National Academies Press: OpenBook

Safety of U-Turns at Unsignalized Median Openings (2004)

Chapter: Chapter 2 - Literature Review

« Previous: Chapter 1 - Introduction
Page 5
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 5
Page 6
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 6
Page 7
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 7
Page 8
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 8
Page 9
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 9
Page 10
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 10
Page 11
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 11
Page 12
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 12
Page 13
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 13
Page 14
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 14
Page 15
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 15
Page 16
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 16
Page 17
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 17
Page 18
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 18
Page 19
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 19
Page 20
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 20
Page 21
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 21
Page 22
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 22
Page 23
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 23
Page 24
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 24
Page 25
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 25
Page 26
Suggested Citation:"Chapter 2 - Literature Review." National Academies of Sciences, Engineering, and Medicine. 2004. Safety of U-Turns at Unsignalized Median Openings. Washington, DC: The National Academies Press. doi: 10.17226/13768.
×
Page 26

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.

5CHAPTER 2 LITERATURE REVIEW The objectives of the literature review are as follows: • To document current knowledge of the safety effect of installing nontraversable medians on multilane highways, • To identify key issues related to the increase of U-turns at unsignalized median openings, and • To document the safety effect of installing or removing median openings. The literature review includes the following issues related to the safety and operation of U-turns at unsignalized median openings: • Location of median openings, • Spacing of median openings, • Safety of median openings, • Median width, • Median opening length, • Safety effects of median treatments, • Safety effects of increasing U-turn maneuvers through use of nontraversable medians, • Left-turn lanes, • Median acceleration lanes, • Loons (i.e., paved aprons opposite median openings to assist larger vehicles in completing U-turn maneuvers), • Sight distance at median openings, • Indirect left-turn maneuvers, • Access management, • Spacing between access points, and • Effects of adjacent traffic signals. Literature related to each of these topics is summarized below. LOCATION OF MEDIAN OPENINGS The growing number of multilane highways with raised or depressed medians and without access control has created the need to provide median openings, or crossovers, at various locations along such facilities to permit vehicles to reach abut- ting property or reverse their direction of travel. Median open- ings, however, may also become points of increased conges- tion and accident exposure. Turbulence in traffic flow created by vehicles turning on or off high-speed roadways causes undesirable acceleration and deceleration maneuvers. There- fore, if traffic safety on multilane highways is to be preserved, the location of median openings must be given careful con- sideration. Some factors that influence median opening loca- tions include the following: • Spacing between median openings, • Stopping sight distance, • Intersection sight distance, • Operating speeds, • Length of turn lanes, • Right-turn conflict overlap, and • Size and type of traffic generator. A committee of the Institute of Transportation Engineers developed a list of factors to consider in locating median openings (1). These included the potential number of left turns into driveways, length of frontage along the street right- of-way line of the property proposed to be served, distance of proposed opening from adjacent intersections or other openings, length and width of the left-turn storage lane as functions of the estimated maximum number of vehicles to be in the lane during peak hours, and traffic control. The committee noted the need to consider circuitous routing and added intersection turns that may be caused by closing a median opening. Research for the Florida Department of Transportation (2) found that the overall reductions in the number of median openings along the roads studied in their analysis resulted in accident rate reductions, despite the increased through traffic flow and higher density of traffic flow per median opening. The Florida research, performed for four-lane and six-lane roadway sections, also found that the reduction in conflict points can improve traffic flow characteristics without increased risk of accidents at the remaining median openings. Policies recommended at the national level for geometric design of median openings are presented in the AASHTO A Policy on Geometric Design of Highways and Streets, com- monly known as the Green Book (3). The Green Book states that “median openings on divided highways with depressed or raised curbed medians should be carefully considered. Such openings should only be provided for street intersec- tions or for major developments.” Regarding the location of median openings, the Green Book recommends that median

openings designed to accommodate vehicles making U-turns only are needed on some divided highways in addition to openings provided for cross and left-turning movements. Separate U-turn median openings may be needed at the fol- lowing locations: • Locations beyond intersections to accommodate minor turning movements not otherwise provided in the inter- section or interchange area. The major intersection area is kept free for the important turning movements, in some cases obviating expensive ramps or additional structures. • Locations just ahead of an intersection to accommodate U-turn movements that would interfere with through and other turning movements at the intersection. Where a fairly wide median on the approach highway has few openings, U-turns are necessary for motorists to reach roadside areas. Advancing separate openings to accom- modate them outside the intersection proper will reduce interference. • Locations occurring in conjunction with minor cross- roads where traffic is not permitted to cross the major highway but instead is required to turn right, enter the through traffic stream, weave to the left, U-turn, and then return. On high-speed or high-volume highways, the dif- ficulty of weaving and the long lengths involved usu- ally make this design pattern undesirable, unless the volumes intercepted are light and the median is of ade- quate width. This condition may occur where a crossroad with high-volume traffic, a shopping area, or other traf- fic generator that needs a median opening nearby and additional median openings would not be practical. • Locations occurring where regularly spaced openings facilitate maintenance operations, policing, repair service of stalled vehicles, or other highway-related activities. Openings for this purpose may be needed on controlled- access highways and on divided highways through unde- veloped areas. • Locations occurring on highways without control of access where median openings at optimum spacing are provided to serve existing frontage developments and at the same time minimize pressure for future median openings. A preferred spacing at 400 to 800 m (0.25 to 0.50 mi) is suitable in most instances. Fixed spacing is not necessary, nor is it fitting in all cases, because of variations in terrain and local service needs. SPACING OF MEDIAN OPENINGS Increasing the spacing between median openings improves arterial flow and safety by reducing the number of conflicts and conflict points per mile, providing greater distance to anticipate and recover from turning maneuvers, and provid- ing opportunities for use of turn lanes (4). It is increasingly recognized that spacing standards for unsignalized access points should complement those for signalized access points. 6 The Green Book makes the following recommendations on the spacing of median openings (3): • Spacing between median openings should be adequate to allow for introduction of left-turn lanes. • Median openings should reflect street or block spacing and the access classification of the roadway. • Full median openings should be consistent with traffic signal spacing criteria. • Spacing of openings should be consistent with access management classifications of criteria. Research reported in NCHRP Report 348 (5) indicates that several states have set median opening spacing criteria that range from 100 to 800 m (330 to 2,640 ft). These criteria are mainly applicable in suburban and rural environments. The report also presents minimum desired spacing of unsignal- ized median openings at driveways as a function of speed. This spacing ranges from 113 m (370 ft) at 48 km/h (30 mph) to 278 m (910 ft) at 88 km/h (55 mph). In addition, the report suggests the following guidelines be considered for the spac- ing and design of median openings on divided roadways: • The spacing of median openings for signalized drive- ways should reflect traffic signal coordination require- ments and the storage space needed for left turns. • The spacing of median openings for unsignalized drive- ways should be based on a roadway’s function or access level and the environment in which the roadway is located (e.g., rural) and should be conducive to signalization. • Median openings for left-turn entrances should be spaced to allow sufficient storage for left-turning vehicles. • Median openings at driveways could be subject to clo- sure where volumes warrant signals, but signal spacing would be inappropriate. • Median openings should be set back far enough from nearby signalized intersections to avoid possible interfer- ence with intersection queues, and storage for left turns must be adequate. TRB Circular 456 (6) indicates that median openings gen- erally should relate to the street or block spacing. Thus, where cross-streets are placed at regular intervals, these intervals will influence median opening spacing. The Circular recom- mends that access points on both sides of the road should be aligned on undivided highways. Where this is not possible, sufficient left-turn storage should be provided by establish- ing a minimum offset distance. Driveways should be offset from median openings by the following: • At least 60 m (200 ft) when two low-volume traffic gen- erators are involved, • The greater of 60 m (200 ft) or the established median opening spacing interval when one major traffic gener- ator is involved, and

• At least two times the established median opening spacing interval when two major traffic generators are involved. NCHRP Report 375 (7) found that very few state high- way agencies have formal policies on the minimum spac- ing between median openings. Those agencies that do have criteria generally use a spacing between median openings in the range from 0.4 to 0.8 km (0.25 to 0.5 mi). The Florida DOT Median Handbook (8) identifies the fol- lowing factors that should be considered in determining the spacing of median openings: • Deceleration length, • Queue storage, • Turn radius, and • Perception/reaction distance. Based on consideration of all of these factors, Florida has identified a 330-m (1,070-ft) spacing between median open- ings as being a realistic minimum for urban arterials. SAFETY OF MEDIAN OPENINGS Much of the safety research of median openings has been conducted at intersections on divided highways. These studies generally document the safety performance of the intersection as a whole, but usually do not provide specific information on the safety performance of the median opening area by itself. Furthermore, safety research on midblock median openings (i.e., median openings not located at intersections or drive- ways) is limited and usually addresses a “system” or combi- nation of intersections and midblock median openings, such as for indirect left-turn treatments. However, even research on divided highway intersections that is broader in scope than just the median opening area can provide valuable informa- tion on the safety performance of median openings. Two published research studies have specifically addressed the safety performance of divided highway intersections, one in California and one in Ohio. A 1953 California study by McDonald (9) developed relationships between the number of accidents and traffic volume at divided highway intersections. This study was based on the accident experience over periods of 6 months to 5 years at 150 at-grade intersections on 290 km (180 mi) of rural divided expressways on the state highway system in California. Most of the intersections were two-way stop-controlled intersections, with stop control on the minor road and no control on the divided highway, although a few signalized intersections were included. The analysis of these California data found that low- crossroad-volume intersections have higher accident rates per crossroad vehicle than do high-crossroad-volume intersec- tions. The following relationship between the number of acci- dents and the traffic volume at divided highway intersections was developed: 7 (1) where N = expected number of intersection accidents per year Vd = ADT volume entering the intersection from the divided highway (veh/day) Vc = ADT volume entering the intersection from the cross- road (veh/day) This finding is evidence that concentrating cross-traffic at a few locations, by closing low-volume-crossroad intersections and providing frontage roads, may effectively reduce the number of intersection accidents. The Ohio study, conducted by Priest (10) in 1964, included 3 years’ worth of accident data for 316 at-grade intersections on divided highways with partial or no control of access. Most of the intersections were unsignalized; however, the author does not explicitly state the type of traffic control used at the intersections studied. Priest, like McDonald, also found that accident frequency is more sensitive to the crossroad traffic volume than to the divided highway traffic volume. Figure 2 illustrates the relationship between the average num- ber of accidents per year and traffic volume at divided high- way intersections developed by Priest. In a 1967 study of divided urban and rural highways in North Carolina, Cribbins et al. (11) found that median open- ings do not necessarily experience high accident rates under conditions of low volumes, wide medians, and light roadside development. However, as traffic volumes and roadside devel- opment increase, the frequency of median openings does affect accident potential significantly. Specifically, Cribbins et al. found the following relationships: • As traffic volumes increase, median openings experience a sharp increase in accident frequency. When combined with intensive roadside development, this increase in accident frequency becomes even more pronounced. • Signalization of median openings does not necessarily reduce accident experience under high-volume condi- tions, but it makes the traffic flow more efficiently by dis- tributing time for each movement. • As roadside development increases, and crossovers of any type are permitted, accidents will increase. In another study of multilane highways in North Carolina, Cribbins et al. (12) used the same data to determine the effects of selected roadway and operational characteristics on accidents on multilane highways. Cribbins et al. correlated eight highway characteristics with all injury accidents: median width, speed limit, traffic volume, level of service, access point index, intersection openings per mile, signalized openings per mile, and median openings per mile. The following conclu- sions were drawn from the analysis: N V V= 0 00078. 3 d0.455 c0.633

• The number of median openings, excluding intersections, affected the accident rate significantly. • The two roadway characteristics having the least effect on the accident rate were median width and speed limit. • Whenever storage lanes are installed at openings, the median-opening accident rate is no longer significantly affected by (1) the number of openings excluding inter- sections, (2) median width, (3) speed limit, or (4) ADT. Research by Harwood et al. in NCHRP Report 375 (7) found that the median width and median opening length have a strong influence on the safety performance of median open- ings. These issues are addressed in the next two sections of this report. MEDIAN WIDTH The safety and operational effects of median width at sig- nalized and unsignalized intersections were evaluated exten- 8 sively by Harwood et al. in NCHRP Report 375: Median Inter- section Design (7). However, this evaluation addressed median width for median openings at intersections—not median open- ings at driveways or median openings used solely for U-turns. NCHRP Report 375 used two separate approaches to address the relationship between median width and accidents at divided highway intersections: an accident study and a field observa- tional study. The traffic accident analysis of divided highway intersec- tions was conducted for NCHRP Report 375 with a statewide database of accident, geometric, traffic control, and traffic volume data for state highways in California. The findings of the analysis concerning median width are as follows: • At rural, four-leg, unsignalized intersections, accident frequency decreases as median width increases. • At rural, three-leg, unsignalized intersections, no statis- tically significant relationship exists between accident frequency and median width. Figure 2. Average number of accidents per year related to traffic volume at divided highway intersections (10).

• At urban/suburban, four-leg, unsignalized intersections, accident frequency increases with increasing median width over the range of median widths from 4 to 24 m (14 to 80 ft). • At urban/suburban, three-leg, unsignalized intersections, the intersection accident frequency increases with increas- ing median width. The field observational study in NCHRP Report 375 inves- tigated the effect of median width on three types of undesir- able driving behavior as commonly observed in the median opening area at intersections on divided highways: • Encroachment on through lanes by vehicles in the median opening area, • Side-by-side queuing of vehicles in the median opening area, and • Angle stopping by vehicles in the median opening area. Figure 3 illustrates side-by-side queuing of vehicles in an unsignalized median opening. NCHRP Report 375 reached the following conclusions concerning the effect of median width on accidents and unde- sirable driving behavior at unsignalized intersections: • At rural unsignalized intersections, the frequency of both accidents and undesirable driving behavior decreases as the median width increases. • At suburban unsignalized intersections, the frequency of both accidents and undesirable driving behavior increases as the median width increases. Based on these findings, NCHRP Report 375 recommended that rural unsignalized intersections should have medians that are as wide as practical, as long as the median is not so wide that approaching vehicles on the crossroad cannot see both roadways of the divided highway. At suburban unsignalized intersections, by contrast, medians should generally not be wider than necessary to provide whatever left-turn treatment is selected. At specific intersections where substantial turn- 9 ing and crossing volumes of large vehicles (such as school buses or trucks) are present, highway agencies may find it appropriate to select an appropriate median width to store a design vehicle of that type safely in the median. One issue of interest to the research for NCHRP Report 375 was how drivers making opposing left turns are influ- enced by the median width. Specifically, it was hypothesized that, at intersections with narrow medians, drivers making opposing left turns tend to turn in front of one another and, at intersections with wide medians, drivers making opposing left turns turn behind one another. Although no quantitative information exists on the median width at which drivers cease to turn in front of one another and begin to turn behind one another, an analysis of rural, unsignalized intersections found that, at intersections with median widths of less than 15 m (50 ft), vehicles making opposing left turns tend to turn in front of one another. In contrast, at intersections with median widths of more than 15 m (50 ft), vehicles making opposing left turns tend to turn behind one another. A similar pattern was found for suburban, unsignalized intersections with median widths of less than 15 m (50 ft). There were no comparable suburban, unsignalized intersections to verify whether the same turn- behind behavior observed at the rural, unsignalized intersec- tions with median widths of more than 15 m (50 ft) occurred at similar suburban, unsignalized intersections. A 1964 Ohio study by Priest found that, except at very low volume levels, intersection accident rates decrease with increasing median width (10). However, the difference in acci- dent rates between medians less than 6 m (20 ft) wide and medians 6 to 12 m (20 to 39 ft) wide is greater than the differ- ence in accident rates between medians with widths of 6 to 12 m (20 to 39 ft) and medians with widths of 12 m (40 ft) or more. These results are illustrated in Figure 4. Figure 3. Side-by-side queuing at unsignalized median opening. Figure 4. Variation of annual accident frequency at divided highway intersections as a function of median width and exposure index (the product of the ADTs of the intersecting roadways) (10).

A 1977 Purdue University study by Van Maren (13) devel- oped relationships between geometric and traffic volume vari- ables and accident experience at divided highway intersec- tions. Van Maren found no statistically significant relationship between median width and intersection accident rate. The author speculated that this finding may have resulted because of the limited range of median widths (9 to 18 m or 30 to 60 ft) that were evaluated. However, this range includes most of the rural divided nonfreeways that have been built by highway agencies since the 1950s, including current practices. Research sponsored by the Michigan Department of Trans- portation involved the collection and analysis of data for 1,503 km (934 mi) of Michigan state highways (14). Acci- dents on divided highway segments were compared with highway segments (mostly five-lane) with two-way left-turn lanes (TWLTLs). The analysis of the accident data with respect to the width of the median did not show any mean- ingful differences for divided highway segments that did not have traffic signals. The research found that divided high- ways with traffic signals may have lower accident rates with wider medians. However, the data were insufficient for con- clusive findings on this issue. The Florida Median Handbook (8) suggests that the appro- priate median width is a function of the purpose which the median is to serve in a particular application, such as the following: • Separation of opposing traffic streams, • Pedestrian refuge, • Left turn to side street, • Left turn out of side street, • Crossing vehicles, • U-turns, and • Aesthetics and maintenance. Table 1 presents a summary of standards and recommenda- tions for median widths, as presented in the Florida Median Handbook. The handbook recommends that extremely wide medians are needed to accommodate U-turn maneuvers by all design vehicles other than passenger cars. 10 MEDIAN OPENING LENGTH The only literature found on the relationship of median opening length to safety is the research by Harwood et al. in NCHRP Report 375: Median Intersection Design (7). That report addressed the effect of median opening length on unde- sirable driving behavior. Most undesirable driving behavior at divided highway intersections arises from the competition for limited space on the median roadway between drivers travel- ing through the median in the same direction. NCHRP Report 375 found that the frequency of undesirable driving behavior increases as median opening length increases at rural inter- sections and decreases as median opening length increases at suburban intersections. SAFETY EFFECTS OF MEDIAN TREATMENTS (RAISED/DEPRESSED/FLUSH/TWLTL) The treatment of roadway medians influences the safety and operational experience of a roadway as well as the access provided to adjacent developments. The four major types of median treatments are as follows: • Raised median—A raised median is a nontraversable median separated from the traveled way by curbs. Raised medians are used where it is desirable to separate traffic traveling in opposite directions and limit left-turn move- ments. The area within the median can be either con- crete or turf. • Depressed median—A depressed median is a nontra- versable turf median that separates traffic in opposite directions of travel and limits left-turn movements. As the name implies, a depressed median usually slopes away from the roadway to provide proper drainage. A depressed median has no curbs; the median is typically separated from the traveled way by pavement markings and shoulders. • Flush median—A flush median is a paved area, at the same grade as the traveled way, that may be marked as a median or as a center two-way left-turn lane (see below). Roadway type Speed Median width m (ft) Reconstruction Project 40 mph or less 5.0 (15.5) Minimum Reconstruction Project 45 mph 6.0 (19.5) Minimum Reconstruction Project 50 mph 7.0 (22.0) Minimum Four-lane highways with medians expecting significant U-turns and directional median openings with excellent positive guidance All 9.0 (30.0)—single left turns 12.6 (42.0)—dual left turns Recommended Six-lane highways with medians expecting significant U-turns and directional median openings with excellent positive guidance All 7.0 (22.0)—single left turns 10.6 (34.0)—dual left turns Recommended TABLE 1 Minimum and recommended median widths (8)

• Two-way left-turn lane (TWLTL)—A TWLTL is a cen- ter lane used for left turns from both directions of travel. At intersections, there is often a transition to conven- tional left-turn treatments. The literature on these median treatments is extensive. Sev- eral NCHRP reports and other sources present safety compar- isons of alternative median treatments. Research by Bonneson and McCoy in NCHRP Report 395: Capacity and Operational Effects of Midblock Left- Turn Lanes (15) considered the relative traffic operational and safety performance of cross-section for arterials and highways that are undivided, divided by a median, or divided by a center TWLTL. Table 2 presents a comparison of these three alternative cross-sections, indicating which cross-section is preferred with respect to operational, safety, access, and other factors. NCHRP Report 395 reviewed the relative safety performance of arterials with different cross-sections. Table 3 summarizes the safety performance of these cross- sections as reported by the following key sources in the liter- ature: Bowman and Vecellio (16),Chatterjee et al. (17), Parker (18), Squires and Parsonson (19), McCoy and Ballard (20), Walton and Machemehl (21), and NCHRP Report 282 (22). NCHRP Report 420: Impacts of Access Management Tech- niques (4) presents a summary of individual studies that have analyzed the safety benefits of replacing TWLTLs with non- traversable medians on undivided highways. Eleven studies were reviewed: some where the benefits were based on before- 11 and-after studies of the same roadway and some comparing accident rates for the two basic types of roads. The accident rate comparisons from the various studies are summarized in Table 4. In 15 out of the 16 comparisons shown in Table 4, the accident rates were reduced when a nontraversable median was installed in place of a TWLTL. NCHRP Report 420 concluded that nontraversable medians appear safer than TWLTLs. NCHRP Report 282: Multilane Design Alternatives for Improving Suburban Highways (22) presents a comparison of the safety, operational, and cost characteristics of selected multilane design alternatives for use in suburban areas. Advan- tages and disadvantages of each alternative are provided to assist in the selection of the most appropriate design for a given condition. The report states that the four-lane divided design alternative is best suited for use on major arterials with high volumes of through traffic and less than 45 drive- ways per mile. The five-lane TWLTL design alternative is most appropriate for suburban highways with commercial development, driveway densities greater than 45 driveways per mile, low-to-moderate volumes of through traffic, high left-turn volumes, and/or high rates of rear-end and angle accidents associated with left-turn maneuvers. Thus, NCHRP Report 282 does not make a blanket statement about the rel- ative safety of nontraversable medians and TWLTLs, but indicates that each has appropriate applications. NCHRP Report 330: Effective Utilization of Street Width on Urban Arterials (23) evaluated various alternative strate- gies for reallocating the usage of street width without chang- TABLE 2 Comparison of effects of three alternative cross-sections with differing midblock left-turn treatment types (15)

TABLE 3 Comparison of safety performance of alternative midblock cross-section as reported by studies in the literature (15)

TABLE 4 Synthesis of safety experience comparing TWLTLs with nontraversable medians by percent difference (4)

ing the total curb-to-curb width. Table 5 presents the advan- tages and disadvantages of four-lane divided roadways and five-lane roadways with TWLTLs. Research for the Florida Department of Transportation was performed for five roadway segments within the Central Florida area that underwent median modifications (2). Sev- eral of these segments also had other improvements, such as the addition of a through or auxiliary lane. The research results showed that the introduction of medians can greatly reduce collision potential and injuries. These reductions were found to occur as a result of the decrease in the number of conflict points. The research noted that conflict points are numerous along roadways with a continuous TWLTL. Con- flict reductions with an associated decrease in collision poten- tial may be achieved by either reducing the number of median openings or adding a median. The FHWA sponsored research to quantify the safety effect of raised curb, TWLTL, and undivided cross-sections on vehicles and pedestrians (16). A total of 32,894 vehicle and 1,012 pedestrian accidents were analyzed from 234.8 km (145.9 mi) of unlimited access arterials in three large metro- politan areas. The research found that streets with raised medians in both central business districts (CBDs) and subur- ban areas had lower pedestrian accident rates than TWLTLs and undivided arterials. In suburban areas, arterials with raised- curb medians were found to have significantly lower accident rates than TWLTLs for rear-end, right-angle, and left-turn col- lisions. Raised-curb medians also were found to have signif- icantly lower accident rates than undivided cross-sections for right-angle collisions. The research results also indicated that, in both CBDs and suburban locations, raised-curb medi- 14 ans had lower injury accident rates than either the TWLTL or undivided cross-sections. Research sponsored by the Michigan Department of Trans- portation involved the collection and analysis of data for 1,503 km (934 mi) of Michigan state highways (14). Accidents on divided highway segments were compared with highway segments (mostly five-lane) with TWLTLs. The divided high- way segments in Michigan generally have directional U-turn median crossovers that are also used for the indirect movement of left-turning traffic. Divided highway segments were found to have lower accident rates than TWLTLs for nearly every type of accident. The total accident rate (for all accident types) for the divided highway segments was approximately 50 per- cent of the total accident rate for highways with a TWLTL. Divided highways that exclusively have directional U-turn median crossovers were found to have approximately the same accident rate as divided highways that have conven- tional (bidirectional) median crossovers for unsignalized sec- tions of highways. Signalized divided highways with direc- tional crossovers were found to have about 50 percent of the accident rate of similar facilities with conventional median openings. However, the size of the data sample for this issue does not support conclusive findings. The comparison between raised medians and TWLTLs has been the focus of numerous other research studies. Accident data assembled by Chatterjee et al. (17) and by Parker (18) indicate that raised-curb median segments have lower accident rates than TWLTL segments. Walton and Machemehl (21) developed accident prediction equations for roadway seg- ments with TWLTLs. Not enough data were available to develop comparable equations for segments with raised medi- Design alternative Advantages Disadvantages Four-lane divided roadways 1. Provides additional lanes to increase capacity for through traffic movement 2. Reduces rear-end and angle accidents associated with left-turn maneuvers 3. Provides physical separation to reduce head- on accidents 4. Provides a median refuge area for pedestrians 1. Required street width may not be available 2. Increased delay to left- turning vehicles 3. Indirect routing required for large trucks 4. Lack of operational flexibility due to fixed median Five-lane roadways with TWLTLs 1. Provides additional lanes to increase capacity for through traffic movement 2. Reduces delay to through vehicles caused by left- turning vehicles 3. Reduces frequency of rear- end and angle accidents associated with left-turn maneuvers 4. Provides spatial separation between opposing lanes to reduce head-on accidents 5. Increases operational flexibility 1. Required street width may not be available 2. No refuge area in median for pedestrians 3. May generate safety problems at closely spaced driveways and intersections TABLE 5 Advantages and disadvantages of four-lane divided roadways and five-lane roadways with TWLTLs alternatives (23)

ans. Bonneson and McCoy (24) found raised-curb median treatments to be associated with fewer accidents than TWLTLs, especially for streets with average daily traffic (ADT) volumes exceeding 20,000. To determine the opera- tional and safety effects of replacing a TWLTL with a raised median, the Gwinnett County Department of Transportation in Georgia compared the accident experience before and after construction of a raised median (25). An accident analy- sis determined that retrofitting a TWLTL with a 254-mm (10-in) concrete raised median reduces accidents. Squires and Parsonson (19) concluded that, on high-volume road- ways, nontraversable medians have a lower crash experience than roadways with continuous TWLTLs. An analysis by Glennon et al. (26, 27, 28) found that TWLTLs have higher accident rates than raised medians where frequent driveways are found in combination with high arterial street volumes. Glennon et al. also found the raised median to be a more effective technique under higher traffic volumes. Margiotta and Chatterjee (29) performed a safety analysis of TWLTLs and raised medians. The study concluded that medians are generally safer than TWLTLs, but certain conditions exist where TWLTLs would have a more favorable safety experi- ence. Regression analysis found that driveway density is an important contributor to accident rates for medians, but not for TWLTLs. SAFETY EFFECTS OF INCREASING U-TURN MANEUVERS THROUGH USE OF NONTRAVERSABLE MEDIANS The installation of a nontraversable median may prevent many direct left-turn movements previously accessible to motorists, forcing those motorists to use indirect routes and, thus, increasing the volume of U-turn maneuvers. The effect of this increase in U-turn volumes on the safety of the roadway is not well understood. However, several studies have identi- fied issues related to U-turns at unsignalized median openings. Research by Gluck et al. in NCHRP Report 420: Impacts of Access Management Techniques (4) documents the safety and operational experience in several states where directional U-turn median openings have replaced conventional median openings. The states reported that closing conventional median openings and replacing them with directional median openings improves safety. Specifically, NCHRP Report 420 indicates that eliminating direct left turns from driveways and replacing them with indirect U-turn maneuvers results in a 20-percent reduction in accident rate. U-turn crossovers were found to have roughly one-half of the accident rates of roads with TWLTLs. The operational analysis in NCHRP Report 420 identified several operational benefits of direc- tional versus conventional median openings: shorter travel times, less delay, and increased capacity. The report states that right turns followed by U-turns can provide comparable, if not shorter, travel times than direct left turns from drive- 15 ways under heavy volume conditions when the diversion dis- tances are generally less than 0.8 km (0.5 mi). In another study, conducted by Kach (30) in Michigan, the safety performance of directional median openings was com- pared with that of conventional median openings to deter- mine the safety benefits that can be attributed to prohibiting left turns from the minor road. The mean intersection-related accident rates for directional median openings were found to be 15 percent lower than for conventional median openings. Similarly, the rates for intersection-related injury accidents were 30 percent lower for directional median openings than for conventional median openings. The study also showed substantial reductions in right-angle, rear-end, left-turn, and head-on accidents. In general, the results of this study indi- cate that directional median openings, where left turns are prohibited on the minor road, carry higher volumes at a lower accident rate than conventional median openings, where all turns are permitted. Levinson et al. analyzed the safety benefits of directional versus conventional median openings as a function of traf- fic signal density for 123 segments constituting 364 km (226 mi) of divided highway in Michigan (31). The authors reported that directional median openings have one-third the accident rate of TWLTLs and about two-thirds the accident rate of conventional median openings. The operational ben- efits also included increased capacity, reduced travel times, and improved signal coordination. In a paper presented at the 1996 annual meeting of the Transportation Research Board (32), Maki reported the safety results of replacing conventional median openings with direc- tional median openings in Michigan. The Michigan DOT was experiencing capacity problems on an arterial because of interlocking left turns within the conventional median open- ings at major intersections. The results of replacing four con- ventional median openings with directional median openings were significant in reducing the number of crashes, particu- larly right-angle crashes. The average number of accidents per year was reduced from 32 to 13—a decrease of about 61 percent. Angle collisions were reduced by 96 percent, sideswipe collisions by 61 percent, and rear-end collisions by 17 percent. Injury accidents decreased by 75 percent. Although the Florida Median Handbook (8) does not address specific safety issues related to U-turn maneuvers, it provides guidance on where a U-turn median opening should be con- sidered. First, it lists several indicators that a U-turn median opening should be considered in advance of a signalized intersection: • Level-of-service problems exist at the intersection. • Heavy left-turn movements are present at the signal. • Heavy conflicting right-turn movements are present at the intersection. • Gaps in oncoming traffic would be beneficial at a sepa- rate U-turn opening.

• There is sufficient space to separate the signalized inter- section and the opening. The Handbook also provides three design options to accommodate U-turns: (1) wide medians, (2) median “bulb- out” or loon, and (3) flare-out (jughandles). LEFT-TURN LANES Vehicles turning left from a multilane highway may pose safety and operational problems at median openings. They not only increase conflicts with and delays to other vehicles, but also pose a major safety problem with the large speed dif- ferential between left-turning and through vehicles. The FHWA National Highway Institute (NHI) has devel- oped a short course on Access Management (33). The man- ual for the course recommends installing left-turn lanes at existing median openings to • Allow turning vehicles to clear the through traffic lane with an acceptable speed differential, • Provide queue storage without interference with through traffic, • Reduce rear-end collisions and sideswipes, and • Increase capacity and decrease delay. The course also recommends increasing the length of an existing left-turn bay at all existing median openings where • Deceleration in the through traffic lane results in an unde- sirable speed differential between left-turning vehicles and through traffic. • The left-turn queue exceeds the length of the full-width left-turn lane. Left-turn lanes are often installed at median openings to accommodate high left-turning volumes. NCHRP Report 420 (4) summarizes the following safety benefits of left-turn lanes: • They remove the turns from the through travel lanes, thus reducing rear-end collisions. • They improve the visibility of oncoming traffic for vehi- cles turning left, thus reducing right-angle collisions. Installation of left-turn lanes has been the focus of many research studies. Various safety-related factors have been documented based on the type of intersection (e.g., signal- ized, unsignalized, and four-leg) where the left-turn treat- ment was implemented. Although many of these studies focus on left-turn treatments on two-lane highways, the safety rela- tionships can be useful for a broader range of roadway types, including divided arterials. 16 Parker et al. (34) determined that the addition of left-turn lanes at rural intersections along two-lane highways can reduce the potential for passing-related accidents. On urban four-lane roadways, McCoy and Malone (35) found that installation of left-turn lanes reduced rear-end, sideswipe, and left-turn accidents. Foody and Richardson (36) found that accident rates decreased by 38 percent with the addition of a left-turn lane at signalized intersections and by 76 per- cent at unsignalized intersections. Hauer (37) reported that left-turn channelization reduced accidents to varying degrees, depending on the intersection configuration; and Gluck et al. (4) reported accident rate reductions ranging from 18 to 77 percent as a result of the installation of left-turn lanes. A report by ITE indicates that median deceleration and storage lanes installed at intersections generally provide sig- nificant safety and operational benefits (38). Agent (39) per- formed an accident analysis of unsignalized intersections in Lexington, Kentucky, and found that the left-turn accident rate was 77 percent lower for intersections with left-turn lanes than intersections without left-turn lanes. Cribbins et al. (11) also reported that the number of rear-end collisions is less where storage lanes are provided. When implemented with additional safety measures, left- turn lanes have been found to be very effective in increasing safety. Hauer (37) reported that the provision of left-turn lanes at unsignalized intersections, when combined with instal- lation of curbs or raised medians, reduced accidents by 70, 65, and 60 percent in urban, suburban, and rural areas, respec- tively. When the channelization was painted, rather than raised, accidents decreased only by 15, 30, and 50 percent in urban, suburban, and rural areas, respectively. At signalized intersections, installation of left-turn channelization accom- panied by a left-turn signal phase reduced accidents by 36 per- cent; however, without the left-turn phase, accidents decreased only by 15 percent. At unsignalized intersections, findings of a California study indicate greater reductions in accidents with the use of a left-turn lane in a raised median than with painted left-turn lanes (40). Similarly, Lacy (41) found that a left-turn lane, when coupled with several other safety improvements, reduced accident frequency by 35 percent and accident sever- ity by 80 percent. Dale (42) found that installation of a traf- fic signal and left-turn channelization at intersections along rural two-lane highways reduced the total number of acci- dents by 19.7 percent, while the installation of a traffic sig- nal without any channelization reduced the total number of accidents by only 6 percent. Several predictive models and accident modification fac- tors have been developed that indicate left-turn lanes have a positive effect on safety. Maze et al. (43) developed a model that predicted a reduction in left-turn accident rate of 5.5 per- cent as a result of the installation of a left-turn lane with per- mitted signal phasing and a reduction of approximately 35 percent from installation of a left-turn lane with protected/ permitted signal phasing. Vogt (44) developed a model for a

four-leg rural intersection of a four-lane major road with STOP-controlled two-lane minor roads, which yielded an acci- dent reduction factor for total accidents of 38.4 percent as a result of the installation of a left-turn lane along the major road. In another study, Harwood et al. (45) developed algorithms to predict the expected safety performance of rural two-lane highways. The prediction algorithms combined elements of historical accident data, predictions from statistical models, results of before-and-after studies, and expert judgments made by experienced engineers. As part of the research, an expert panel of engineers developed accident modification factors (AMFs) for specific geometric design and traffic control fea- tures. AMFs are used in the accident prediction algorithms to represent the effects of safety of the respective features. The base value of each AMF is 1.0. Any feature associated with a higher accident experience than the base condition has an AMF value greater than 1.0, and any feature associated with lower accident experience than the base condition has an AMF value less than 1.0. In developing AMFs for the installation of left-turn lanes on the major-road approaches to intersections on two-lane rural highways, the expert panel reviewed various sources of information related to the accident reduction effectiveness of left-turn lanes. However, the panel did not find any well- designed before-and-after studies. Therefore, the panel com- bined results from several sources and developed AMFs for left-turn lanes, which are presented in Table 6. The AMFs rep- resent a judgment by the panel. The panel estimated that instal- lation of a left-turn lane along one major approach reduces intersection-related accidents by 18 to 24 percent, depending on the type of traffic control and the number of legs, and instal- lation of left-turn lanes along both major approaches to a four- leg intersection reduces intersection-related accidents by 33 to 42 percent, depending on the type of traffic control. Not all studies, however, have shown that left-turn lanes reduce accidents. Bauer and Harwood (46) found that left- turn lanes had a statistical association with higher frequen- cies of both total multiple-vehicle accidents and fatal and injury multiple-vehicle accidents. However, this result was not advanced by the authors as a basis for policy because the direc- tions of specific effects in predictive models often represent the surrogate effects of other variables, rather than the true effect of the variable of interest. At unsignalized intersections, McCoy and Malone (35) determined there was a significant 17 increase in right-angle collisions with installation of a left-turn lane. However, at unsignalized intersections on rural two-lane highways, McCoy et al. (47) found no significant difference in rear-end and left-turn collision rates between intersections with and without left-turn lanes. Poch and Mannering (48) also found some situations in which accidents of specific types increased with installation of left-turn lanes. NCHRP Report 348 (5) indicates that, although turning lanes may be required for some or all access locations to major activity centers, they are not always required for smaller developments. The report cites reference materials, such as the Highway Capacity Manual (49), that should be consulted for information to help guide the decision of whether turn lanes are needed. An emerging issue in the design of left-turn channelization is the restriction in sight distance that opposing left-turn vehi- cles cause one another. As an indication of this safety problem, David and Norman (50) determined that for ADT volumes between 10,000 and 20,000, four-leg intersections with oppos- ing left-turn lanes had more accidents than those without. A potentially effective countermeasure for safety problems where opposing left-turn lanes are present is to eliminate the sight restrictions by offsetting the left-turn lanes, as shown in Figure 5. NCHRP Report 375 reviewed the safety perfor- mance of a limited set of tapered and parallel offset left-turn lanes and found no safety problems (7). Both McCoy et al. (51) and Joshua and Saka (52) developed procedures to com- pute the amount of offset required for clear sight lines. How- ever, no evaluations of the accident reduction effectiveness of offset left-turn lanes have been found. Although offset left-turn lanes have been used primarily at signalized median openings, they have been used by at least two agen- cies at unsignalized median openings. Offset left-turn lanes are a potential concern because they may make U-turn maneu- vers more difficult to complete because they move the starting point for the U-turn maneuver closer to the opposing roadway. This potential problem is not addressed in the literature. In another study, Harwood et al. (53) conducted a before- and-after evaluation of the safety effects of providing left- and right-turn lanes for at-grade intersections. Geometric design, traffic control, traffic volume, and traffic accident data were gathered for 280 improved intersections, as well as for 300 similar intersections not improved during the study Number of major-road approaches on which left-turn lanes are installed Intersection type Intersection traffic control One approach Both approaches Three-leg intersection STOP signa Traffic signal 0.78 0.85 – – Four-leg intersection STOP signa Traffic signal 0.76 0.82 0.58 0.67 a STOP signs on minor-road approach(es). TABLE 6 Accident modification factors for installation of left-turn lanes on the major- road approaches to intersections on two-lane rural highways (45)

period. The research developed quantitative safety effective- ness measures for installation design improvements involving added left-turn lanes and added right-turn lanes. The research concluded that added left-turn lanes are effective in improv- ing safety at signalized and unsignalized intersections in both rural and urban areas. More specifically, the research con- cluded that • Installation of a single left-turn lane on a major-road approach would be expected to reduce total intersection accidents at rural unsignalized intersections by 28 per- cent for four-leg intersections and by 44 percent for three-leg intersections. • At urban unsignalized intersections, installation of a left- turn lane on one approach would be expected to reduce accidents by 27 percent for four-leg intersections and by 33 percent for three-leg intersections. • At four-leg urban signalized intersections, installation of a left-turn lane on one approach would be expected to reduce accidents by 10 percent. 18 Based on the results of this study, the AMFs for turn lanes in Table 6 have been revised as shown in Table 7. MEDIAN ACCELERATION LANES Median acceleration lanes are increasingly used at inter- sections on high-speed divided highways. They provide vehicles turning left onto a divided highway from a minor road with a path to accelerate to an appropriate speed before entering the through travel lanes. Median acceleration lanes provide both safety and operational benefits in that the enter- ing vehicles do not cause vehicles on the major roadway to decelerate substantially. Median acceleration lanes can allow a full median open- ing to operate with some of the characteristics of a directional median opening. Figure 6 illustrates a typical divided high- way intersection with median acceleration lanes. In NCHRP Report 375 (7) four intersections with one or more median acceleration lanes were studied with accident field data. These studies found that median acceleration lanes Figure 5. Examples of offset left-turn lanes (3).

can enhance the operation of intersections on divided high- ways. In particular, median acceleration lanes reduce the likelihood that vehicles making a left turn from a crossroad approach will need to stop in the median. Encroachment on the through lanes of a divided highway with a narrow median is a particular problem when larger vehicles are forced to stop in the median opening area. NCHRP Synthesis of Highway Practice 281: Operational Impacts of Median Width on Larger Vehicles (54), recom- mends the provision of median acceleration lanes to mini- mize the likelihood that larger vehicles will be required to stop in the median opening area. Median acceleration lanes normally allow vehicles turning left onto the divided high- way to proceed without stopping, except when conflicting traffic is present in the median opening area at the same time. Furthermore, the acceleration lane permits larger vehicles, which accelerate slowly, to enter the through lanes of the divided highway at a higher speed. This should minimize the potential for collisions between through and turning vehicles. In 1982, ITE conducted a survey of highway and traffic agencies in Canada and the United States concerning current use of median acceleration lanes and TWLTLs (38). Given 19 that median acceleration lane use was found to be limited, there was insufficient information for a quantitative analysis. However, some of the advantages and disadvantages reported by the agencies include the following: Advantages: • Median acceleration lanes reduce delays when traffic vol- umes are high. • Median acceleration lanes provide higher merging speeds. • Median acceleration lanes are useful if the acceleration lane is long enough to allow a safe merge. • Median acceleration lanes reduce accidents. Disadvantages: • It is difficult to merge from median acceleration lanes because of blind spots. • Median acceleration lanes are not used properly by drivers. • Median acceleration lanes create anxiety to through traffic. Number of major-road approaches on which left-turn lanes are installed Intersection type Intersection traffic control One approach Both approaches Three-leg intersection STOP signa Traffic signal 0.56 0.85b – – Four-leg intersection STOP signa Traffic signal 0.72 0.82b 0.52 0.67b a STOP signs on minor-road approach(es) b based on results in Reference 45 TABLE 7 Accident modification factors for installation of left-turn lanes on the major- road approaches to intersections on two-lane rural highways (53) Figure 6. Typical divided highway intersection with median acceleration lanes (7).

• Median acceleration lanes create conflicts. • Median acceleration lanes are unexpected and unfamil- iar to drivers. • The benefits of median acceleration lanes do not warrant the construction costs. The agencies also stated that median acceleration lanes are most effective at high-speed T-intersections on rural roads. Median acceleration lanes can improve the operation of directional median openings by helping U-turning vehicles to accelerate and merge with traffic on the through roadway. There are no data on whether median acceleration lanes at conventional median openings create additional conflicts for drivers making U-turn maneuvers. LOONS TO ASSIST LARGER VEHICLES IN COMPLETING U-TURN MANEUVERS A common problem associated with the use of directional crossovers for indirect left turns is the difficulty of larger vehicles to negotiate U-turns along cross-sections with nar- row medians. This situation often affects the operation and safety of commercial vehicles that typically require more space in order to perform a U-turn maneuver. One possible solution to this problem is the construction of a loon. Loons are defined as expanded paved aprons opposite a median crossover. Their purpose is to provide additional space to facilitate the larger turning path of commercial vehicles along narrow medians. Figure 7 presents a typical loon design. The genesis of the term “loon” is not clear, but it appears to be coming into common use. Loons appear to have been used at directional median openings, but the concept may be applicable to conventional median openings as well. A study by Sisiopiku and Aylsworth-Bonzelet (55, 56) evaluated the operation, placement, and safety of existing loons at directional median openings in western Michigan. The Michigan DOT has placed several loons along a 47-km (29-mi) corridor of divided roadway to facilitate the larger turning radii of commercial vehicles performing indirect left turns. Field data (including geometrics, posted speed limits, sign types and location, and traffic control) and 5 years’ of accident data were collected for the analysis. Results of the study indicate that directional crossovers with loons experi- 20 enced a high percentage of fixed-object and sideswipe crashes. Specifically, the following safety concerns were found at loons: • Fixed-object crashes with delineator posts, sign posts (in the median and along the mainline), and guardrail; • Sideswipe crashes involving vehicles merging into main- line traffic from the loon; • Sideswipe crashes involving mainline traffic attempting to use the right-turn lane and crashing with U-turning vehicles that turned from the crossover into the loon and then proceeded directly into the right-turn lanes; and • Commercial vehicles backing up and parking within the crossover. An operational analysis concluded that loons provide com- mercial vehicles with the extra pavement necessary to com- plete the U-turn maneuver required by indirect left-turns along narrow medians. Use of advance warning signs to improve driver expectancy is recommended. Finally, the authors present guidelines for the design and placement of loons. SIGHT DISTANCE AT MEDIAN OPENINGS Intersection sight distance (ISD) is an important design and operational consideration at all intersections, but may be even more important at divided highway intersections, includ- ing unsignalized median openings, where the median may increase the ISD requirements or may contain sight obstruc- tions that reduce the ISD. U-turn maneuvers should not be encouraged at locations with limited sight distance. Both NCHRP Report 383: Intersection Sight Distance (57) and NCHRP Report 375 (7), identify situations where ISD requirements for divided highway intersections may differ from undivided highway intersections. NCHRP Synthesis of Practice 281: Operational Impacts of Narrow Medians on Larger Vehicles (54) identifies sight distance as an important issue in determining locations where U-turns by larger vehi- cles should be permitted or encouraged. The Florida Median Handbook (8) also addresses sight distance issues at median openings. NCHRP Report 383 (57) presents revised ISD models that have been adopted by AASHTO and incorporated into the Figure 7. Typical loon design at a directional median opening (55, 56).

2001 Green Book (3). Addressing the unique nature of inter- sections on divided highways, the report states that these intersections may have substantial sight-distance concerns for left-turning vehicles. Despite the provision of stopping sight distance (SSD) along each roadway, sight obstructions in the median could limit left-turn sight distance. Further- more, opposing left-turn vehicles on divided highways may be aligned in such a way that they become sight obstructions to one another, blocking the view of oncoming traffic on the major road. The sight restrictions created by opposing left- turn vehicles can be minimized by the use of parallel and tapered offset left-turn lanes. NCHRP Report 375 (7) recognizes that ISD at divided highway intersections is complicated by the presence of the median on the major road, which may increase the ISD requirements at some intersections or may contain sight obstructions that reduce the ISD. The Green Book (3) con- siders ISD to be adequate when drivers at, or approaching, an intersection have an unobstructed view of the entire inter- section and of sufficient lengths of the intersecting highways to permit them to anticipate and avoid potential collisions. Adequate ISD requires unobstructed sight distance along both approaches of both intersecting roadways, as well as across the clear sight triangles. Adequate clear sight triangles are required both for drivers approaching an intersection where they are not required to stop and for drivers who are stopped at an intersection waiting to proceed safely to cross a major roadway or to turn left or right onto a major roadway. ISD requirements for crossing and turning maneuvers at divided highway intersections are generally increased with median width until the median becomes wide enough to store a vehicle. If the median is wide enough to store a vehicle, then the intersection operates as two separate intersections, because drivers can cross the near roadway and stop in the median, if necessary, before crossing or turning into the far roadway. In this case, the sight distance requirements of the intersections with the two roadways of the divided highway can be deter- mined separately. NCHRP Synthesis of Practice 281 (54) discusses alterna- tive improvement techniques that can be implemented to mit- igate the problems encountered by larger vehicles at divided highway intersections with narrow medians. When sight dis- tance for left-turn vehicles is limited by opposing through vehicles, this report recommends the following mitigation techniques at unsignalized intersections: • Offset opposing left-turn lanes by moving them laterally within the median. • Prohibit left turns from the major road. • Close the median opening. • Require indirect left-turn movements. The Florida Median Handbook (8) acknowledges that crossing and turning maneuvers onto a divided highway from a minor road or driveway can be performed as two separate 21 operations. The stopped vehicle must first have adequate sight distance to depart from a stopped position and cross traffic approaching from the left. The crossing vehicle may then stop in the median prior to performing the second oper- ation. The second move requires the necessary sight distance for vehicles to depart from the median, to turn left into the crossroad, and to accelerate without being overtaken by vehi- cles approaching from the right. The handbook also presents recommended sight distance values for U-turns at unsignalized median openings—these values are provided here in Table 8. Finally, the Florida Median Handbook (8) discusses sight distance issues related to opposing left-turn vehicles and sug- gests that vehicles turning left from opposing left-turn lanes restrict each other’s sight distance unless the lanes are suffi- ciently offset. A positive offset of 0.6 m (2 ft) is recommended when the opposing left-turn vehicle is a passenger car and 1.2 m (4 ft) when the opposing left-turn vehicle is a truck. INDIRECT LEFT-TURN MANEUVERS Indirect left-turn maneuvers include the use of jughandle roadways before a crossroad, loop roadways beyond a cross- road, and directional median openings beyond a crossroad. Indirect left-turn treatments enable drivers to make left turns efficiently on divided highways, including highways with rel- atively narrow medians. The Michigan and New Jersey DOTs have used indirect left-turn treatments extensively; other state highway agencies have used them occasionally (7). Increas- ingly, Florida is limiting unsignalized median openings to left turns from the arterial roadway; hence, drivers wishing to turn left from a driveway must turn right and then make a U-turn or use some other alternative route. Design policies concerning Speed (mph) Sight distance (ft) 35 520 40 640 45 830 50 1,040 55 1,250 60 1,540 Speed (km/h) Sight Distance (m) 60 160 70 200 80 260 90 380 100 470 Assumptions: Design vehicle = passenger vehicle Reaction time = 2.0 sec Extra time needed in the U-turn maneuver U-turn vehicle begins acceleration from 0 mph only at the end of the U-turn movement Values are based on speed/distance/acceleration figures from the 1990 AASHTO Green Book 50-ft clearance factor TABLE 8 Sight distance for U-turns at unsignalized median openings (8)

indirect left-turn treatments are addressed in the AASHTO Green Book (3). Indirect left turns and U-turns are discussed on pages 709 through 716 of the Green Book (3). Several design alterna- tives are presented. Figure 8 presents a jughandle-type ramp or diagonal roadway that intersects a secondary crossing road- way. The driver exits via the jughandle-type ramp and makes a left turn onto the crossroad. For a U-turn maneuver, the driver makes an additional left turn onto the divided highway. Figure 9 shows an at-grade loop that may be considered when the jughandle-type ramp would require costly right-of- way. Other factors favoring the at-grade loop include verti- cal alignment and grading costs. 22 Figure 10 illustrates a design that provides for indirect left turns to be made from the right, via separate turning road- ways connected to a crossroad. Such arrangements eliminate left turns from the through lanes and provide storage for left- turning vehicles not available on the highway itself. The left- turning vehicles are able to cross the main highway with lit- tle extra travel time. Figure 11 presents an indirect left turn for two arterials where left turns are heavy on both roads. Because lack of storage for left turns from the minor road would cause con- gestion, left turns from the minor road are prohibited. Left- turning traffic turns right onto the divided road and then makes a U-turn at a one-way crossover located in the median Figure 8. Jughandle-type ramp with crossroad (3). Figure 9. At-grade loop (surface loop) with crossroad (3).

of the divided road. Auxiliary lanes are highly desirable on each side of the median between the crossovers for storage of turning vehicles. In a series of ITE articles (58, 59), Hummer described seven unconventional left-turn design alternatives for urban and suburban arterials. The alternatives share two major prin- ciples: (1) reduce delay to through vehicles and (2) reduce and separate the conflict points at intersections. Hummer and Reid recently reviewed five of the seven alternatives—the median U-turn, bowtie, superstreet, jughandle, and continuous flow intersection—and summarized new information about each (60). After presenting the advantages and disadvantages of each alternative, the authors suggest when analysts should consider each alternative during feasibility studies and func- tional designs. NCHRP Synthesis 281 (54) presents a discussion of indirect left-turns by larger vehicles. The report states that although the denial of left-turn access by a raised median is likely to increase U-turn demand at nearby median openings, it is also likely that some larger vehicles will use indirect routes that do not involve a U-turn maneuver to reach their destination. Such routes may involve going around the block or may 23 incorporate an entirely different route from origin to destina- tion, so that the larger vehicle can make a right turn into the driveway at its destination. Where the median width of a divided highway at a median opening is narrow, no left-turn lane is provided, and the opposing traffic flow is high, drivers of larger vehicles that want to make a left turn may recognize that the median opening does not have sufficient size to accommodate their vehicle and that stopping in a through traffic lane to wait for a gap in opposing traffic leaves them potentially exposed to rear-end collisions. In this situation, drivers of larger vehicles may proceed to the next major intersection to complete a U-turn maneuver or may use an indirect route to their destination, just as they would if no median opening were provided. There are no generally applic- able estimates concerning how much delay to larger vehicles may result from such indirect routings. NCHRP Report 420 (4) reports an estimated 20-percent reduction in accident rate by replacing direct left turns from driveways with right-turn/U-turn treatments. Table 9 sum- marizes the differences in accident rate at three unsignalized locations where direct left turns were replaced by indirect left turns. Levinson et al. (31) present the safety and operational ben- efits of prohibiting left turns at signalized intersections along divided arterials in Michigan and installing directional U-turn crossovers downstream. Key features of the indirect left-turn treatments include the following: • Two-phase signal operation at the major intersection where all left turns are prohibited; • Directional U-turn crossovers for left turns located about 200 m (660 ft) on each side of the signalized intersection; • Right-turn lanes on the major and minor roads; • Left-turn lanes in the median of the major road for U-turn crossovers; • Coordination of signals in each direction of travel along the major road to ensure progressions; and Figure 10. Special indirect left-turn designs for traffic leaving highway with narrow median (3). Figure 11. Indirect left turn through a crossover (3).

• Minor-road intersections that are unsignalized become two T-intersections, so there are no direct unsignalized crossings of the median. The safety and operational benefits included lower accident rates, increased capacity, and reduced travel times. Recently, the Florida DOT sponsored a study to evaluate the safety effects of replacing full median openings with directional median openings, resulting in the indirect left- turn treatment that forces drivers to make a right turn fol- lowed by a U-turn at a midblock U-turn lane (61). Over 250 sites were evaluated in this study, including 125 sites involv- ing right turns followed by U-turns and 133 sites involving direct left turns. A cross-sectional comparison was used to measure the safety effects. The cross-sectional comparison method compares the crash rates of sample sites with two dif- ferent egress designs: (1) direct left turns and (2) right-turns/ U-turns. If the average crash rate of right-turn/U-turns is less than that of direct left turns at a certain statistical level of sig- nificance, it is presented that right-turn/U-turn movements could improve safety conditions. An assumption behind this comparison is that all the traffic patterns and geometric con- ditions remain consistent during the study period. The com- parison concluded that on six-lane divided arterials with large traffic volumes, high speeds, and high driveway/side-street access volumes, the implementation of the right-turn/U-turn treatment leads to a statistically significant reduction in total crash rate (26.4-percent reduction) as compared with direct left turns. The injury/fatality crash rate for right-turn/U-turns is significantly less than that of direct left turns (32.0-percent reduction). For eight-lane arterials, replacing direct left-turn openings with right-turn/U-turn openings leads to a reduction in the crash rate while the four-lane group results in an increased crash rate; however, the results for the four-lane and eight-lane groups were not statistically significant because of small sample sizes. ACCESS MANAGEMENT The safe and efficient operation of the highway system depends heavily on the effective management of access to adjacent developments. Access management is generally understood to preserve the flow of traffic on the surrounding roadways, maintain mobility, and improve safety. A consid- 24 erable amount of literature focuses on access management issues, and several studies have evaluated the relationship between safety and access management. Although none of these studies specifically address the safety at unsignalized median openings, the research that has established relation- ships between access density and safety can be useful in this study as well. The research performed for NCHRP Report 348 (5) inves- tigated and documented the state of the art in access control and the broader concept of access management. The report defines access management as “. . . providing (or managing) access to land development while simultaneously preserv- ing the flow of traffic on the surrounding road system in terms of safety, capacity, and speed.” The report also defines the overall concept of access management, reviews current practices, and sets forth basic policy, planning, and design guidelines. The guidelines include possible legislative changes and enforcement procedures, as well as strategic design and operating guides. NCHRP Report 420 (4) presents methods to predict and analyze the safety and traffic operational effects of selected access management techniques for different roadway variables and traffic volumes. Over 200 roadway segments, involving more than 37,500 accidents, were analyzed in detail. Acci- dent rates were derived for various spacings and median types. Key findings related to accident density and safety at unsignal- ized intersections were as follows: • Accident rates rise as the density of unsignalized access connections per mile increases. • The number of affected through vehicles traveling in the curb lanes increases as high-volume driveways are spaced closer together. The likelihood of spillbacks across a driveway rises with either an increase in the traffic vol- umes entering driveways and/or the driveway density. • Access spacing or setback distances on arterial roadways near freeway interchanges are generally inadequate for the weaving and left-turn storage movements that must be accommodated. A planning and access management guide (62) for Florida cities and counties presents two recommendations related to the location of driveways: Location Treatment Difference in accident rate US-1, Florida Driveway left turns replaced by right- turn/U-turn –22% Michigan Bi-directional crossover replaced by directional crossover +14% Michigan TWLTL replaced by directional crossover –50% TABLE 9 Accident rate differences—U-turns as alternate to direct left turns (4 )

• Construction of driveways along acceleration and decel- eration lanes and tapers is discouraged because of the potential for vehicular weaving conflicts. • Driveways across from median openings shall be con- solidated whenever feasible to coordinate access at the median opening. In a research study by Lall et al. (63), guidelines were developed for an access management program for the Ore- gon DOT. An analysis was performed on a 47-km (29-mi) section of Oregon Coast Highway 9 to determine the rela- tionship between access density and accident experience and severity. The analysis demonstrated a relationship between frequency of accidents and density of access points. The results showed that the number of accidents increased as the number of access points increased along the highway. Brown and Tarko developed impact models to predict crash frequencies based on the geometric and access control charac- teristics of a roadway (64). Negative binomial regression mod- els were developed to predict the total number of crashes, number of property-damage-only crashes, and number of fatal and injury crashes. The significant factors included density of access points, proportion of signalized access points, presence of an outside shoulder, presence of a TWLTL, and presence of a median with no openings between signals. The results indi- cated that access control has a beneficial effect on safety. The need to address the safety effects of U-turns at unsignal- ized median openings is a direct result of increased attention to access management. Highway agencies are installing more raised medians on arterials in response to access management guidelines. Median installation generally increases U-turn vol- umes and necessitates effective design of unsignalized median openings. SPACING BETWEEN ACCESS POINTS Access spacing is a key element of access management. An earlier portion of Chapter 2 addressed the effect on safety of the spacing between median openings. However, the spacing of access points between median openings is also an impor- tant aspect of access management. Access points introduce conflicts and friction into the traffic stream. Vehicles entering and leaving the main roadway often slow the through traffic, and the difference in speeds between through and turning traf- fic increases accident potential. The Green Book (3) states that “Driveways are, in effect, intersections . . . The number of crashes is disproportionately higher at driveways than at other intersections; thus their design and location merit spe- cial consideration.” It is believed that increasing the spacing between access points improves arterial flow and safety by reducing the number of conflicts per mile, by providing greater distance to anticipate and recover from turning maneuvers, and by providing opportunities for use of turn lanes. NCHRP Report 420 (4) presents an extensive summary of the safety research and experience associated with access 25 spacing. The various studies point to one consistent finding: an increase in the number of access points translates into higher accident rates. The specific relationships vary, reflect- ing differences in road geometry, operating speeds, and drive- way and intersection traffic volumes. Figure 12 shows the composite accident rates as presented in NCHRP Report 420. In addition to the review of safety studies, a comprehensive safety analysis was performed using accident data from eight states. Accident rates were derived for various unsignalized access spacings and median types. The analysis showed that accident rates increase with total access points per mile in urban and rural areas. The authors concluded that in urban and suburban areas, each additional access point (or driveway) on undivided highways increases the annual accident rate by 0.07 to 0.11 accidents per million veh-km (0.11 to 0.18 accidents per million veh-mi) traveled. Each additional access point on highways with TWLTLs or nontraversable medians increases the annual accident rate by 0.06 to 0.08 accidents per million veh-km (0.09 to 0.13 accidents per million veh-mi) traveled. In rural areas, each additional access point (or driveway) increases the annual accident rate by 0.01 to 0.04 accidents per million veh-km (0.02 and 0.07 accidents per million veh-mi) traveled on undivided highways and on highways with TWLTLs or nontraversable medians, respectively. TRB Circular 456 (6) presents a compilation of the cur- rent state and local practices for designing streets and high- ways from an access management perspective. The circular illustrates the basic considerations for spacing standards and Figure 12. Composite accident rates as a function of access point density (4).

guidelines and describes current state, county, and local spac- ing requirements. Among the various agencies, there is little consensus on unsignalized intersection spacing; however sight distance requirements and driver response times are key param- eters. The authors recognize the need for additional research on unsignalized spacing and corner clearance criteria and their applicability in various urban, suburban, and rural settings. Finally, the circular presents some established traffic engi- neering and roadway design and planning principles related to unsignalized access spacing: • Limit the number of conflicts. • Separate basic conflict areas. • Reduce interference with through traffic resulting from turns into or out of a site. • Provide sufficient spacing between at-grade intersections. • Maintain progressive speeds along arterials. • Provide adequate on-site storage areas. NCHRP Report 348 (5) suggests that access spacing guide- lines be keyed to allowable access levels, roadway speeds, and operating environments. The guidelines should apply to new developments and to significant changes in the size and nature of existing developments. It goes on to add that the guidelines do not have to be consistent with existing prac- tices; because of historical conditions, access to land parcels that do not conform to the spacing criteria may be necessary when no alternative reasonable access is available. In addi- tion, NCHRP Report 348 observes that research and prac- tices have not identified any clear method of establishing spacing standards for unsignalized intersections and that, moreover, many proposed guidelines have never been imple- mented. Standards may be based on speed, stopping sight distance, roadway function, type of traffic generator, or other considerations. The report presents guidelines for unsignal- ized driveway spacing that are based on speed, access level, size of activity center, and environment (e.g., urban). In gen- eral, spacing increases as the size of the activity center and operating speed increase. For example, for a minimum use activity in an urban area on a low-speed roadway with a high degree of access allowed, the spacing could be about 15 m (50 ft). In a rural area for a major activity on a high-speed roadway with allowable access more limited, the spacing could be about 150 m (500 ft). In a manual for the NHI short course on Access Manage- ment (33), various conditions that should be considered in the determination of unsignalized access spacing are presented: 26 • Stopping sight distance, • Intersection sight distance, • Maneuver distance, • Right-turn conflict overlap, • Maximizing of egress capacity, and • Corner clearance. In a paper addressing the effect of street spacing on scale (65), Levinson compares the spacing and design of city streets and suburban highways, identifies the strengths and weaknesses of each, and suggests spacing guidelines for var- ious urban and suburban environments. The paper demon- strates how the provision of arterial streets at closer intervals improves access opportunities, reduces traffic concentrations where these streets meet, and allows reduced cross-section. In another paper by Levinson (66), a method is presented for predicting the safety of arterial roads based on arterial traffic volumes, access road volumes, and access density. The procedure applies the long-established relationship between intersection accidents and the product of conflicting traffic volumes. Safety indices are provided that relate only to the change in access density; these indices are generally consis- tent with those reported in NCHRP Report 420 (4). The indices also show that the increase in accidents is equal to the square root of the increase in access density. EFFECTS OF ADJACENT TRAFFIC SIGNALS The effects of adjacent traffic signals on the operation of median openings and U-turn maneuvers include queue spill- backs that block a median opening and the influence of the adjacent signal on available gaps in traffic. The Florida Median Handbook (8) addressed queue spill- backs and states that median openings should not be placed across regularly forming queues from neighboring intersec- tions. Median openings placed too closely to an intersection cause both safety and operational problems. The safety prob- lem is that when these queues build, “good Samaritans” might allow vehicles in the median opening through the queue with- out an adequate gap in the adjacent lane, creating a potential collision with a vehicle moving freely in the adjacent lane. A traffic operational problem is that when the queue in the through lane extends past the median left-turn lane, vehicles wanting access to the median opening are trapped in the queue and are unable to move into the turn bay until the queue advances.

Next: Chapter 3 - Current Design Policies and Practices of Highway Agencies »
Safety of U-Turns at Unsignalized Median Openings Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s National Cooperative Highway Research Program (NCHRP) Report 524: Safety of U-Turns at Unsignalized Median Openings includes recommended guidelines for locating and designing unsignalized median openings, and a methodology for comparing the relative safety performance of different designs.

  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!