Head-On Collisions

Description of Strategies

Objectives

The objectives for reducing the number of head-on fatal crashes are to

  • Keep vehicles from encroaching onto the opposite lane,
  • Minimize the likelihood of crashing into an oncoming vehicle, and
  • Reduce the severity of crashes that occur.

These objectives are similar to those cited for ROR crashes (emphasis area 15.1; see Volume 6 of this report, the ROR guide). Reduction in the severity of crashes is covered in that emphasis area. Specific strategies include improving the design of roadside hardware (e.g., bridge rails) and the design and application of barrier and attenuation systems.

Exhibit V-1 summarizes the objectives and strategies available to help system managers meet them.

EXHIBIT V-1
Objectives and Strategies for Addressing Head-On Crashes

Objectives Strategies

18.1 A—Keep vehicles from encroaching into opposite lane

18.1 A1—Install centerline rumble strips for two-lane roads (T)a

18.1 A2—Install profiled thermoplastic strips for centerlines (T)

18.1 A3—Provide wider cross sections on two-lane roads (E)

18.1 A4—Provide center two-way, left-turn lanes for four- and two-lane roads (T)

18.1 A5—Reallocate total two-lane roadway width (lane and shoulder) to include a narrow "buffer median" (T)

18.1 B—Minimize the likelihood of crashing into an oncoming vehicle

18.1 B1—Use alternating passing lanes or four-lane sections at key locations (T)

18.1 B2—Install median barriers for narrow-width medians on multilane roads (T)

a For an explanation of (T), (E), and (P), see the next page.

Explanation of Strategy Types

The strategies in this guide were identified from a number of sources, including the literature, contact with state and local agencies throughout the United States, and federal programs. Some of the strategies are widely used, whereas others are used at state or even local levels. Some have been subjected to well-designed evaluations to prove their effectiveness. However, it was found that many strategies, including some that are widely used, have not been adequately evaluated.

The implication of the widely varying experience with these strategies, as well as the range of knowledge about their effectiveness, is that the reader should be prepared to exercise caution in many cases before adopting a particular strategy for implementation. To help the reader, the strategies in the AASHTO guides have been classified into three types, each identified by a letter:

  • Tried (T): Those strategies that have been implemented in a number of locations, and may even be accepted as standards or standard approaches, but for which there have not been found valid evaluations. These strategies—while in frequent, or even general, use—should be applied with caution, carefully considering the attributes cited in the guide, and relating them to the specific conditions for which they are being considered. Implementation can proceed with some degree of assurance that there is not likely to be a negative impact on safety and very likely to be a positive one. It is intended that as the experiences of implementation of these strategies continues under the AASHTO Strategic Highway Safety Plan initiative, appropriate evaluations will be conducted so that effectiveness information can be accumulated to provide better estimating power for the user, and the strategy can be upgraded to a "proven" (P) one.
  • Experimental (E): Those strategies that have been suggested and that at least one agency has considered sufficiently promising to try on a small scale in at least one location. These strategies should be considered only after the others have proven not to be appropriate or feasible. Even where they are considered, their implementation should initially occur using a very controlled and limited pilot study that includes a properly designed evaluation component. Only after careful testing and evaluations show the strategy to be effective should broader implementation be considered. It is intended that as the experiences of such pilot tests are accumulated from various state and local agencies, the aggregate experience can be used to further detail the attributes of this type of strategy, so that it can be upgraded to a "proven" one.
  • Proven (P): Those strategies that have been used in one or more locations and for which properly designed evaluations have been conducted that show the strategies to be effective. These strategies may be employed with a good degree of confidence, but understanding that any application can lead to results that vary significantly from those found in previous evaluations. The attributes of the strategies that are provided will help the user judge which strategy is the most appropriate for the particular situation.

All but one of the strategies detailed in this guide are classified as "tried." That means that although they may be widely used, sufficient evidence of their effectiveness is not available to say that they have been proven to be effective.

Strategies may fall into two categories—those implemented over extended sections of highway or those implemented at selected spot locations. Just as the objectives for head-on crashes are similar to ROR crashes, so too are the strategies. Strategies common to both include

  • Enhanced delineation of sharp curves;
  • Improved highway geometry, especially for horizontal curves, including design elements such as curvature, superelevation, and widening through the curve;
  • Better pavement markings;
  • Skid-resistant pavement surfaces;
  • Improved shoulders to prevent ROR over-recovery, including paving, eliminating edge drops, and improving shoulder slopes; and
  • Rumble strips to slow vehicles on approaches to hazardous locations.

A discussion of these strategies is included in Volume 6 of this report, the ROR guide.

While some of the strategies presented in this guide may be appropriate for reducing head-on crashes on urban roadways, no strategies specifically for urban roads are included in this guide.

Related Strategies for Creating a Truly Comprehensive Approach

The strategies listed above and described in detail below are those considered unique to this emphasis area. However, to create a truly comprehensive approach to the highway safety problems associated with this emphasis area, related strategies should be included as candidates in any program planning process. These related strategies are of five types:

  • Public Information and Education (PI&E) Programs—Many highway safety programs can be effectively enhanced with a properly designed PI&E campaign. The traditional emphasis with PI&E campaigns in highway safety is to reach an audience across an entire jurisdiction or a significant part of it. However, there may be reason to focus a PI&E campaign on a location-specific problem. While this is a relatively untried approach compared with areawide campaigns, use of roadside signs and other experimental methods may be tried on a pilot basis.

    Where the application of PI&E campaigns is deemed appropriate, this guide is usually in support of some other strategy. In such a case, the description for that strategy will suggest this possibility (see the attribute area for each strategy entitled "Associated Needs"). In some cases, specialized PI&E campaigns are deemed unique for the emphasis area and are explained in detail within the guide. In the future, additional guides may exclusively address the details regarding PI&E strategy design and implementation. When that occurs, the appropriate links will be posted online at http://safety.transportation.org.
  • Enforcement of Traffic Laws—Well-designed, well-operated law-enforcement programs can have a significant effect on highway safety and must be an element in any comprehensive highway safety program. It is well established, for instance, that an effective way to reduce crashes (and their severity) resulting from driving under the influence (DUI) or driving without using seat belts is to have jurisdictionwide programs that enforce an effective law against such behavior. When that law is vigorously enforced with well-trained officers, the frequency and severity of highway crashes can be significantly reduced.

    Enforcement programs, by the nature of how they must be performed, are conducted at specific locations. The effect (e.g., lower speeds, greater use of seat belts, reduced impaired driving) may occur at or near the specific location where the enforcement is applied. This effect can often be enhanced by coordinating the effort with an appropriate PI&E program. However, in many cases the impact of enforcement is areawide or jurisdictionwide. The effect can be either positive (i.e., the desired reductions occur over a greater part of the system) or negative (i.e., the problem moves to another location as road users move to new routes where enforcement is not applied). Where it is unclear how the enforcement effort may impact behavior, or where it is desired to try an innovative and untried method, a pilot program is recommended.

    As with PI&E campaigns, where the application of enforcement programs is deemed appropriate, this guide often supports some other strategy. In such cases, the description for that strategy will suggest this possibility (see the attribute area for each strategy entitled "Associated Needs"). In some cases, where an enforcement program is deemed unique for the emphasis area, the enforcement program will be explained in detail. In the future, additional guides may exclusively address the details regarding enforcement strategy design and implementation. When that occurs, the appropriate links will be posted online at http://safety.transportation.org.
  • Strategies to Improve Emergency Medical and Trauma System Services—Treatment of injured parties at highway crashes can have a significant impact on the level of severity at which and length of time during which an individual spends treatment. This is especially true when it comes to timely and appropriate treatment of severely injured persons. Thus, a basic part of a highway safety infrastructure is a well-based and comprehensive emergency care program. While the types of strategies that are included here are often thought of as simply support services, they can be critical to the success of a comprehensive highway safety program. Therefore, for this emphasis area, an effort should be made to determine if there are improvements that can be made to this aspect of the system, especially for programs focused upon location-specific (e.g., corridors) or area-specific (e.g., rural area) issues. As additional guides are completed for the AASHTO plan, they may address the details regarding the design and implementation of emergency medical systems strategies. When that occurs, the appropriate links will be posted online at http://safety.transportation.org.
  • Strategies Directed at Improving the Safety Management System—The management of the highway safety system is foundational to success in making the highway safety system safer. A sound organizational structure, as well as an infrastructure of laws, policies, etc., to monitor, control, direct, and administer a comprehensive approach to highway safety should be in place.

    It is important that a comprehensive program not be limited to one jurisdiction, such as a state DOT. Local agencies are often responsible for the majority of the roadway system and know its related specific safety problems. As additional guides are completed for the AASHTO plan, they may address the details regarding the design and implementation of strategies for improving safety management systems. When that occurs, the appropriate links will be posted online at http://safety.transportation.org.
  • Strategies Detailed in Other Emphasis Area Guides—As mentioned in the previous section, many of the strategies in Volume 6 of this report (the ROR guide) are also effective in addressing head-on collisions. It is important that such strategies, effective for more than one emphasis area, are connected to all other related emphasis areas. As more guides are created, related links will be posted online at http://safety.transportation.org.

    Comprehensive approaches are facilitated by involvement of as many of the potential stakeholders as possible. Appendix 12 provides a list of candidate types of stakeholders to consider including in the planning and implementation of a program.

Objective 18.1 A—Keep Vehicles from Encroaching into Opposite Lane

Strategy 18.1 A1—Centerline Rumble Strips for Two-Lane Roads (T)

General Description
Centerline Rumble Strips Implemented in Maryland

EXHIBIT V-2
Centerline Rumble Strips Implemented in Maryland

Centerline rumble strips are similar to shoulder rumble strips. The purpose of rumble strips is to alert drivers who may inadvertently stray or encroach into opposing lanes. While this is a relatively new treatment, it has been implemented by some states, including Minnesota, Pennsylvania, Colorado, Delaware, Maryland, California, Washington, and Virginia. Other states, like Kansas, are still evaluating the use of centerline rumble strips in experimental installations.

Although there is no standard design, the rumble strip is generally wider than the center markings, extending into the travel lane by 5 in. to as much as 1.5 ft. In some states, the strips are continuous along the centerline; in others, they alternate with a smooth gap.

Since centerline rumble strips do not require changes in the overall cross section of the roadway, they would be compatible with other strategies such as shoulder rumble strips and horizontal curve improvements. This strategy, although fairly widely used, has not been sufficiently evaluated to be considered "proven."

EXHIBIT V-3
Strategy Attributes for Centerline Rumble Strips on Two-Lane Roads (T)

Technical Attributes

Target

Drivers of vehicles who unintentionally cross the centerline.

Expected Effectiveness

Although some literature sources report on the effectiveness of shoulder rumble strips, use of centerline rumble strips is relatively new and so there are very few reports concerning its success or failure. However, the findings below indicate that centerline rumble strips are a promising countermeasure to reduce ROR collisions, and further study of their effectiveness is warranted.

In Pennsylvania, the major testing program involving implementation of centerline rumble strips began in 2001, and thus formal evaluation is not yet possible. Pennsylvania installed centerline rumble strips on a 6-mile section of Route 322 in 1993. In the 6 years following the installation, there were no fatal accidents. However, centerline rumble strips were only one part of a treatment package that included separating lanes by a narrow "buffer median" and other modifications. Centerline rumble strips were also installed on some roadways in Delaware. The 1995 Annual Report on Highway Safety Improvement Programs for the State of Delaware reported a reduction of head-on collisions where centerline rumble strips were installed. The results are tabulated below:

Head-On Crashes on a Two-Lane Rural Highway in
Delaware Before and After Use of Centerline Rumble Strips

 

Head-On Crash Frequency

Severity of Crash

36 Months Before

24 Months After

Fatal

6

0

Injury

14

12

Damage only

19

6

  Total

39

18

Crashes per month

1.1

0.76

NCHRP Report 440 (Fitzpatrick et al., 2000) describes the installation of centerline rumble strips as part of "the improvements that could correct driver behavior in a manner that would reduce fatal head-on accidents." The improvements were made to a 20-mile rural two-lane segment in California. For nonpassing sections of the roadway, centerline double yellow strips were replaced with a 16-in. (40.6-cm)-wide rumble strip and raised profile traffic striping. Additional centerline improvements included raised markings for passing sections, and shoulder treatments included rumble strips and raised markings. A limited accident review using 34 months of before data and 25 months of after data did show a reduction in accidents. On average, 4.5 accidents occurred per month in the before period and 1.9 accidents per month occurred in the after period.

These studies appear to involve "high-crash sites." Due to the "regression to the mean" bias, the estimates of effectiveness are probably inflated to some degree. Thus, there remains a need for well-designed before/after studies that can produce more accurate results of effectiveness.

Centerline rumble strips have not been sufficiently evaluated to be considered a proven strategy. However, this strategy was tried and accepted in a number of applications. Additionally, there have been no significant findings of negative effects from the use of centerline rumble strips.

Keys to Success

To be effective, centerline rumble strips must be implemented over a continuous length of facility. It may not be cost-effective to implement this strategy on all undivided road sections. Therefore, a key to success is identifying the characteristics of the roadway (traffic volume, speed, alignment quality, cross section) for which rumble strips may be expected to have the greatest positive effect.

Potential Difficulties

Shoulder rumble strips have either real or perceived drawbacks such as difficulty with snow removal, additional shoulder maintenance requirements, and undesirable noise levels. States not using rumble strips may have concerns about these effects. However, states that use rumble strips (on the roadway shoulder or otherwise) have not reported any additional maintenance requirements as long as the rumble strips are placed on pavement that is in good condition. This pitfall may make centerline rumble strips an expensive countermeasure if targeted implementation is not achieved (i.e., if an agency tried to implement rumble strips everywhere) and measurable benefits are not accomplished. In a related vein, an effective implementation strategy may be to deploy centerline rumble strips in conjunction with resurfacing or reconstruction projects. This may, however, forestall the overall systemwide benefits sought by AASHTO over the short term.

There is the possibility of adverse effects on motorcycling. Note, however, that Pennsylvania has worked with motorcycle groups, and no major concerns were raised by these groups.

Finally, it is possible that the use of a centerline rumble strip might have some negative operational effects by inhibiting passing maneuvers (due to the look and noise of the strip). However, states currently using these rumble strips have not reported such problems (e.g., Washington, Minnesota, Pennsylvania).

Appropriate Measures and Data

In an evaluation of centerline rumble strip programs, process measures would include the number of road miles or number of hazardous locations where rumble strips are installed. Process measures may also include the aspect of exposure—number of vehicle miles of travel exposed to centerline rumble strips.

Impact measures will include the number (or rate) of head-on crashes reduced at these locations, along with any change in total crashes. Another measure may be public acceptance, including complaints from roadway users and even nonusers adjacent to the road.

Accident data, traffic volume data, and roadway data will be required to identify appropriate sites for installation.

Associated Needs

Since this countermeasure is relatively new (unlike shoulder rumble strips), there may be a need for public information to explain the function of the treatment in order to address any public concerns or potential misunderstandings. Such campaigns may address the concerns of motorcyclists and the education of the motoring public regarding the effects of centerline rumble strips on passing maneuvers.

Organizational and Institutional Attributes

Organizational, Institutional and Policy Issues

These strategies will be implemented by state and local roadway agencies, and it does not appear that extra coordination with other agencies or groups is needed. Since proof of effectiveness will be what sells this treatment in the long term, this effort will be most effective when reporting and analyzing head-on crashes becomes routine. It will also be most effective when a framework and a methodology exist targeting the implementation to the most appropriate sites. The framework and methodology will include institutionalizing centerline rumble strips within an agency's design standards and policies, as appropriate.

Issues Affecting Implementation Time

This low-cost strategy does not involve reconstruction and would not involve the environmental process or right-of-way acquisition. Rumble strips can be implemented quickly, certainly within a year once a site or highway is selected. Programmatic implementation may take slightly longer, depending on availability of necessary traffic, crash, and roadway data within an agency. Incorporation of centerline rumble strips as part of an agency's design practice for new construction or resurfacing can occur quickly (within 1 year).

Costs Involved

Costs will vary depending on whether the strategy is implemented as a stand-alone project or incorporated as part of a reconstruction or resurfacing effort already programmed. Including rumble strips as part of a resurfacing project offers the opportunity for lowest cost implementation. Some recent cost figures are given in Appendix 11.

Training and Other Personnel Needs

There appear to be no special personnel needs for implementing this strategy. States would either use agency personnel or contractors.

Training of state safety engineers on the attributes, benefits, and applicability of centerline rumble strips would be necessary. Training regarding actual installation of the rumble strips would depend on whether the agency has been using retrofitted rumble strips on freeways or other roadways. If not, either agency personnel or contractor personnel would need to be trained in proper installation techniques.

Legislative Needs

None identified.

Other Key Attributes

None Identified

Information on Agencies or Organizations Currently Implementing this Strategy
Washington, Maryland, Minnesota, Virginia, and Pennsylvania DOTs have provided information on current use. Kansas has provided information on plans for installation. Washington reports it is using raised discs, raised pavement markers, or plastic strips to get the rumble effect. These are being used continuously along a route in both passing and no-passing zones, and passing maneuvers do not seem to be affected. Washington also reported that drivers are using the raised discs and markers as a guide when the centerline is not visible due to winter weather. Maryland has used them primarily on access-controlled highways. For example, the "treatment package" on Maryland Route 90 includes a wider center yellow line combined with raised pavement markers and rumble strips. Information on Maryland's program can be found in Appendix 1. Minnesota has implemented strips on a limited basis without any current problems. Virginia installed centerline rumble strips on approximately 2 miles of two-lane roadway and found no initial problems with the installation. Information on Virginia's program can be found in Appendix 2. Pennsylvania reports both public acceptance and no centerline obscuring in a pilot effort there. In Kansas, installation of centerline rumble strips was to begin in summer 2001. Information on Kansas's program can be found in Appendix 3.

Strategy 18.1 A2—Profiled Thermoplastic Stripes for Centerline

General Description
This treatment has been used for centerlines on two-lane roads by at least two states—California and Texas. Standard plans used in Texas are included as Appendix 4. Both states use this treatment for sections where passing is not permitted. North Carolina has used this treatment to mark both sides of a two-way, left-turn lane on a multilane roadway. The 6-in. yellow profile thermoplastic stripes are installed on a 15-mile section of U.S. 158.

This treatment provides an audible/tactile effect, but it is less noticeable for larger vehicles, especially trucks. This effect is similar to that experienced with raised pavement markers with short spacing. While the audible/tactile effect can be advantageous, its principal benefit is apparently the longer visibility distance provided at night, especially during wet conditions, when compared with standard pavement markings. However, as with standard raised pavement markers, this treatment would be limited to areas where there is little or no snow, as snow plow blades will easily scrape off the stripe.

This strategy, although used in several states, has not been sufficiently evaluated to be considered "proven."

EXHIBIT V-4
Strategy Attributes for Profiled Thermoplastic Stripes for Centerlines (T)

Technical Attributes

Target

Motorists who unintentionally cross the centerline of a roadway.

Expected Effectiveness

Although both California and Texas have used this treatment on significant mileage of roadways, there has not been any formal evaluation. The anecdotal information is that this treatment provides good visibility of the centerline during night conditions and that even the mild audible and tactile effect is useful for reminding motorists to keep right of the centerline. There is no known evaluation of changes in accidents or in changes in any driver performance measure with the use of this treatment for a centerline. Caltrans has been using this treatment on an experimental basis since 1993, but no longer considers it experimental.

This treatment has not been sufficiently evaluated to be considered a proven strategy. However, there have been no significant findings of negative effects of this strategy.

Keys to Success

While no agency has identified specific guidelines on where this treatment could be applied, it would seem that it may be a reasonable option for two-lane rural roads under the following conditions:

  • Snow removal is not required.
  • Sections of no-passing zones are relatively long.
  • Volume levels and head-on related crash experience do not justify centerline rumble strips or some other more costly treatment.
  • Resurfacing or other pavement maintenance activities that would cause removing the treatment are not scheduled for at least 3 years.
  • Areas are of higher-than-normal rainfall.

Additionally, Caltrans has used this strategy as an incremental improvement while a more cost-intensive project is being designed and funded.

Potential Difficulties

There are no significant obstacles or difficulties in using this treatment. There have not been any adverse effects reported for motorcyclists. Application of the patterns illustrated above is easily accomplished with standard thermoplastic machines. One other cautionary note about its use is that it may not be suitable for open-graded or seal-coated surfaces.

Appropriate Measures and Data

In an evaluation of this treatment, process measures would be the same as for rumble strips—the number of road miles or number of locations where this treatment is installed. The exposure measure would be the vehicle miles of travel exposed to the treatment.

Performance measures are also the same as for rumble strips and include the number (or rate) of head-on crashes reduced at the locations. Public acceptance and preferences from road users would also be an appropriate measure.

Associated Needs

There should not be any special needs, such as a public information since this treatment is not likely to surprise motorists. Raised pavement markers are fairly common and this treatment is similar to that application.

Organizational and Institutional Attributes

Organizational, Institutional, and Policy Issues

Since this strategy is similar to normal pavement-marking practices, vis-à-vis standard flat thermoplastic stripes and raised pavement markers, there should not be any special issues. Those jurisdictions that use this treatment will want to conduct their own evaluations to establish whether the treatment is a cost-effective strategy.

Issues Affecting Implementation Time

Implementation of this treatment can be almost immediate, which makes it an appealing initial strategy to address an identified head-on crash problem on two-lane roads.

Costs Involved

When considering cost on a life-cycle basis, the thermoplastic stripe becomes less costly than standard paint striping. Its use also will result in a reduction of potential conflicts with maintenance vehicles operating under traffic because of the reduced frequency of application needed. Some example costs for 2002 are given in Appendix 10.

Training and Other Personnel Needs

There are no special personnel or training needs for implementing this strategy. States vary as to whether or not they use their own forces for applying thermoplastic markings.

Legislative Needs

None identified.

Other Key Attributes

None Identified

Information on Agencies or Organizations Currently Implementing this Strategy
Caltrans seems to be the state most active with this application. Caltrans has applied this treatment—as both centerlines and edgelines—on a significant mileage of two-lane rural roads that were identified as safety problem corridors. While no specific evaluation has been conducted, Caltrans managers think that the treatment has helped improve the safety of the corridor. For several years, the treatment was considered experimental; it is now considered a standard treatment under the guidelines discussed above. See Appendix 8 for further details.

Strategy 18.1 A3—Two-Lane Highways with Wide Cross Sections

General Description
Even though there is a broad range of definitions for the "Super Two" design in both the United States and Europe, most definitions are characterized by wider cross sections. These designs include wider lanes, wider full-strength shoulders, and high-speed alignment with 100-percent passing sight distance. A common design for the United States includes 14-ft travel lanes, 10-ft shoulders, and a design speed of 70 mph. The design may also include alternating passing lanes and sometimes two-way left-turn lanes.

These designs have been proposed by some DOTs as a substitute for more costly conversion to or construction of four-lane divided roadways, while offering primarily capacity and level of service improvements. Although the "Super Two" design is not promoted as a safety improvement, it has been acknowledged that safety benefits might be expected for these designs when compared with conventional two-lane rural highways. Moreover, agencies employing this strategy may spend less than the alternative (reconstructing a corridor to a four-lane divided facility).

The combination of alignment and cross section is intended to minimize the potential adverse effects of cross-centerline conflicts and to reduce excursions onto the roadside. However, it is noted that such designs employ high design speeds (75 mph, or 120 km/h). As such, their application would be incompatible with other strategies intended to lower speeds and reduce crash severity.

Implementation of these wider cross sections involves reconstruction of a road or construction on new alignment. As such, this strategy is clearly among the higher cost of those considered. Moreover, implementing the design change typically entails the environmental process and often includes right-of-way acquisition, both of which by their nature typically require a substantial timeframe. From project inception through planning, design, and construction, a 5- to 10-year timeframe is typical for such projects.

Given the direction by AASHTO to focus on low-cost, short-term-oriented strategies, the use of these designs would appear to be outside the scope of the AASHTO Strategic Highway Safety Plan. However, it is included here since it is a new approach to highway design, having apparent safety benefits, being considered by states. States may wish to consider this strategy within the context of an overall review of their design process for their two-lane system.

The design is currently being used in an experimental or pilot mode. There has not been sufficient experience with this design, nor has the design been adequately evaluated, for the design to be used in any significant way. Pilot testing should be employed before this design is used widely. As results of current evaluations become available, the efficacy of this strategy will become better known.

EXHIBIT V-5
Strategy Attributes for "Wider Cross-Section" Designs (E)

Technical Attributes

Target

Drivers on two-lane roads susceptible to head-on crashes. (It could apply to other types of crashes as well.)

Expected Effectiveness

There are very limited North American data available on the safety effectiveness of these wider cross-section designs. The Iowa DOT, a proponent of the "Super Two" design, estimates that overall safety performance will be between that of a four-lane divided road and that of a typical two-lane highway. Some European studies, however, have not indicated the same positive effects. These two examples provided will need further experimentation and review. Therefore, this strategy is a likely candidate for pilot implementation only at this time.

Keys to Success

The keys to success will be sound evaluations that can define the safety-related effectiveness. This suggests the need for prototype studies conducted under carefully monitored conditions using a properly designed evaluation study. Other keys to success include establishing a set of acceptable and proven design standards to promote use of the safest design, developing state "champions" to sell the idea to other states if effective, and incorporating the design into the standard design process. Other keys will include gaining acceptance by the public and local stakeholders, many of which have expressed concerns over the "Super Two" concept, preferring instead four-lane divided highway improvements.

Potential Difficulties

The major pitfalls are cost and time to implement. This strategy involves complete reconstruction, unlike some other strategies. Note also that this strategy would be impractical in geographies with difficult terrain (mountainous and heavily rolling) due to the cost of providing 100-percent passing sight distance and high design speeds. Finally, if the higher-speed design is implemented without increases in roadside clear zone, there may be an increase in serious run-off-road crashes.

Appropriate Measures and Data

In the evaluation of these wider cross-section designs, process measures would include the number of road miles or number of hazardous locations where these cross sections are implemented. Impact measures would include the number of head-on crashes reduced at these locations, along with any other changes in total crashes (either positive or negative).

If Super Twos or other wider cross-section designs are shown to improve safety in relation to other alternatives (e.g., four-lane divided or four-lane undivided designs), there will also be a need for developing better "targeting" tools—guidelines for when the design is the best alternative.

Associated Needs

Since this is a new design that will "compete with" traditional multilane designs, which the public feels are good, there will be a need for a program to inform the public why this design is being implemented.

Where the designs have been proposed, public reaction has been mixed or negative. The public and business communities tend to lobby for or support projects that reconstruct or replace a two-lane road with a multilane highway. The latter are viewed positively in terms of the high and consistent level of service. Super Twos and other wider cross-section designs are not viewed favorably in this respect.

Agencies must be careful not to send mixed messages to the public regarding Super Two and other wider cross-section designs. DOTs often sell the safety benefits of multilane highways in rural areas; thus, their substitution of Super Two designs for the traditional multilane design may not be understood or may be viewed as accepting a lesser level of safety. Indeed, there is no evidence that such designs are as safe as multilane divided designs, and they are not likely to be as safe given the lack of a median dividing opposing flows.

Organizational and Institutional Attributes

Organizational, Institutional, and Policy Issues

Since this is a new design for major realignments, its implementation depends on an innovative state design agency. Since implementation more than likely includes the environmental process, a cooperative planning process involving federal and state regulatory agencies as well as the state DOT will be needed.

Issues Affecting Implementation Time

Implementation would greatly depend on the project limits, the specific circumstances described above, and funding available. This would probably involve between 5 and 10 years.

Costs Involved

Because this strategy involves complete reconstruction, costs can vary greatly depending on the specific needs at each site and the length of the wider cross section. In general, the costs of constructing a two-lane rural highway with high-quality geometry are on the order of $1­3 million per mile, exclusive of right-of-way costs.

Training and Other Personnel Needs

There appear to be no special personnel or training needs for implementing this strategy.

Legislative Needs

None identified.

Other Key Attributes

Until proven otherwise, this design should be considered inferior to certain other construction strategies (construction of four-lane divided highways) that would be expected to produce lower fatal crash experience. Thus, there would appear to be a clear tradeoff between levels of safety and construction and right-of-way costs.

Information on Agencies or Organizations Currently Implementing this Strategy
Iowa, Minnesota, Washington, and Kansas have used wider cross-section designs in certain locations. Usage is relatively new in each state. While Iowa, Minnesota, and Kansas have not noted any major problems, Washington noted significant problems with subsequent head-on crashes and converted the design to one involving alternating passing lanes. Information on Washington's, Iowa's, and Minnesota's programs can be found in Appendix 5, Appendix 6, and Appendix 7, respectively. Finally, Texas has recently asked the Texas Transportation Institute at Texas A&M to study this design for possible future use in Texas.

Strategy 18.1 A4—Center Two-Way, Left-Turn Lanes on Four-Lane and Two-Lane Roads

General Description
This strategy involves the development of two-way, left-turn lanes (TWLTLs) on existing roadways. It can be accomplished either by the conversion of four-lane undivided arterials to three-lane roadways with a center left-turn lane or by the more conventional reconstruction of a two-lane road to include the TWLTL. Since the latter could be a costly conversion because it may require new right-of-way, the four-lane road conversion is considered more appropriate to the AASHTO emphasis on low-cost alternatives. However, where right-of-way cost is not a major consideration, the inclusion of TWLTLs on existing two-lane roads may be an even more effective treatment for head-on collisions since more of such collisions would likely occur on two-lane roads than on four-lane roads.

The development of TWLTLs is usually for traffic operations rather than safety concerns. TWLTLs are usually implemented to improve access. When they are used in response to a safety concern, the use is traditionally to reduce driveway-related turning and rear-end collisions. However, because studies have also indicated a positive effect on head-on crashes, the strategy is included here. The principle behind the use of TWLTLs in this context is to provide a buffer between opposing directions of travel. The strategy is intended to reduce head-on crashes by keeping vehicles from encroaching into opposing traffic lanes through the use of the buffer.

It is also noted that the conversion of urban four-lane undivided streets to two through lanes with a TWLTL is now often referred to as a "road diet" or "street diet." There is a growing body of information on these urban conversions, which are usually done to better accommodate pedestrians, bicyclists and other road users through the resulting decreases in travel speeds and less hazardous street crossings (see Knapp and Giese, 2001, for example). However, since head-on collisions are a more significant problem in rural areas, this section focuses on rural conversions/additions to TWLTLs.

While there will be very few passing-related, head-on crashes on undivided four-lane roads, head-on collisions can result from a vehicle inadvertently leaving its lane and crossing the centerline or from a vehicle attempting to turn left across oncoming traffic. (While this crash could be termed a "turning" crash, it is sometimes termed "head-on," and the severity can be the same as a head-on crash.) The strategy can reduce head-on collisions in two ways. First, turning-related, head-on crashes are made from the left-turn lane rather than a through lane, giving the driver a more protected (and thus less pressured) location to make judgments concerning acceptable gaps in the oncoming flow. Second, the center left-turn lane provides a "clear zone" or median between opposing vehicles, allowing drivers to leave their lanes but to recover and return safely.

This strategy, although fairly widely used, has not been sufficiently evaluated to be considered "proven." It should be compatible with other strategies for head-on and ROR crashes (e.g., shoulder rumble strips, improved pavement markings, and horizontal curve improvements). With respect to the driving public, it should be particularly helpful to local drivers who must turn left across traffic into private or business driveways. It may also be compatible with improving the safety of unsignalized intersections and improving the safety of pedestrians by reducing the number of traffic lanes to be crossed.

EXHIBIT V-6
Strategy Attributes for Conversion of Four-Lane Undivided Arterials to Three-Lane TWLTL Designs

Technical Attributes

Target

Drivers crossing the centerline of an undivided multilane roadway inadvertently and drivers attempting to make a left turn across oncoming traffic. Roads to be targeted would be those where traffic volumes make it feasible to reduce the number of lanes to two through lanes (generally 15,000 vehicles per day [vpd] or less), or where alternative parallel routes exist for diversion of traffic.

Expected Effectiveness

A number of publications provide indications of the effectiveness of this strategy. NCHRP Report 282 (Harwood, 1986) and NCHRP Report 420 (Gluck et al., 1999) looked at access management strategies including three-lane designs. But as with other countermeasures, the road sections used in the analysis were not restricted to rural two-lane roads. In fact, NCHRP Report 282 uses traffic volume, speed, spacing of intersections, number of access points, on-street parking, and area population to identify road sections labeled as suburban arterial highways. More analysis specific to quantifying the effectiveness of cross-sectional design alternatives on rural two-lane roads is needed. Nevertheless, some of the statistics and conclusions in those reports are summarized below to show the effectiveness of such roadway conversions on other roadway types.

NCHRP Report 282 uses data from California and Michigan to establish a database for suburban highways. Using this database, accident rate estimates (accidents per million vehicle miles) were obtained for various multilane design alternatives. The table below displays the average accident rate for nonintersection accidents. The report also gives further adjustment factors to these rates based on truck percentage, number of driveways per mile, and shoulder width. These adjustment factors are not displayed below.

Basic Accident Rates for Five Design Alternatives (Accidents per Million Vehicle Miles)

 

Type of Development

Design Alternative

Commercial

Residential

Two-lane, undivided

2.39

1.88

Three-lane, TWLTL

1.56

1.64

Four-lane, undivided

2.85

0.97

Four-lane, divided

2.90

1.39

Five-lane, TWLTL

2.69

1.39

From this table, the commercial three-lane TWLTL (3T) design accident rate is 45 percent less than the four-lane undivided (4U) design accident rate. Thus, 45 percent is a reasonable estimate for the accident reduction effectiveness of altering a 4U design to a 3T design. However, engineering judgment and specific location attributes should also be considered when estimating the accident reduction benefit.

Two other study results indicate that there seems to be a roadway volume threshold where the increase in delay is minimal compared with the resulting accident reduction. A 1978 study found that installing a 3T design on a highway with an existing 4U design and an average daily traffic (ADT) of 16,000 vpd resulted in an increase in delay because of the reduction of through lanes. However, a 1981 study conducted on a facility with a lower traffic volume found no increase in delay and a substantial reduction in accidents.

NCHRP Report 420 states that highway facilities with TWLTLs had accident rates that were, overall, roughly 38 percent less than those experienced on undivided facilities (Gluck et al., 1999). This percentage is calculated using statistics from 12 studies conducted since 1970. The three studies that evaluated the conversion of four-lane undivided facilities to three-lane facilities with TWLTLs specifically calculated the following reductions: 9 percent decrease in accident rate, 28 percent decrease in accident rate, and 40 percent decrease in number of accidents.

A study by Fitzpatrick and Balke (1995) compared the accident rates of rural, four-lane roadways with TWLTLs and similar roadways with flush medians. They found no significant difference in the accident rates of the two median treatments in rural areas, when the roadways had comparable speed limits and driveway densities. Although this study looked at four-lane roadways and not two-lane roadways with TWLTLs or flush medians, there would seem to be a similarity between the two. Therefore, the crash experience of this design would likely be comparable to a roadway with a flush "buffer median." Refer to Strategy 18.1 A5 for more information on buffer medians.

In recent work related to the Interactive Highway Safety Design Module for two-lane rural roads, an expert panel developed a series of accident modification factors (AMFs) based on a critical review of available literature. The AMF for adding a TWLTL was based on the review by Hauer (1999) and was found to be a function of access point density (i.e., APD, the number of driveways and unsignalized intersections per mile) and the number of access-point-related crashes. The addition of the turn lanes was found to be beneficial only where there were more than 5 access points per mile (3 points per kilometer). For higher densities, the AMF (which can be multiplied by the existing mean number of total crashes on the section) is as follows:

AMF = 1 ­ 0.07 PAP PLT/AP,

Where

PAP = access-point-related crashes as a proportion of total crashes

= (0.0047APD ­ 0.0024APD2) / (1.199 + 0.0047APD + 0.0024APD2)

PLT/AP = left-turn crashes susceptible to correction by the TWLTL as a proportion of access-point-related crashes. This is estimated as 0.5.

In summary, there are a number of studies summarizing the effects of a conversion from the current cross-section geometry to inclusion of a TWLTL. However, this strategy cannot be considered a proven strategy because there are no truly valid estimates of the effectiveness of such conversions based on sound before/after studies for a two-lane road. There have been no significant findings of negative effects of this strategy. Precise estimates of effectiveness should be developed in well-designed evaluations of pilot conversions.

Keys to Success

Successful application of this strategy will be based on the ability to identify road segments with sufficiently high speeds and traffic volumes such that serious head-on crash frequencies are significant, yet traffic volumes are low enough that operation with one through lane in each direction is feasible. If inappropriate sites are selected for this countermeasure, illegal passing in the TWLTL may occur as a result of eliminating the passing possibility. It may also require communicating the benefits of such improvements to the public and business owners, who may view such improvements as reducing the capacity of the roadway.

Potential Difficulties

Using this strategy in locations with traffic volumes that are too high could result in diversion of traffic to routes less safe than the original four-lane design. It may also result in congestion levels that contribute to other crashes. While a crash reduction might occur on the modified roadway, this diversion could result in an overall increase in crashes in the system.

Other pitfalls concern public and business owners' acceptance of a plan that appears to reduce the highway's traffic-carrying capacity. Also, the relative safety of unsignalized intersections along the route may be influenced adversely. By consolidating through traffic into one lane in each direction, the potential exists to decrease the number of gaps and increase the amount of risk taking by drivers queued at stop signs on minor approaches. Although the number of gaps may be reduced, thereby increasing the risk, the gaps will become less complex since drivers are now dealing with one lane in each direction and the TWLTL can be used as a refuge and for acceleration. Thus, the potential pitfall of a decrease in the number of gaps may be countered by the decrease in the complexity of the gap.

Appropriate Measures and Data

In implementation evaluations, process measures would include the number of road miles or number of hazardous locations where such conversions are implemented.

Impact measures will include the number of head-on crashes reduced at these locations, along with any change in total crashes.

Associated Needs

A public information effort would be needed to educate both the normal commuting drivers and the roadside businesses concerning the effects of this strategy.

Organizational and Institutional Attributes

Organizational, Institutional and Policy Issues

This strategy would be implemented by the state DOT or local highway agency, and it does not appear that extra coordination with other agencies would be needed. However, since this would affect access to neighborhood businesses and residences (and should be advantageous to both), a planning/public hearing process that involves these groups should help facilitate treatment implementation.

No immediate policy changes are required. Should this strategy prove effective, DOT and other agencies may find it advantageous to adjust their design policies to incorporate three-lane designs as alternatives to four-lane undivided roadways for certain traffic conditions.

Issues Affecting Implementation Time

Implementation would require more time than in some low-cost treatments. However, assuming no new right-of-way would be required, environmental analyses and documentation would be minimal and readily accomplished, and the modification would require only new lane markings. This strategy could be implemented within a short time period.

Costs Involved

Costs should be relatively low since this strategy involves only re-marking of existing pavement and minor improvements to markings and signalization at intersections.

Training and Other Personnel Needs

There appear to be no special personnel or training needs for implementing this strategy.

Legislative Needs

None identified.

Other Key Attributes

None Identified

Information on Agencies or Organizations Currently Implementing this Strategy
Iowa DOT is implementing this strategy on several urban arterial streets and on short sections of main highways through local communities. While the research team is not aware of other states developing such center-turn lanes through modification or addition in rural areas, this basic TWLTL design has been used for a number of years.

Strategy 18.1 A5—Reallocation of Total Two-Lane Width (Lane and Shoulder) to Include a Narrow "Buffer Median"

General Description
Head-on fatalities are affected both by the number of vehicles that cross the centerline and by the speed of oncoming vehicles. A particularly effective strategy might be affecting both factors by reallocating the existing cross section—narrowing lanes to encourage slower speeds while incorporating a narrow buffer median between opposing flows. For example, a high-speed rural two-lane roadway with a cross section consisting of 12-ft lanes and 10-ft paved shoulders could be restriped to provide narrower shoulders (e.g., 8 ft) or slightly narrower lanes (e.g., 11 ft), with the difference forming a 6-ft flush median divider. The median could include milled-in centerline rumble strips to help prevent inadvertent crossings.

Any number of combinations of lane and shoulder width could be considered, producing a median separation of 1 ft to 6 ft, and additional paved shoulders could be added if needed. Recent research suggests that total roadway width (lanes and shoulders) influences the safety of two-lane roads. Review of existing empirical evidence by Hauer indicates that the safety performance of roadways with 11-ft lanes is about as good as or slightly better than that of roadways with 12-ft lanes (Hauer, 2001). Indeed, in Europe, the concept of "optical narrowing" of lanes is employed to slow travel speeds.

The strategy would be most effective where reallocation of roadway width could be accomplished without degrading the roadside clear zone. As noted in the section concerning current implementation following the table below, this strategy has only been implemented in limited cases and should be considered experimental at this point. However, as will be noted there, both Pennsylvania and Maryland have implemented similar strategies with promising results. Information on Maryland's program can be found in Appendix 1.

This strategy, although tried in several locations, has not been sufficiently evaluated to be considered "proven." This design should be compatible with shoulder rumble strips, curve improvements, and other two-lane treatments (except shoulder widening). However, this strategy would be incompatible with ROR strategies if its implementation required narrowing the shoulder or otherwise reducing the quality of the roadside.

EXHIBIT V-7
Strategy Attributes for Reallocation of Total Width to Provide Median Buffer

Technical Attributes

Target

Drivers of vehicles who unintentionally cross the centerline.

Expected Effectiveness

In development of this guide, evaluation studies documenting the effectiveness of the narrow buffer medians without barriers on two-lane roads were not found. The strategy appears promising, however, given Pennsylvania's experience with a similar treatment of converting a two-lane roadway section with a passing lane to a two-lane roadway with a buffer median, including centerline rumble strips.

Maryland did implement a buffer-median treatment, including a median guardrail on sections of two access-controlled, two-lane state routes. The before/after analyses indicated a reduction in total and opposite direction (head-on) crash rates of approximately 50 percent. However, it must be noted that these sections were "high-accident" sections prior to the treatment, meaning that the studies could be subject to "regression to the mean" bias. In addition, the number of head-on crashes studied was usually quite small, ranging from one to five per year. Finally, positive improvements in severity were not measured.

Clearly, these results are promising; however, additional evaluation is needed before this strategy can be considered a proven strategy. Measurable effectiveness would seem to be achievable. Additionally, there have been no significant findings of negative effects of median buffers.

Keys to Success

The initial key to success would be experimental use and careful evaluation to determine effectiveness, cost, and adverse consequences. If effective, there will need to be effort toward identifying the characteristics of the roadway (traffic volume, speed, alignment quality, cross section) for which this buffer median may be expected to have the greatest positive effect.

Potential Difficulties

Given that this is a new and somewhat unproven design, the major difficulties will be objections raised by highway agencies such as the potential of lessening the public's willingness to pass, difficulty with snow removal, loss of shoulders for disabled vehicles, and undesirable noise levels (associated with the rumble strips). A sample of state DOT representatives who reviewed this concept suggested that the buffer median would be much less effective if only delineated by paint. They suggested the use of rumble strips in the median at a minimum. However, this was expressed opinion, rather than based upon any solid evaluation. No studies regarding the effectiveness of buffers with either treatment have surfaced.

There is also a concern that traffic operations would be worsened if the buffer (with rumble strips) inhibits passing maneuvers or if passing zones are eliminated because of buffer painting. Again, experience is needed to see if this is a problem.

Finally, Iowa has experienced some problems with vehicles using the buffer in preparation for turning left and being sideswiped by passing vehicles. Thus, left-turn lanes should be considered at intersections, and the design should be carefully studied if implemented in locations with high numbers of left turns into driveways.

Since there is cost involved with both restriping and rumble strip placement, this measure may be more cost-effectively targeted to the "highest risk" locations rather than to large sections of the two-lane network. Otherwise, the benefits may not exceed the costs.

Appropriate Measures and Data

In program implementation evaluations, process measures include number of road miles or number of hazardous locations where buffer-medians are installed. They may include the aspect of exposure—the number of vehicle miles of travel exposed to medians.

Impact measures will include the number (or rate) of head-on crashes reduced at these locations, along with any change in total crashes.

The strategy will be most effective when a framework and a methodology exist to target the implementation to the most appropriate sites. This will require more detailed analysis of locations with excessive head-on crashes (rather than just total crashes).

Associated Needs

It would be expected that the driving public's view of this strategy would be similar to their view of centerline rumble strips. These may be viewed negatively by motorcyclists. Given the proposed width of the buffer median and the fact that it would include wider centerline rumble strips, it could also be viewed more negatively by the drivers of other vehicles. Thus, there would be the need for public information to explain the function of the treatment in order to address any public concerns or potential misunderstanding.

Organizational and Institutional Attributes

Organizational, Institutional and Policy Issues

These strategies will be implemented by state and local roadway agencies, and it does not appear that extra coordination with other agencies or groups is needed.

Implementing this policy will in many cases run counter to DOT design policy. Conventional thinking that is institutionalized in design policy is that wide lanes are safer and thus preferred over narrow lanes. Narrow lanes are normally considered as design exceptions and avoided. The design exceptions process is rigorously followed in most states and is driven in part by safety concerns and risk management involving potential tort claim actions.

Other related policy issues include the established relationship between lane width and capacity, and state policies emphasizing provision for minimum levels of service.

Acceptance and widespread implementation of this strategy will in many cases occur only with revision to a state's design policy. Acceptance of such a policy change is most likely if the safety benefits are clearly understood and articulated. It would seem that an effective design policy would promote this design treatment for specific, limited circumstances related to the safety concerns of head-on crashes and need for speed reductions.

On a more positive note, there is growing concern and interest in design measures to proactively reduce speeds for certain conditions. Examples may include roadway approaches to small towns or through rural communities. Design agencies thus are beginning to show an interest in measures that effectively reduce speed without changing the fundamental character of the highway.

Issues Affecting Implementation Time

This low-cost strategy does not involve reconstruction, the environmental process, or ROW acquisition. Restriping can be accomplished quickly, but widening of paved shoulders would take longer.

Costs Involved

Costs will vary depending on whether the strategy is implemented as a stand-alone project or whether it is incorporated as part of a reconstruction or resurfacing effort already programmed. Including the necessary striping and rumble strips as part of a resurfacing project offers the opportunity for lowest cost implementation.

In Pennsylvania, the cost to install rumble strips is about $2 a foot ($10,000 a mile), including traffic control while installing. If the strip pattern used in these buffer-medians is wider than the normal centerline rumble strip, then the cost will increase. It also will be increased by the cost of lane restriping.

Training and Other Personnel Needs

There appear to be no special personnel needs for implementing this strategy. States would use either agency personnel or contractors. However, training of state safety and design engineers on the attributes, benefits, and applicability of this treatment would be necessary. Conventional wisdom is that wider lanes are safer, hence overcoming initial skepticism toward a safety improvement that narrow lanes will require special training. Training regarding actual installation of rumble strips accompanying the median buffer will depend on whether the agency has been using retrofitted rumble strips on freeways or other roadways. If not, either agency personnel or contractor personnel will need to be trained in proper installation techniques.

Legislative Needs

None identified.

Other Key Attributes

None Identified

Information on Agencies or Organizations Currently Implementing this Strategy
The DOTs in Iowa, Pennsylvania, and Maryland have implemented this or similar strategies on certain roadway segments. As noted above, Iowa noted some sideswipe problems due to vehicles using the median to stop while turning left (perhaps converting some rear-end collisions to sideswipes) and noted the need for careful transition design to left-turn lanes at intersections. This strategy appears to be similar to a treatment used by Pennsylvania on a 6-mile section of Route 322 in 1993. There, both centerline rumble strips and a narrow "buffer-median" were installed, and in the 6 years since installation, there have been no fatal accidents. Maryland DOT implemented a narrow (4­10 ft) buffer median with median guardrail on two limited-access, two-lane routes (MD 140 and MD 90). Limited before/after results indicated large reductions in head-on and other crash types. More information on Maryland's program can be found in Appendix 1.

Objective 18.1 B—Minimize the Likelihood of Crashing into an Oncoming Vehicle

Strategy 18.1 B1—Alternating Passing Lanes or Four-Lane Sections at Key Locations

General Description
This strategy involves improving two-lane locations that experience many passing-related collisions. It involves constructing either alternating passing lanes or short four-lane sections that allow passing for both flows. While the treatment is designed to reduce passing-related, head-on crashes (a relative low percentage of all head-on crashes), it should also positively affect nonpassing head-on collisions at the treated sections since the passing lanes would provide extra "clear zone" for vehicles inadvertently leaving their through lanes. It may also affect other types of crashes such as rear-end crashes involving a turning vehicle, since the passing lane provides some protection for the left-turning vehicle.

This strategy would be more expensive and take longer to construct since it requires lane construction and would usually require additional right-of-way. However, it is less expensive than full-scale realignment or reconstruction and thus would appear to fit within AASHTO's current framework for this effort.

The construction of passing lanes would be compatible with the other head-on and ROR strategies. They would not be feasible for routes with narrow right-of-way where additional right-of-way cannot be purchased or is too expensive. This strategy may also be compatible with aggressive driving strategies such as mitigating congestion and minimizing frustration with drivers wanting but unable to execute passing maneuvers.

This strategy, although fairly widely used, has not been sufficiently evaluated to be considered "proven."

EXHIBIT V-8
Strategy Attributes for Providing Alternating Passing Zones on Two-Lane Highways (T)

Technical Attributes

Target

(1) Vehicles involved in head-on collisions on undivided two-lane roads during passing maneuvers. (2) Vehicles involved in nonpassing head-on crashes.

Expected Effectiveness

Based upon work by Harwood and St. John (1984), Rinde (1977), and Nettleblad (1979), as reviewed by the expert panel that developed two-lane Accident Modification Factors for the Interactive Highway Safety Design Model, a one-way passing lane could reduce total (not just head-on) crashes by 25 percent for the length of the installation. Based upon the same review by Harwood and St. John, the short four-lane sections allowing passing in both directions simultaneously are estimated to reduce total crashes by 35 percent for the length of the passing zone. It is noted that both of these estimates may be somewhat inflated since the effect of simultaneous shoulder treatments could not be completely removed in the effectiveness evaluation.

NCHRP Report 440 summarizes the result of a study that examined how sections of roadway treated with passing improvements compared with untreated sections (Fitzpatrick et al., 2000). The results, shown below, are for two-lane roads in rural or suburban areas and the reductions range from 25 percent to 40 percent. However, the report cautions that readers should use engineering judgment regarding the crash effects of these alternatives because crash experience varies greatly depending upon specific traffic and site characteristics.

Percent Reduction in Crashes for Four Design Alternatives
 

Type of Crash

Design Alternative

Total Crashes

Fatal and Injury Crashes

Passing lanes

25

30

Short four-lane section

35

40

Turnouts

30

40

Shoulder use section

*

*

* No known significant effect.

Passing lanes are known to have traffic operational effects that extend 5 to 13 km (3 to 8 miles) downstream of the passing lane. It might be presumed that these operational effects provide analogous safety benefits over a similar length of highway. However, since this effect has not been quantified, it is not included in the estimate.

It is important to carefully consider the cautions given in each of the studies cited above. Although these studies seem to indicate a reduction in crashes, more analysis of such treatments is required before a statistically sound measure of the effectiveness can be developed. This treatment has not been sufficiently evaluated to be considered a proven strategy. However, there have also been no significant findings of negative effects of this strategy.

Keys to Success

The effectiveness of this strategy would be maximized on routes with significant unmet demand for safe passing. This would include routes with a range of driving speeds (e.g., hilly terrain with significant truck traffic) and a horizontal and vertical alignment providing few safe-passing opportunities. Thus, one key to success is the agency's ability to identify these locations.

Potential Difficulties

This strategy would require significant construction and may involve right-of-way costs.

Some states that have implemented three-lane passing sections (i.e., a passing lane in one direction) have noted a potential problem with downhill passing across the yellow "no passing" centerline. This occurs when a vehicle in the no-passing, downhill flow can see a safe passing gap and passes across the "no-passing" centerline. Pennsylvania notes that in some states, there is an enforcement issue in that the police may not consider crossing the double-yellow centerline as sufficient proof of an infraction. In addition, some states allow this passing maneuver legally. It can lead to driver confusion. State DOT staff interviewed indicated that they did not feel that allowing such passing was a good idea. The Michigan DOT is currently adding a fourth lane to downhill sections that have a passing lane for one direction to eliminate the potential problem.

Minnesota has implemented some short four-lane passing sections and has noted some problems with left-turning traffic at intersections. Minnesota DOT staff suggests transitioning the four-lane sections back to two-lanes prior to an intersection and possibly implementing a left-turn lane at the intersection.

Appropriate Measures and Data

In implementation evaluations, process measures would include the number of road miles, or number of hazardous locations, where such passing zones are constructed. Impact measures would include the number of head-on and total crashes reduced at these locations.

Traffic operations data and crash information will be needed to define sections with unmet safe-passing demand.

Associated Needs

Unless this is a new treatment in a given state, there should not be significant public "training" needs. However, since the public assumes that conversion to a multilane configuration is always better, a public information program explaining the benefits and relative costs of this treatment could be helpful in selling the implementation.

Organizational and Institutional Attributes

Organizational, Institutional and Policy Issues

These strategies will be implemented by state and local roadway agencies, and it does not appear that extra coordination with other agencies or groups is needed. No new policy efforts are required.

Issues Affecting Implementation Time

Since this would require reconstruction (and usually right-of-way acquisition), the time required would be longer than for other less-involved strategies.

Costs Involved

The costs involved would depend on the number and length of locations treated and whether right-of-way must be purchased. In general, this strategy would be significantly more costly than some others (e.g., rumble strips, marking and delineation), but less costly than conversion to four lanes, or to a Super Two design.

Training and Other Personnel Needs

There appear to be no special personnel needs for implementing this strategy, since state or contractor personnel would implement it. Since this treatment is not new, it should not require significant training other than providing information to design and construction engineers concerning the benefits of the treatment.

Legislative Needs

None identified.

Other Key Attributes

None Identified

Information on Agencies or Organizations Currently Implementing this Strategy
A number of states have employed passing lanes. More specifically, both the Washington and Pennsylvania DOTs have implemented passing lanes and wider shoulders to allow passing at certain locations. Michigan has added passing and climbing lanes to increase capacity (rather than as a safety treatment) for a number of years. Because of the earlier noted potential problem with downhill passing across the yellow "no-passing" centerline, Michigan is currently adding a fourth lane to downhill sections that have a passing lane for one direction. Therefore, Michigan can provide guidelines on these installations. Minnesota has installed some short four-lane passing sections, and it can provide guidelines on the length of the passing lanes based on AASHTO information.

Strategy 18.1 B2—Median Barriers for Narrow Medians on Multilane Roads

General Description
Median Barrier Application in Massachusetts
EXHIBIT V-9
Median Barrier Application in Massachusetts
This strategy involves providing barriers on multilane roads with narrow or no medians. Barriers can be rigid (e.g., concrete median barrier, guardrail) or semi-rigid (e.g., cable barrier). The treatment would be designed to prevent head-on collisions from occurring. The treatment is also used on high-speed, two-lane roads during construction (e.g., during freeway reconstruction, both directions of traffic are often shifted to one roadway, with temporary barriers provided between the opposing traffic).

The strategy is primarily applicable in the rural or outlying suburban environment where speeds are higher and the need for median openings for intersections and driveways are less than in urban areas. Arterials, expressways, and full freeways are candidates for treatment.

The strategy would apply to roadways that may have experienced significant traffic growth and increased serious crashes since original construction. In many cases, original design assumptions on speed and traffic demands resulted in decisions to forego median barriers.

Finally, this design may be incompatible with other strategies designed to minimize ROR crashes. A continuous median barrier is itself a potential hazard. As many as 30 percent of high-speed barrier impacts produce injuries and fatalities.

This strategy, although fairly widely used, has not been sufficiently evaluated to be considered "proven."

EXHIBIT V-10
Strategy Attributes for Providing Median Barriers on Multilane Roads (T)

Technical Attributes

Target

Drivers of vehicles who unintentionally cross the centerline.

Expected Effectiveness

There are various studies that describe the crashworthiness of a number of different median barriers. The points below summarize the quantitative results for two studies identified in NCHRP Synthesis of Highway Practice 244 (Ray and McGinnis, 1997).

  • A study performed in the late 1960s using New York data found that the injury and fatality rate for weak-post barriers was 10 percent versus the 20 percent rate for strong-post barriers.
  • The results below evaluate the performance of median barriers using Longitudinal Barrier Special Studies (LBSS) data. The findings support conventional wisdom that the barriers that allow more lateral deflection result in less severe collisions. However, the percentage of vehicles being redirected, snagging, or penetrating increases with the weak-post system.

Strategy Attributes for Providing Median Barriers on Multilane Roads

 

Median Barrier Type

 

Weak-Post

Strong-Post

Concrete

Other

Injury or Fatality (%)

8.8

17.5

16.2

11.5

Redirect (%)

82

88

91

78

Snag (%)

12

5

0

7

Penetrate (%)

3

5

5

15

The statistics listed above highlight a few types of median barriers but do not give insight regarding the safety of a roadway with and without a median barrier. Statistically sound studies are still needed to produce effectiveness measures and to consider this strategy a proven strategy. However, this strategy was tried and accepted in a number of applications. Additionally, there have been no significant findings that the effects of striking the barrier will have a worse result than the head-on collision it is designed to prevent.

Keys to Success

Initial keys to success would be experimental use and careful evaluation to determine effectiveness, cost, and adverse consequences. If effective, effort will be directed at identifying the characteristics of the roadway (traffic volume, speed, alignment quality, cross section) for which median barriers may be expected to have the greatest net positive effect. In addition, many barriers require clear area behind the rail for deflection. If this alternative is to address narrow medians, then careful consideration must be made to ensure that the median barrier installed has the necessary clear area. NCHRP Synthesis of Highway Practice 244 reports that weak-post cable barriers require about 11 ft (3.3 m) behind the rail. Such a requirement limits the use of this barrier on narrow median roadways.

Potential Difficulties

Many states are reluctant to implement barriers given the uncertain net safety performance and maintenance problems created. In some cases, implementation of concrete barriers would require closing up of the median and construction of expensive closed drainage systems. Less expensive cable barrier systems also are viewed as requiring high levels of maintenance.

DOT staff from both Michigan and Minnesota has expressed some concern with possible end treatments at intersections—i.e., what attenuator can be used in these narrow medians. Minnesota also identified some potential problems with sight distance decreases at intersections due to the presence of the barrier.

Finally, barriers in northern climates present snow removal and storage problems, particularly where medians are narrow.

Appropriate Measures and Data

In implementation evaluations, process measures include number of road miles or number of hazardous locations where median barriers are installed. They may include the aspect of exposure—the number of vehicle miles of travel exposed to medians.

Impact measures will include the number (or rate) of head-on crashes reduced at these locations, along with any change in total crashes.

The strategy will be most effective when data and an analysis methodology exist to target the implementation to the most appropriate sites—a methodology that identifies sites based on head-on rather than total crashes.

Associated Needs

There do not appear to be any special needs. There should not be a need for any public information and education, since the motoring public is familiar with median barriers.

Organizational and Institutional Attributes

Organizational, Institutional and Policy Issues

These strategies will be implemented by state and local roadway agencies, and it does not appear that extra coordination with other agencies or groups is needed.

If the state does not have a policy defining median width and/or traffic characteristics where barriers are to be installed, one may be needed. Most states use median barrier warrants similar to those published in AASHTO's Roadside Design Guide. Such warrants may need to be adjusted or refined.

Issues Affecting Implementation Time

This moderate cost strategy would in many cases be readily implementable within a 1- to 3-year period after site selection. Barrier design and placement within existing narrow medians would require no right-of-way, a minimal environmental process, and generally one construction season.

Costs Involved

Costs will vary depending on whether the strategy is implemented as a stand-alone project or whether it is incorporated as part of a reconstruction or resurfacing effort already programmed. However, NCHRP Synthesis of Highway Practice 244 summarized survey results of 39 states to quantify the typical installation cost for roadside and median barriers (Ray and McGinnis, 1997). See Appendix 9.

When evaluating the cost-effectiveness of a particular type of median barrier, it may be more appropriate to consider the life-cycle cost of the barrier. This takes into account both the life span and the expected maintenance cost of the barrier. Both of these items are important considerations in overall cost-effectiveness.

Training and Other Personnel Needs

There appear to be no special personnel or training needs for implementing this strategy, assuming that an agency deploys its standard barrier treatments. States would either use agency personnel or contractors.

Legislative Needs

None identified.

Other Key Attributes

None Identified

Information on Agencies or Organizations Currently Implementing this Strategy
State DOTs currently using or exploring this strategy include Minnesota, Washington, Pennsylvania, Maryland, Michigan, Massachusetts, and Iowa. In addition, the city of Seattle, Washington, has installed movable median barriers in very narrow medians to reduce head-on crashes. Seattle chose the movable barrier because of the barrier's narrow footprint. A number of states, including California and Michigan, are conducting studies and implementing median barrier projects on high-volume roads. Indications from the work in California are that barriers appear to be warranted at greater median widths than where used previously, suggesting that the magnitude of the utility of this strategy may be significant.

Key References

Alexander, H. B., and P. A. Pisano. "An Investigation of Passing Accidents on Two-Lane, Two-Way Roads." Public Roads, Volume 56, Issue 2, 1992.

"The Effectiveness and Use of Continuous Shoulder Rumble Strips." Accessed July 10, 2000. http://safety.fhwa.dot.gov/fourthlevel/pro_res_rumble.effect.htm.

Fitzpatrick, K., and K. Balke. "Evaluation of Flush Medians and Two-Way, Left-Turn Lanes on Four-Lane Rural Highways." Transportation Research Record 1500, Transportation Research Board of the National Academies, Washington, D.C., 1995, pp. 146­152.

Fitzpatrick, K., K. Balke, D. W. Harwood, and I. B. Anderson. NCHRP Report 440: Accident Mitigation Guide for Congested Rural Two-Lane Highways, Transportation Research Board of the National Academies, Washington, D.C., 2000.

Gluck, J., H. S. Levinson, and V. Stover. NCHRP Report 420: Impacts of Access Management Techniques, Transportation Research Board of the National Academies, Washington, D.C., 1999.

Harwood, D. W., and A. D. St. John. Passing Lanes and Other Operational Improvements on Two-Lane Highways. Report No. FHWA-RD-85-028, Federal Highway Administration. July 1984.

Harwood, D. W., F. M. Council, E. Hauer, W. E. Hughes, and A. Vogt. Prediction of the Expected Safety Performance of Rural Two-Lane Highways. Report No. FHWA-RD-99-207, Federal Highway Administration. December 2000.

Harwood, D. W., and C. J. Hoban. Low Cost Methods of Improving Traffic Operations on Two-Lane Roads, Report No. FHWA-IP-87-2, Federal Highway Administration. January 1987.

Harwood, D. W. NCHRP Report 282: Multilane Design Alternatives for Improving Suburban Highways, Transportation Research Board of the National Academies, Washington, D.C., 1986.

Harwood, D. W., F. M. Council, E. Hauer, W. E. Hughes, and A. Vogt. Prediction of the Expected Safety Performance of Rural Two-Lane Highways. Report No. FHWA-RD-99-207 Federal Highway Administration. December 2000.

Hauer, E. "Lane Width and Safety: Review and Interpretation of Published Literature." Accessed May 23, 2001. http://www.roadsafetyresearch.com.

Hauer, E. "Two-Way Left-Turn Lanes: Review and Interpretation of Published Literature." Unpublished (author can be contacted at the University of Toronto, Department of Civil Engineering). 1999.

Knapp, K. K., and K. Giese. Guideline for the Conversion of Urban Four-Lane Undivided Roadways to Three-Lane Two-Way Left-Turn Lane Facilities, Final Report. Center for Transportation Research and Education, Iowa State University, April 2001. Available at http://www.ctre.iastate.edu/reports/4to3lane.pdf.

Mohamedshah, Y. M. "Investigation of Passing Accidents Using the HSIS Data Base." Public Roads, Volume 56, Issue 2, 1992.

Nettelblad, P., "Traffic Safety Effects of Passing (Climbing) Lanes: An Accident Analysis Based on Data for 1972­1977," Meddelande TU 1979-5, Swedish National Road Administration. 1979.

Ray, M. H., and R. G. McGinnis. NCHRP Synthesis of Highway Practice 244: Guardrail and Median Barrier Crashworthiness, Transportation Research Board of the National Academies, Washington, D.C., 1997.

Rinde, E. A. Accident Rates vs. Shoulder Width. Report No. CA-DOT-TR-3147-1-77-01, California Department of Transportation. 1977.