Click for next page ( 26


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

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

OCR for page 25
27 Table 10 Example Framework for Selecting One-Way Traveled Way Widths Metric U.S. Customary Traveled Way Width (m) Traveled Way Width (ft) Undivided Divided Undivided Divided Facility Type Highway Highway Highway Highway Lanes per Direction One Two One Two One Two One Two Constraint along neither 3.01 6.02,3 3.3 6.63 101 202,3 11 223 traveled way edge Edge Conditions Traveled Way Constraint along one 3.31 6.32,3 3.6 6.93 111 212,3 12 233 traveled way edge Constraint along both 3.61 6.62,3 3.9 7.23 121 222,3 13 243 traveled way edges 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). midblock crashes, while unpaved fixed-object offsets had a mixed Lord and Bonneson (2007) examined the safety perfor- safety effect [of] decreasing roadside crashes but slightly [increas- mance of rural frontage road segments. Their findings sug- ing] midblock crashes. To understand better the reasons for these findings, the study then examined roadside crash site locations for gested that wider lane and shoulder widths are associated tree and utility pole crashes. [His conclusion was] that the major- with a reduction in segment-related collisions. In addition, ity (between 65% and 83%) [of crashes] did not involve random the data suggest that the presence of edge marking has a midblock encroachments, as currently assumed, but instead significant impact on the safety performance of rural two- involved objects located behind both driveways and side streets along higher-speed urban arterials. [He stated that], collectively, way frontage roads. However, the magnitude of crash reduc- these findings [suggested] that most urban roadside crashes were tion resulting from marking presence was significant and not the result of random error but were instead systematically believed to overstate the true benefit of such markings. They encoded into the design of the roadway. The study concluded by distinguishing between random and systematic driver errors and developed a safety performance function and three CMFs by discussing strategies for eliminating systematic error while from a statistical model that was estimated through data col- minimizing the consequences of random error. lected on rural frontage road segments. The variables they Table 11 Recommended CMFS for Average Shoulder Width 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: 1 CMFs are for all crashes and all severities. 2 The 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.

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

OCR for page 25
29 Table 13 Recommended Crash Modification Factor for Shoulder Rumble Strips TREATMENT: Add Shoulder Rumble CMF Level of Predictive Certainty: Medium-High Strips METHODOLOGY: CRASH TYPE STUDIED AND ESTIMATED EFFECTS Before-After with Comparison Sites REFERENCE: CMF Griffith (1999) Number of All Freeways (Rural and Urban) (std. Improved Sites error) STUDY SITES: All Single-Vehicle Run-Off-Road 0.82 Crashes (0.07) Included 55 treatment sites and 55 55 matched comparison sites from rural and Injury Single-Vehicle Run-Off-Road 0.87 urban freeways in Illinois. Crashes (0.12) The treatment sites covered 196 miles Rural Freeways of rural freeway and 67 miles of urban freeway. All Single-Vehicle Run-Off-Road 0.79 Crashes (0.10) The treatment sites were not selected 29 on the basis of crash history; thus, there Injury Single-Vehicle Run-Off-Road 0.93 was no selection bias. Crashes (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 14 Recommended Crash Modification Factor for Centerline Rumble Strips TREATMENT: Add Centerline Rumble CMF Level of Predictive Certainty: Medium-High Strips METHODOLOGY: CRASH TYPE STUDIED AND ESTIMATED EFFECTS Empirical Bayes Before-After REFERENCE: CMF Persaud et al. (2003) Number of Crash Type (All Severities) (std. Improved Sites error) STUDY SITES: 0.86 All Crashes (0.05) Crash and traffic volume data were 98 collected for 98 treatment sites, Frontal/Opposing-Direction Sideswipe 0.79 consisting of 210 miles, where centerline Crashes (0.12) rumble strips had been installed on rural two-lane roads in the states of California, Crash Type (Injury Crashes) Colorado, Delaware, Maryland, Minnesota, Oregon, and Washington. 0.85 All Crashes (0.08) The average length of the treatment 98 sites was 2 miles, and the traffic volumes Frontal/Opposing-Direction Sideswipe 0.75 ranged from 5,000 to 22,000 vpd. Crashes (0.15) COMMENTS: The reference group of sites was developed from HSIS data for the states The authors note that the results cover a wide range of geometric of California, Washington, and conditions, including curved and tangent sections and sections with and Minnesota.1 Additional data were without grades. acquired from Colorado for SPF The results include all rumble strip designs (milled-in, rolled-in, calibration for the Colorado sites. 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). 1 The 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.

OCR for page 25
30 ties and by 48.6% for fatal and serious injury crashes. They possible to maximize safety benefits. They also stated that the also found that crashes involving asleep or fatigued drivers safety benefits of CLRS for roadways on horizontal curves were reduced by 75.3% (72.6% for fatal and serious injury and on tangent sections are for practical purposes the same. crashes) where CLRS were installed. Their data showed With regard to rumble strip design, researchers concluded that on a horizontal curve the rate of fatal and serious injury that shoulder rumble strip patterns for freeways and other crashes was almost twice as high for those lane departures to roadways where bicyclists are not expected be designed to the outside of a curve than to the inside of the curve, but that produce sound level differences between 10 to 15 dBA in CLRS were equally effective countermeasures for crashes in the passenger compartment; for other roadways, the recom- both directions, with reductions of about 35%. The research- mended sound level difference was 6 to 12 dBA. Similarly, ers recommended that WSDOT's current guidance continue they recommended that CLRS patterns be designed to pro- to be implemented to reduce cross-centerline collisions. The duce sound level differences in the range of 10 to 15 dBA researchers also recommended that investment priority be in the passenger compartment, except near residential or given to locations with AADT less than 8,000, combined urban areas where consideration would be given to design- lane/shoulder width of 12 to 17 ft, and posted speed of 45 to ing CLRS to produce sound level differences in the range of 55 mph. With consideration of available funding, investment 6 to 12 dBA in the passenger compartment. 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. Treatments for Edge of Roadway Torbic et al. (2009) conducted NCHRP Project 17-32, the Although the use of curbs is discouraged on high-speed road- objectives of which were to investigate the safety effective- ways because of their potential for "tripping" a skidding vehi- ness and optimal placement and dimensions of shoulder and cle into a rollover condition, "they are often required because CLRS. NCHRP Report 641, which documents the project's of restricted right-of-way, drainage issues, access control, and activities, "provides guidance for the design and applica- other curb functions." Highway agencies have typically tried tion of shoulder and centerline rumble strips as an effective to reduce problems caused by curbs by offsetting the curb crash reduction measure, while minimizing adverse effects from the travel way as far as possible, using different curb for motorcyclists, bicyclists, and nearby residents." Using the shapes and using a barrier in combination with the curb. results of previous studies and the research conducted under this project, "researchers developed" safety effectiveness esti- Plaxico et al. (2005) undertook research to develop design mates for shoulder rumble strips on rural freeways and rural guidelines for using curbs and curbbarrier combinations two-lane roads and for CLRS on rural and urban two-lane on roadways with operating speeds greater than 60 km/h roads. Their estimates with associated standard errors (SE) (37.3 mph). The research team reviewed published literature were as follows: and conducted computer simulation methods to gain infor- mation on the nature of typical designs and crashes of curb Urban/Rural Freeways--Rolled shoulder rumble strips: systems. Results from computer simulations were used to 18% reduction in single-vehicle run-off-road (SVROR) determine which type of curbs were safe to use on higher- crashes (SE = 7) speed roadways and the proper placement of barrier with 13% reduction in SVROR fatal and injury (FI) crashes respect to the curb. They also conducted full-scale crash tests (SE = 12). to validate the computer simulations. The results of the study Rural Freeways--Shoulder rumble strips: were then synthesized to develop guidelines for the use of 11% reduction in SVROR crashes (SE = 6) curbs and curbbarrier systems. The researchers' recommen- 16% reduction in SVROR FI crashes (SE = 8). dations included the following: Rural Two-Lane Roads--Shoulder rumble strips: 15% reduction in SVROR crashes (SE = 7) Any combination of a sloping-faced curb that is 150 mm 29% reduction in SVROR FI crashes (SE = 9). (6 in.) or shorter and a strong-post guardrail can be used Urban Two-Lane Roads--CLRS: where the curb is flush with the face of the guardrail up 40% reduction in total target (head-on and opposite- to an operating speed of 85 km/h (52.8 mph). direction sideswipe) crashes (SE = 17) Guardrails installed behind curbs are not to be located 64% reduction in FI target crashes (SE = 27). closer than 2.5 m (8.2 ft) for any operating speed in excess Rural Two-Lane Roads--CLRS: of 60 km/h (37.3 mph). Upon striking the curb, the vehicle 9% reduction in total crashes (SE = 2) bumper may rise above the critical height of the guardrail 12% reduction in FI crashes (SE = 3) for many road departure angles and speeds in this region, 30% reduction in total target crashes (SE = 5) making vaulting the barrier likely. A lateral distance of at 44% reduction in FI target crashes (SE = 6). 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- The NCHRP 17-32 research team added that shoulder sion and bumper have returned to their normal position, rumble strips should be placed as close to the edgeline as impacts with the barrier would proceed successfully.