, indicating a 3% to 6% crash reduction for 12-ft lanes compared with 10-ft lanes. • For 34-ft total paved width, 11-ft lanes provided the optimal safety benefit. The CMF for 11-ft lanes was 0.78 compared with the 10-ft baseline. • For 36-ft total paved width, 11- or 12-ft lanes provided the optimal safety benefit. The CMF was 0.95 for 11- and 12-ft lanes compared with the 10-ft baseline. These results applied, in general, to rural, two-lane roads with traffic volumes greater than 1,000 vehicles per day and posted speeds of 25 mph or greater. Although 12-ft lanes appeared to be the optimal design for 26- to 32-ft total paved widths, 11-ft lanes performed equally well or better than 12-ft lanes for 34- to 36-ft total paved widths. The Highway Safety Manual (AASHTO 2010) provides CMFs for lane width on two-lane highway segments, which are presented in Table 8. The base value for the lane width CMF is 12 ft. For lane widths with 0.5-ft increments that are not depicted specifically in Table 8, a CMF value can be interpolated because there is a linear transition between the various AADT effects. A corresponding chart is also pro- vided as a figure in the HCM. number of lanes Kononov et al. (2008) explored the relationship between safety and congestion on urban freeways by examining the shape of the safety performance functions (SPFs). SPFs are crash prediction models that relate traffic exposure, measured in AADT, to safety, measured in the number of accidents over a unit of time (e.g., accidents per mile per year). They found that to that point “the focus of most SPF modeling efforts had been on the statistical technique and the underlying prob- FIGURE 4 Open priced dynamic shoulder lane (Credit: Minnesota Department of Transportation). AADT (vehicles per day) Lane Width <400 400 to 2,000 >2,000 9 ft or less 1.05 1.05 + 2.81 × 10-4 (AADT – 400) 1.50 10 ft 1.02 1.02 + 1.75 × 10-4 (AADT – 400) 1.30 11 ft 1.01 1.01 + 2.50 × 10-5 (AADT – 400) 1.05 12 ft or more 1.00 1.00 1.00 Source: AASHTO (2010) . Note: The collision types for which this CMF is applicable include single-vehicle run-off-road and multiple-vehicle head-on, opposite-direction sideswipe, and same-direction sideswipe crashes. TABLE 8 CRASH MODIFICATION FACTORS FOR LANE WIDTH ON ROADWAy SEGMENTS

25 ability distribution, with only limited consideration given to the nature of the phenomenon itself.” Their relationship of safety to the degree of congestion suggested that safety deteri- orated with the degradation in the quality of service expressed through the level of service. Their assessment was that prac- titioners generally believed the additional capacity afforded by additional lanes was associated with more safety, but how much safety and for what time period were generally not con- sidered. Comparison of SPFs of multilane freeways suggested that adding lanes may initially result in a temporary safety improvement that disappears as congestion increases. They found that total as well as injury and fatal crash rates increased with AADT and that it was significantly safer to travel on urban freeways that operate at level-of-service (LOS)-C or better during the peak period than on more congested facili- ties. As AADT increased, the slope of SPF, described by its first derivative, became steeper, reflecting that crashes were increasing at a faster rate than would be expected from a free- way with fewer lanes. As the number of lanes increased, so did the opportunity for drivers to maneuver around slower traffic. Increased maneuverability tended to increase the aver- age speed of traffic, but at the same time it increased the speed differential and the number of crashes related to lane changes, such as sideswipes and rear-end crashes. road diet Huang et al. (2002) investigated the effects on crashes and injuries through conversion of an undivided four-lane road to three lanes and a TWLTL, also known as a “road diet.” They reviewed before-and-after crash data from 12 road diet sites and 25 comparison sites in California and Washington, and found that “the percent of road diet crashes occurring during the ‘after’ period was about 6% lower than that of the matched comparison sites.” However, a separate analy- sis in which a negative binomial model was used to control for possible differential changes in ADT, study period, and other factors indicated no significant treatment effect. Crash severity was virtually the same at road diets and compari- son sites, whereas there were some differences in crash type distributions between road diets and comparison sites, they found none between the “before” and “after” periods. They concluded that conversion to a road diet should be made on a case-by-case basis in which traffic flow, vehicle capacity, and safety are all considered. They also recommended that the effects of road diets be further evaluated under a variety of traffic and roadway conditions. Pawlovich et al. (2006) used a Bayesian approach to evalu- ate the effects of the “road diet” on crashes in Iowa. Their meth- odology incorporated both monthly crash data and estimated volumes for 30 sites—15 treatment and 15 comparison— for more than 23 years (1982 to 2004). Their results indicated a 25.2% reduction in crash frequency per mile and an 18.8% reduction in crash rate. The authors stated that their results from the Iowa study fit practitioner experience and agreed with another Iowa study that used a simple before-and-after approach on the same sites. NCHRP Report 617 (Harkey et al. 2008) presents the findings of a research project to develop CMFs for traffic engineering and Intelligent Transportation System improve- ments. One such improvement was the “road diet.” Researchers estimated the change in total crashes owing to the conversion and use Empirical Bayes methodology to compare the results with previous studies. They reviewed geometric, traffic, and crash data for 45 treatment sites and 347 reference sites in Iowa, Washington, and California, and found significant effects on crashes. Their recommendations for CMFs are shown in Table 9. resurfacing The research team on NCHRP Project 3-56 (Harwood et al. 2003a) developed a process for allocating resources to maxi- mize the effectiveness of 3R projects in improving safety and traffic operations on the nonfreeway highway network. They developed a program called the Resurfacing Safety Resource Allocation Program (RSRAP) designed to allow highway agencies to maximize the cost-effectiveness of the funds spent on 3R projects by improving safety on nonfreeway facilities while maintaining the structural integrity and ride quality of the highway pavement. To do this, their process considered: • “A specific set of highway sections that are in need of resurfacing either at the present time or within the rela- tively near future; • A specific set of improvement alternatives for each can- didate site, including doing nothing, resurfacing only, and various combinations of safety improvements for the site; and • A limit on the funds available for improvements to the set of highway locations.” The RSRAP procedure considers other treatments in addi- tion to resurfacing, such as lane width changes, turning lane improvements, and shoulder widening. Among their find- ings, the research team concluded that: • Resource allocation methods provided an effective method for highway agencies to decide when safety improvements are to be made in conjunction with pave- ment resurfacing projects. • For a given set of sites, resource allocation methods provided an optimal mix of resurfacing treatments with and without accompanying safety improvements that provided greater benefits than any fixed strategy. • Resurfacing without accompanying geometric improve- ments may cause a small, short-term increase in acci- dents resulting from increased speeds; however, the evidence for this effect was conflicting. An optional feature in the RSRAP software allowed the user to

26 include this short-term effect if desired. The increase in accidents following resurfacing was assumed to occur only at sites with existing lane widths of less than 11 ft and existing shoulder widths of less than 6 ft. work Zone considerations Changes in lane width, particularly lane constrictions, are often used in conjunction with lane shifts, lane closures, and shoulder closures. The authors of NCHRP Report 581 (Mahoney et al. 2007) discussed some aspects of lane width for designers to consider in work zones on high-speed roadways, defined as those with free-flow speeds of 50 mph or more. They mention that it is common practice to reference “travel lane width” as the key lane constriction decision variable. However, operations in one travel lane can be influenced by operations in adjacent lanes. Additionally, adjoining travel lanes occasion- ally have different widths; therefore, it may be more appro- priate for design guidance to address traveled way width. For example, they suggested that a 10-ft travel lane adjacent to a 12-ft travel lane is generally more desirable than a 10-ft travel lane adjacent to a travel lane of the same width. Although their desirable traveled way width resulted in 12-ft travel lanes, 11-ft lane widths were common in work zones, and lanes nar- rower than 10 ft were generally not used for work zones on high-speed roads. They offered the information presented in Table 10 as an example framework to determine minimum traveled way width in a work zone on a high-speed roadway. shOulders width NCHRP Report 633 (Stamatiadis et al. 2009) presented rec- ommendations for CMFs for shoulder width and median width for four-lane roads with 12-ft lanes. The authors’ rec- ommended CMFs for average shoulder width are shown in Table 11. Recommendations for median width CMFs are provided in the section on medians elsewhere in this chapter. FHWA’s Highway Design Handbook for Older Driv- ers and Pedestrians (Staplin et al. 2002) recommends that “for horizontal curves on two-lane nonresidential facilities that have 3 degrees of curvature, the width of the lane plus the paved shoulder be at least 5.5 m (18 ft) throughout the length of the curve.” The Handbook’s authors cite previous research stating that “older drivers, as a result of age-related declines in motor ability, have been found to be deficient in coordinating the control movements involved in lanekeep- ing, maintaining speed, and handling curves.” Dumbaugh (2006) conducted an analysis of roadside safety in urban areas, looking specifically at three treatments: . . . widening paved shoulders, widening fixed-object offsets, and providing livable-street treatments. [His] model results indicated that of the three strategies, only the livable-streets variable was consistently associated with reductions in roadside and midblock crashes. Wider shoulders were found to increase roadside and TREATMENT: Convert Undivided Four-Lane Road to Three-Lane and TWLTL (Road Diet) CMF Level of Predictive Certainty: High METHODOLOGY: Empirical Bayes Before–After Crash Type Studied and Estimated Effect REFERENCE: NCHRP Project 17-25 research results State/Site Characteristics CrashType Number of Treated Sites CMF (std. error) STUDY SITES: • 15 urban locations in Iowa with a mean length of 1.02 miles, a minimum and maximum length of 0.24 and 1.72 miles. AADT after conversion ranged from 3,718 to 13,908. • 30 urban locations from Washington and California studied previously with a mean length of 0.84 miles, a minimum and maximum length of 0.08 and 2.54 miles. AADT after conversion ranged from 6,194 to 26,376. Iowa Predominately U.S. and state routes within small urban areas (average population of 17,000) Total Crashes 15 15 miles 0.53 (0.02) California/Washington Predominately corridors within suburban areas surrounding larger cities (average population of 269,000) Total Crashes 30 30 miles 0.81 (0.03) All Sites Total Crashes 45 45 miles 0.71 (0.02) FOOTNOTES: 1Huang et al. (2002). 2Pawlovich et al. (2006). COMMENTS: • The study conducted was a reanalysis of data from two prior studies.1,2 • The reanalysis of the Washington/California data indicated a 19% decrease in total crashes. The reanalysis of the Iowa data showed a reduction of 47% in total crashes. If the characteristics of the treated site can be defined on the basis of road and area type (as shown above), the CMFs of 0.53 and 0.81 should be used. Otherwise, it is recommended that the aggregate CMF of 0.71 be applied. Source: Harkey et al. (2008). TABLE 9 RECOMMENDED CRASH MODIFICATION FACTOR FOR ROAD DIET TREATMENT

27 midblock crashes, while unpaved fixed-object offsets had a mixed safety effect [of] decreasing roadside crashes but slightly [increas- ing] midblock crashes. To understand better the reasons for these findings, the study then examined roadside crash site locations for tree and utility pole crashes. [His conclusion was] that the major- ity (between 65% and 83%) [of crashes] did not involve random midblock encroachments, as currently assumed, but instead involved objects located behind both driveways and side streets along higher-speed urban arterials. [He stated that], collectively, these findings [suggested] that most urban roadside crashes were not the result of random error but were instead systematically encoded into the design of the roadway. The study concluded by distinguishing between random and systematic driver errors and by discussing strategies for eliminating systematic error while minimizing the consequences of random error. Lord and Bonneson (2007) examined the safety perfor- mance of rural frontage road segments. Their findings sug- gested that wider lane and shoulder widths are associated with a reduction in segment-related collisions. In addition, the data suggest that the presence of edge marking has a significant impact on the safety performance of rural two- way frontage roads. However, the magnitude of crash reduc- tion resulting from marking presence was significant and believed to overstate the true benefit of such markings. They developed a safety performance function and three CMFs from a statistical model that was estimated through data col- lected on rural frontage road segments. The variables they Metric U.S. Customary Traveled Way Width (m) Traveled Way Width (ft) Facility Type Undivided Highway Divided Highway Undivided Highway Divided Highway Lanes per Direction One Two One Two One Two One Two Tr av el ed W ay Ed ge C on di tio ns Constraint along neither traveled way edge 3.0 1 6.02,3 3.3 6.63 101 202,3 11 223 Constraint along one traveled way edge 3.3 1 6.32,3 3.6 6.93 111 212,3 12 233 Constraint along both traveled way edges 3.6 1 6.62,3 3.9 7.23 121 222,3 13 243 Notes: 1. Values apply only when all of the following conditions are met: low truck volumes, all curve radii equal or exceed 555 m (1,820 ft); and anticipated 85th-percentile speeds are less than or equal to 80 km/h (50 mph). If any of the three conditions is not met, add 0.3 m (1 ft) to the base value. 2. Values apply only to roadways carrying moderate truck volumes where all curve radii equal or exceed 555 m (1,820 ft). If either condition is not met, add 0.3 m (1 ft) to the base value. 3. Values shown apply to two-lane, one-way traveled ways. For constricted two-way traveled ways, consider separation of opposing directions using (1) additional traveled way width, (2) channelizing devices, or (3) a traffic barrier. To use this exhibit, first determine the traveled way edge conditions. “Constraint” refers to the presence of an imposing feature, such as a feature that results in “shying away” at the edge of the traveled way. Temporary barriers are a common constraint feature. Next, identify the type of facility (undivided or divided) approaching the work zone. Using this information and the number of travel lanes through the work zone, determine the base (i.e., unadjusted) value within the appropriate cell. Superscripted numerals indicate the note numbers that should be referenced to determine appropriate adjustments, if any, to the base value. For traveled ways with edge constraint, the distances indicated are measured to the face of the constraining features (i.e., the offset is included in the tabulated or adjusted dimension). Values lower that those obtained from this method may be appropriate for very low exposure (i.e., traffic volume, constricted lane segment length, and duration of operation). Source: Mahoney et al. (2007). TABLE 10 ExAMPLE FRAMEWORK FOR SELECTING ONE-WAy TRAVELED WAy WIDTHS Average Shoulder Width (ft) Category 0 3 4 5 6 7 8 Undivided 1.22 1.00 0.94 0.87 0.82 0.76 0.71 Divided 1.17 1.00 0.95 0.90 0.85 0.81 0.77 Source: Stamatiadis et al. (2009). Notes: 1CMFs are for all crashes and all severities. 2The average shoulder width for undivided highways is the average of the right shoulders; for divided, it is the average of left and right shoulder in the same direction. TABLE 11 RECOMMENDED CMFS FOR AVERAGE SHOULDER WIDTH

28 found to have significant correlation with crash frequency included lane width, paved shoulder width, and, for two-lane frontage roads, edge marking delineation. Their SPFs and CMFs did not consider crashes that would be attributed to the ramp-frontage road terminal or the frontage road–crossroad intersection. Moreover, they did not consider crashes on the main lanes that may indirectly be related to wrong-way travel down an exit ramp. The Highway Safety Manual (AASHTO 2010) provides CMFs for shoulder width and shoulder type, which are pre- sented in Table 12. The base value of shoulder width and type is a 6-ft paved shoulder. Operational and safety treatments Rumble Strips Multiple studies have examined the effects of both shoulder and centerline rumble strips (CLRS). Information and find- ings for both types are presented in this section. Although not all roadway departure collisions can be attributable to drowsy driving, research shows that a large percentage of them are. Morena (2003) distinguishes between run-off-road and a subset of drift-off-road collisions. Whereas run-off- road crashes can occur for many reasons (loss of control, swerving to avoid another vehicle or object, icy roadway conditions, etc.), drift-off-road crashes are solely attributed to drowsy or inattentive drivers. The FHWA Rumble Strip website estimates that 40% to 60% of single-vehicle crashes on rural freeways are actually drift-off-road crashes. In examin- ing Michigan roadway data, Morena arrived at a much lower percentage of 16%, in part because nearly half (48%) of the run-off-road collisions in that state occurred on snowy or icy roadways and an additional 9% occurred on wet roadways. Persaud et al. (2003) investigated installation of rumble strips along the centerlines of undivided rural two-lane roads to warn or alert distracted, fatigued, or speeding motorists whose vehicles were susceptible to crossing the centerlines and encroaching into opposing traffic lanes. They analyzed data for approximately 210 mi of treated roads in seven states using an Empirical Bayes before–after methodology. Overall, they found that crashes at treated sites were reduced 14% and injury crashes were reduced by an estimated 15%. All frontal and opposing-direction sideswipe crashes were reduced by an estimated 21%, and those crashes involving injuries were reduced by an estimated 25%. All of the reduc- tions were determined to be statistically significant. Among the improvements investigated for CMFs in NCHRP 617 (Harkey et al. 2008) were shoulder and CLRS. The recommendations from that report are shown in Tables 13 and 14. NCHRP Synthesis 339 (Russell and Rys 2005) summa- rized the state of the practice on CLRS, examining design practices, installation, configuration, dimensions, and vis- ibility. The synthesis addressed the need for guidance on warrants, benefits, successful practices, and concerns (e.g., external noise and the reduced visibility of centerline striping material). The report also discussed pavement deterioration, ice buildup in the grooves, adverse impact on emergency vehicles, and the effect of CLRS on bicyclists. Particular attention was paid to available before-and-after installa- tion crash data to document the safety aspects of CLRS and the availability of policies, guidelines, warrants, and costs regarding their use and design. The authors did not find reli- able evidence of negative effects of CLRS, but they deter- mined that adequate data were not yet available to make definitive conclusions for a number of the issues listed. They noted that there was no standard nationwide design of CLRS and no conclusive studies had been conducted on mainte- nance issues. They did conclude that there was a definite pos- sibility that CLRS milled over the centerline could increase or accelerate deterioration of the typical centerline pavement joint and they recommended that, at a minimum, CLRS be installed only in good pavement. In 2006, the Washington State DOT (WSDOT) imple- mented policy for installing CLRS on undivided highways and invested in funding strategies for those installations. WSDOT subsequently conducted a study (Olson et al. 2011) to evaluate the effectiveness of CLRS under a variety of traf- fic and geometric conditions, in an effort to develop better guidance on when to use rumble strips to address various collision types. They determined that cross-centerline col- lisions have been reduced by 44.6% for all injury severi- AADT (vehicles per day) Shoulder Width <400 400 to 2,000 >2,000 0 ft 1.10 1.10 + 2.50 × 10-4 (AADT – 400) 1.50 2 ft 1.07 1.07 + 1.43 × 10-4 (AADT – 400) 1.30 4 ft 1.02 1.02 + 8.125 × 10-5 (AADT – 400) 1.15 6 ft 1.00 1.00 1.00 8 ft or more 0.98 0.98 + 6.875 × 10-5 (AADT – 400) 0.87 Source: AASHTO (2010). Note: The collision types for which this CMF is applicable include single-vehicle run-off-road and multiple-vehicle head-on, opposite-direction sideswipe, and same-direction sideswipe crashes. TABLE 12 CRASH MODIFICATION FACTORS FOR LANE WIDTH ON ROADWAy SEGMENTS

29 TREATMENT: Add Shoulder Rumble Strips CMF Level of Predictive Certainty: Medium-High METHODOLOGY: Before-After with Comparison Sites CRASH TYPE STUDIED AND ESTIMATED EFFECTS REFERENCE: Griffith (1999) All Freeways (Rural and Urban) Number ofImproved Sites CMF (std. error) STUDY SITES: • Included 55 treatment sites and 55 matched comparison sites from rural and urban freeways in Illinois. • The treatment sites covered 196 miles of rural freeway and 67 miles of urban freeway. • The treatment sites were not selected on the basis of crash history; thus, there was no selection bias. All Single-Vehicle Run-Off-Road Crashes 55 0.82 (0.07) Injury Single-Vehicle Run-Off-Road Crashes 0.87 (0.12) Rural Freeways All Single-Vehicle Run-Off-Road Crashes 29 0.79 (0.10) Injury Single-Vehicle Run-Off-Road Crashes 0.93 (0.16) COMMENTS: • Results for all freeways based on yoked comparison analysis; results for rural freeways based on comparison group method using 29 of the treatment sites. Results could not be developed for urban sites separately. • An analysis of multi-vehicle accidents showed the rumble strips to have no effect on such accidents. • The CMF is not applicable to other road classes (two-lane or multilane). Source: Harkey et al. (2008). TABLE 13 RECOMMENDED CRASH MODIFICATION FACTOR FOR SHOULDER RUMBLE STRIPS TREATMENT: Add Centerline Rumble Strips CMF Level of Predictive Certainty: Medium-High METHODOLOGY: Empirical Bayes Before-After CRASH TYPE STUDIED AND ESTIMATED EFFECTS REFERENCE: Persaud et al. (2003) Crash Type (All Severities) Number ofImproved Sites CMF (std. error) STUDY SITES: • Crash and traffic volume data were collected for 98 treatment sites, consisting of 210 miles, where centerline rumble strips had been installed on rural two-lane roads in the states of California, Colorado, Delaware, Maryland, Minnesota, Oregon, and Washington. • The average length of the treatment sites was 2 miles, and the traffic volumes ranged from 5,000 to 22,000 vpd. • The reference group of sites was developed from HSIS data for the states of California, Washington, and Minnesota.1 Additional data were acquired from Colorado for SPF calibration for the Colorado sites. All Crashes 98 0.86 (0.05) Frontal/Opposing-Direction Sideswipe Crashes 0.79 (0.12) Crash Type (Injury Crashes) All Crashes 98 0.85 (0.08) Frontal/Opposing-Direction Sideswipe Crashes 0.75 (0.15) COMMENTS: • The authors note that the results cover a wide range of geometric conditions, including curved and tangent sections and sections with and without grades. • The results include all rumble strip designs (milled-in, rolled-in, formed, and raised thermo-plastic) and placements (continuous versus intermittent) that were present. • The CMF is not applicable to other road classes (multilane). Source: Harkey et al. (2008). 1The Highway Safety Information System (HSIS) is a multistate safety database that contains crash, roadway inventory, and traffic volume data for a select group of states and is sponsored by the FHWA. TABLE 14 RECOMMENDED CRASH MODIFICATION FACTOR FOR CENTERLINE RUMBLE STRIPS

30 ties and by 48.6% for fatal and serious injury crashes. They also found that crashes involving asleep or fatigued drivers were reduced by 75.3% (72.6% for fatal and serious injury crashes) where CLRS were installed. Their data showed that on a horizontal curve the rate of fatal and serious injury crashes was almost twice as high for those lane departures to the outside of a curve than to the inside of the curve, but that CLRS were equally effective countermeasures for crashes in both directions, with reductions of about 35%. The research- ers recommended that WSDOT’s current guidance continue to be implemented to reduce cross-centerline collisions. The researchers also recommended that investment priority be given to locations with AADT less than 8,000, combined lane/shoulder width of 12 to 17 ft, and posted speed of 45 to 55 mph. With consideration of available funding, investment priorities, and site-specific conditions it was the research team’s opinion that the installation of CLRS be pursued for all highways that comply with design guidance. Torbic et al. (2009) conducted NCHRP Project 17-32, the objectives of which were to investigate the safety effective- ness and optimal placement and dimensions of shoulder and CLRS. NCHRP Report 641, which documents the project’s activities, “provides guidance for the design and applica- tion of shoulder and centerline rumble strips as an effective crash reduction measure, while minimizing adverse effects for motorcyclists, bicyclists, and nearby residents.” Using the results of previous studies and the research conducted under this project, “researchers developed” safety effectiveness esti- mates for shoulder rumble strips on rural freeways and rural two-lane roads and for CLRS on rural and urban two-lane roads. Their estimates with associated standard errors (SE) were as follows: • Urban/Rural Freeways—Rolled shoulder rumble strips: – 18% reduction in single-vehicle run-off-road (SVROR) crashes (SE = 7) – 13% reduction in SVROR fatal and injury (FI) crashes (SE = 12). • Rural Freeways—Shoulder rumble strips: – 11% reduction in SVROR crashes (SE = 6) – 16% reduction in SVROR FI crashes (SE = 8). • Rural Two-Lane Roads—Shoulder rumble strips: – 15% reduction in SVROR crashes (SE = 7) – 29% reduction in SVROR FI crashes (SE = 9). • Urban Two-Lane Roads—CLRS: – 40% reduction in total target (head-on and opposite- direction sideswipe) crashes (SE = 17) – 64% reduction in FI target crashes (SE = 27). • Rural Two-Lane Roads—CLRS: – 9% reduction in total crashes (SE = 2) – 12% reduction in FI crashes (SE = 3) – 30% reduction in total target crashes (SE = 5) – 44% reduction in FI target crashes (SE = 6). The NCHRP 17-32 research team added that shoulder rumble strips should be placed as close to the edgeline as possible to maximize safety benefits. They also stated that the safety benefits of CLRS for roadways on horizontal curves and on tangent sections are for practical purposes the same. With regard to rumble strip design, researchers concluded that shoulder rumble strip patterns for freeways and other roadways where bicyclists are not expected be designed to produce sound level differences between 10 to 15 dBA in the passenger compartment; for other roadways, the recom- mended sound level difference was 6 to 12 dBA. Similarly, they recommended that CLRS patterns be designed to pro- duce sound level differences in the range of 10 to 15 dBA in the passenger compartment, except near residential or urban areas where consideration would be given to design- ing CLRS to produce sound level differences in the range of 6 to 12 dBA in the passenger compartment. Treatments for Edge of Roadway Although the use of curbs is discouraged on high-speed road- ways because of their potential for “tripping” a skidding vehi- cle into a rollover condition, “they are often required because of restricted right-of-way, drainage issues, access control, and other curb functions.” Highway agencies have typically tried to reduce problems caused by curbs by offsetting the curb from the travel way as far as possible, using different curb shapes and using a barrier in combination with the curb. Plaxico et al. (2005) undertook research to develop design guidelines for using curbs and curb–barrier combinations on roadways with operating speeds greater than 60 km/h (37.3 mph). The research team reviewed published literature and conducted computer simulation methods to gain infor- mation on the nature of typical designs and crashes of curb systems. Results from computer simulations were used to determine which type of curbs were safe to use on higher- speed roadways and the proper placement of barrier with respect to the curb. They also conducted full-scale crash tests to validate the computer simulations. The results of the study were then synthesized to develop guidelines for the use of curbs and curb–barrier systems. The researchers’ recommen- dations included the following: • Any combination of a sloping-faced curb that is 150 mm (6 in.) or shorter and a strong-post guardrail can be used where the curb is flush with the face of the guardrail up to an operating speed of 85 km/h (52.8 mph). • Guardrails installed behind curbs are not to be located closer than 2.5 m (8.2 ft) for any operating speed in excess of 60 km/h (37.3 mph). Upon striking the curb, the vehicle bumper may rise above the critical height of the guardrail for many road departure angles and speeds in this region, making vaulting the barrier likely. A lateral distance of at least 2.5 m (8.2 ft) is needed to allow the vehicle suspen- sion to return to its pre-departure state. Once the suspen- sion and bumper have returned to their normal position, impacts with the barrier would proceed successfully.

31 • For roadways with operating speeds of 70 km/h (43.5 mph) or less, guardrails may be used with sloping- face curbs no taller than 150 mm (6 in.) as long as the face of the guardrail is located at least 2.5 m (8.2 ft) behind the curb. • In cases where guardrails are installed behind curbs on roads with operating speeds between 71 and 85 km/h (44.1 and 52.8 mph), a lateral distance of at least 4 m (13.1 ft) is needed to allow the vehicle suspension to return to its pre-departure position. Once the suspen- sion and bumper have returned to their normal position, impacts with the barrier would proceed successfully. At these speeds, guardrails may be used with sloping- face curbs of 100 mm (4 in.) in height or less as long as the face of the guardrail is located at least 4 m (13.1 ft) behind the curb. • At operating speeds greater than 85 km/h (52.8 mph), guardrails are to only be used with 100-mm (4-in.) or shorter sloping-faced curbs, and they would be placed so that the curb is flush with the face of the guardrail. Operating speeds above 90 km/h (55.9 mph) require that the sloping face of the curb must be 1:3 or flatter and must be no more than 100 mm (4 in.) in height. • Curbs are to only be used on higher-speed roadways when concerns about drainage make them essential to the proper maintenance of the highway. NCHRP Report 600C (Campbell et al. 2010) discusses the potential safety ramifications of shoulder edge drop-offs, which typically arise from tire rutting erosion, excessive wear, or resurfacing. Guidelines for treating these locations are offered for purposes of design practices. The report authors cited a previous study by Graham and Glennon (1984), which stated that vertical or near-vertical shoulder drop-off heights in work zones that exceeded the indicated values in Table 15 war- rant consideration for drop-off treatment or traffic control. The original source table also contained drop-off height thresholds that were greater than 3 in.; however, these were changed in the NCHRP report to reflect a more conservative assessment of other related driver performance data on driver encounters with drop-offs of various heights (Hallmark et al. 2006). One potential treatment is a wedge-shaped application of asphalt; when placed between the roadway and the shoulder, the material can help drivers recover from the shoulder to the driving surface. NCHRP Report 600C advises that the asphalt material needs to be compacted to increase strength; otherwise the material will break apart over time owing to forces of overrunning vehicles and runoff water. A specific application of this treatment, called “Safety Edge” (shown in Figure 5), is being developed by FHWA, as discussed by Hallmark et al. (2006). An evaluation of this treatment by Graham et al. (2010) indicated small but positive results in crash reduction at 56 of 81 treated sites. Their results indi- cated that for all two-lane highway study sites in two states, the best estimate of the treatment’s effectiveness was a reduc- tion in total crashes of approximately 5.7%. The results were not statistically significant, but they were generally positive. work Zone considerations There are a number of ways in which shoulders may be used under work zone traffic control conditions. The authors of NCHRP Report 581 (Mahoney et al. 2007) discuss some con- siderations for designers in the use of shoulders in work zones on high-speed roadways. They state that adoption of a work zone design speed may be appropriate for the evaluation of superelevation and sight distance. Because the shoulders will be part of a permanent high-speed roadway, no horizontal or vertical alignment decisions are generally needed. Temporary work zone features can affect sight distance, and it was rec- ommended that the design be developed and evaluated from that perspective. If the shoulder being used to carry traffic is on a horizontal curve, the magnitude and direction of its cross slope would be compared with the superelevation require- ment, and the agency’s typical work zone policy for super- elevation would be applied. They added that the designer is to also consider the adequacy of the shoulder in terms of structure (ability to carry the vehicle loads) and surface con- ditions (friction and smoothness), with particular attention to the presence and placement of shoulder rumble strips. mediAns NCHRP Report 633 (Stamatiadis et al. 2009) presented rec- ommendations for CMFs for shoulder and median width for four-lane roads with 12-ft lanes. The authors made the Speed (mph) Drop-Off Height (inches) for a Lane Width of 12 ft 11 ft 10 ft 9 ft 30 3 3 3 2 35 3 3 2 1 40 3 2 1 1 45 2 1 1 1 >50 1 1 1 1 Adapted from Graham and Glennon (1984). TABLE 15 VERTICAL DROP-OFF HEIGHT WARRANTING TRAFFIC CONTROL FOR VARIOUS LANE WIDTHS Existing Unpaved Shoulder Existing Pavement Asphalt Overlay For a 30˚ Safety Edge, = 30˚ FIGURE 5 Illustration of FHWA’s Safety Edge Treatment (Hallmark et al. 2006).

32 assumption that median width had no effect on single-vehicle crashes, so their recommended CMFs for average median width, shown in Table 16, are for multi-vehicle crashes. Rec- ommendations for shoulder width CMFs are provided in the section on shoulder width elsewhere in this chapter. Tarko et al. (2007) investigated the impact of median designs on crash frequency. They analyzed data collected in eight participating states using negative binomial regres- sion and before-and-after studies, and they examined crash severity using a logit model. The results of their analyses quantified the separate effects of changes in median geom- etry for single-vehicle, multiple-vehicle same-direction, and multiple-vehicle opposite-direction crashes. They concluded that results were significantly different for the various classes of crash types, indicating that reducing the median width without adding barriers (even if the remaining median width is still reasonably wide) increases the severity of crashes, particularly opposite-direction crashes. Further, they found that reducing the median width and installing concrete bar- riers eliminated opposite-direction crashes but doubled the frequency of single-vehicle crashes, increased crash severity, and tended to lessen the frequency of same-direction crashes. rOAdside horizontal clearance The developers of the Roadside Safety Analysis Program included encroachment frequency curves (shown in Fig- ure 6) and adjustment factors to increase encroachment rates on horizontal curves and vertical grades (shown in Figure 7) (Mak and Sicking 2003). The developers found three pre- vious studies on encroachment data: Hutchinson and Ken- nedy (1966), Cooper (1980), and Calcote et al. (1985). The Hutchinson and Kennedy study involved observation of wheel tracks on medians of rural Illinois Interstate highways in the mid-1960s. Cooper conducted a similar encroachment study in Canada in the late 1970s. This research involved weekly observations of wheel tracks on grass-covered road- sides of rural highways of various functional classes. The data collection periods were during summer months on highways with speed limits between 80 and 100 km/h. Calcote et al. attempted to overcome the major problems with both the Cooper and the Hutchinson and Kennedy studies, but they “still did not offer an effective method to distinguish between controlled and uncontrolled encroachments.” An overwhelming majority of the encroachments recorded involved vehicles “moving slowly off the roadway for some distance and then returning into the traffic stream without any sudden changes in trajectory,” which could be caused by “a fatigued or distracted driver drifting off the roadway, or a con- trolled driver responding to roadway or traffic conditions.” Average Median Width (ft) Category 10 20 30 40 50 60 70 80 Multi-Vehicle 1.00 0.91 0.83 0.75 0.68 0.62 0.57 0.51 Source: Stamatiadis et al. 2009. TABLE 16 RECOMMENDED CMFs FOR AVERAGE MEDIAN WIDTH ON DIVIDED ROADWAyS FIGURE 6 Encroachment rates used in Roadside Safety Analysis Program (Mak and Sicking 2003). FIGURE 7 Encroachment frequency adjustment factors for curvature (Wright and Robertson 1976).

33 Roadside Safety Analysis Program developers selected the Cooper encroachment data for use in their encroachment rate–traffic volume relationships, because the Cooper data are more recent, constitute a larger sample size, and are believed to be of better quality than the Hutchinson and Kennedy data. The developers then incorporated adjustment factors based on previous studies (Wright and Robertson 1976; Perchonok et al. 1978) that compared roadway characteristics with fatal single- vehicle run-off-road crashes, with the underlying assump- tion that differences in roadway characteristics between the fatal crash sites and the comparison sites are correlated with the occurrence of these fatal crashes. They cited studies that showed that crash rates on horizontal curves and vertical grades were significantly higher than those on tangent sections, and they assumed by extension that encroachment rates would also be similarly affected by horizontal curves and vertical grades. The developers also stated their belief that the adjust- ment factors overstated the effects of curvature on encroach- ment rates, but represented the best information available at the time of the study. NCHRP Project 16-04 (Dixon et al. 2008) was initiated to develop design guidelines for safe and aesthetically pleasing roadside treatments in urban areas and a toolbox of effective roadside treatments to balance the safety and mobility needs of pedestrians, bicyclists, and motorists, and accommodate community values. In fulfilling the first of those objectives, researchers recommended the following guidelines for road- side treatments: • Where possible at curb locations, provide a lateral off- set to rigid objects of at least 6 ft from the face of the curb and maintain a minimum lateral offset of 4 ft. • At lane merge locations, do not place rigid objects in an area that is 10 ft longitudinally from the taper point. This will result in a 20-ft object-free length at the taper point. The lateral offset for this 20-ft section should be consistent with the lane width, typically 12 ft. • Although many auxiliary lanes, such as bus lanes or bicycle lanes, have low volumes and may be included as part of a clear zone in the urban environment, higher- speed auxiliary lane locations, such as extended length right-turn lanes, are common locations for run-off-road crashes. A lateral offset of 6 ft from the curb face to rigid objects is preferred, and a 4-ft minimum lateral offset should be maintained. • At locations where a sidewalk buffer is present, such as in Figure 8, rigid objects are not to be located in a buffer area with a width of 3 ft or less. For buffer widths greater than 3 ft, lateral offsets from the curb face to rigid objects are to be maintained with a minimum offset of 4 ft. At these wider buffer locations, other frangible objects can be strategically located to help shield any rigid objects. • Rigid objects should not be located in the proximity of driveways, and care is to be taken to avoid placing rigid objects on the immediate far side of a driveway. In addi- tion, objects are not to be located within the required sight triangle for a driveway. safety treatments Volume 3 of NCHRP Report 500 (Neuman et al. 2003a) discusses modifying the clear zone in proximity to trees to reduce crashes. This strategy involves any change to the sideslope or roadside clear zone designed to reduce the likelihood of tree crashes by increasing the chances that a [run-off-road] (ROR) vehicle can successfully recover without striking a tree. While both tree removal and shielding strategies modify the roadside, this strat- egy may be implemented in a variety of ways, such as flattening or grading sideslopes, regrading ditch sections, adding shoulder improvements, or providing protective plantings on the roadside. [The authors state that] the cost to modify the roadside is often considerably higher than tree removal and guardrail installation; however, applying this strategy on specific curves or short tangent sections of roadway may help manage the costs. The authors of NCHRP Report 500 add that this strategy has been proven to reduce the severity of ROR crashes and rollover crashes. Although they identified no specific studies that related to only trees, much work has been completed on the benefits of improving the geometry of the roadside to allow vehicles to recover when they encroach on the roadside. summAry Of Key findings This section summarizes key findings from the research noted in this chapter. This is an annotated summary; conclu- sions and recommendations are those of the authors of the references cited. Allocation of traveled way width • The benefits of 2+1 roads in Europe validated a recom- mendation for their use in the United States to serve as an intermediate treatment between an alignment with periodic passing lanes and a full four-lane alignment. FIGURE 8 Example of buffer between sidewalk and street (Credit: Marcus Brewer, Texas Transportation Institute).

34 Such 2+1 roads are most suitable for level and rolling terrain, with installations to be considered on roadways with traffic flow rates of no more than 1,200 veh/hr in a single direction. The use of a cable barrier as a separa- tor is discouraged, and major intersections should be located in the buffer or transition areas between oppos- ing passing lanes, with the center lane used as a turning lane (Potts and Harwood 2003). • Passing activity on 2+1 roads was greatest at the beginning of the segments and the greatest benefits of decreased pla- tooning and increased safety occurred within the first 0.9 mi of a passing lane segment (Gattis et al. 2006). • Most passing on Super 2 passing lanes occurs within the first mile of a passing lane, so additional length may be less useful than additional lanes in a Super 2 corridor, particularly at lower volumes. Designers should avoid intersections with state highways and high-volume county roads within passing lanes, consider terrain and right-of-way in determining alignment and placement of passing lanes, avoid the termination of passing lanes on uphill grades, and discourage passing lane lengths longer than 4 mi (Brewer et al. 2011). • TWLTLs could be used as a strategy to reduce head-on collisions on two-lane roads (Neuman et al. 2003b). lane width • Researchers investigating the relationship between lane width and safety on urban and suburban arterials found no general indication that the use of lanes narrower than 12 ft on urban and suburban arterials increased crash frequencies. They suggested that geometric design policies should provide substantial flexibility for use of lane widths narrower than 12 ft (Potts et al. 2007b). • Lane widths of 11 or 12 ft provide optimal safety ben- efit for common values of total paved width on rural two-lane roads. Although 12-ft lanes appear to be the optimal design for 26- to 32-ft total paved widths, 11-ft lanes perform equally well or better than 12-ft lanes for 34- to 36-ft total paved widths (Gross et al. 2009). road diet • Road diet crashes occurring during the period after instal- lation were about 6% lower than that of matched compar- ison sites. However, controlling for possible differential changes in ADT, study period, and other factors indicated no significant effect of the treatment. Crash severity was virtually the same at road diets and comparison sites. Conversion to a road diet should be made on a case-by- case basis in which traffic flow, vehicle capacity, and safety are all considered (Huang et al. 2002). • The effects of the road diet on crashes in Iowa, account- ing for monthly crash data and estimated volumes for treatment and comparison sites, resulted in a 25.2% reduction in crash frequency per mile and an 18.8% reduction in crash rate (Pawlovich et al. 2006). shoulder width • For horizontal curves on two-lane nonresidential facili- ties that have 3 degrees of curvature, the width of the lane plus the paved shoulder should be at least 5.5 m (18 ft) throughout the length of the curve (Staplin et al. 2002). • Wider lane and shoulder widths are associated with a reduction in segment-related collisions on rural front- age road segments (Lord and Bonneson 2007). rumble strips • Crashes at approximately 210 mi of undivided rural two-lane roads treated with CLRS were reduced by 14% and injury crashes by an estimated 15%. All fron- tal and opposing-direction sideswipe crashes were reduced by an estimated 21%, and those crashes involv- ing injuries by an estimated 25%. All of the reductions were determined to be statistically significant (Persaud et al. 2003). • Crash data on roads treated with CLRS or shoulder rumble strips revealed noticeable crash reductions on all classes of roads (rural and urban two-lane roads and freeways). Shoulder rumble strips should be placed as close to the edgeline as possible to maximize safety benefits. The safety benefits of CLRS for roadways on horizontal curves and on tangent sections are for practi- cal purposes the same (Torbic et al. 2009). shoulder edge treatments • Plaxico et al. (2005) made the following recommenda- tions on design guidelines for using curbs on roadways with operating speeds greater than 60 km/h (37.3 mph): – Any combination of a sloping-faced curb that is 150 mm (6 in.) or shorter and a strong-post guardrail can be used where the curb is flush with the face of the guardrail up to an operating speed of 85 km/h. – Guardrails installed behind curbs are not to be located closer than 2.5 m (8.2 ft) for any operating speed in excess of 60 km/h (37.3 mph). – For roadways with operating speeds of 70 km/h (43.5 mph) or less, guardrails may be used with sloping-face curbs no taller than 150 mm (6 in.) as long as the face of the guardrail is located at least 2.5 m (8.2 ft) behind the curb. – Where guardrails are installed behind curbs on roads with operating speeds between 71 and 85 km/h (44.1 and 52.8 mph), a lateral distance of at least 4 m (13.1 ft) is needed to allow the vehicle suspension to return to its pre-departure position. – At operating speeds greater than 85 km/h (52.8 mph), guardrails are only to be used with 100-mm (4-in.) or shorter sloping-faced curbs, and be placed so that the curb is flush with the face of the guardrail. Operating

35 speeds above 90 km/h (55.9 mph) require that the sloping face of the curb must be 1:3 or flatter and must be no more than 100 mm (4 in.) in height. • The “Safety Edge” treatment produced small but posi- tive results in crash reduction at 56 of 81 treated sites. For all two-lane highway study sites in two states, the best estimate of the treatment’s effectiveness was a reduction in total crashes of approximately 5.7%. The results were not statistically significant, but they were generally positive (Hallmark et al. 2006). roadside • Where possible at curb locations, provide a lateral offset to rigid objects of at least 6 ft from the face of the curb and maintain a minimum lateral offset of 4 ft (Dixon et al. 2008). • At lane merge locations, do not place rigid objects in an area that is 10 ft longitudinally from the taper point. The lateral offset for this 20-ft section is to be consistent with the lane width, typically 12 ft (Dixon et al. 2008). • A lateral offset of 6 ft from the curb face to rigid objects is preferred for higher-speed auxiliary lane locations, such as extended length right-turn lanes, and a 4-ft minimum lateral offset is to be maintained (Dixon et al. 2008). • At locations where a sidewalk buffer is present, rigid objects are not to be located in a buffer area with a width of 3 ft or less. For buffer widths greater than 3 ft, lateral offsets from the curb face to rigid objects must be maintained with a minimum offset of 4 ft. At these wider buffer locations, other frangible objects can be strategically located to help shield any rigid objects (Dixon et al. 2008). • Rigid objects are not to be located in the proximity of driveways, and care should be taken to avoid placing rigid objects on the immediate far side of a driveway. In addition, objects should not be located within the required sight triangle for a driveway (Dixon et al. 2008).

Next: Chapter Five - Intersections »
Recent Roadway Geometric Design Research for Improved Safety and Operations Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s National Cooperative Highway Research Program (NCHRP) Synthesis 432: Recent Roadway Geometric Design Research for Improved Safety and Operations reviews and summarizes roadway geometric design literature completed and published from 2001 through early 2011, particularly research that identified impacts on safety and operations.

The report is structured to correspond to chapters in the American Association of State Highway and Transportation Officials’ A Policy on Geometric Design of Highways and Streets, more commonly referred to as the Green Book.

NCHRP Synthesis 432 is an update of NCHRP Synthesis 299 on the same topic published in 2001.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!