Guidelines for Selection of Speed Reduction Treatments at High-Speed Intersections: Supplement to NCHRP Report 613 (2008)

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.

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page2-1 2 Preliminary Findings 2.1 SPEED CONSIDERATIONS 2.1.1 Introduction Improving safety has long been a goal of the FHWA, NHTSA, AASHTO, ITE, and safety advocacy groups such as the Insurance Institute for Highway Safety. With over three million intersections (300,000 signalized) and nearly 9,000 intersection fatalities per year in the United States, improving safety at roadway junctions has received increased focus and energy. High- speed environments increase the potential for more fatalities and injury severity for motorists and non-motorists alike. While much is often assumed regarding the role of speed at intersections, little data exist that isolate the effects of speed on overall intersection performance (safety, operations, and ability to serve all modes). Speed is a product of many roadway and intersection features and, in turn, speed affects the performance of roadway facilities and the quality of adjacent environments. Speed reduction does not necessarily guarantee safety, operational, or environmental benefits; rather, the specific conditions of an intersection must be considered to determine what speeds are desirable for that particular location and environment. Speed may be deemed “excessive” when drivers do not have sufficient time to react to and safely navigate around interruptions in the flow of traffic or to adapt their operations to the current conditions at an intersection. Excessive speeds generally result when environmental and operational elements are incompatible, sending motorists a mixed message about appropriate behavior. Excessive speed may result when a driver misinterprets the tasks needed to operate safely. In some cases, excessive speed may be a deliberate result of driver attitude, risk assessment, and behavior. The conditions at an intersection may require an operating speed that is slower than required by the conditions of the adjacent roadway segments. Defining the intersection influence area and the transition area is necessary to identify the area within which speed reduction treatments are needed. The following section focuses on the role of speed in an intersection environment and discusses the ways in which speed affects intersection performance and the adjacent environment. It details ways in which speed is affected by roadway design and elements of the adjacent environment and highlights some physical conditions and user characteristics that may make an intersection particularly sensitive to speed. 2.1.2 Intersection/Segment Relationship Defining the influence area of an intersection is fundamentally necessary to differentiate between reducing speeds on the segment rather than the intersection proper. As described in the Guidelines for Selection of Speed Reduction Treatments at High-Speed Intersections, which was developed as part of this research effort, intersections can be defined by geometric and operational influence areas:

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-2 • Geometric: The location where the typical section of the roadway segment is modified to create the intersection features. These modifications include tapers for adding or dropping lanes approaching and departing from the intersection. • Operational: The area that is influenced by traffic operations, including queuing, lane changing, merging, and vehicle acceleration/deceleration capabilities. This operational influence area could be independent of the geometric influence area and can change by time of day, season, or other conditions. Speed transition needs should be considered between a roadway segment and the intersection influence area to allow drivers the opportunity to react to changing conditions and adjust their speed accordingly. This could potentially include a change in the roadway cross section (such as adding curbs and landscaping or via a “gateway” treatment) or simply providing adequate sight distance from the upstream segment to the intersection’s geometric or operational influence area. The length needed for the transition area will vary depending on the total desired speed reduction and the operating speeds in upstream segments. Exhibit 2-1 schematically depicts the roadway segment and intersection speed relationships. Exhibit 2-1 Roadway Segment and Intersection Speed Relationships In some cases the design speeds of the adjacent roadway segments are appropriate for an intersection. In other cases the intersection characteristics and driver workload vary and a reduced speed may be desirable. The need for speed reduction at intersections can be considered in the following general conditions: • The posted speed of the segment is higher than the desired speed of the intersection approach (e.g., the intersection approach is stop-controlled, or a transition from a rural to a more urbanized environment occurs at the intersection). • The posted speed of the segment is the same as the desired speed of the intersection approach; however, drivers exceed the posted speeds. • The posted and operating speeds at the segment and intersection are reasonable; however, potential conflicts at the intersection (e.g., diverging or merging

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-3 maneuvers, crossing traffic, or queues) require drivers to be especially alert to the need to respond to these potential conflicts. Stop-controlled intersection approaches will fall under the first condition, while uncontrolled and yield-controlled approaches may fall under any of these conditions. A stop-controlled condition requires operations that are independent of the roadway design speed (i.e., a tangent intersection approach on level grade with unlimited sight distance has an unlimited design speed regardless of the posted speed approaching the intersection). A yield condition may operate as a stop control during some time periods, depending on traffic flows. In other cases, a yield condition may be virtually free flow and the operating speeds for this movement may be dictated by the roadway approach geometry (i.e., a separate turn lane, turning roadway width, and turning roadway radius). 2.1.3 Designing for Appropriate Speeds In general terms, good roadway geometric design provides a sufficient level of mobility and land-use access for motorists, bicyclists, pedestrians, and transit while maintaining an appropriate degree of safety. Higher-speed roadways are typically provided in locations where travel time and mobility are priority needs. Speed is often used as a performance measure to evaluate the effectiveness of highway and street designs, with higher speeds generally associated with longer trips and lower speeds generally associated with shorter trips or with facilities that have more frequent access. Posted speeds frequently correlate with these intended uses. High- speed facilities serve key network needs and it is not always appropriate to expect reduced speeds at intersections. Environmental and operational indicators should be in place to provide drivers with a consistent message about the potential for conflict so they are best able to select an appropriate speed. The goal is to provide geometric street designs that look and feel like the roadway’s intended purpose. Because drivers choose their speed based on what they see on the roadway ahead, calling a driver’s attention to roadway features that present a potential risk that may not be apparent provides increased opportunities to avoid conflicts. Drivers who perceive potential risk can better adapt their driving behavior to roadway conditions. A facility’s design speed is a fundamental design criterion that affects three-dimensional roadway design parameters (plan, profile, and cross-section). Aside from roundabouts, where entry speeds of about 25 mph are specifically attained through geometric design, there is no common intersection design speed; it is typically assumed to be that of the roadway segment. Although designers generally seek speed and operational consistency, intersection operations (e.g., queuing, deceleration, turning vehicles) and/or geometry may create localized conditions that require reduced speeds. At intersections, drivers must perceive and comprehend a greater variety of situations than they need to while driving through the high-speed roadway segment that precedes the intersection. Intersection conditions should be considered independently from the adjacent roadway segments. For new facilities, this means ensuring that an intersection’s operational and geometric elements are appropriately configured. The existing geometric and operational elements of an intersection

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-4 should be assessed when considering appropriate actions for retrofit projects. Such a design philosophy and approach can produce geometric conditions that are more likely to result in operating speeds consistent with driver expectations and commensurate with the roadway’s function. This, of course, does not account for impaired or overly aggressive motorists or conditions such as adverse weather. Ideally, drivers should operate their vehicles in a manner appropriate for the conditions for as long as those conditions prevail. 2.1.4 Factors Affected by Speed High-speeds serve key network functions; however, operating speeds inconsistent with the prevailing conditions may adversely affect environmental quality and safety or require a larger- than-desirable facility size. The effects of speed are most pronounced at intersections where friction between competing movements is concentrated. The friction may manifest itself as decelerating or accelerating vehicles, queues, crossing traffic, or traffic yielding to crossing pedestrians. Any of these kinds of friction create the potential for conflicts. The ways in which speeds through intersections affect intersection operations, environmental quality, and safety are discussed below. 2.1.4.1 Facility Size Drivers’ perception-reaction time are essentially constant; therefore, higher speeds require drivers to understand their driving tasks farther in advance of intersections, compared to slower- speed environments. This means that the distance between driver decision points must increase as speeds increase. Higher-speed facilities require larger clear zones and flatter horizontal and vertical curves. In addition, fundamental stopping-sight-distance dimensions increase with speed, leading to flatter and more open roadways. Attaining appropriate sight distances affects horizontal and vertical alignments as well as such cross-sectional features as cuts, fills, and landscaping. For example, using intersection sight distance values from the AASHTO Green Book for the case of a right turn from a stop or a crossing maneuver, the intersection-sight-distance value for 45 mph is 430 feet. This value increases to 575 feet when speeds are 60 mph. In urban environments, sight-distance needs can affect building setbacks, on-street parking locations, and other design elements such as the locations of street furniture and landscaping. The transition zones within which drivers are required to slow down as they approach an intersection need to be longer when a greater change in speed is required. This affects the length of alignment tapers, bay tapers, and the deceleration components of turn lane designs. 2.1.4.2 Quality and Comfort of the Roadway Environment The function of roadways and intersections must be balanced with the needs of adjacent land

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-5 uses to both maintain environmental quality and to provide necessary mobility. High-speed intersections can create a barrier to the mobility of non-auto users crossing the facility. The increased noise levels and intense environments that result from proximity to motor vehicles can create discomfort for pedestrians and bicyclists who travel parallel to the facility. Adjacent neighborhoods and businesses may also experience adverse effects from high-speed traffic, such as tire and engine noise, and may benefit from buffer treatments. Intersections near schools, hospitals, or other concentrations of pedestrians, particularly elderly, young, and disabled people, may be particularly sensitive to high speeds. 2.1.4.3 Safety The relationship between speed and intersection safety is a critical concern for transportation professionals. Road safety is often divided into three constituent elements: exposure (increases with the number of conflicting movements), risk (increases with increased traffic volume), and consequence (which increases with speed). Reducing speed at intersections has the potential to improve consequence, although it has little relationship to exposure or risk. There is a decisive relationship between speed and crash severity; while the relationships between speed and crash frequency are less clear. The physical relationship between mass and energy explains that higher speeds and larger speed differentials create the potential for higher-severity crashes. As speeds increase, the energy from the mass of the vehicles increases. Studies of modern multilane roundabouts illustrate the relationship between speed and crash severity: total crash rates and frequency may stay the same after an intersection is converted to a roundabout, but the slower speeds help reduce the severity of crashes. A variety of intersection traffic conditions create large speed differentials, increasing the potential for severe crashes. For example: • An unbalanced distribution of traffic for a given number and arrangement of lanes creates high differentials in speeds between vehicles traveling in the same direction. For example, a channelized right-turn lane may operate under near free-flow conditions while the adjacent travel lanes experience queuing. • An intersection or approach with extensive queuing, where the back of the queue is a significant distance from the intersection proper. In these conditions, vehicle queues in turn lanes may exceed the storage length provided, requiring vehicles to decelerate in the through travel lanes. • Uniform traffic congestion at an approach could lead to queues that extend outside an intersection approach’s typical available sight distance. In these conditions, vehicle queues extending beyond the intersection’s geometric influence area may require that drivers decelerate in advance of visual cues of the impending intersection. No research was found that identified a relationship between crash frequency and mean or 85th-

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-6 percentile speed. However, many studies have found that the likelihood of being involved in a crash increases with deviation from the mean speed of traffic on the facility (Taylor, 1965) (Hauer, 1971). If the facility and intersection design provide adequate sight distance and appropriate user expectancy of potential conflicts, there should be adequate space and time to react to and avoid crashes. If, however, speeds are excessive for the facility and intersection design, there will be insufficient space and time to avoid crashes, and a higher crash frequency may result. High-speed travel may affect crash avoidance because faster moving vehicles travel farther during the typical reaction time needed for a driver to avoid a potential hazard. In addition, the greater a vehicle’s speed, the less time there is for other motorists, bicyclists, or pedestrians to react to and avoid that vehicle. Thus, while there is no research to support the common assumption that reducing speeds will reduce crash frequency, reducing speed variation may achieve this. The relationships between intersection speed and safety are complex and it cannot be assumed that reduced speeds will result in a safety improvement. 2.1.4.4 Traffic Operations Traffic operations refers to the performance of a roadway facility and is typically measured in terms of capacity, travel time, delay, number of stops, and queuing. Although vehicle speed directly correlates to the motorists’ perceived level of service along an arterial or highway segment, it does not have a significant effect on the traffic operational performance of an individual intersection. Furthermore, vehicle speed through the influence area of intersections has little effect on overall travel time. 2.1.4.5 Capacity Vehicle speed is not a primary determinant of intersection capacity – the number of vehicles that the intersection can process in a given time period. For the case of a minor-street left-turn movement from a two-way stop-controlled intersection, research performed as part of NCHRP 3- 46 (Capacity and Level of Service at Unsignalized Intersections) found that speed did not have a significant effect on driver’s critical gap – which corresponds directly to the capacity of the stop- controlled movement. For a signalized approach, the capacity of an intersection is a function of the saturation flow rate of the approaching lanes. Saturation flow rate is defined as the flow rate at which previously queued vehicles can traverse an intersection approach under prevailing conditions. As previously queues vehicles are starting from a stopped position, a vehicle’s speed through an intersection is not relevant in the calculation. 2.1.4.6 Travel Time Higher overall travel speeds along an arterial result in lower travel time and an improved level of

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-7 service. Vehicle speed within the influence area of intersections does not generally have a significant influence on overall travel time. For example, given an intersection with a total influence area of 1000 feet, the travel time difference between a vehicle traveling at 50 miles per hour versus a vehicle traveling at 30 miles per hour through the intersection is less than 10 seconds, assuming an uncontrolled approach or a “green light” without interfering queues. 2.1.5 Factors that Affect Speed A driver’s selection of a safe speed and path is determined by his or her judgments, estimates, and predictions based on highway characteristics, traffic, and traffic control devices. (NCHRP 3- 50(2)). Roadway design elements, environment, traffic type, and other factors help drivers determine an appropriate speed. Some elements affect traffic flow directly, while others can influence driver behavior by contributing to the visual complexity (or simplicity) of the roadway edge. The design and characteristics of an intersection proper affect speed at the intersection as do the design and characteristics of the roadway facility and adjacent segments. This section presents a variety of human, vehicle, and roadway characteristics that affect in drivers’ speed choice. 2.1.5.1 Roadway Facility Design and Characteristics Intersections are often relatively infrequent occurrences on high-speed facilities and drivers may expect that they can operate at a consistent speed. Without clear indications of the need to reduce speed and without adequate transition distance within which to do so, drivers will navigate the intersection area at speeds they deem appropriate for the adjacent roadway segments. The chosen speed may or may not be appropriate for the actual conditions at the intersection. The characteristics of the roadway segment prior to an intersection affect speeds at the intersection. Exhibit 2-2 provides a list of roadway facility factors that may affect speeds on intersection approaches. These features may influence driver behavior or vehicle operations and result in speed changes. Many of these relationships are derived from relationships documented in NCHRP Report 504: Design Speed, Operating Speed, and Posted Speed Practices.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-8 Exhibit 2-2 Roadway Facility Characteristics That May Affect Intersection Speed Intersection Variable Potential Relationship to Speed Facility Type Speeds tend to be higher on higher-order facilities. Speeds tend to be slightly lower when a raised median or no median is provided than when a depressed median or a two-way left-turn lane is present. Limited access, low signal density, unimpeded visibility, and viaducts may promote high speeds. Roadway Characteristics Wide shoulders, medians, and overall pavement widths are associated with higher speeds. Lane widths, horizontal/vertical geometry, sight distance, curbs, and bike lanes may influence measured speeds and desired speeds. Conflicts and Friction As the distance between points of friction (driveways, intersections, pedestrian crossings, lane drops) increases, speeds increase to a point and then plateau. Posted Speed Posted speed and 85 th -percentile speed increase or decrease together. Roadside Environment Higher speeds occur in rural and undeveloped areas compared to urban or undeveloped areas. Lower speeds occur in areas with higher levels of pedestrian activity. Pavement Type and Condition Poor, cracked, or uneven pavement and joint details may slow travel speeds, while smooth pavement may allow faster speeds. The absence of centerline or edge line markings is associated with lower speeds. Transition The intersection location in relation to the roadway segment (tangential, curvilinear, flat, mountainous) may influence measured and desired speeds. 2.1.5.2 Speed Adaptation Drivers often underestimate their speeds, particularly in the medium- and high-speed ranges. Thus, excessive speed is not always a conscious decision. In some cases, excessive speed can be attributed to speed adaptation. The speed adaptation hypothesis states that the perceived speed of one’s vehicle will be lower than the actual speed if the driver has recently operated the vehicle at a higher speed. Speed adaptation may contribute to excessive speeds in transition areas between rural and built environments, or between access-controlled or other high-speed facilities and street environments that have driveways, multiple intersections, and non-motorized users. Drivers who have adapted to higher speeds may not appreciate the need to slow down at intersections. These drivers may have attained a feeling of comfort or safety that may not be appropriate for the potentially changing conditions at a high speed intersection. 2.1.5.3 Intersection Design and Characteristics The physical characteristics of the intersection proper affect speed as do the changing conditions at the intersection such as the lighting and congestion patterns. Exhibit 2-3 summarizes many intersection characteristics that may influence drivers’ speed choice. Exhibit 2-3 originates from relationships documented in NCHRP Report 504: Design Speed, Operating Speed, and Posted Speed Practices.

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-9 Exhibit 2-3 Intersection Characteristics That May Affect Intersection Speed Intersection Variable Potential Relationship to Speed Traffic Control /Approach Type A signalized intersection, a stop-controlled intersection, and a yield-controlled intersection require different driver tasks and operating speeds. Wayfinding Complex intersection maneuvers tend to reduce speeds, especially for unfamiliar drivers. Visual Complexity Roadside development, pedestrians, bicyclists, signage, and environmental elements that interest drivers may affect speed. Roadside Impedances Roadside parking, bus stops, and vehicle loading zones may interrupt traffic flow near intersections and effect speed reductions. Lane Drops Drivers may slow down to make lane changes. Increased lane densities may also reduce speeds. Merging Through drivers may need to slow down to create gaps for vehicles entering the traffic stream. Sight Distance Drivers may travel at higher speeds through intersections without sight-distance constraints. Sight distance restrictions can induce a small reduction in speeds, although only for the faster vehicles. Lighting Inadequate lighting may not allow drivers to perceive and react in advance of an intersection. Traffic Conditions Congestion, queuing, directional distribution, and low volumes influence speeds through intersections. 2.1.5.4 Drivers and Vehicles Different types of drivers will choose different speeds at intersections. Commuters and other familiar drivers may tend to drive faster than infrequent users. Driver age and attitude also influence speed through intersections. Additionally, transit vehicles, heavy vehicles, and recreational vehicles may have especially slow turning speeds, requiring them to travel slower than general traffic to maneuver through an intersection. 2.1.5.5 Weather Conditions Weather conditions may affect driver behavior and speeds. Drivers may use caution and drive slower during snow, ice, rain, fog, or dust than they do in clear and sunny conditions. 2.1.6 Identifying Conditions Potentially Sensitive to Speed This section identifies a variety of conditions, elements, and data that may indicate that an intersection is particularly senstive to speed. In some cases these are contextual considerations related to the intersection configuration, location or environment, in other cases the the senstivity might be attributed to specific users or user characteristics at that location. Field conditions may also provide insights about an intersection’s senstivity to speed. This might include observing

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-10 user behavior or evaluating traffic and safety data. 2.1.6.1 Common Conditions There are a variety of conditions that may be associated with a heightened sensitivity to speed. Many of these are related to the characteristics that affect speed identified in Table 2-2. If drivers are alert to the characteristics of an intersection that make lower speeds desirable, they may slow down. If, however, they are not alert to these characteristics it can create a condition that is sensitivite to speed. Examples include: • Intersections that are difficult to detect – horizontal or vertical curvature of an intersection approach may make it difficult to detect. • Intersections within a corridor with a variety of changing contexts (land uses, design philosophies) – drivers need to be warned of the increased need to respond to pedestrians, buses, driveway traffic, or other interruptions in traffic flow. • Intersections that link high-order and low-order roadway segments – gateways between rural and urban areas. • Intersections with complex geometry or irregular route continuity – require a high driver workload to comprehend or navigate through the intersection. • Intersections proximate to concentrations of sensitive or high-risk populations – children, elderly, or disabled • An approach with limited sight distance – roadside obstacles, sunlight at certain times of day • Intersections with high-speed differentials or limited acceptable gaps at certain times of day – long queues for one or more approach lanes. 2.1.6.2 Observed Field Conditions Field observations of an intersection’s operating conditions may provide an opportunity to identify intersections sensitive to speed. 2.1.6.3 Crash Avoidance Patterns Skid marks on a roadway are often an indicator that drivers do not have sufficient time to react to interruptions in the flow of traffic. The direction and location of skid marks may be useful cues for deciphering operating conditions. Skid marks may indicate that drivers are traveling faster than their ability to perceive and react to an intersection condition (i.e., a single stopped vehicle turning left, a vehicle that is accelerating or decelerating, or the back of a queue that extends beyond the perceived area of the intersection).

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-11 2.1.6.4 Driver Behavior Undesirable driver behavior, such as lack of compliance with signals or stop signs, may indicate that excessive speeds approaching an intersection do not give drivers enough time to see and react to traffic control. The speeds may be the result of adaptation from the prior segment with drivers unaware of their speed and required stopping distances. Desired speeds might be attained if drivers had assistance in transitioning from one segment to the other. 2.1.6.5 Congestion Patterns Congestion patterns can increase the intersection influence area creating sensitivity to speed. Congestion patterns that create high speed differentials (such as those listed in Section 2.3.4), that are associated with increased pedestrian or transit use, or that reduce the availability of acceptable gaps for side street traffic, may indicate a particular sensitivity to speed. 2.1.6.6 Speed Data Speed data can provide insights about effective roadway conditions. For example, the measured 85th-percentile speed, mean speed, and speed variance can be compared to the implied design speed based on the AASHTO Green Book’s intersection-sight-distance and stopping-sight- distance criteria. Comparing measured speeds to the posted speed could help identify differences between desired operations (as indicated by a posted speed) and actual conditions. 2.1.6.7 Crash Data A high crash rate at an intersection may indicate any number of operational or geometric issues and does not, in isolation, indicate excessive speeds. A closer analysis of the crash types, locations, time of day, weather conditions, and other factors is useful for understanding the role of speed, if any, in the crashes. Crash patterns that may be associated with excessive speeds include: • Frequent rear-end crashes – Drivers not anticipating the location of the back of a queue • Frequent run-off-road crashes – Drivers avoiding conflicts in the roadway proper • Angle crashes – Drivers accepting gaps that are too small A close analysis of the crash data is necessary to determine if these patterns could also be attributed to other factors such as driver inattention or impairment, geometric conditions (alignment, sight distance), or other conditions that do not meet a driver’s expectations. 2.1.7 Highway Safety Information System (HSIS) One of the most comprehensive sources available to analyze crash characteristics and trends

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-12 associated with specific geometric roadway and traffic volume conditions is the Highway Safety Information System (HSIS) database. The HSIS database was developed by the Federal Highway Administration and is currently maintained by the University of North Carolina Highway Safety Research Center. The database includes detailed crash data from nine participating states. The HSIS data files vary from state to state because they are reflective of the particular data elements and characteristics that each state deems relevant for use in managing the safety issues involved in its highway system. As part of the development of the FHWA guide, Signalized Intersections: An Informational Guide, a detailed analysis was conducted of the California HSIS database. The California database includes the greatest breadth of data elements and characteristics for signalized intersections of the participating states. The objective of the safety analyses was to measure the relative risk and safety effectiveness for various treatments and characteristics of signalized intersections. The interim report developed for the FHWA Signalized Intersection Guide project titled “Task A – Critical Review of Literature and Comparison of Risk Levels Associated with Different Treatments at Signalized Intersections” (Kittelson & Associates, Inc., Synectics Inc., and CH2M Hill, March 2002) provides an analysis of relative safety risk for various roadway geometric and traffic design characteristics such as: • Intersection type • Traffic control type • Mainline number of lanes • Mainline left-turn channelization Crash summaries were also provided for various thresholds of average daily traffic volumes. Since the scope of the risk analysis conducted for the FHWA Signalized Intersection Guidelines project was to evaluate whether any risk differentials exist among signalized intersections due to different roadway-related characteristics and attributes (e.g., operations and geometrics), driver- related causes of collisions such as ‘speeding’ were not part of the study. Also, the project included intersections of both lower speed and higher speed intersections, not just high-speed intersections. Thus, the findings from the analysis are not relevant to NCHRP 3-74. The HSIS database for California and Minnesota could be further investigated to identify the characteristics of crashes at intersections where excessive speeding was listed as a contributing factor. In addition, an analysis could be performed to identify safety characteristics of intersections based on the posted speed and design speed of the approaching roadways. In principal, one could conduct an analysis to identify whether any particular intersection type experiences over-representation of collisions where the speeding is the major factor. However, study results would have several limitations due to missing data in the database. It is not recommended that the HSIS databases be explored further, given the limitations that will be associated with the results and the time and effort that will be expended in obtaining the data

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-13 from HSIS, organizing the data, and setting up the analysis specs, and given the little return expected in advancing our knowledge and understanding of intersection types and their characteristics that are most affected by speeding and unsafe speed-related factors. A detailed HSIS database investigation and statistical analysis is outside of the scope of NCHRP 3-74 and was not be conducted as part of this work effort. However, future research projects might focus on or include a robust evaluation of HSIS data to better understand the safety relationship between posted and operating speeds at intersections. 2.1.8 Summary Reduced speeds do not guarantee safety, operational, or environmental benefits at a high-speed intersection. Speed is a product of many features, including the adjacent roadway segment, the user and vehicle type, and the general environmental context. Intersection approach speeds can affect a facility’s safety and performance. Speed, when considered as a design criterion or consideration, can affect the roadway design while also influencing the environmental context. Understanding how speed affects intersection conditions and how those conditions affect speed is considered first to evaluate and select an appropriate speed reduction treatment. 2.2 TREATMENT DESCRIPTIONS 2.2.1 Overview This section provides a summary of the findings reported in published literature on potential speed-reduction treatments. Some of the reported treatments originate from applications for roadway segments, while others are specific intersection applications. Once the intersection characteristics described in Section 3 have been assessed, the user can use the information in the chapter to help determine which treatment(s) may be appropriate at a specific location. This section contains an overview of each treatment, as well as more detailed descriptions, summaries of applicability and pertinent considerations such as maintenance, discussions of potential layouts and designs, and summaries of the documented effectiveness of each treatment in reducing speeds and improving safety. The treatments discussed are: • Dynamic warning signs • Transverse pavement markings • Transverse Rumble strips • Longitudinal Rumble strips • Wider longitudinal pavement markings • Roundabouts • Approach reverse curvature

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-14 • Splitter islands • Speed tables • Reduced lane width • Visible shoulder treatments • Roadside Design Features In some cases, more than one treatment may be appropriate for a given intersection. To apply a treatment, users select a target speed and determine the location upstream of the intersection where sufficient distance is provided for drivers to react to potential conflicts related to the intersection when traveling at the target speed. A transition area is needed upstream of this target speed location to allow drivers to decelerate from their segment speed. The transition length and target speed location help determine an appropriate treatment layout. The information provided in this section should be considered in tandem with appropriate local engineering practices and professional judgment related to the site-specific situation. 2.2.2 Dynamic Warning Signs Overview Documented applications for high-speed intersections Three test sites: one in Washington and two in Texas Function Alert drivers of the need to reduce their speed, encourage earlier deceleration Applicability Rural unsignalized, in advance of abnormal roadway features Design Variations Speed activation, message, image, size, location Secondary Effects, Considerations Driver workload, power supply 2.2.2.1 Applicability and Considerations Dynamic warning signs are placed along the side of the roadway prior to a location that requires reduced speed. The signs are activated by vehicles that exceed a predetermined speed (typically in excess of the posted speed limit) or by potential vehicle conflicts at the intersection. This type of sign is not intended to enforce the speed limit, rather it is assumed that drivers will reduce their speeds once they are brought to their attention (Maze, 2000). Although most commonly applied in work zones, these signs have potential to be applied at intersections. Dynamic warning signs have been used at curve approaches, in work zones, and in other locations that require reduced speeds.

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-15 Dynamic warning signs have also been used to warn drivers of conflicting cross traffic at intersections. Called “Collision Countermeasure Systems” (CCS) in this application, they are used to reduce side-impact crashes at rural non-signalized intersections. On low-volume rural highways with a history of high intersection crash rates, dynamic warning signs of this type are a cost-effective alternative to a conventional traffic signal (Hanscom, 2001). Dynamic warning sign systems typically combine a radar device or pavement loop detectors with a variable message sign. The system measures the speed of the approaching vehicle or detects a potential vehicle conflict. The system provides a message to drivers who are traveling at excessive speeds or whose movements are in potential conflict with another vehicle. Examples of the types of messages displayed on the variable message sign include, “Slow Down,” “XX mph Curve Ahead,” “Your Speed XX mph” or “Traffic Ahead” (Torbic et al., 2004; Hanscom, 2001). Some systems are designed to provide messages to only a certain type of vehicle, such as trucks. In these cases, the dynamic warning systems involve weigh-in-motion devices, loop detectors, and height detectors (Torbic et al., 2004). One of the challenges for implementing dynamic warning signs is to determine the maximum safe speed above which the sign will be activated. Vehicle loads, suspension, vehicle size, tires and variable weather conditions can affect this speed (Torbic et al., 2004). The maximum safe speeds selected by both the Washington and Texas for the NCHRP 3-74 research test sites were higher than the posted speed limit. Dynamic warning signs can vary in complexity, therefore, some systems can be installed in a very short period of time, while others may take years to implement. The radar and video equipment, loop detectors, and weigh-in-motion detectors can expect the same maintenance issues as when placed at conventional signalized intersections or other locations. When dynamic warning sign equipment and posts are placed outside of the pedestrian and bicycle zones, they are not expected to impact multimodal users.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-16 Exhibit 2-4 Dynamic Warning Sign (Photo: WSDOT) Exhibit 2-5 Dynamic curve-warning sign (Photo: TXDOT) Exhibit 2-6 Dynamic warning sign used as a Collision Avoidance System (Photo: PennDOT)

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-17 2.2.2.2 Treatment Layout/Design Dynamic warning signs placed in a location that provides adequate advance warning time for drivers to reduce their speed appropriately. The placement of a dynamic warning sign should be determined from estimated perception-reaction time, deceleration, and the stopping sight distance in advance of the intersection. The sign placement will reflect the desired target speed location and the driver deceleration rates from the operating speed of the upstream segment. Vertical and horizontal alignments must provide a clear sight line for the radar or video equipment. A power source is also necessary (Torbic et al., 2004). Constraints such as steep slopes, bridges, and power supplies should also be considered to determine the appropriate location for the sign. If a radar unit is included as part of the dynamic sign treatment, it must be placed to detect drivers’ speeds in advance of the sign. 2.2.2.3 Speed Effects Dynamic warning signs reduced speeds significantly at the three high-speed intersection approaches tested through the NCHRP 3-74 project. Speed data was collected at three locations on each intersection approach. A mean speed reduction of 1.7 mph was observed at the sign locations after a three-month acclimation period. At the perception-reaction data collection data collection locations, mean speeds were reduced by 2.3 mph. At the accident-avoidance data collection locations mean speeds were reduced by 2.8 mph. During a study of warning signs installed at two high-speed intersection sites in Texas, Ullman and Rose recorded an initial reduction of 3 to 4 mph at both sites, with one site sustaining the lowered speeds after the novelty effect of the sign diminished (2003). Studies examining the effectiveness of speed displays at rural interstate work zones in South Dakota and Virginia found that the speed monitoring displays reduced mean vehicle speeds by 1.4 to 4 mph within the work zones and reduced the number of vehicles speeding through the work zone by 20 to 40 percent. Maze concluded that the placement of the signs can impact the effectiveness of this treatment. (Maze, 2000) A study of three dynamic curve-warning systems installed on ramps in Virginia and Maryland showed that the three systems significantly influenced truck speeds—drivers reduced their speed if they had exceeded the maximum safe speed. (CH2M Hill, 2004) (Torbic et al., 2004) In the field evaluation of the CCS installed in Prince William County, Virginia, it was observed that the CCS resulted in lower intersection approach speeds and longer projected times-to- collision. (Hanscom, 2001) 2.2.2.4 Safety Effects No documentation of the safety benefits of dynamic warning systems at intersection approaches was found; however, the study of dynamic curve warning systems in Maryland and Virginia

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-18 identified that prior to installation there had been 10 truck rollover crashes and after three years in operation, no rollover crashes had been reported. (CH2M Hill, 2004) 2.2.3 Transverse Pavement Markings Overview Documented applications for high-speed intersections Eight test sites in Oregon Function Improve visibility, driver attention Applicability Visual cue reinforcing changing conditions and need to reduce speed Design Variations Peripheral or full transverse lines Secondary Effects, Considerations Long-term effectiveness, driver familiarity 2.2.3.1 Applicability and Considerations As defined by the Manual on Uniform Traffic Control Devices (MUTCD), transverse pavement markings are “pavement markings that are generally placed perpendicular to and across the flow of traffic” (FHWA, MUTCD, 2003). Peripheral transverse lines involve bars only at the edge of travel lane, instead of bars extending across the travel lane. Transverse chevrons are painted geometric arrows that converge to give the illusion of speed (Griffin, 1995). Transverse pavement markings are commonly used in speed management to reinforce the need to reduce speed or to warn drivers of an approaching condition that may require vehicular maneuvers. Common applications of transverse pavement marking locations include approaches to traffic circles and intersections, horizontal curves, construction areas, bridges, and freeway off ramps (Griffin, 1995). Transverse pavement markings have also been placed at locations where excessive speed was a contributor to crashes and other speed reduction treatments had not been effective (Agent, 1980). Exhibit 2-7 Peripheral Transverse pavement markings Peripheral transverse lines require less maintenance than transverse lines that extend across the travel lane because most drivers do not track over the markings (Human Factors North, Inc.,

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-19 2002). Transverse pavement markings have no known impact to multimodal users—it is not expected that multimodal users would be adversely affected by transverse pavement markings. 2.2.3.2 Treatment Layout/Design Transverse pavement markings should be placed in a location that provides adequate advance warning time for drivers to reduce their speed appropriately. The placement of transverse pavement markings that is optimal to reduce speeds on intersection approaches has not been quantified. The AASHTO Green Book values for deceleration may provide a starting point in locating transverse pavement markings. Transverse pavement lines may be installed in several small clusters on a high-speed intersection approach. The number of lines, their spacing, and the distance from the intersection proper should be determined through a combination of a review of field conditions, driver sight lines and desired response, and local practice and judgment. As guidance, the MUTCD suggests, “…because of the low approach angle at which pavement markings are viewed, transverse lines should be proportioned to provide visibility equal to that of longitudinal lines” (FHWA, MUTCD, 2003). The appropriate location and length of the lines depends on the speed at which the driver is traveling, the deceleration rate, and the driver’s perception-reaction time (Agent, 1980). Appropriate installation points may be selected to reinforce other new or existing treatments or features on the intersection approach such as warning signs, shoulder markings, parking space markings, and pavement legends (“SLOW DOWN,” “REDUCE SPEED,” etc.), and at the intersection proper, lane-use legends, stop lines, crosswalks lines, and others. The transverse pavement markings have the potential to draw additional attention to those signs and markings and to encourage drivers to reduce their speeds as they approach the intersection. Other appropriate locations for treatment installation may include the stopping sight distance for the approach speed, or a point where the roadway environment changes, such as at the point of tangency or at a driveway. The spacing of the transverse bars and transverse chevrons may reflect the desired travel speeds of the vehicles on the roadway. Some applications of transverse bars have used a rate of advancement of two stripes per second. In some applications transverse bars are spaced progressively closer together at an increasing rate as the driver travels along the roadway. When applied in this way, the bars are referred to as “optical speed bars.” The intent is that the reduced spacing gives the driver the perception of acceleration causing the driver to slow down; however, there is no data to support this claim (Agent, 1980). 2.2.3.3 Speed Effects After a 90 day acclimation period, transverse pavement markings were found to reduce speed marginally at the four high-speed intersection approaches tested through the NCHRP 3-74 project. Overall, the markings reduced mean speeds by 0.6 mph (standard error of 0.3 mph). Additionally, at the perception-response time data collection locations, transverse pavement

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-20 markings were found to reduce mean speeds by 0.9 mph (standard error of 0.4 mph). Studies of segment applications of transverse pavement markings have reported reduced mean and 85th-percentile speeds on the order of 20 to 30 percent. (Katz, 2003) (Griffin, 1995) At the Newbridge Roundabout in Scotland, transverse pavement markings resulted in reduced mean and 85th percentile speeds. The overall mean speed throughout the day was reduced by approximately 23 percent, and the overall 85th percentile speed throughout the day was reduced by approximately 30 percent. (Katz, 2003) In a study conducted on US Highway 60 in Meade County, Kentucky, transverse bar pavement markings were placed prior to a sharp curve with a high crash rate. This study revealed that the treatment became less effective as a speed reduction technique as drivers became familiar with the treatment. Furthermore, the long-term effects during the nighttime were less than the long- term effects during the day (Agent, 1980). Research by Godley et al. (2000) using a driving simulator found that transverse lines with both constant and reducing spacing (optical speed bars) reduced speeds. The research determined that speed perception was not influenced by the decreased spacing of the lines. Additionally, Godley found that the peripheral transverse lines induced speed reduction almost as effectively as full transverse lines. No studies have been found that evaluate the effectiveness of the transverse chevron markings with respect to speed. 2.2.3.4 Safety Effects No published data was found to address the effects of transverse pavement markings on safety at conventional intersections. Safety improvements associated with segment or roundabout applications of transverse pavement markings were reported by each study referenced below. In 1993, a study in Osaka, Japan, reported that converging chevron pavement markings placed on the Yodogawa Bridge was more effective than conventional signing in helping to prevent crashes at a high crash location. No crashes resulting in injuries occurred in the two years after the chevron markings were installed. In the past, the bridge had a history of crashes causing injuries and fatalities (Griffin, 1995). At the Newbridge Roundabout in Scotland, transverse pavement markings reduced the number of reported crashes from 14 in the year prior to installation to 2 in the 16 months after installation (Katz, 2003). A study conducted on US Highway 60 in Meade County, Kentucky, found crashes were reduced after transverse pavement markings were installed. During the previous six years, an average of eight crashes occurred at this location each year, and speed was identified as a contributing factor in 75% of those crashes. During the year after installation, three crashes were reported, with one attributed to high speed (Agent, 1980). Data for additional years after treatment installation were not included in the study.

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-21 In another study that reduced crashes, transverse lines were applied at 42 approaches to roundabouts. Each approach had a minimum of 3.2 km (2 miles) of uninterrupted road to allow drivers to adapt to the high-speed environment. During the two-year period after the transverse lines were installed on the approaches, speed were reduced by 57 percent. The before-and-after studies conducted at all 42 approaches also indicated that the number of crashes decreased from 96 to 47. A follow-up study conducted at seven of the sites, four years after installation, indicated that the treatment continued to be effective with speed-related crashes showing a significant decline. (Human Factors North, Inc., 2002) 2.2.4 Transverse Rumble Strips Overview Documented applications for high-speed intersections Five test sites in Texas Function Provide audible and tactile warning to encourage deceleration and reduce comfortable speed Applicability Generally STOP-controlled approaches; also toll plazas, horizontal curves, and work zones Design Variations Paved, rolled, or milled; raised or depressed; painted or not; cluster spacing Secondary Effects, Considerations Noise, motorcycle and bike impacts 2.2.4.1 Applicability and Considerations Rumble strips are raised or grooved patterns installed on the roadway travel lane or shoulder pavements. The texture of rumble strips is different from pavement and produces both an audible warning and physical vibration when vehicle tires pass over them (FHWA Research and Tech., 2007). Rumble strips can be installed to warn drivers of an upcoming need to act, such as the need to stop at a traffic signal, slow down at an intersection, change lanes in a work zone, or steer back into the travelway. Their purpose is to provide a motorist with an audible and tactile warning that their vehicle is approaching a decision point of critical importance to safety. While rumble strips warn drivers that some action may be necessary, they do not identify what action is appropriate. The driver must use visual cues to decide what type of action is appropriate. Thus, rumble strips serve only to supplement, or call attention to, information that reaches the driver visually. In many cases, the objective of a transverse rumble strip is to call attention to a specific traffic control device, such as a STOP AHEAD–sign. Transverse rumble strips are placed perpendicular to the direction of travel, and are designed to reduce the deceleration rate and the potential for sudden braking, skidding, and loss of control. The most common use of rumble strips is on intersection approaches controlled by a stop sign, but they have also been used on approaches to signalized intersections, especially for isolated

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-22 signals on high-speed roadways where drivers may not expect the presence of a signal. Transverse rumble strips are generally installed on approaches to intersections of expressways, rural highways, and parkways. Transverse rumble strips have also been placed prior to toll plazas, horizontal curves, and work zones (FHWA Research and Tech., 2007). The noise impacts of transverse rumble strips may make them a poor choice for locations where pedestrians or others will spend time adjacent to the treatment. Transverse rumble strips have the potential to adversely affect bicycles and motorcycles due to the vibrations generated, startle drivers as they cross over the rumble strips, and cause drivers to maneuver quickly to avoid the in-lane rumble strips. Rumble strip design varies by state and depends on the type of facility to which the treatment is being applied Exhibit 2-8 Transverse Rumble Strips in Wheel Path (TTI) Exhibit 2-9 Transverse Rumble Strips Exhibit 2-10 Transverse Rumble Strips Across the Entire Travel Lane (Photo: Corkle et al.)

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-23 2.2.4.2 Treatment Layout/Design Rumble strips should be placed so that either the upcoming decision point, or a sign that identifies the action that may be required, is clearly visible as the driver passes over the rumble strip. Rumble strip locations should be selected to provide adequate advance warning time for drivers to take the potentially required action. Values for deceleration in the AASHTO Green Book provide a starting point for locating the strips. Transverse rumble strips can be installed to cover either the entire width of the travel lane or just the width of a vehicle’s wheel path (Corkle et al., 2001). Exhibit 2-10 shows transverse rumble strips that stretch across the entire width of the driving lane. Exhibit 2-8 illustrates transverse rumble strips that only cover a vehicle’s wheel path, enabling drivers familiar with the area to straddle the rumble strips to avoid driving over them. Transverse rumble strips may be installed in several small clusters on a high-speed intersection approach. The number of strips, their spacing, and the distance from the intersection proper should be determined through a combination of a review of field conditions, driver sight lines and desired response, and local practice and judgment. Appropriate installation points may be selected to reinforce other new or existing treatments or features such as warning signs. The markings have potential to draw additional attention to those warning signs and to encourage drivers to reduce their speeds as they approach the intersection. Other appropriate locations for installation may include the stopping sight distance for the approach speed, or at a point where the roadway segment environment changes, such as at a point of tangency or at a driveway in advance of the intersection. To accommodate bicycles on roadways with rumble strips, agencies should provide a smooth, clear, paved surface wide enough for a bicycle to travel comfortably to the right of the rumble strips. Some agencies have begun to redesign rumble strips to make them safer for bicycles. A skip pattern enhances rumble strip safety for bicycles, providing cyclists the opportunity to enter and exit the bike path without having to cross over the rumble strips (Walls, 1999). There are a variety of types of rumble strips that vary in their installation methods, shape, size, and amount of noise and vibration produced (FHWA Safety, 2007). Rolled rumble strips must be installed when constructed or reconstructed shoulder surfaces are compacted. Formed rumble strips are appropriate for Portland Cement concrete shoulders and involve grooves or indentations formed into the concrete surface during the finishing process (Elefteriadou et al., 2001). Raised rumble strips are strips of material that adhere to new or existing surfaces. Raised and surface-mounted rumble strips can easily be removed by snowplows, causing the need for replacement. Thus, the use of raised rumble strips is usually restricted to warmer climates (Morgan, 1997). Some agencies paint over rumble strips to make them more visible (FHWA Research and Tech., 2004). Milled continuous shoulder rumble strips are easy to install on existing and new pavement, maintain the integrity of the pavement structure, and produce more noise and vibration than other types of rumble strips (FHWA Safety, 2004).

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-24 2.2.4.3 Speed Effects At the high-speed intersection approaches tested through NCHRP 3-74, rumble strips produced statistically significant speed reductions at the perception-response time data collection location, where a mean speed reduction of 1.3 mph (standard error of 0.5 mph) was observed. At each site this data collection point was beyond the rumble strips, and about 250 feet upstream of the intersection. Overall, no statistically significant speed reduction was observed at the data collection point where the rumble strips were installed or at the accident-avoidance data collection point (roughly 100 feet upstream of the intersections). Transverse rumble strips have been used with mixed results. On stop-controlled approaches, transverse rumble strips have repeatedly resulted in more gradual deceleration by drivers and increased the percentage of drivers making a full stop at the stop-sign by about 30 percent (Kermit and Hein, 1962) (Owens, 1967) (Khotari, 1992) (Zaidel et al., 1986) (Harder et al., 2001). However, it has also increased speed variance. (Morgan, 1997) The research conducted by Owens found an increase in speed variance on the intersection approach, indicating that some drivers slowed down more than others. University of Minnesota research involving a driving simulator concluded that drivers started to slow down and finish braking at the same time with and without rumble strips, but braking occurred earlier with rumble strips. It was also found that drivers brake more and brake earlier with full-width rumble strips than with wheel-track rumble strips (Harder et al., 2001). Miles, et al. found that transverse rumble strips produce mostly small changes in traffic operations at both horizontal curve and stop-controlled intersection sites. This research concluded that the treatment was not successful in significantly reducing approach speeds to an intersection (2005). 2.2.4.4 Safety Effects NCHRP Synthesis 191 (Harwood, 1993) summarized ten before-and-after studies that investigated the safety effectiveness of transverse rumble strips. The synthesis concluded that transverse rumble strips may effectively reduce crashes, although more rigorous evaluation is needed to better understand the safety effects (Harwood, 1993). However, a 2005 TTI study found that transverse rumble strips produced marginal safety benefits at best, only slightly reducing erratic lane movements (Miles, 2005).

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-25 2.2.5 Longitudinal Rumble Strips Overview Documented applications for high-speed intersections None Function Provide audible and tactile warning to reduce comfortable speed and minimize off-the-road and crossover crashes Applicability Rural undivided highways Design Variations Paved, rolled, or milled; raised or depressed; painted or not; cluster spacing Secondary Effects, Considerations Noise, motorcycle and bike impacts 2.2.5.1 Applicability and Considerations Like transverse rumble strips, longitudinal rumble strips are raised or grooved patterns installed on the roadway travel lane or shoulder pavements to warn drivers of an upcoming need to act. Longitudinal rumble strips are placed parallel to the direction of travel and may be located in the centerline or along the shoulder. Refer to Section 4.4 for more information on rumble strips. Longitudinal rumble strips are most commonly used to reduce head-on, sideswipe, and run-off- road crashes along roadway segments. The treatment may be a useful speed management treatment for high-speed intersections that exhibit these crash patterns. Bicycle and motorcycle impacts should be considered as the treatment for intersection applications is designed. Continuous shoulder rumble strips are the most common type of longitudinal rumble strip. These are placed on the roadway shoulder to help prevent drivers from running off the road and are generally used along expressways, interstates, parkways, or two-lane rural roadways (FHWA Safety, 2004). Continuous shoulder rumble strips are typically installed with breaks or gaps only at exits and entrances to ramps and at street intersections or major driveways. This allows drivers and bicyclists to maneuver near the intersections and driveways without having to cross over the rumble strips. Centerline rumble strips are generally located either are placed on either side of the centerline (Exhibit 2-11), along the width of the centerline pavement markings (Exhibit 2-12), or extend slightly into the travel lane (Exhibit 2-13). Centerline rumble strips are generally installed to reduce head-on and sideswipe crashes along undivided roadways. Their primary function is to use tactile and auditory stimulation to alert inattentive or drowsy drivers that their vehicles are encroaching on the opposing lane. Centerline rumble strips may also discourage drivers from cutting across the inside of horizontal curves.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-26 Exhibit 2-11 Centerline Rumble strips (Photo: MnDOT) Exhibit 2-12 Centerline Rumble Strips on the Centerline Pavement Markings (Photo: Torbic et al.) Exhibit 2-13 Centerline Rumble Strips Extending into the Travel Lane (Photo: Torbic et al.) 2.2.5.2 Treatment Layout/Design The placement and dimensions of longitudinal rumble strips on high-speed intersection approaches should be determined through a combined review of field conditions, driver sight lines and desired response, and local practice and judgment. The distance that the treatment extends from the intersection proper should be related to stopping sight distance and/or the distance required to achieve the desired deceleration at a comfortable

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-27 deceleration rate. The distance should also be selected to work in concert with other treatments or features, such as warning signs. The markings have the potential to draw additional attention to those warning signs and to encourage drivers to reduce their speeds as they approach the intersection. The dimensions of the rumble strips will vary by application. If the treatment is being used to narrow the functional width of the approach lane, the treatment should extend beyond the existing striping. Consideration of truck traffic is likely to constrain the amount of narrowing that may occur. 2.2.5.3 Speed Effects No documentation was found regarding the speed effects of longitudinal rumble strip applications in segment or intersection locations. However, it is expected that longitudinal rumble strips have the greatest potential to reduce speeds when they are used to narrow the functional width of the roadway. The Federal Highway Administration is currently conducting research into these applications; however, this study applied multiple treatment types at a given location and therefore, the direct effects of the rumble strips are not documented. At non-intersection locations on two-lane highways, one study found that the presence of centerline rumble strips had no effect for sites with 11- and 12-foot lanes (Porter et al, 2004). 2.2.5.4 Safety Effects Safety improvements have been documented at locations where continuous or shoulder rumble strips were installed along roadway segments; however, there has been no documentation to show the results of this specific application at intersection approaches. Shoulder rumble strips have shown to be a very inexpensive and effective treatment for reducing the number of run-off-the-road crashes. Reports from Maine, New York, and California have reported run-off-road crash reductions of 20 to 72 percent after continuous shoulder rumble strips were installed (Corkle et al., 2001). Similarly, centerline rumble strips have been shown to improve safety along undivided highways. A study analyzing the safety effectiveness of centerline rumble strips along approximately 210 miles of treated roads in seven states found a 14-percent reduction for all injury crashes combined, as well as a 25-percent reduction for frontal and opposing-direction sideswipe injury crashes (Persaud et al., 2004).

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-28 2.2.6 Wider Longitudinal Pavement Markings Overview Documented applications for high-speed intersections None Function Increase intersection visibility, attract driver attention to intersection ahead Applicability Older drivers, lack of driver expectancy Design Variations Width, length, reflectivity Secondary Effects, Considerations Speed may increase with increased visibility 2.2.6.1 Applicability and Considerations Many departments of transportation increase the width of pavement markings to improve the visibility of centerline, lane line, and edge line striping and can provide added guidance to drivers from greater distances (Gates and Hawkins, 2002) (Hutchins, email). Although no documentation for wider longitudinal pavement marking applications at intersection approaches was found, this treatment may be an effective speed reduction treatment for some high-speed intersections because it may increase driver awareness of the presence of an intersection and help reinforce the need for drivers to operate differently at the intersection than they in the roadway segment. The most common reasons that jurisdictions apply wider longitudinal pavement markings are to improve visibility, to assist older drivers, and to reduce crashes. Many departments of transportation have policies to apply wider longitudinal pavement markings to routes of a certain roadway classification—most commonly access-controlled highways. Some jurisdictions have policies that require using wider edge lines on all state routes while others install this treatment exclusively at hazardous locations that could benefit from greater driver visibility. (Gates and Hawkins, 2002) Wider longitudinal pavement markings assist peripheral vision by improving the peripheral signal, thus decreasing driver workload and increasing driver comfort and performance. The largest visibility benefit is seen with older drivers, whose visual and cognitive capabilities decline with age (Gates and Hawkins, 2002). The higher cost of wider longitudinal pavement markings may be somewhat offset by greater durability. Pavement lighting systems are an alternative to wider longitudinal pavement markings that serve the same function, and may have similar effects. This treatment does require a power source that may require maintenance or replacement over time.

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-29 Exhibit 2-14 Examples of wider longitudinal pavement markings in France 2.2.6.2 Treatment Layout/Design The placement and dimensions of wider longitudinal pavement markings on high-speed intersection approaches should be determined through a combination of a review of field conditions, driver sight lines and desired response, and local practice and judgment. The distance that the treatment extends from the intersection proper should be related to stopping sight distance and/or distance to achieve the desired deceleration at a comfortable deceleration rate. The distance should also be selected to work in concert with other treatments or features such as in-pavement markings at pedestrian crosswalks, lane drops or adds, or warning signs. The markings have potential to draw additional attention to those warning signs and to encourage drivers to reduce their speeds as they approach the intersection. The standard width for longitudinal pavement markings is four inches. Wider pavement markings generally range from 5 to 10 inches wide. 2.2.6.3 Speed Effects No specific information was found to describe the impacts of wider longitudinal pavement markings on speed at intersection approaches or roadway segments. The increased visibility and comfort associated with wider pavement markings could lead to increased speed in some applications. While this treatment may not directly affect reduced speeds, it may increase driver awareness of an impending intersection thereby indirectly reducing speeds if drivers perceive a greater risk. The Federal Highway Administration is conducting demonstration projects in Alaska and Tennessee to evaluate the impacts and effectiveness of increasing the width of pavement marking edge lines from 4 inches to 6 inches. The report is due by June 30, 2009. 2.2.6.4 Safety Effects There is no conclusive data associated with the crash-reduction effects of wider longitudinal pavement markings. Cottrell (1988) conducted a before-and-after evaluation of wider pavement edge lines on rural two-lane highways in Virginia and found no evidence to indicate any safety benefit from their

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-30 use. There has been conclusive evidence to suggest that wider longitudinal pavement markings are easier for drivers to see, which can contribute to roadway safety (Gates, Chrysler, and Hawkins, 2002). The greatest improvement in visibility is achieved at night. 2.2.7 Roundabouts Overview Documented applications for high-speed intersections Widely used throughout US, UK, France, and Australia Function Intersection geometry reduces speeds and conflict points Applicability Many types for varying applications Design Variations Type, size, lanes, geometry Secondary Effects, Considerations Operations, right-of-way, access, horizontal and vertical geometry, driver expectancy 2.2.7.1 Applicability and Considerations A roundabout is a type of circular intersection with specific design and traffic control features to ensure that travel speeds on the circulatory roadway are typically less than 30 mph (50 km/h). These features include channelized approaches and geometric curvature (FHWA, Roundabouts: An Informational Guide, 2000). The following qualities distinguish a roundabout from other circular intersections: • Yield control is provided on all entries • Right-of-way is given to circulating vehicles • Pedestrians can cross only the roundabout approaches (behind the yield line) • Parking is not allowed within the circulatory roadway or at the entries • Vehicles circulate in a counter-clockwise direction Roundabouts have been designed to improve some of the safety and operational deficiencies that occur in other types of circular intersections, such as traffic circles and rotaries. The specific design features of a roundabout can reduce speed reduction and the number and severity of collisions.

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-31 Exhibit 2-15 Single-lane roundabout (Photo: Oregon DOT) Roundabouts are appropriate for locations with a high crash frequency or severity, intersections where queues need to be minimized, intersections with irregular geometry, intersections that need to accommodate U-turns, and areas with a large amount of right-of-way available. Pedestrians are accommodated at a roundabout by crossings through splitter islands located around the perimeter of the roundabout. Roundabouts have potential to be difficult for visually impaired pedestrians to navigate. NCHRP 3-78, “Crossing Solutions at Roundabouts and Channelized Turn Lanes for Pedestrians with Vision Disabilities,” will provide additional information on the impacts roundabouts have on visually impaired pedestrians. At single-lane roundabouts, bicyclists have the option to mix with traffic or use the pedestrian crossings. At multi-lane roundabouts, bicycle paths should be separate and designated from the circulatory roadway, for example, by providing a shared bicycle-pedestrian path. Landscaping placed on the center island of a roundabout may need to be maintained to keep the intersection aesthetically pleasing and to manage intersection sight distance. A maintenance program should be developed for the landscape design of a roundabout (FHWA, Roundabouts: An Informational Guide, 2000). 2.2.7.2 Treatment Layout/Design Roundabouts can be adapted to a variety of user and vehicle types based on the environment, number of lanes, and space used by the intersection. Identifying the proper dimensions of the key design features, including an appropriate design vehicle, is critical to ensure proper operations and safety for all users at roundabouts (FHWA, Roundabouts: An Informational Guide, 2000). To distinguish roundabouts from other types of circular intersections, key design features such as a central island, splitter islands, circulatory roadway, yield lines, pedestrian crossings and, in some cases, an apron, have been defined and are shown in Exhibit 2-16.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-32 Exhibit 2-16 Key Roundabout Design Features (FHWA, Roundabouts: An Informational Guide, 2000) A mountable apron may be designed around the perimeter of the central island to provide additional width required for tracking through a single-lane roundabout. Large vehicles may use the entire circulatory roadway at double-lane roundabouts to track and maneuver (FHWA, Roundabouts: An Informational Guide, 2000). Single-lane, four-leg roundabouts have a typical daily service volume of approximately 20,000 entering vehicles per day. Two-lane, four-leg roundabouts have a typical daily service volume of approximately 40,000 entering vehicles per day (FHWA, Roundabouts: An Informational Guide, 2000). 2.2.7.3 Speed Effects Roundabouts have the potential to lower speeds to allow drivers more time to react to potential conflicts. Traffic measured at 43 locations revealed that the geometry yields 85th-percentile entry speeds to between 13 and 17 mph (NCHRP 572). 2.2.7.4 Safety Effects One of the largest benefits of roundabout installation is the overall improvement to intersection safety. Research on crash reduction for conversions of all types of intersections to roundabouts (55 sites studied) found a 35-percent reduction in all crashes and 76-percent reduction in injury crashes. Specifically for rural two-way stop-controlled intersections that have been converted to roundabouts (10 sites studied), research found a 72-percent reduction in all crashes and 87- percent reduction in injury crashes (NCHRP 572). The crash reduction is achieved because there are fewer conflict points at roundabouts than at conventional intersections and because speeds are reduced (FHWA, Roundabouts: An Informational Guide, 2000). Single-lane and multi-lane roundabouts have different operational characteristics that affect their

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-33 relative safety benefits. Although the number of crashes and the severity of injuries were reduced at both single-lane and multi-lane roundabouts, a smaller reduction in total crashes was observed at the multi-lane roundabouts. Both single-lane and multi-lane roundabouts exhibit similar reductions in injury crashes (NCHRP 572). Roundabouts improve the safety of bicyclists by slowing automobiles to speeds closer to bicycle speeds (thus reducing high-speed conflicts) and by simplifying turning movements for bicycles (FHWA, Roundabouts: An Informational Guide, 2000). Crashes involving bicycles comprise approximately 1 percent of all reported crashes at roundabouts (NCHRP 572). Roundabouts also decrease the risk of severe pedestrian collisions, due to the reduced vehicle speeds and a reduced number of conflict points for pedestrians. A Dutch study conducted at 181 roundabouts found a 73-percent reduction in all pedestrian crashes and an 89-percent reduction in injury crashes for pedestrians (FHWA, Roundabouts: An Informational Guide, 2000). Crashes involving pedestrians comprise approximately 1 percent of all reported crashes at roundabouts (NCHRP 572). Bicycles do not experience the same safety benefit at roundabouts as vehicles and pedestrians. In some cases, roundabouts can be less safe for bicycles than conventional intersections despite the speed reduction benefits (FHWA, Roundabouts: An Informational Guide, 2000). 2.2.8 Approach Curvature Overview Documented applications for high- speed intersections Roundabout approaches only Function Introduce geometry to induce slower speeds Applicability New facilities, rural highways Design Variations Curve geometry Secondary Effects, Considerations Right-of-way, grading, driver workload, truck movements 2.2.8.2 Applicability and Considerations Approach curvature is a geometric design treatment that can be used at high-speed intersection approaches to force a reduction in vehicle speed though the introduction of horizontal deflection. As shown in Exhibit 2-17, approach reverse curvature consists of successive curves with progressively smaller radii. Research and applications of approach curvature have focused on roundabouts. However, this geometric design treatment has potential to be applied to conventional intersections.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-34 Exhibit 2-17 Approach Reverse Curvature (FHWA, Roundabouts: An Informational Guide, 2000). Approach curvature consists of successive curves with progressively smaller radii. Appropriately designed curve radii and length of curve can reduce vehicle speeds at the entry of a roundabout or conventional intersection, as well as reduce certain types of crashes. If successive reverse curves are installed in high-speed environments, introducing other speed reduction treatments prior to the approach curvature may reduce the speed changes needed between successive curves (Arndt, 2000). The use of approach curvature at downhill approaches is discouraged. Experience with approach curvature suggests that this geometric treatment should be used in conjunction with reduced speed limit signs or advisory speed signs. Sight distance should also be considered at the intersection approaches to ensure appropriate visibility of pedestrians and bicycles crossing the intersection approach or traveling through the intersection. Excessive sight distance should be discouraged to focus driver attention to the immediate roadway. Raised medians to reinforce speed reduction may become problematic for maintenance, especially in areas subject to snow plowing. 2.2.8.3 Treatment Layout/Design The length and curve geometry should be determined from the upstream segment operating speed and the target speed at the intersection. Design control information from the AASHTO Green Book joined with published documentation from Australian applications at roundabouts may provide design guidance. It may be appropriate to introduce the curved geometry concurrently with additional treatments such as warning signs. The dominant feature of approach curvature treatments is the radius of the curve. When designing successive reverse curves, each curve must be visible before the next curve to allow drivers adequate time to adjust speeds (Arndt, email). The length of the approach curvature affects speed reduction. Adequate curve length must be provided to discourage drivers from traveling into adjacent lanes; however, excessive curve lengths can result in an increase of single-vehicle crashes (Arndt, 2000). Super-elevation and side friction also affect vehicles traveling through the curvature. To provide required super-elevation on each curve, short tangent sections can be designed between each successive curve (Arndt, email).

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-35 Similar to roundabout design, approach curvature should be designed with an appropriate “design vehicle” to ensure that all trucks, buses, and emergency vehicles may traverse the approach without encroaching on the shoulder or adjacent sidewalks. 2.2.8.4 Speed Effects Current information about approach reverse curvature and how it impacts speed and safety relates exclusively to roundabouts. Although no published research or testing results have been found that describe the impacts on speed and safety at conventional intersection approaches, similar safety benefits might be realized by applying this treatment to conventional intersections. The geometric design of roundabouts shows that the curve radius directly impacts vehicle speeds. As the radius of the curve decreases, a larger reduction in vehicle speed is required to negotiate the curve. Exhibit 2-18 shows the speed-radius relationship for curves with +0.02 and – 0.02 super-elevations. Exhibit 2-18 Speed-Radius Relationship (FHWA, Roundabouts: An Informational Guide, 2000). 2.2.8.5 Safety Effects Although the majority of research and testing includes approach curvature at roundabouts, similar safety benefits might be realized by applying this type of geometric treatment to conventional intersections. Approach curvature that includes successive curves reduces crashes at a roundabout and minimizes single-vehicle crashes. A Queensland, Australia, study found that reducing the change in 85th percentile speed on successive curves to 12 mph reduced single-vehicle and sideswipe crashes. Although decreasing the curve radius on an approach can generally minimize rear-end crashes, it may also increase the potential for single-vehicle crashes. (FHWA, Roundabouts: An

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-36 Informational Guide, 2000) 2.2.9 Splitter Islands Overview Documented applications for high-speed intersections Roundabout intersection approaches and conventional intersection approaches in New Zealand and France Function Slow, direct, and separate conflicting traffic Applicability Stop- or yield-controlled approaches Design Variations Length, geometry, landscaping Secondary Effects, Considerations Splitter islands can provide refuge for pedestrians crossing at the intersection 2.2.9.1 Applicability and Considerations A splitter island is a raised or painted area on an intersection approach used to separate entering and exiting traffic, to deflect and slow entering traffic, and to provide refuge for pedestrians crossing the road in two stages (FHWA, Roundabouts: An Informational Guide, 2000). These islands can increase driver awareness by channelizing the traffic on the minor approaches. In some cases, these treatments are called throat or fishtail islands; these are shown in Exhibit 2- 19. The geometry of fishtail islands introduces deflection at the approach, thereby reducing vehicle speeds at the intersection. Exhibit 2-19 Throat and fishtail islands While splitter islands have generally been used at roundabout approaches, this treatment has been applied to the minor approaches of T-intersections and two-way stop-controlled intersections in France and New Zealand. Exhibit 2-20 depicts a photo of a splitter island at a T- intersection in France. Similar to a splitter island at a roundabout, splitter islands at conventional intersections can benefit pedestrians crossing at the intersections. The island shortens the crossing distance, creates a refuge area, and allows pedestrians to cross the approach in two stages if necessary. FHWA has sponsored a study to determine the effectiveness of combining splitter islands on

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-37 minor approaches with other types of speed reduction treatments such as centerline rumble strips or pavement markings on major approaches at two-way stop-controlled intersections on rural highways. Exhibit 2-20 Splitter Island at a T-intersection in France 2.2.9.2 Treatment Layout/Design The length and geometry of splitter islands vary significantly. Splitter islands can either take the form of a cross-section change including either visually or physically narrowing the travel way or creating horizontal deflection. In Exhibit 2-21, Splitter Island A creates a cross-section change without directly creating horizontal deflection. Splitter Island B conceptually creates horizontal deflection. Form A could present a visual clue to drivers to slow before the intersection. Form B could potentially reduce speed by creating a deflected path, similar to a roundabout through movement. Exhibit 2-21 Splitter Islands Exhibits 2-22 and 2-23 depict schematic layouts of signing and striping for teardrop splitter islands. The oval shape of the island aids in reducing vehicle approach speeds and increases visibility of the intersection. This splitter island can be mountable to allow larger vehicles to maneuver through the intersection. UK standards recommend that no obstacles, such as signs or signals be placed on the splitter island; that the color of the median be uniform (no striping), but different than the pavement color to offer increased day and night visibility; and that the splitter islands be constructed of a mineral treatment rather than turf or grassy compositions (Ministry of

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-38 Equipment, 1998). In US applications, a sign is usually mounted on the splitter island to increase its visibility . Exhibit 2-22 Example layout of a splitter island at a T-intersection Exhibit 2-23 Example layout of a splitter island at a 4-legged intersection 2.2.9.3 Speed Effects Much of the information about splitter islands and how they impact safety and speed relates specifically to roundabouts. There is little published research or test results to describe the impacts of approach splitter islands on speed or safety at conventional intersection approaches. Splitter islands create deflection at roundabouts, which reduces the speeds of vehicles traveling through the intersection. 2.2.9.4 Safety Effects The limited literature that is available on splitter islands shows that there is a safety benefit to applying this treatment to conventional intersections. However, there is no information to indicate to what extent the safety effects of splitter islands are due to speed reduction and to what extent they are due to the separation between traffic moving in opposite directions. A study of 134 intersections in New Zealand found that installing throat or fishtail islands resulted in crash reductions of 30 to 60 percent for fatal, injury, and pedestrian crashes during

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-39 both the daytime and nighttime. The majority of the reduction in injury crashes involved vehicles crossing paths at the intersection. Most of the treated intersections were four-legged intersections in urban areas on local roads (LTSA) (FHWA, 2006). Implementations in France and New Zealand resulted in a total crash reduction of 30 percent, and a reduction of angle and crossing crashes by 30 percent (FHWA, 2006). Although no evidence was found to support this claim, it is possible that installing these types of islands at intersections could increase collisions with obstructions (LTSA). 2.2.10 Speed Tables and Plateaus Overview Documented applications for high-speed intersections None Function Create vertical deflection, Alert drivers of the need to slow down Applicability Stop-controlled approaches Design Variations Geometry, materials Secondary Effects, Considerations Snow removal, driver expectancy 2.2.10.1 Applicability and Considerations Speed tables are speed humps with a flat section on top, allowing the entire wheelbase of a passenger car to rest on the raised, flat section. The flat section of the speed table can be aesthetically treated with decorative surface material or constructed with brick or other textured materials. Speed tables have gently sloped ramps on both ends, which allows slightly higher vehicle speeds and a smoother transition than speed humps (ITE, 2007).

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-40 Exhibit 2-24 Speed table combined with textured pavement in Naples, FL (Photo: Fehr & Peers) Exhibit 2-25 Speed table combined with striping and colored concrete in Charlotte, NC (Photo: Fehr & Peers) A plateau is an alternative speed table design that has been used in the Netherlands and installed on facilities with speeds ranging from 60 km/h (35mph) to 80km/h (50mph). In applications of this treatment, plateaus are generally installed on all intersection approaches and the minor approaches are yield-controlled. Advance warning devices are also recommended on all approaches. (Schermers et al., 2001) Although speed tables are generally applied on low-speed facilities in the United States, they may have applications on approaches to high-speed intersections where low speeds are desired. Speed tables can provide benefits for multimodal users when user needs are considered in design and application. Speed tables that are used at crosswalks can eliminate the need for curb ramps, increase pedestrian visibility, and provide more sidewalk area for multimodal users. Detectable warnings are recommended for speed tables used at crosswalks to accommodate visually impaired users (ITE, 2007). When designed with a gentle rise, speed tables have little impact on bicyclists (Cochituate Rail Trail, 2007). Vertical deflection devices such as speed tables are sometimes prohibited on emergency response routes because of the friction they create. Snow and ice removal may require special attention when plowing roadways. Speed tables should be constructed downstream of storm sewer inlets and should have a tapered edge along the curb line to allow for drainage (ITE, 1997). Additional or new street lighting may be necessary to improve the visibility of speed tables. 2.2.10.2 Treatment Layout/Design Speed tables and plateaus should be placed at a location where vehicles will not abruptly

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-41 encounter the speed table at a high speed. The 1993 ITE Guidelines for the Design and Application of Speed Humps suggests that the first speed table in a series be placed no more than 200 feet from a stop sign or horizontal curve, and not within 250 feet of a traffic signal. Both horizontal and vertical sight distance should be considered to determine the placement of a speed table. Many jurisdictions have standard placement guidelines for the location of speed tables (Hallmark, 2002). The most distinguishing design features of a speed table are the flat section in the middle of the table, the ramps on each end, and the aesthetic treatments often applied to the top. The entire speed table is typically 22 feet long in the direction of travel, with a 10-foot flat section in the middle and six-foot ramps on both ends. The vertical deflection of a speed table typically ranges from three to four inches, but has been designed for a height of six inches in some cases. The ramps on each side of the speed table typically have a parabolic or linear shape. (ITE, 2007) Plateaus should be located approximately 50 to 100 meters (165 to 325 feet) from the intersection and designed for a 40 km/h (25 mph) design speed (Schermers et al., 2001). 2.2.10.3 Speed Effects Much of the information about the effectiveness of speed tables relates to roadway segments with operating speeds less than 45 mph. No information was found that describes the affect speed tables have on speeds at intersections on high-speed facilities. Most research has shown that speeds were reduced after speed tables were installed. A study in Gwinnett County, Georgia, in which 43 speed tables were installed, found that speeds were reduced by an average of 9 mph (Bretherton, 2003). Another study reported an average 18- percent decrease in the 85th-percentile travel speeds for a sample of 58 test sites (Fehr & Peers, 2004). However, in some cases speed tables may be too gentle to mitigate excessive speed. A test site in Fort Lauderdale, Florida, found no speed reduction (ITE Traffic Calming, 2007). When designed to the typical dimensions, speed tables have an 85th-percentile speed of 25 to 30 mph (ITE, 1999). While the speed reduction potential of speed tables is not well known, the ITE study noted above suggests that the provision of speed tables results in an 85th-percentile speed of 25 to 30 mph. Thus, if a speed table is used at an intersection, the speed reduction on the intersection approach may be from the 85th-percentile mid-block speed to 25 to 30 mph. It is expected that most intersections where speed tables have been used have moderate mid-block or approach speeds. Speed tables located at intersections and mid-block locations have been shown to reduce speeds and collisions as well as lower traffic volumes. 2.2.10.4 Safety Effects Speed tables located at intersections and mid-block locations have been shown to reduce speeds and collisions as well as lower traffic volumes. Some studies have reported that collisions have been reduced by an average of 45 percent on streets treated with speed tables at particular test

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-42 sites (ITE, 2007; Fehr & Peers, 2004). Speed tables have potential to increase safety on the treated road by lowering volumes and reducing the probability of a crash. The study conducted in Gwinnett County showed a 7-percent reduction in traffic volumes based on 24-hour traffic counts collected on 18 streets before and after the speed tables were installed (Bretherton, 2003). ITE reported a 12-percent reduction in traffic volumes on roadways with speed table applications, depending on the availability of alternative routes (ITE, 2007). 2.2.11 Reduced Lane Width Overview Documented applications for high- speed intersections None Function Heighten driver attention, narrow the available lane width Applicability Expect that narrowing the lane width on the intersection approach will produce similar effects to segment applications Design Variations Can involve re-striping only, or narrowing the paved section Secondary Effects, Considerations Use with caution with heavy truck traffic, multilane facilities, and curvilinear alignments 2.2.11.1 Applicability and Considerations Reduced lane width could come in two basic forms: lanes narrowed with paint or a physically reduced roadbed width. Lane narrowing benefits using paint may be negated by the relative open field of vision. Narrow roadbeds physically constrain the cross-section but may have secondary impacts. Lane widths that are considered “reduced” tend to range from 9 to 12 feet. Exhibit 2-26 Reduced Lane Width – Painted

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-43 Exhibit 2-27 Reduced Lane Width – Road Bed Lane-width reductions have been used in both temporary and permanent applications. Permanent lane width reductions that reduce the overall pavement width are expected to have the greatest potential to reduce speeds because both the pavement width and the striping contribute to drivers’ perceptions of the roadway environment. Lane-width reductions have been documented for work zone applications and low-speed urban and residential locations. Lane widths on multi-lane facilities, roadways with curvilinear alignments, and facilities with high heavy truck or transit vehicle use should be reduced with caution, particularly if the width of the paved section will be reduced. Reducing lane widths can negatively impact bicyclists if bicycle lanes or wide curb lanes are not maintained, but it can also improve bicyclist conditions if the additional space is used to provide or widen bicycle lanes. Reducing lane width can provide other positive effects such as space for other roadway features (such as medians, and curbside parking), space for roadside features (such as sidewalks and clear zones), and reduced interference with existing roadside development (Zegeer, 2002). Reduced lane widths improve pedestrian conditions at intersections by reducing crossing distances and exposure time. Excessive crossing distances increase pedestrian exposure time, increase the potential of vehicle-pedestrian conflicts, and may add to vehicle delay. At signalized intersection approaches, reducing lane widths and thereby the pedestrian crossing distance provides more flexibility for the intersection’s signal timing. 2.2.11.2 Treatment Layout/Design NCHRP 3-72: “Lane Widths, Channelized Right Turns, and Right-turn Deceleration Lanes in Urban and Suburban Areas” evaluated the operational, speed, and safety effects of reduced lane widths on roadway segments and intersections. The research findings indicate that reducing lane widths to less than 10 feet may not be advisable on four-lane undivided arterials or on the approaches to four-leg stop-controlled intersections. Additionally, it suggests that it may not be advisable to reduce lane widths to less than 9 feet on four-lane divided arterials. These findings are consistent with the AASHTO Policy on Geometric Design of Highways and Streets, which recommends lane widths of 10 to 12 feet on urban and suburban arterials. Redesigned lanes can be reduced using concrete barriers (for short-term solutions), curbs, or standard striping. The effects of reducing the width of pavement on turning radii at intersections should be evaluated to ensure that adequate radii are available to accommodate the design vehicle.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-44 2.2.11.3 Speed Effects Research related to the speed, safety, and operational effects of reduced lane widths provides inconsistent results, indicating that the relationships are complex and difficult to evaluate without considering other elements of the intersection or roadway environment. Research and analysis conducted for NCHRP 3-72, which included both high-speed and low- speed facilities, made the following conclusions regarding the speed effects of reduced lane widths: • Reduces mid-block speeds on four-lane arterials (average lane-width reduction of 2.7 feet associated with average speed reduction of 4 mph) • Reduces driver comfort on higher-speed facilities • May decrease capacity due to reduced saturation flow rates, although the calculated reductions were about half of the reductions suggested in the HCM • May increase capacity at signalized intersections due to decreases in pedestrian crossing times Research that evaluates how effectively reduced lane widths reduce speeds in work zones found that the treatment was most effective on two-lane rural roads and found that a 7-percent reduction in speed was achieved by reducing lane widths to 11.5 and to 12.5 feet (Benekohal, 1992). In low-speed environments, research has concluded that speed is not consistently related to lane width (Gattis, 1999). It is also likely that wider lanes induce faster travel speeds that may increase crash risk and increase crash severity. Earlier editions of the Highway Capacity Manual (HCM) suggested that wider lanes on multi-lane highways also increase capacity and, therefore, reduce following distances (TRB, 1985). However, HCM editions since 1985 have indicated that wider lanes, up to 12 feet, on multi-lane highways increase free-flow speeds but do not increase capacity and, therefore, do not reduce vehicle headways (TRB, 1994; TRB, 2000). 2.2.11.4 Safety Effects The documented safety effects of reduced lane widths generally address segment applications, although some intersection approach data may be included. Despite inconsistent data, concern remains that narrower lanes may be accompanied by reduced safety. The two principal aspects to the potential link between lane width and safety are: • Wider lanes increase the average separation between vehicles moving into adjacent lanes and may provide a wider buffer to accommodate small, random deviations from drivers’ intended paths; and

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-45 • Wider lanes may provide more room for correction maneuvers by drivers in near- crash circumstances (Hauer, 2000). Research and analysis conducted for NCHRP 3-72 found no decrease in safety at mid-block locations or intersection approaches with narrow lanes except in these limited cases: • Four-lane undivided arterials with lane widths less than 10 feet • Four-lane divided arterials with lane widths less than 9 feet • Four-leg stop-controlled intersections with lane widths less than 10 feet Research by Harwood et al. developed accident modification factors (AMFs) for accident types related to cross-section elements (including opposite direction, and single-vehicle run-off-road) for two-lane highways. The AMFs range from 1.05 to 1.5 for 9- to 11-foot lanes compared to 12- foot lanes, indicating that two-lane facilities with narrower lanes would be expected to experience higher crash rates for certain crash types (Harwood, et al., 2000). The AMF of 1.50 for 9-ft lanes implies that a roadway with 9-ft lanes would be expected to experience 1.5 times as many accidents as a roadway with 12-ft lanes. However, some studies indicate that lanes wider than 12 feet may also be associated with higher crash frequencies. 2.2.12 Visible Shoulder Treatments Overview Documented applications for high- speed intersections None Function Heighten driver attention, narrow the visual width of the roadway Applicability Transition areas, roadways with existing shoulders, roadways in rural areas with no sight line limitations Design Variations Many Secondary Effects, Considerations Depending on the materials, bicyclists using the shoulder may be impacted 2.2.12.1 Applicability and Considerations Visible shoulder treatments are used to change the appearance of the paved area to either aesthetically blend roadway facilities into the surrounding landscape or to create a contrast between the shoulder and travel lane to heighten drivers’ awareness of the roadway. Visible shoulder treatments can consist of shoulder pavement markings, pavement coloring, shoulder composition, and shoulder rumble strips. If introduced in advance of the intersection, these treatments may be used to alert drivers of the change from roadway segment to the intersection influence area.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-46 Exhibit 2-28 shows an example of a visible shoulder treatment that used brick. Exhibit 2-28 Example of a shoulder composition application (Photo: Dan Burden) The applicability of visible shoulder treatments depends on the presence or absence of existing shoulders and the ability to acquire right-of-way if needed. Generally, visible shoulder treatments are most cost efficient for intersection approaches with existing shoulders. Shoulder coloring can have an effect on the thawing and freezing characteristics of the shoulder pavement. Shoulders colored black retain heat longer while light-colored shoulders have the potential to freeze more easily. Though not documented, colored treatments could fade or deteriorate because of snow removal (Straub, 1969). Consideration should be given to the effects of textured shoulder treatments on bicyclists. 2.2.12.2 Treatment Layout/Design Shoulder dimensions are typically presented in the roadway design typical sections. In cases where the shoulder exists or is planned, the special treatment area might be limited to the vicinity of the intersection influence area to help differentiate between the roadway segment and the impending intersection. 2.2.12.3 Speed Effects No documentation was found that describes the effect visible shoulder treatments have on speed or safety at high-speed intersection approaches. No documentation was found that describes multimodal impacts. 2.2.12.4 Safety Effects No documentation has been found that describes the safety benefits of visible shoulder treatment applications at high-speed intersection approaches.

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-47 2.2.13 Roadside Design Features Overview Documented applications for high-speed intersections None Function Reinforce transitioning environment, draw attention to multimodal users, reduce comfortable approach speeds Applicability Many applications Design Variations Roadside design features, gateways, and landscaping Secondary Effects, Considerations Sight distance 2.2.13.1 Applicability and Considerations The roadway environment can influence drivers’ perceptions of the road and provide safety benefits when implemented appropriately. Landscaping, cross-sectional changes, and gateways are three characteristics of the roadway environment that can affect the speed and safety of a roadway while providing aesthetically pleasing surroundings. Landscaping is typically planted along the roadside, set back from the edge or within a center median. Landscape improvements should not be planted in an area that obstructs signs, sightlines, or the visibility of motorists, bicyclists or pedestrians. A roadway’s cross section includes travel lanes, shoulders, curbs, drainage channels, side slopes, and clear zones. Other cross-sectional elements include sidewalks, bike lanes, barriers, medians, and frontage roads. Specific design guidelines for each of these elements generally depend on surrounding environment, adjacent land uses, and roadway facility type. Gateways are prominent physical features that inform drivers they are transitioning into a new roadway environment and can include landscaping, lighting, signage, or physical structures. Exhibit 2-29 Landscaped median treatment (Photo: FHWA FHD)

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-48 Exhibit 2-30 Gateway Maintaining landscaping treatments can be costly and difficult. Maintenance tasks can be labor and infrastructure intensive, and include mowing, pruning branches and shrubs, watering, and fertilizing. In addition, removing and replanting vegetation or trees can be a significant investment, especially during the first few years until the vegetation is established. Maintenance tasks require a minimum clear space of three feet from the edge of a travel lane to allow maintenance crews to work without being in danger of passing vehicles (Mok, 2006). Gateways are typically installed as an entrance to a city, neighborhood, or downtown area. The intent of these treatments is to create a welcoming physical feature that informs motorists that they are about to enter an activity area where lower speeds are desirable. Gateways are also placed at the exits of these same areas, typically to inform drivers that they are leaving an activity area. 2.2.13.2 Treatment Layout/Design Drivers must have a clear sight distance to view other vehicles approaching or pulling out at intersections, and landscape heights and locations should not impede visibility. The Ontario Ministry of Transportation conducted a study on transition zones near intersections; however, the results of this study are not yet available. 2.2.13.3 Speed Effects Much of the information about how the roadway environment impacts safety and speed relates to roadway segments. No information has been found to describe the speed reduction benefits associated with roadway environment changes at intersection approaches. Cross-sectional changes can influence the way a driver perceives the road and can affect driver behavior. A roadway with several lanes, wide shoulders, and clear zones gives the driver a feeling of openness, thus increasing the impression that they can drive fast. A narrow road with horizontal curves, steep slopes, or even a cliff on the side of the road, induces drivers to slow down. Changing the appearance of the road primarily impacts unfamiliar drivers. A study of 30 test sites in Canada reported 85th percentile speeds of 30 mph on sites with side friction, such as parking, heavy pedestrian, and bicycle activity; whereas the 85th percentile speed was measured to be 39 mph at sites with simple and open road situations. The posted speed limit at all of the sites was 30 mph. (Smiley, 1999).

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-49 Trees and vegetation can help define the edge of the roadway and slow traffic. Shinar, McDowell, and Rockwell (1974) conducted a study on two-lane rural highways that indicated that trees growing close to the edge of the road caused drivers to maintain lower speeds than on an open stretch of highway. Drivers were requested to maintain a speed of 60 mph, yet, on the stretches of roadway with trees, drivers maintained a speed of 53 mph and on the open stretches of highway drivers maintained 57 mph. (Human Factors North, Inc., 2002) 2.2.13.4 Safety Effects There is information quantifying how the roadway environment impacts safety, and much of the available information relates to roadway segments. Landscaping can be a significant safety hazard when placed within the clear zone or when it obstructs sight distance for drivers pulling out into the roadway. However, when implemented appropriately, it can provide a safety benefit. A before-and-after study conducted by Texas Transportation Institute (TTI) implemented landscape improvements to ten selected Texas Department of Transportation projects. After the landscaping was in place for three years, the mean crash rate per 1 million vehicle miles traveled (VMT) decreased approximately 20 percent. (Mok, 2006) 2.3 COMBINING TREATMENTS Intuition suggests that combining treatments will increase the potential to reduce speed. While there is little research that determines this potential, it is expected that combining treatments will have a benefit up to a point, after which no more speed reduction will occur. This is because drivers will operate at a speed at which they feel comfortable or safe; below this speed, the cumulative application of treatments becomes ineffective. Additional treatments can provide benefits by reinforcing the need to be prepared to slow down, even if additional speed reduction is not observed. An example of combining treatments is a low-cost concept for two-way stop-controlled intersections on high-speed, two-lane rural highways studied by the Federal Highway Administration (FHWA). The concept incorporates a variety of individual treatments, including lane narrowing, splitter islands on each approach, and lateral pavement markings on each side of the traveled way.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-50 Exhibit 2-31 FHWA Low-Cost Treatment Concepts (FHWA, 2006) The objectives of the FHWA Low-Cost Intersection Treatments on High-Speed Rural Roads study were to identify, promote, and evaluate low-cost concepts to reduce speeds at intersections. The research team investigated two concepts that combined various roadway treatments. The first concept reduced the lane width, added rumble strips, and added pavement markings on the major road. The second concept was to install a mountable splitter island with stop signs on the minor road approaches. Exhibit 3-4 shows the proposed concepts (FHWA, 2006). The treatment concepts were implemented in multiple locations in Pennsylvania, New Mexico, and Illinois. The results from this study showed statistically significant speed reduction at all sites. The combination of treatments reduced all vehicle speeds by an average of 3 miles per hour, and reduced the 85th percentile speeds by 4 miles per hour. In addition, testing results for trucks revealed an average speed reduction of 5 miles per hour and a reduction in the 85th percentile speeds of 4 to 5 miles per hour. (FHWA, 2006) The team collected crash data for five years before deployment and will collect data for two years after deployment to determine whether these treatments yield quantifiable results. The team also plans to analyze crash data at the various sites at the project’s end in June 2009 because several years of data are needed to determine the potential effects the treatments had on roadway safety. (FHWA, 2006)

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-51 2.4 SPEED ENFORCEMENT 2.4.1 Overview Enforcement is critical to achieving compliance with speed limits (TRB Special Report 254, 1998). There are perceived advantages and disadvantages to exceeding the posted speed. Perceived advantages may include saving time or having a thrilling driving experience, while disadvantages include the increased risk of a crash or the possibility of being caught. Whether using traditional officer-based enforcement or automated enforcement techniques, the goal of enforcement is to deter drivers from driving over the speed limit by increasing the disadvantages of being caught speeding (ACC and LTSA, 2000). The following discussion summarizes principles of speed enforcement in general and highlights advantages and disadvantages of automated enforcement. 2.4.2 Speed Enforcement Principles Police enforcement mechanisms include specific deterrence and general deterrence. Specific deterrence focuses on an individual driver who is speeding and is intended to influence that individual’s behavior by catching and punishing them. This assumes that once a driver is caught and punished for speeding, that driver will be less likely to speed in the future. General deterrence is focused on the entire driving population. Whether or not there are actual increased enforcement activities, this type of deterrence strives to influence the overall driving population by increasing the public perception that drivers who speed will be caught (ACC and LTSA, 2000). Studies have shown that actual enforcement and deterrence from speeding is most effective when combined with public information campaigns (TRB Special Report 254, 1998). Deterrence is also affected by punishment factors such as the perceived certainty, severity, and immediacy of the punishment (ACC and LTSA, 2000). Traditional speed enforcement involves catching the speeding driver and applying a punishment at the site where speeding occurred. In many cases, speeding drivers are detected with a radar device from a police officer adjacent to the road. The speeding driver is then stopped and issued a penalty depending on the severity of the offence (ACC and LTSA, 2000). This enforcement approach has a short-lived effect in deterring speeding (TRB Special Report 254, 1998). The time or distance that a driver is affected and deterred due to enforcement is referred to as the “halo” effect. The distance halo effect is the distance on either side of the enforcement site over which there is a reduction in speeding behavior. The time halo effect is the amount of time from enforcement activity during which speeds at the enforcement site are reduced. Various studies have determined that the distance and time halo effects are dependent on the enforcement strategy. When enforcement is of high intensity and randomly placed the effects can last up to eight weeks and can extend for a distance of 20 kilometers (ACC and LTSA, 2000). Two approaches of traditional speed enforcement are used: high visibility and low visibility. The high visibility approach aims to increase the overall perceived risk of being caught by deterring drivers from speeding at the enforcement site, whereas the low visibility approach aims to inform

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-52 drivers that enforcement is unpredictable. Whether to use stationary or moving police enforcement is another decision to make when choosing to use police enforcement. Of the speed enforcement programs conducted throughout the United States, the most successful in deterring speeding were those that: • Placed enforcement at specific locations and at times when speeding is most likely to occur; • Made enforcement highly visible to the public; and • Maintained enforcement for more than a year (ACC and LTSA, 2000). 2.4.3 Automated Speed Enforcement Automated speed enforcement consists of a detection device (radar device), a processing unit, and an image-recording device (camera). This technology measures vehicle speed and then takes a picture of a speeding vehicle and driver, including the time and date of the incident. A warning letter or citation is mailed to the registered owner of the vehicle, using the vehicle license plate to identify vehicle ownership (ACC and LTSA, 2000). Some of the advantages to automated speed enforcement are (ACC and LTSA, 2000): • Increases the probability of detection without overextending police resources; • Increases road users’ perception of the risk of getting caught; • Increases fairness of enforcement by eliminating “officer discretion;” • Creates an efficient ticketing process by reducing the number of disputes between motorists and police officers; and • Enforces speeds in locations where patrol vehicles cannot safely or effectively patrol. Some disadvantages to automated speed enforcement include (ACC and LTSA, 2000): • The delay between the offense and the punishment, which is typically two to three weeks. • The speeding driver does not always immediately realize his or her offense has been detected. • Community acceptance varies for this type of enforcement technology. 2.4.4 Speed Enforcement Summary Speed enforcement can be an effective tool for speed management. Automated enforcement has been successfully used in many geographic areas but remains controversial in many

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-53 communities. Speed enforcement could be used in conjunction with other countermeasures or as a single countermeasure. Funding, resource limitations, and prioritizing enforcement locations are real-world variables that can affect its success. There is varied community acceptance of automated enforcement. This study is focused primarily on physical treatments that might be applied. Further exploration of enforcement as a unique speed reduction treatment will not be conducted as part of this work effort. 2.5 STATE AGENCY SURVEY 2.5.1 Introduction This section presents a summary of the responses to the survey questionnaire sent to state highway agencies concerning speed reduction treatments on intersection approaches. The survey solicited specific suggestions on potential treatments, and requested expressions of interest in participating in the treatment evaluation process during Phase II of the project. The survey was designed to allow agencies to share their input about what speed reduction treatments they felt were most important to be evaluated when determining a level of effectiveness. To maximize the number of responses, hard copies of the survey were sent via mail and additional survey copies were also emailed to each contact. A copy of the survey distributed to the state highway agencies is shown in Appendix C. 2.5.2 Survey Recipients The mailing list for the survey included the 50 state highway agencies. The questionnaires for state highway agencies were generally sent to the state traffic engineer. Names and addresses of the state traffic engineers were determined from the membership roster of the AASHTO directory, previous mailing lists, and state DOT web pages. The mailing list used for distributing the survey is shown in Appendix D. 2.5.3 Response Rate A total of 38 responses were received out of the 50 questionnaires that were mailed, for a response rate of 76 percent. Exhibit 2-35 presents a list of the state highway agencies that responded to this survey.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-54 Exhibit 2-32 List of Highway Agencies That Responded to Survey State highway agencies Alabama Mississippi Arizona Missouri California Montana Colorado Nevada Connecticut New Hampshire Delaware New Jersey Georgia New York Hawaii North Dakota Idaho Oklahoma Illinois Oregon Indiana Pennsylvania Iowa Rhode Island Kansas South Carolina Kentucky Tennessee Louisiana Vermont Maine Virginia Maryland Washington Michigan West Virginia Minnesota Wyoming 2.5.4 Summary of Survey Responses The highway agency responses to each question in the survey are summarized below. Where appropriate, the responses are tabulated. 2.4.4.1 Question 1—Type of Highway Agency Represented In Question 1, survey respondents were asked to identify what type of agency they worked for (i.e., state, city, county, other). Since the survey was eventually mailed only to state agencies, 100 percent of the respondents answered accordingly. 2.4.4.2 Question 2—Geometric Speed Reduction Treatments Highway agencies were asked about the types of geometric speed reduction treatments that they use. For agencies responding to the survey, 71 percent indicated that they incorporated some geometric treatment specifically intended to reduce speeds at at-grade intersections. Agencies

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-55 indicating that they used a particular treatment were asked whether or not they found it to be effective and also whether it was currently in use. Exhibit 2-36 presents the responses to this question based on a list of typical treatments given in the survey. In addition to the treatments listed in Exhibit 2-36, one agency currently utilizes curb and gutter cross sections as an alternative treatment to reduce speeds and finds them to be effective. Exhibit 2-33 Highway Agency Use of Geometric Treatments to Reduce Speeds Treatment No. of Agencies Effective In current use Rumble strips 22 82% 20 91% 19 86% Roundabouts 16 59% 15 94% 15 94% Shoulder treatments 9 33% 5 56% 9 100% Roadside treatments 7 26% 4 57% 6 86% Reduced lane widths 5 19% 3 60% 5 100% Reverse horizontal curvature 3 11% 3 100% 3 100% Speed tables 3 11% 2 67% 3 100% Horizontal curvature 2 7% 2 100% 1 50% Roundabout-like treatments 1 4% 1 100% 1 100% NOTE: The percentage of agencies represents the total number of responding agencies; however, the percentage of responses on whether the treatment is effective or in current use corresponds only to the number of agencies using that particular treatment. 2.4.4.3 Question 3—Signing Intended to Reduce Traffic Speeds Exhibit 2-37 summarizes the number and percentage of agencies that use dynamic warning signs in order to reduce traffic speeds on intersection approaches. Approximately 45 percent of the responding state agencies have implemented this general type of treatment. There were several additional signing methods mentioned by respondents beyond those suggested in the survey. These include: • Advance warning signs with flashers (3 respondents) • Flashing yellow beacons on W3-3 (Signal Ahead) signs • Special enforcement area signs • Stopped traffic ahead warning signs with beacons activated by the red signal phase (2 respondents) • Brighter colored yellow warning signs • Intersection warning signs • Intersection warning signs supplemented with advisory maximum safe speed signs

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-56 Four of the nine agencies citing the use of these warning signs also indicated that they were found to be effective and currently in use; except in those cases where they are still being installed on a trial basis and effectiveness data is not available. Exhibit 2-34 Highway Agency use of Signing to Reduce Speeds Type of Sign No. Of Agencies Effective In current use Speed activated dynamic warning signs 6 35% 2 33% 5 83% Conflicting vehicles activated dynamic warning signs 3 18% 2 67% 1 33% NOTE: The percentage of agencies represents the total number of responding agencies; however, the percentage of responses on whether the treatment is effective or in current use corresponds only to the number of agencies using that particular treatment. 2.4.4.4 Question 4—Pavement Markings Intended to Reduce Traffic Speeds Question 4 asked highway agencies about their policies for the use of pavement markings in order to reduce traffic speeds. Twenty-nine percent of the responding agencies indicated a use of at least one of the treatments listed in Exhibit 2-38. Only two agencies gave information on other types of pavement markings used in their jurisdiction that were not mentioned in the survey. Both agencies stated that pavement markings were used to adjust the lane width on intersection approaches; however, neither has found any effectiveness in this method. Exhibit 2-35 Highway Agency use of Pavement Markings to Reduce Speeds Type of Sign No. Of Agencies Effective In current use Transverse pavement markings on the roadway 7 64% 5 71% 7 100% Wider edge line markings 4 36% 0 0% 4 100% Transverse pavement markings on the shoulder 3 27% 2 67% 3 100% Wider centerline markings 3 27% 0 0% 2 67% NOTE: The percentage of agencies represents the total number of responding agencies; however, the percentage of responses on whether the treatment is effective or in current use corresponds only to the number of agencies using that particular treatment. 2.4.4.5 Question 5 Question 5 asked respondents to forward any written policies, specifications, or typical drawings associated with their answers to Questions 2, 3, & 4. One respondent (Maryland) provided the following: • Speed reduction specifications with diagrams for their recommended placement along the roadway

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-57 • Speed reduction sign specifications • Cost estimate table for speed reduction treatments for a specific project they have completed • Report on a study they published, entitled “Optical Speed Bar Application” Another respondent (Virginia) provided a one-page document with the specifications for signing and placing rumble strips on an intersection approach. 2.4.4.6 Question 6—Prior Research on Vehicle Speeds and Geometric Treatments Question 6 asked highway agencies to indicate whether they have conducted formal research on the effects on vehicles speeds of treatments that they identified in Questions 2, 3, and 4. Seven agencies (18 percent) responded affirmatively: • Maine – No specific details were provided. • Maryland – Maryland State Highway Administration conducted a before-and-after study on speed measurements to test the effectiveness of optical speed bars installed at two locations on a test roadway segment. They found that, depending on the time of day, the 85th percentile running speed was reduced by 2.4 to 9 percent. • Minnesota – In 2003, MnDOT performed an evaluation of a rural unsignalized intersection before and after the installation of improved pavement markings that were intended to accentuate the presence of the intersection. The immediate impacts of the study showed no reduction in vehicle speed, but it is yet to be determined if there are fewer crashes in the long-term. • Mississippi – MDOT participated in the NCHRP Pooled Fund Study on Novel Traffic Control Devices. No specific details were provided. • New Hampshire – NHDOT performed before-and-after studies for driver feedback signs (i.e., “Your Speed Is ___”). In addition, they have plans to compare before- and-after crash data for sites where “Stopped Traffic Ahead” signs have been applied. No specific details were provided. • New York – NYSDOT is involved in an FHWA Pooled Fund Study along with Texas and Mississippi to study the effects of pavement markings. The survey respondent was uncertain whether the markings were on the traveled way or on the shoulder, but indicated that all sites were freeway off-ramp approaches. It was noted that MS and TX might be using conventional highways. • Virginia – VDOT conducted research as a part of an evaluation for their Traffic Calming Guide. No specific details were provided.

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-58 2.4.4.7 Question 7—Potential Importance of an Effectiveness Evaluation Question 7 was designed to allow agencies to share their input about what speed reduction treatments they felt were most important to be evaluated to determine a level of effectiveness. This was accomplished by asking respondents to assign a numerical value to each treatment corresponding to its importance on a scale from 1 to 5. A numerical value of 5 corresponded to the “greatest need for evaluation” and a value of 1 corresponded to the “least need for evaluation.” Exhibit 2-39 presents the results in descending order of importance, as ranked by highway agencies. In addition to the average value assigned to each treatment, minimum and maximum value columns are included for each treatment. Exhibit 2-36 Intersection Speed reduction Treatment Importance Rankings Intersection Speed reduction Treatment Average Value Min Value Max Value Rumble Strips in Traveled Way on Approach 3.5 1 5 Dynamic-warning Speed Activated Signs 3.4 1 5 Roundabouts 3.4 1 5 Dynamic-warning Signs Activated by Potentially Conflicting Veh. 3.3 1 5 Transverse Pavement Markings on Roadway 3.1 1 5 Transverse Pavement Markings on Shoulder 2.9 1 5 Wider Edge line Markings 2.6 1 5 Roundabout-like Treatments 2.5 1 5 Shoulder Treatments 2.5 1 5 Reduced Lane Widths 2.4 1 5 Roadside Treatments 2.4 1 5 Wider Centerline Markings 2.3 1 5 Horizontal Curvature on Approach 2.0 1 5 Reverse Horizontal Curvature on Approach 1.9 1 5 Speed Tables 1.7 1 4 Other – Advanced Warning Signs, Flashers (one respondent) 5.0 5 5 2.4.4.8 Question 8—Speed Reduction Treatment Evaluation Criterion Question 8 asked respondents to rank a list of possible criteria for a speed reduction treatment evaluation, according to how important the results obtained by incorporating each criterion would be to their agency. As in the previous question, a response of 5 indicates a “very important” status and a response of 1 indicates that the criterion is “not important.” Exhibit 2-40 presents the results in descending order of importance as ranked by responding agencies. In addition to the average value assigned to each treatment, minimum and maximum value columns

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-59 are included for each treatment. Exhibit 2-37 Intersection Speed reduction Treatment Evaluation Criteria Rankings Intersection Speed reduction Treatment Average Value Min Value Max Value Mean or 85th percentile vehicle speed before vs. after treatment 4.5 1 5 Accident frequency or severity before vs. after treatment 4.4 1 5 Vehicle speed profile 3.9 1 5 Vehicle lateral position 2.2 1 5 2.4.4.9 Question 9—Additional Evaluation Criterion Question 9 asked respondents whether their agency considered any other criteria important enough to be included in an evaluation study of speed reduction treatments at unsignalized intersections. Twelve agencies (32 percent of respondents) provided additional criteria for consideration, including: • Cost of treatments (Benefit/Cost analysis) (3 respondents) • Rural vs. urban locations (2 respondents) • Type of roadway (2 respondents) • Community and political fallout • High vs. low speed intersections • Approach grade • Reduction in the high end of the speed profile; “We’ve seen no change in the 85th percentile speeds but a reduction in the numbers of vehicles traveling 10+ mph above the speed limit.” • Weather effects • Traffic volume effects • Vehicle type effects • 10 mph pace • Effectiveness over extended time periods

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-60 2.4.4.10 Question 10—Recently Installed Speed Reduction Treatments Question 10 asked highway agencies to provide a description of any treatments that have been installed in their jurisdiction in the last three years that they would be willing to have included in a future evaluation. Six agencies (16 percent of respondents) indicated that they have installed such treatments and would be willing to participate in this research. Exhibit 2-41 provides a list of states with the corresponding treatments that they have installed. Exhibit 2-38 Possible Evaluation Locations and Treatments State Treatment(s) Arizona Advance warning signs and flashers Colorado Roundabouts (4 locations) Kansas Wide edge lines Narrow lanes Speed trailers Lane shift Maryland Rumble strips Transverse shoulder markings Wide edge lines Special enforcement area Missouri Flashers activated by vehicle presence Oregon Rumble strips in traveled way 2.4.4.11 Question 11—Speed Reduction Treatments Available for 2005 Study Question 11 asked highway agencies to indicate their interest in participating in an effectiveness evaluation of speed reduction treatments at intersection approaches to be installed in 2005. Exhibit 2-42 presents the treatments that are being sought for study along with those states indicating a willingness to cooperate in such a study. Rumble strips on the traveled way was the treatment most often sited by states as being of interest for the evaluation. Kansas indicated in their response that they would be interested in participating but did not include any specific treatments. Agencies were asked if there were any other treatments that they thought would be worthwhile to consider. In response to this Arizona expressed an interest in having their advance warning signs and flashers studied. Maryland indicated they may have another treatment for consideration, but only provided a contact person.

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-61 Exhibit 2-39 Potential Speed Reduction Treatment Study Locations Treatments Total Pct (%) AZ CO LA IA MD MI MS MO NH NY PA VA WA WY Rumble strips in traveled way 11 73 X X X X X X X X X X X Wider centerline markings 7 47 X X X X X X X Wider edge line markings 7 47 X X X X X X X Dynamic warning signs activated by speed of vehicle 6 40 X X X X X X Transverse pavement markings on roadway 6 40 X X X X X X Reduced lane widths 5 33 X X X X X Roadside treatments 5 33 X X X X X Roundabouts 5 33 X X X X X Shoulder treatments 5 33 X X X X X Transverse pavement markings on shoulder 5 33 X X X X X Dynamic warning signs activated by conflicting vehicles 4 27 X X X X Roundabout-like Treatments 3 20 X X X Horizontal curvature on approach 2 13 X X Reverse horizontal curvature on approach 1 7 X Other 2 13 X X 2.4.4.12 Question 12—Comments Question 12 asked highway agencies to provide additional comments, if desired. Several agencies expressed an interest in this study and reaffirmed the necessity of such research. Those agencies include: • Kansas – “I think that this is very important research as communities want traffic going through their town to go slower. We are looking for good ways to do this.” • Maryland – “This is needed research – the goal is to favorably alter driver behavior, without them consciously knowing it!”

Chapter 2 NCHRP 3-74 Preliminary Findings Selection of Speed Reduction Treatments at High-Speed Intersections Page 2-62 • Oregon – “Oregon is interested in the results of this study, however to date we have little experience with many of the treatments mentioned above. Our most relevant experience is with the installation of a couple of roundabouts in Bend and Astoria as an alternative to signalization.” • California – “We are starting a research effort regarding potentially assessing approaching vehicle speeds and adjusting signal time to reduce or eliminate dilemma zones.” • Missouri – “We are bound by budget when agreeing to participate with (projects). We can only participate within budgetary constraints. Our budget for 2005 has not been completed at this time.” • Vermont – “We will be doing a 3 year study on Dynamic Pavement Marking starting in the spring next year.” 2.4.4.13 Question 13 Question 13 asked the survey respondent to provide their contact information. 2.4.4.14 Question 14—Other Agencies Possibly Interested in the Project In Question 14, respondents were asked to identify other agencies in their state, including cities and counties, they think may be interested in participating in an effectiveness evaluation of speed reduction treatments. Six agencies responded to this question. The additional agencies identified in this question will be contacted in Phase II of the research. A list of responding state agencies along with the recommended local agency is provided below: • Alabama – City of Auburn • Hawaii – City and County of Honolulu • Maryland – City of Trappe Landing • Missouri – Mid-America Regional Council (Kansas City, MO) • Montana – “most cities in Montana” • Oklahoma – Cities of Oklahoma City and Tulsa 2.5.5 Conclusions The 38 responding state agencies provided insight into the current practices being utilized in speed reduction on high-speed intersection approaches. It was reported that geometric treatments (71 percent) are more often used as a speed reduction measure than dynamic signing (45 percent) or pavement marking (29 percent) by state highway agencies. The necessity of this type of research became evident as only 18 percent of reporting agencies indicated that any prior

NCHRP 3-74 Chapter 2 Selection of Speed Reduction Treatments at High-Speed Intersections Preliminary Findings Page 2-63 research on the topic had been undertaken. Rumble strips were identified as having the greatest need for evaluation based on their potential effectiveness. While a wide range of speed reduction criteria were offered by the agencies for an effectiveness evaluation, the mean or 85th percentile vehicle speed, before and after a treatment is installed, was noted most often as an important criterion.