Work Zone Speed Management (2015)

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.

7 PROBLEM STATEMENT SUMMARY AND SCOPE OF THIS SYNTHESIS REPORT This Synthesis of Practice focuses on speed management for work zones on roadways with ordinary (preconstruction) speed limits of 45 mph (70 km/h) and above. Such facilities include freeways, tollways, multilane divided rural highways, and many two-lane and multilane undivided rural highways. Work zone speed management for streets and highways with ordinary speed limits below 45 mph is beyond the scope of this report. For this synthesis, speed management techniques have been organized into several categories: • Engineering Technologies (speed management devices) • Engineering Techniques (changes in the physical or per- ceptual driving environment) • Operational Techniques (using lead vehicles or field per- sonnel to limit traffic speeds) • Traditional “Human” Enforcement Techniques (police officers in cars) • Automated Speed Enforcement • Education and Outreach • Combinations of the above. The purpose of this synthesis report is to identify and com- pile strategies, practices, and technologies that transportation agencies have used to manage speeds and reduce speed-related risks in highway work zones. The information presented in this document has been assembled through: • A review of the research literature on work zone speed management devices. • Evaluation of selected portions of design and policy manuals published by U.S., Canadian, and European highway agencies. • Two surveys of state DOTs. The first survey focused on engineering- and enforcement-related techniques. The second focused on public outreach related to work zone speed management. • Follow-up interviews with selected survey respondents from several U.S. states and one Canadian province. • Compilation of information on state agency work zone public information campaigns and associated metrics. Information discussed includes: • Available data regarding the effectiveness of various methods and devices for managing speed in work zones. Highway work zone safety is of paramount importance to state departments of transportation (DOTs), toll highway authori- ties, and other transportation agencies. Contractors, construc- tion and maintenance workers, law enforcement personnel, engineers, and road users are also important both as stake- holders and as influencers of work zone safety. Toward Zero Deaths: A National Strategy on Highway Safety (TZD Steer- ing Committee 2014) identifies six work zone-related strate- gies for improving highway safety: 1. Improve speed management and enforcement in work zones to reduce the risk of work zone fatalities. 2. Improve work zone design and operations to reduce the risk of work zone fatalities. 3. Educate drivers on safer driving practices in work zones. 4. Educate workers on safety practices in work zones. 5. Educate judges, prosecutors, and law enforcement on . . . risks related to work zones. 6. Enact legislation and implement automated traffic enforcement—including pervasive automated speed enforcement and applications for school and work zones. Work zones affect both safety and mobility, and there are complicated relationships between work zone traffic speeds and overall safety. For example, high-speed traffic may be perceived as dangerous by road workers, while unstable stop- and-go traffic conditions in the work zone can increase speed differentials and the risk of a driver being involved in a back- of-queue crash or sideswipe. Generally speaking, two categories of speeding have been defined (Howard 2008): 1. Excessive speed: exceeding the posted or statutory legal speed limit. 2. Inappropriate speed: driving too fast for the prevailing road and traffic conditions, but within the posted or statutory limits. Work zone speed management serves three primary objectives: 1. Reducing collision risks for drivers and other road users; 2. Protecting the transportation workforce, including law enforcement personnel operating in the work zone; and 3. Providing effective movement of traffic. chapter one INTRODUCTION

8 Appropriate institutional arrangements are required for implementation of work zone speed management tech- niques and can be particularly challenging for some of the enforcement-related techniques discussed in chapter six. As Hyman has noted (2012): Effective work zone management within a transportation agency cuts across organizational boundaries and involves construction, maintenance, safety, and operations personnel. More signifi- cantly, many operational strategies . . . require strong cooperation from many different organizations, such as transportation depart- ments, police, fire, emergency medical services, and towing and recovery. Readers seeking additional information about work zone speed management strategy implementation may find the following resources to be of interest: • NCHRP Report 746: Traffic Enforcement Strategies for Work Zones provides information about the administra- tion of work zone speed enforcement, along with related issues such as determining how much enforcement is required and where to position police vehicles (Ullman et al. 2013). • Institutional Architectures to Improve Systems Opera- tions and Management (SHRP 2 Report S2-L06-RR-1) provides guidance to assess and increase organizational readiness to take on traffic management techniques that require higher levels of internal business process inte- gration and external coordination (Tarnoff et al. 2012). • ISO Standard 39001 Road Traffic Safety (RTS) Manage- ment Systems establishes benchmarks that an organiza- tion (such as a work zone oversight agency, highway maintenance department, or construction contractor) can use to establish and certify ongoing efforts to improve work zone safety or other aspects of roadway safety that the organization is able to influence (ISO 2012). • Toward Zero Deaths: A National Strategy on Highway Safety establishes a shared vision for safer roadways that encourages professional collaboration across organiza- tional, jurisdictional, and ideological boundaries (TZD Steering Committee 2014). • British Traffic Signs Manual, Chapter 8 describes tech- niques for limiting roadworks (work zone) speeds to as little as 10 mph; these techniques may be particularly rel- evant for urban streets and minor roads (DfT 2013a–c). OBJECTIVES OF WORK ZONE SPEED MANAGEMENT A report prepared for the Transportation Association of Canada identified 11 interrelated work zone speed manage- ment issues and objectives (Harmelink and Edwards 2005): 1. Work zone speed management philosophy. 2. Extent to which it is desirable and feasible to reduce work zone speeds. • Summaries of the methods DOTs and their partner law enforcement agencies have utilized to enforce work zone speed, including automated enforcement. • An overview of recent work zone safety public out- reach materials produced by or on behalf of state DOTs. • Examples of combination techniques that have been used by highway agencies. • Case examples illustrating some of the techniques deployed by transportation agencies. To address the report objectives, this document summarizes current research about the effectiveness of various work zone speed management techniques. As such, it is intended to assist transportation agencies in selecting techniques that are appro- priate for each specific work zone, as well as assisting practitio- ners in selecting statewide or programmatic work zone safety techniques (such as public information campaigns) that reflect what is known about why drivers choose unsafe speeds in work zones. The report also discusses some cases where speed man- agement is unlikely to be effective; in such situations agencies must consider other methods for protecting road users and workers, such as diverting traffic to alternate routes or using barriers to isolate the workforce from live traffic. Although they have a secondary effect on speeds, systems for managing work zone queuing (such as dynamic late merge systems) are beyond the scope of this report. Similarly, this report does not address intrusion alarms or other systems for alerting workers to the approach of errant vehicles. Although some speed management techniques (such as public outreach campaigns) are programmatic in nature, engineering- and enforcement-related work zone speed man- agement techniques are often implemented at the highway proj- ect level. A detailed discussion of all safety-related elements of the work zone planning and design process is beyond the scope of this report; however, it can be recognized that the ini- tial selection of speed management techniques usually begins fairly early in the project planning and design process, often as part of the Transportation Management Plan (TMP) or Main- tenance of Traffic (MOT) plan. As each project progresses, its speed management techniques may require adjustment and refinement based on actual field conditions and experience. While the need to manage work zone speeds exists world- wide, U.S. readers may wish to note that under current (2015) policies, FHWA generally allows highway improvement funds to be used to implement work zone speed management (including law enforcement) on federal-aid projects. More- over, the speed management process for federal-aid projects does not end when construction is done: Federal Work Zone Safety & Mobility Rule (23 CFR 630) requires states to con- duct ongoing work zone performance assessment and process reviews by analyzing crash and operational data from multi- ple projects and reviewing randomly selected projects. FHWA states that the assessment results are to be used to improve processes and procedures, data and information resources, and training programs for work zones (FHWA 2005).

9 • When conditions become unstable average speeds may decline sharply, but crash rates may increase as a result of abrupt fluctuations in the running speed. It is also important to recognize that some speed reduction techniques can reduce average running speeds while increas- ing speed variation (differences between the fastest and slow- est vehicles), and this can increase crash frequency. Large speed differentials between the fastest and slowest vehicles are associated with rear-end collision risks. Large differences between the speeds in adjacent lanes (unless they are sepa- rated by barriers) are also undesirable, as are situations that are likely to provoke a breakdown into unstable flow (or oscilla- tion between stable and unstable flow). Such situations can be exacerbated by other adverse factors that are common in work zones, such as limited sight distance, poor visibility, exces- sive glare from natural or artificial lighting (e.g., during night work), or work operations that are distracting to the driver. Khattak and Targa (2004) explored relationships between work zone speeds and crash severity. Their statistical analy- sis included nearly 3,400 North Carolina work zone crashes that occurred in 2000. As shown in Figure 1, the study mod- eled the relationship between posted work zone speed lim- its and the probability that the crash resulted in a casualty (injury or death). Casualty probability increased substan- tially with higher speeds, given a crash. The study’s authors calculated that every 10 mph increase in the authorized speed limit increases the chances that a crash will result in injury by 8% and increases the economic costs of the crash by 15%. Stable Traffic Flow Conditions Stable flow generally occurs under low-to-moderate traffic volumes, making it the most frequently observed operational regime in rural work zones (during most hours of the day), and in urban work zones during off-peak hours. Inappropri- ately high speed is almost certainly a contributing factor in driver and worker casualties in work zones with freely flow- ing traffic. In 2004, WHO and the World Bank co-developed a World Report on Road Traffic Injury Prevention (Peden et al. 2004). The report takes the perspective that highway crashes caused 3. Determining where to apply work zone speed limit reductions. 4. Division of work zone safety responsibilities. 5. Roadway type (urban or rural, high or low speed, freeway or nonfreeway). 6. Driver attitudes to work zone speed reductions. 7. Work zone design. 8. Provision of work zone information to the public, including locations, duration, effects, etc. 9. Credibility of work zone signing and posted work zone speeds. 10. Work zone speed reduction techniques. 11. Work zone speed enforcement. In recent decades there has been gradual change in the approach toward speed management, with increasing empha- sis on persuading drivers to select speeds that offer mobil- ity without compromising safety. Human factors and the interaction between road users, vehicles, and the roadway environment have become a greater consideration, with rec- ognition of the need for systems that anticipate and allow for human error while minimizing the risk of casualties (World Health Organization 2008). These considerations are particu- larly important in work zones because: • The work zone driving environment usually increases the overall cognitive workload for the driver (owing to narrow lanes, temporary traffic control devices, visual distractions, work zone activities, etc.). • The enforceability of the speed limit is often hampered by a lack of physical space suitable for issuing citations and safely re-entering the traffic stream. SPEED AND SAFETY Because of the wide range of traffic conditions found in work zones, the relationships between operating speed and safety are somewhat complicated. A distinction must be made between work zones that are operating under stable traffic flow con- ditions and those where the flow is unstable (“stop-and-go” traffic): • In general, reducing speeds can be expected to improve safety when the traffic flow is stable. FIGURE 1 Predicted probability of a work zone crash resulting in a casualty (injury or death) for various posted speed limits (Khattak and Targa 2004).

10 mation presented to the driver. At higher speeds drivers do not have time to process all of the information that is being gathered visually; the human brain compensates by analyzing only the central part of the image (OECD 2006). As a result, high-speed drivers are less likely to notice objects on the side of the roadway such as work- ers, work vehicles, or construction debris. Unstable Traffic Flow Conditions Unstable traffic flow (stop-and-go traffic) occurs frequently in work zones when the traffic demand exceeds the available capacity. As a result, unstable flow is often associated with urban work zones, especially during peak hours. Unstable flow can also occur in rural work zones, particularly during periods of high demand such as holiday weekends or hours with peak tourist and recreational traffic demand. In both cases, transi- tions from freely flowing traffic upstream to stop-and-go con- ditions in the work zone can be hazardous, particularly if the speed change is abrupt, inconsistent with driver expectations, or occurs under conditions that limit visibility. While excessive speed (exceeding the speed limit) is often the main concern in work zones that are operating under stable traffic flow conditions, inappropriate speed (driving too fast for prevailing conditions) contributes to crashes and near-misses in work zones with stop-and-go traffic. Unstable flow (or oscillation between free-flow and unstable traffic operations) is likely to increase the risk of rear-end crashes and same-direction sideswipes. Two fairly common crash scenarios are: • When traffic becomes congested, running speeds within a lane can change rapidly. Some drivers overaccelerate as speeds increase, and then brake sharply when the by excessive and inappropriate speed are a preventable pub- lic health problem that has become one of the leading causes of injury and death worldwide. The report makes a number of general observations about speed-safety relationships: • Higher speed reduces the time available for stopping and crash avoidance. For example, a car proceeding at 30 mph typically requires approximately 43 ft to stop, whereas one traveling at 25 mph can generally stop in less than 30 ft. Stated somewhat differently, in this instance the 25% increase in speed results in about a 50% longer stopping distance. • Speed magnifies driver error and increases crash risk. An increase in average speed of 1 km/h (0.6 mph) typi- cally results in a 3% higher risk of a crash involving injury, with a 4% to 5% increase for crashes that result in fatalities. Conversely, a 1 km/h decrease in travelling speed can be expected to reduce crashes by 2% to 3%. • Speed increases impact severity when a collision does occur. For car occupants involved in a crash with an impact speed of 50 mph, the likelihood of death is approx- imately 20 times what it would have been at an impact speed of 20 mph. • Speed reduces crash survivability for pedestrians, bicy- clists, and unprotected workers. As shown in Figure 2, pedestrians have been shown to have a 90% chance of survival when struck by a car travelling at 20 mph, but less than a 50% chance of surviving an impact at 30 mph. Pedestrians and unprotected road workers have almost no chance of surviving an impact at 50 mph. A closely related point is raised in a 2006 report from the Organization for Economic Cooperation & Development: • Speed reduces a driver’s effective field of vision. As vehi- cle speed increases, so does the amount of visual infor- Source: Interdisciplinary Working Group for Accident Mechanics (1986); Walz et al. (1983); Swedish Ministry of Transport (2002). FIGURE 2 Probability of fatal injury for a pedestrian colliding with a vehicle (OECD/ ECMT 2006; Speed Management: http://www.internationaltransportforum.org/Pub/ pdf/06Speed.pdf).

11 caused by the road work extend upstream of the signs. The accuracy of the information reported by law enforcement on the “check the box” sections of crash report forms has also been questioned; for example, comparison of detailed Illinois crash narratives with statistical crash abstracts found that 65% of work zone crashes were miscoded (Raub et al. 2001). Minor crashes were less likely to be correctly attributed to the work zone than severe ones. Crash location was a factor: in general, crashes occurring in the work activity area were cor- rectly identified as construction zone crashes, but those occur- ring in the approach, transition, or exit were less likely to be properly coded. Based on the narratives, approximately 40% of all crashes occurred in the approach or taper, which is also where speed-related crashes were most prevalent. Taken as a whole, these results suggest that property-damage-only and minor-injury crashes are probably underrepresented in statis- tical summaries for most U.S. states, making the total number of work zone crashes appear to be lower than what is reported and the severity of a “typical” work zone crash appear to be worse than it is. In 1998, ARROWS, a European work zone safety study completed an international review of accident studies (Dimitropoulos et al. 1998). The study concluded that work zones typically have higher crash rates than equivalent sec- tions without roadwork. They went on to note that “studies on road user behavior in work zones reveal that speeding, abrupt deceleration and inadequate distances from preceding vehicles occur frequently in road work zones. Such behav- ior is reasonably characterized as high-risk behavior and assumed to influence traffic safety negatively.” Citing Ger- man and British studies, the researchers noted that approxi- mately 60% of daytime work zone crashes were rear-end collisions, with the remainder comprised primarily of side- swipes (both collision types are likely to be exacerbated by unstable traffic flow conditions: abrupt speed changes can result in rear-end collisions and speed differentials between lanes can encourage abrupt lane changing maneuvers and the temptation to attempt to merge into small gaps). At night, col- lisions with fixed objects were of particular concern and were typically associated with inappropriate vehicle speeds. Crash rates were generally higher for short-duration work zones and those utilizing full (rather than partial) contraflow. The ARROWS report also found that: A cause of real concern regarding driver behavior at road work zones is the fact that drivers believe they take sufficient caution, choose the right speed and decelerate properly. Experimental studies have shown that the majority of drivers in fact approach road work zones driving too fast for the circumstances, and usu- ally well above the posted speed limit. Moreover, they do not decelerate until just before an abrupt change in the conditions (for example, a crossover point), and then in an extremely abrupt manner. A 2006 study at 23 locations in Kentucky supports the ARROWS conclusions: drivers do reduce their speeds in work zones, but not to the extent desired by transportation speed drops (Kemer 2009). A driver who misperceives the required deceleration has an increased risk of hitting the rear end of the vehicle ahead. • In work zones on multilane roadways, highly aggres- sive drivers may attempt to exceed the prevailing speed by making frequent, abrupt lane changes into the fast- est moving lanes. If an aggressive driver misjudges the headway or the traffic speed in the destination lane (or someone fails to yield to the aggressive driver owing to inattention or a blind spot), same-direction sideswipe may occur. There is a difference between objective safety (the actual number of crashes) and subjective safety (peoples’ perception of traffic crash risks). An unintended consequence of unstable traffic flow is that workers may perceive an improvement in their personal safety (owing to lower speeds in the adjacent lanes), while drivers are probably less safe than they would be under stable flow conditions. Anecdotal evidence from the work zone engineering community suggests that contractors occasionally attempt to destabilize the traffic flow (e.g., by unnecessarily narrowing the travel lanes) to achieve this per- ceived benefit. Such actions by contractors may be based on flawed logic: “the problem is that if motorist safety is reduced in work zones, worker safety is also reduced, because the traf- fic crashes that occur often spill over into the work areas and put workers at risk” (Harmelink and Edwards 2005). DRIVER SPEEDING AND SAFETY IN WORK ZONES Work zone crashes are a significant problem in the United States and worldwide, and speed is often cited as a contributing fac- tor. In 2012, the Fatality Analysis Reporting System (FARS), maintained by the NHTSA, recorded 30,800 fatal motor vehi- cle crashes in the United States, of which 547 (1.7%) were reported to have occurred in work zones. Among the work zone fatalities, speeding was indicated as a contributing factor in 192 (35.1%). While FARS tracks only fatal crashes, fatalities represent only a small proportion of all work zone crashes. For exam- ple, in Wisconsin 1,675 work zone crashes were reported in 2012, of which only 6 (0.35%) resulted in a fatality (WisDOT 2014a). FHWA has reported that, for work zone crashes that occurred in the United States in 2010, 0.6% were fatal crashes, 30% were injury crashes, and 69% were property damage only crashes (FHWA 2012). It is important that data from state and national databases be interpreted carefully, because reporting errors potentially bias both the number and severity of reported work zone crashes. Both academic research and investigative journal- ism (McIntire and Orr 2009) suggest that work zone crashes are underreported, in part because many police agencies do not consider crashes that occur upstream of the road work ahead signage to be work zone crashes, even when queues

12 limits of the temporary traffic control zone [as defined by the U.S. Manual on Uniform Traffic Control Devices (MUTCD)]. It concluded that during that during that 8-year period, 962 work- ers were killed at road construction sites, representing approx- imately 2.2% of all fatal occupational injuries in the United States. Not all of these “occupational” deaths were highway construction workers, 13% were truck drivers who were just passing through the site. Construction work has many hazards; a significant portion of the deaths occurred when workers were hit by construc- tion equipment, struck by materials that were being moved, or became involved in incidents directly related to the construc- tion such as trench collapses, falls, contact with live electri- cal wires, or similar hazards. Nevertheless, the BLS analysis shows that over the 8-year period 153 workers were hit at least once by a car, van, tractor-trailer, bus, or motorcycle. In other words, from 2003–2010, an average 19 highway workers per year were killed each year by traffic in work zones in the United States. Workers were flagging or performing other traffic control duties in 92 cases. Of these, 20 workers were reported as wear- ing reflective or brightly colored clothing (such as vests) to increase visibility. Only 32 of the workers were employed as flaggers; the remaining 60 worked in other occupations such as laborers, maintenance workers, and operating engineers. BLS noted the following other transportation-related deaths incurred by road construction workers: • Five workers were killed when a bucket truck they were in was struck by another vehicle. In each case, the worker fell from the bucket truck. • Five workers were killed when they fell from a truck while setting up or removing traffic control devices such as signs and cones. • Three workers were killed when the mobile equipment being used by the worker was struck by a train. Case Example 1: Agency and Legislative Response to Highway Worker Fatality in Saskatchewan Saskatchewan is a predominantly rural Canadian province that bor- ders the U.S. states of North Dakota and Montana. On Friday, August 23, 2012, Ashley Richards was working as a flag person on a road con- agencies. As indicated in Table 1, at the control sites without work zones the 85th percentile speed was 6.6 mph above the 65 mph posted limits. Although typical work zone sign- age resulted in an 8.8 mph speed reduction, the 85th per- centile speed remained 7.8 mph above the posted 55 mph work zone speed limit. In the Kentucky study, full compliance occurred only when police were present (Pigman et al. 2006). This finding has implications for many aspects of work zone speed management; for example, work zone public information campaigns often ask drivers to “reduce speed in work zones”; however, it is quite likely that a large majority of drivers believe they already comply with this instruction and, as a result, the campaign may not provoke the intended behavioral change. Additional numerical examples of poten- tial voluntary speed reductions can be found in Table 13 in chapter eight. WORKER SAFETY Collisions with road workers were noted as being “of special importance” by the ARROWS report. Vehicles that strike road workers have a high public profile and are often mentioned in the public outreach materials published by state DOTs. Records of on-the-job fatalities at the California DOT (Caltrans) show that “errant drivers” caused 49 of the 91 Caltrans employee deaths (54%) that occurred from 1971–2013 (contractor employees are not included) (Caltrans). Highway worker casualties sometimes serve as a call to action for improving work zone safety. As discussed in more detail in Case Example 1, Saskatchewan’s Ministry of High- ways & Infrastructure came under pressure to improve work zone safety after a speeding driver killed a young highway worker in the summer of 2012. The province is now imple- menting many of the recommendations developed in response to the crash, including “simplified” work zone signage to clar- ify when workers are present, installation of temporary rum- ble strips and gateway treatments at work zone approaches, increased fines for work zone speeding, increased police enforcement, and automated speed enforcement. In 2013, the U.S. Bureau of Labor Statistics (BLS) pub- lished an analysis of fatal occupational injuries at road con- struction sites, based on 2003–2010 data (Pegula 2013). The study focused on fatalities that occurred within the formal Situation Speed Limit (mph) Observed Speed (MPH) 50th percentile 85th percentile Not in Work Zone 65 67.8 71.6 Work Zone: No Activity 55 62.7 67.7 Work Zone: Active, Typical Signs 55 57.5 62.8 Work Zone: Active, Double Fine Signs Only 55 57.8 62.2 Work Zone: Active, Double Fine Signs, Police 55 53.8 57.3 Work Zone: Active, Double Fine Signs, Radar Box, Police 55 54.8 56.2 Source: Pigman et al. (2006). TABLE 1 OBSERVED TRAFFIC SPEEDS ON KENTUCKY FREEWAYS WITH AND WITHOUT WORK ZONES

13 stepped up immediately. In September 2012, the RCMP launched a provincewide safety blitz to catch speeders in construction zones, which involved officers dressed as construction workers observing drivers in the work zone and communicating with downstream officers to inter- cept speeders. In September and October RCMP officers issued more than 400 work zone speeding tickets according to media reports. With an existing statutory work zone speed limit of 60 km/h (approximately 35 mph) for all sites where workers are present (includ- ing freeways and mobile operations) reducing speed limits was not an option; however, by late October the Saskatchewan government announced plans for three statutory changes: 1. Work zone speeding fines were increased to triple the ordinary fine. A work zone violation at 70 km/h (approximately 45 mph) now results in a fine of CAD $300 (about U.S. $280). Speeding at 120 km/h (approximately 75 mph) results in a fine of CAD $798 (about U.S. $740). 2. Highway Transport Patrol officers who previously enforced only truck safety and weight regulations were authorized to issue work zone speeding tickets. 3. A 5-year pilot program for automated speed enforcement was put in place. The system began operation in July 2013 and makes it possible to issue speeding citations by mail, without having to intercept speeders in the work zone. The province’s Ministry of Highways and Infrastructure also responded with several engineering measures in preparation for the 2013 construction season: • Signage was modified to indicate more clearly when workers are present (which activates the statutory triple-fines provision). A new sign was also added to make it easier for drivers to see where the construction area ends. (In a November 2013 online sur- vey of 804 drivers conducted for the Ministry of Highways & Infrastructure by a private firm, 86% of respondents agreed that “there has been an improvement in clarity for work zone signing on provincial highways.”) • Certain signs at approaches to long-term work zones were doubled-up to improve their visibility. Specifically, on all four- lane roadways (divided and undivided) the roadwork ahead, speed limit, and flagger ahead signs are now placed on both the left and right sides of the roadway approaching the work zone. • Temporary rumble strips were specified contractually for the approaches to most projects lasting 5 days or more (79% of respondents to the November 2013 survey agreed that “rumble strips alerted me to important information.”) • Recognizing that Saskatchewan is a prairie province where wide- open spaces encourage fast driving, a gateway treatment based on a Manitoba design was implemented at work zone approaches on rural freeways, as shown in Figure 3 The barricade-like design is used on higher-volume highways for projects lasting 5 days or struction crew on Highway 39 about 5 miles north of Midale, a small town about an hour north of the U.S. border. According to media reports, it was Richards’ first full day on the job; the day before she had taken a flagperson training course and also had some supervised training on the jobsite. At 5:30 p.m., about 45 minutes into her shift, the 18-year-old was struck from behind and killed by an SUV. The story received extensive media attention and sparked public outcry: Richards and her fiancé Ben Diprose had recently moved to Saskatchewan to get a fresh start; Diprose was working on the site as an asphalt truck driver and witnessed the crash, and Richards was preg- nant with the couple’s child. In a radio interview Diprose said, “She was bleeding to death in my arms and there was nothing I could do.” The SUV driver, 44-year-old Keith Dunford of Regina, was arrested at the scene. In October 2012, following an investigation by the RCMP (Royal Canadian Mounted Police, Canada’s national police force), Dunford was charged with two offences (criminal negligence caus- ing death and dangerous operation of a motor vehicle causing death); however, a series of legal actions resulted in the postponement of a trial until August 2015. Although details of the crash were not released by the police prior to the trial, in a radio interview Diprose said that Dunford hit Richards while attempting to pass a line of vehicles that were stopped by the flag- ging operation. According to Diprose, Dunford told him he did not see Richards because he was looking for a paper he dropped. Although Dunford did not testify, a police interview recorded 2 hours after the crash was played at the trial. “I wasn’t paying attention, I must admit,” Dunford said on the tape. “I was looking at my paperwork.” Witnesses also testified that prior to the crash Dunford passed two semi-trucks in the work zone, despite the presence of two no-passing signs. Prosecu- tors noted that Dunford had driven through the work zone previously and a police officer testified that there was no indication Dunford was intoxicated or high. Cpl. Jeff Burnett, a collision analyst, testified that Richards’ body was found 54 m (177 ft) from the estimated point of impact, and that Dunford was going 82 to 99 km/h (51 to 62 mph). The work zone speed limit was 60 km/h (35 mph). Although the legal process took more than 3 years, Saskatchewan’s political and administrative leadership responded to the incident quickly. On the Wednesday following the crash, Saskatchewan Pre- mier Brad Wall (equivalent to a Governor in the United States) posted two Twitter messages that were soon relayed by other media outlets: 12:29 p.m.: “Angry to learn from owner of constr co. where Ashley worked that drivers are still not obeying orange zone laws. SLOW DOWN. NO PASSING.” 12:32 p.m.: “Have asked Hwys and Justice Ministers to work with police, stakeholders to canvass any and all ideas to improve orange zone safety.” In coordination with the RCMP, the Saskatchewan Ministry of Justice announced that patrols and ticketing in work zones would be FIGURE 3 Gateway treatment for entrance to Saskatchewan rural highway work zone (Saskatchewan MHI 2013).

14 time a driver exceeds the speed limit without apparent consequences. • Although speed is a factor in a very high percentage of serious and fatal crashes, many drivers underestimate these risks. As a result, drivers tend to think more about the risk of being penalized for speeding than about the risk of being involved in a speed-related crash. • Most drivers consider themselves to be above average in terms of skill. A number of surveys conducted in various countries around the world demonstrate that up to 90% of drivers believe they are an above-average, low-risk driver. • Many drivers regard speed limits as arbitrary and do not fully understand the greater risks associated with even small increases in speed. • In some cases, commercial drivers feel pressure to drive faster to increase their income or meet company pro- ductivity goals. Some large U.S. trucking companies are aware of this issue and equip their vehicles with speed limiters; however, this is often not the case for other fleets such as taxis, shuttle vans, small trucking compa- nies, or independent truck drivers. Toward Zero Deaths: A National Strategy on Highway Safety (TZD Steering Committee 2014) emphasizes the impor- tance of integrating knowledge gained through the social sciences with the more traditional approaches to highway safety. As illustrated in Figure 4, the report developed a new Traffic Safety Culture Model (TSC) that augments the long- standing focus on engineering, enforcement, and education. The report states: The TSC model focuses on how social factors in a culture influence how people prioritize traffic safety and accept traffic safety strategies. That is, the TSC model assumes that behav- iors related to traffic safety performance are . . . influenced by our culture. Therefore, it is difficult to achieve sustainable improvements in traffic safety until we understand these pro- cesses and create a culture in which everyone values traffic safety and works to enhance it. By operating and integrating TSC programs across multiple levels [we can] achieve effec- tive and sustainable improvements in road user behavior. One example of the cultural dimension of speeding is raised by the Global Road Safety Partnership report (Howard et al. 2008): Very small increments of speed in excess of speed limits are a major factor in increasing crash risk on the network, especially if it is a behavior that is widely practiced by the driving popula- tion. Over time, low-level speeding can become the accepted behavior of drivers and they will expect to drive at a higher level until or unless they encounter some enforcement. If low level speeding is widespread and is more than 2 or 3 km/h (1 to 2 mph) above a posted speed limit, there may be the need to apply tougher standards to speed enforcement than those that currently exist. For example, some jurisdictions allow drivers to travel up to 15 km/h (10 mph) over the limit before being given an infringement notice. This results in the de facto speed limit becoming 15 km/h over the posted limit. The increase in crash risk as a consequence can be large. Responses to the Engineering and Enforcement survey con- ducted for this report indicate that speeding tolerance in the more. It consists of three horizontal boards, positioned starting at the break point of the shoulder and extending 12 ft down the sideslope on each side of the roadway. The spacing of the boards converges to create an exaggerated sense of perspective and heighten the sense that the roadway is narrowing. Each bar has orange and black stripes facing traffic approaching the work zone and reflective white and black stripes facing traffic leaving the work zone. • Contractual provisions were strengthened to ensure that con- tractors promptly remove work zone speed limit signs when they are not needed. According to Marla Muhr of the Ministry of Highways & Infrastructure, “Often contractor compliance with taking down signs when workers were not present was not good, so the public does not feel the signs are meaningful; the public is skeptical and may not slow down until they are immediately in the vicinity of the workers.” Saskatchewan’s 2014 work zone safety public information cam- paign featured a television advertisement that appears to be loosely based on the Ashley Richards crash. In the video, a woman, late for work, is seen kissing her adolescent son at the breakfast table. In the next scene she is approaching a work zone at 110 km/h, impatiently tapping her fingers on the steering wheel. She sees the 60 km/h speed limit sign, but only slows to 80. When she looks at the flagger, it is her son. She cringes and brakes abruptly, then realizes that the flagger is actually an adult man. He waves at her, and she waves back sheepishly. The ad closes with the tag line, “Imagine how fast you would drive in a work zone if someone you loved was there.” The full results of Saskatchewan’s experience with these work zone speed management techniques were not available prior to the prepara- tion of this synthesis report, and it may ultimately be difficult to separate the effects of the individual components of this combination strategy. Saskatchewan’s situation has inherent challenges: with a small number of law enforcement personnel spread across more than 16,000 miles of provincial highways it is not possible to have police present at all work zones. Although automated speed enforcement augments the traditional patrols, the province’s three automated speed enforcement devices are shared among numerous construction projects. The statutory speed limit is 60 km/h (approximately 35 mph), but achieving full compli- ance is difficult; according to Muhr, observed work zone speeds are closer to 80 km/h (50 mph) for most projects and around 70 km/h (45 mph) for bridge projects. Nevertheless, the public appears to be accepting the changes implemented in response to the death of Ashley Richards and press coverage continues to be pro-safety. After attending a speech where officials announced the implementation of automated speed enforcement, Diprose was quoted as saying, “It means quite a bit because I wouldn’t want anybody else to go through the same thing I went through. I wouldn’t wish this on my worst enemy.” References: (CBC News 2012a-e; CTV 2121; Saskatchewan MOJ 2012; SGI 2010 CBC News 2013; Insightrix 2013; Saskatchewan MHI 2013, 2014a,b; Global News 2014; Government of Saskatchewan 2014; M. Muhr, Saskatchewan Ministry of Highways & Infrastruc- ture 2014; Wilson 2014; Discover Weyburn.com 2015; CBC News 2015a, b). SOCIAL, PSYCHOLOGICAL, AND CULTURAL FACTORS AFFECTING SPEEDING The Global Road Safety Partnership developed a comprehen- sive speed management handbook on behalf of WHO (Howard et al. 2008). The document discusses several social and psychological factors that contribute to speeding: • Travelling at higher speeds offers the immediate “reward” of a faster trip. This benefit is reinforced each

15 major wrong by exceeding the work zone speed limit (Bolling and Nilsson 2001). In an interview conducted for this synthesis report, a Saskatchewan official identified an interaction between con- tractor and driver attitudes that has had the effect of promot- ing work zone speeding in the past. The province’s statutes establish a mandatory 60 km/h (approximately 35 mph) work zone speed limit when workers are present. Marla Muhr of the Saskatchewan Ministry of Highways & Infrastructure expressed concern that the credibility of this limit was under- mined by contractors’ ongoing failure to remove or cover the signs when the workforce leaves the site. Citing the need for contractor cooperation and compliance Muhr said, “For too many years, the [work zone speed limit] signs did not mean anything” (M. Muhr, Saskatchewan Ministry of Highway & Infrastructure, personal communication, 2014). The TSC model suggests that anti-safety behaviors such as those cited by Rush and Muhr are linked to personal and organizational values, beliefs, frames-of-mind, norms, and attitudes. The TZD report notes that the term safety culture has “a long history in organizations in which safe operations are critical, such as in nuclear power plants, commercial aviation, and healthcare.” It asserts that long-term internal and external improvement in safety outcomes is achievable through a sys- tematic approach that emphasizes shared values, “cohesion, trust, and willingness to engage with the community” (TZD Steering Committee 2014). Social science research suggests that punishments and rewards have equal value in promoting socially coopera- tive behavior, as long as they are administered equitably and are seen as promoting collective interests (rather than the self-interests of the administrator) (Balliet et al. 2011). The Toward Zero Deaths report suggests that, where feasible, positive approaches to safety improvement are more likely to be accepted than negative or punitive approaches. Examples cited in the report include rewarding good choices and using humor to demonstrate the value of safety. “Rather than hav- ing agencies dictate appropriate behaviors, the intent is to United States varies from state to state, ranging from less than 5 mph to more than 10 mph above the posted speed limit. Con- sequently, a posted work zone speed limit of 55 mph is actually enforced as 60 to 65 mph (or more) depending on the locality. Another indicator of the variability of public attitudes toward speeding is the state-to-state differences in the statu- tory fines leveled for speed infractions. According to a NHTSA summary (NHTSA 2011), as of February 2010 the maximum fine for speeding ranged from $50 in Tennessee to $2,500 in Virginia. The maximum jail time imposed by the states for speed law violations varies dramatically, from 15 days to one year. Some states impose additional penalties if the incident results in a worker casualty. Comparatively, as of February 2013, France levied uniform national fines that range from €68 to €3750 (approximately U.S. $120 to $5,250) depending on the severity of the infraction; in addition to these fines, drivers who exceed the speed limit in France by more than 50 km/h (30 mph) are subject to a prison sentence of up to 3 months, a 3-year driving license suspension, and confiscation of their vehicle. Frustration with entrenched behaviors that negatively impact work zone safety is evident in many of the information sources identified for this synthesis report. For example, the authors of a Canadian work zone speed management report wrote that, “Drivers are unwilling to slow down and seem to resent construction and maintenance delays” (Harmelink and Edwards 2005). Similarly, in an interview conducted for this project, Virginia DOT’s (VDOT’s) David Rush noted that, “[Often] work zone speed limit compliance only occurs when police are present” (D. Rush, VDOT, personal commu- nication, 2014). Research supports this point of view: A 1990 survey of drivers at a rural work zone on I-57 in central Illi- nois found that although 79% said the posted 45 mph speed limit was about right, only 59% reported that they drove at or below this limit. More than one-third of the drivers admitted to speeding through the work zone (Benekohal et al. 1990). A 2001 Swedish survey of drivers who were cited for work zone speed violations found that very few reported that they had been in a rush; most believed they had not done anything FIGURE 4 Descriptive and predictive model of key concepts that define traffic safety culture (TSC) and their relationship with behavior and crash risk (TZD Steering Committee 2014).

16 if there are merges within the Activity Area. Multiple- vehicle involvement can be high because of a lack of escape routes (especially if lane width is constrained by temporary concrete barriers). MEASUREMENT OF WORK ZONE SPEEDS AND SPEED REDUCTIONS It is important that the numerical values of work zone speed reductions reported for techniques discussed in this synthe- sis report be interpreted with some caution. Many variables affect work zone speeds, even when special speed manage- ment techniques are not in use. Variations in free-flow speed, time of day, traffic volume, and truck percentage can com- plicate work zone speed studies (Chen et al. 2007). Traffic volumes influence the extent to which the speeds of indi- vidual vehicles are independent or result from car-following phenomena (Porter and Mason 2008). Characteristics of the work zone itself [such as the number of available lanes, sur- face condition, vertical and horizontal geometry, and type of delineation (e.g., concrete barrier vs. drums)] also influence work zone speeds and can change rapidly as construction progresses. As a result, researchers face inherent challenges in ensuring that observed work zone speed changes are attrib- utable to the devices and techniques under study, and not to external factors. One consideration when interpreting reported speed reduc- tions is that the magnitude of the observed speed reduction is likely to depend on the severity of the speeding upstream. For example, an Illinois study comparing automated speed enforcement at two sites found greater reductions in the Chicago area than in the St. Louis area, primarily because speeding was more prevalent at the Chicago site (Benekohal et al. 2010). Work zone speed reductions have been measured in a number of ways by various researchers depending on the characteristics of the work zone, the type of speed manage- ment technique that is being studied, and the available obser- vational equipment and personnel. Many studies focus on the average speed of the vehicles in the traffic stream, while others emphasize 85th percentile speeds or speed variation (often reported as the standard deviation of the speeds of individual vehicles). Relationships between these measures are not always well-established or consistent, and may dif- fer for cars and heavy trucks (Porter and Mason 2008). Some studies compute speed reductions using simple meth- ods, whereas others use sophisticated statistical analysis. These methodological differences can make it difficult to make direct comparisons of the effectiveness of speed man- agement techniques. Examples of some methods that have been used to evaluate work zone speeding countermeasures include: • The difference between the speed upstream of the work zone and the speed within the work zone where the countermeasure was deployed. create the motivation within the driving population to partner with highway safety agencies to achieve mutual goals” (TZD Steering Committee 2014). An unconventional example of a positive approach toward preventing speeding comes from an automated enforcement pilot project in Sweden, which rewarded drivers who did not speed through a business dis- trict with a portion of the fines collected from the violators (Volkswagen 2010). During the brief experiment, average speeds went down 22% according to a video produced by the project sponsor. The video quotes a driver as saying, “This is a really positive thing. Drive legally and earn money. Perfect.” ELEMENTS OF A WORK ZONE The MUTCD (FHWA 2009) divides work zones into four areas, as shown in Figure 5: • Advance Warning Area (including shoulder taper) where traffic is told what to expect ahead. • Transition Area (including upstream taper) where traf- fic moves out of its normal path. • Activity Area where work takes place. • Termination Area (including downstream taper) where traffic resumes normal operations. Slightly different terminology is used in other countries to describe these four areas. There is a general consensus that improving compliance with speed limits and reducing inappropriate driving speeds are not easy tasks, that speed management remains one of the biggest challenges facing safety practitioners, and that a concerted, long-term, multidisciplinary approach is neces- sary (Peden et al. 2004). These challenges are compounded by the differing conditions in various parts of a work zone: • In Advance Warning and Transition areas the goal is generally to encourage drivers to slow to the work zone speed limit or make a gradual reduction to the speed associated with a lateral shift, lane drop, or queue. One study found that approximately 40% of crashes occur in advance warning and transition areas, primarily as a result of drivers approaching lane drops and queued traffic at high speeds (Raub et al. 2001). Based on video observations, the same study found that approximately 5% of drivers approached at speeds that were high rela- tive to the queue. • After entering the Activity Area drivers often attempt to increase their speeds; therefore, a separate set of tech- niques may be necessary to achieve sustained speed reduction throughout the entire work zone (especially for longer work zones). Raub et al. (2001) found that approximately 60% of crashes occur in Activity and Exit areas, caused mainly by sudden driver maneuvers, inappropriate following distance, and driver distrac- tion. Although speed-related crashes are less common in these areas compared with the upstream portions of the work zone, sudden slowing can occur, particularly

17 The location where speeds are measured is another consid- eration: some speed management measures have only a local- ized effect, while others are effective over a longer distance. When comparing the results of work zone speed studies it is also necessary to consider the observational methods and equipment used to compute the speeds. Measurement accuracy • The difference between the speed within the work zone with and without the countermeasure. • Spot speed observations at several locations in and adja- cent to the work zone. • Speeds a predetermined distance upstream of a flagger station during two-way one-lane operations. FIGURE 5 Elements of a work zone as defined in the U.S. Manual on Uniform Traffic Control Devices (MUTCD) (FHWA 2009).

18 of the speed and axle spacing calculations (owing to the decrease in the significance of any error in tube length or tube spacing)” (Mendigorin et al. 2003a, b). A spacing of 1 m (39 in.) had historically been used in the study area. Quadrupling the spacing required changes in the processing software. The study also found poor speed accuracy in congested traffic condi- tions (a frequent occurrence in some work zones). The study noted that in addition to previously documented issues with high-speed vehicles causing “reflections” (false pulses potentially interpreted as axle hits), “quite surprisingly . . . vehicles at very low speed were gen- erating reflections.” 3. Side-fired radar. Permanent or semi-permanent radar units mounted at the roadside have been used for a num- ber of work zone speed studies. The term “side-fire” indicates that the units are positioned with the radar beam perpendicular to the traffic stream, an unfavor- able vantage point for gathering speed data. Table 2 shows the manufacturer-claimed speed accuracy for three common side-fire radar units. 4. Bluetooth Vehicle Re-Identification. Bluetooth is a short-range wireless telecommunications technol- ogy that is widely used in mobile phones, headsets, electronic games, and other consumer devices. Each Bluetooth device is capable of emitting a unique serial number called the media access control (MAC) address. Traffic speed data can be collected using detectors that scan for MAC addresses; these detectors are typically mounted at the roadside or in the median. When a Blue- tooth device is observed, the detector records the MAC address and a timestamp. Addresses presumed to be associated with vehicles are then matched (re-identified) with data from other locations. If the distance between the Bluetooth detectors is known, travel time can be imputed from the time difference between the observa- tions. In principle, similar information can be gleaned from other types of wireless devices such as WiFi; how- ever, as of 2014 most commercially available detection products are based on Bluetooth (Chitturi et al. 2014). When appropriate filtering settings are used Bluetooth “is capable of providing reliable and high-quality ground truth travel time data on highways” (Haghani et al. 2010). is particularly important when the speed management tech- nique under study is expected to yield fairly small speed reduc- tions (i.e., 2 to 3 mph). Four of the most frequently used methods for work zone speed measurement are: 1. Point speed observations. Typically, in this method an observer is stationed near the roadway with a police-type portable Doppler radar or Lidar device. The observer points the device at individual vehicles (parallel or at a slight angle to oncoming traffic) and records their speeds. This method is considered highly accurate, but has some limitations: it is labor-intensive, so the num- ber of samples that can be obtained is usually limited, and it is not always feasible to have observers on-site during the full range of work zone traffic conditions. In high-volume, multilane situations it may be necessary to position observers on an overpass to obtain data from all lanes. 2. Pneumatic counters. Single-hose pneumatic traffic counters were developed for traffic volume studies nearly a century ago (Hogentogler 1923). Some mod- ern pneumatic counters are also capable of gathering traffic speed data, which requires installing two air hoses (or “tubes”) across the roadway. Counters that support this function typically incorporate a digital controller that records time-stamped observations for each axle hit; if the hose spacing is known, the speed of individual vehicles can be imputed from the time differential between hits on the two hoses. The accuracy of the speed computation from pneu- matic counters is dependent on correct installation of the hoses and proper selection of detector settings. Systematic errors can occur if the hoses are not pre- cisely parallel to each other or are not perpendicular to the vehicle path. Excessive hose slack, bends, and kinks can also affect results. It is important that the hose spacing used in the speed computation soft- ware match the actual field spacing very closely; for example, if the expected hose spacing is 36 in. and the actual spacing is 37.5 in., a speed of 48 mph would be attributed to a vehicle whose actual speed is 50 mph (a –4% error). An Australian study found that increasing the hose spacing to 4 m (13.1 ft) “increased accuracy Product Per-Direction AverageSpeed Accuracy Per-Lane Average Speed Accuracy Per Vehicle Average Speed Accuracy Autoscope RTMS G4 Not stated Not stated ±10% Wavetronix SmartSensor HD ±3 mph ±3 mph ±5 mph for 90% of measurements Wavetronix SmartSensor V ±5 mph ±10 mph Not stated Sources: Image Sensing Systems (2012); Wavetronix LLC (2014). TABLE 2 SPEED MEASUREMENT ACCURACY OF SIDE-FIRE RADAR UNITS AS STATED BY THE MANUFACTURERS

19 Finally, it is necessary to note that some of the speed man- agement techniques included in this synthesis report have been tested only on a limited basis (e.g., small sample size, small number of sites, limited diversity of sites), resulting in some uncertainty. Site-specific field conditions can also affect the observed speed reduction; for example, a Kansas test of optical speed bars was hampered by lack of contrast between the bar markings and the pavement (Meyer 2004). Similarly, geometric design details (such as the lateral shift rate of chicanes) are likely to affect the observed speed reduction. In contrast to the point-speed data gathered by pneu- matic and radar-based speed data collection devices, each individual Bluetooth observation represents a vehicle’s average speed between a pair of detectors. Therefore, if one detector is installed at the approach to a work zone and another is placed at the work zone’s termination, data indicative of the overall speeds in the work zone can be gathered. Consequently, Bluetooth is well-suited to evaluation of techniques intended to influence speed throughout the work zone.