National Academies Press: OpenBook

Safe and Aesthetic Design of Urban Roadside Treatments (2008)

Chapter: Chapter 2 - State-of-the-Art Summary

« Previous: Chapter 1 - Background
Page 5
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 5
Page 6
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 6
Page 7
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 7
Page 8
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 8
Page 9
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 9
Page 10
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 10
Page 11
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 11
Page 12
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 12
Page 13
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 13
Page 14
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 14
Page 15
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 15
Page 16
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 16
Page 17
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 17
Page 18
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 18
Page 19
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 19
Page 20
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 20
Page 21
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 21
Page 22
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 22
Page 23
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 23
Page 24
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 24
Page 25
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 25
Page 26
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 26
Page 27
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 27
Page 28
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 28
Page 29
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 29
Page 30
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 30
Page 31
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 31
Page 32
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 32
Page 33
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 33
Page 34
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 34
Page 35
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 35
Page 36
Suggested Citation:"Chapter 2 - State-of-the-Art Summary." National Academies of Sciences, Engineering, and Medicine. 2008. Safe and Aesthetic Design of Urban Roadside Treatments. Washington, DC: The National Academies Press. doi: 10.17226/14171.
×
Page 36

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.

5Urban areas present unique challenges to the roadway de- signer. Urban and regional stakeholders need a transportation network that allows them to accomplish their travel objec- tives with a minimum amount of travel delay and to have these travel demands met on a road network that is both operationally efficient and safe. While the transportation profession has made dramatic advancements toward meeting safety and mobility mandates, it is critical that the function of the street system complement the adjacent land use and balance the needs of all users while maintaining the safest possible transportation facility. Of par- ticular interest is the design of roadsides—the area between the shoulder (or curb) and the edge of the right-of-way (1). The roadside is a common location for pedestrian activity, utility placement, landscaping, transit stops, driveway place- ment, mailbox placement, and placement of a variety of other roadside features typical of the urban environment. Urban roadside environments can range from dense downtown zones with on-street parking to high-speed zones with motor vehicle operational priorities. Given the importance of the roadside environment to the quality of urban life, it is unsurprising that urban residents and stakeholders often seek to have the roadside designed in a manner that enhances the quality of the urban environment. Commonly requested functional roadside elements include sidewalks, street trees, and street amenities such as seating. Requested aesthetic elements include public art and special paving materials. Placing these roadside elements in a way that enhances urban roadside safety is the focus of this liter- ature review. Overview of Roadside Crash Statistics In 2005, over 6.2 million crashes occurred on U.S. roadways. Almost 1.9 million of these crashes involved an injury, and 39,189 people were fatally injured (2). Of particular concern are crashes that involve a vehicle that leaves the roadway. Of the 6.2 million crashes in 2005, run-off-road crashes accounted for 0.95 million, or about 15 percent of the total. While run-off-road crashes happen less frequently than other types of crashes, they are often severe. Although run-off-road crashes accounted for only 15 percent of all crashes in 2005, they accounted for 32.2 percent of the total fatal crashes in that year (see Table 1). Examining fatal crashes by the first harmful event illustrates the magnitude of specific roadside hazards. Of the fatal crashes occurring in 2005, 39 percent involved collisions between motor vehicles. Rollover crashes and collisions with fixed objects—two kinds of crashes that are associated with the roadside environment—made up 11 percent and 32 percent, respectively, of fatal crashes in 2005. The highest percentage of fixed-object crashes was in the category of tree/shrub, with tree or shrub impacts accounting for slightly more than 3,200 crashes, or roughly 8 percent of all fatal crashes. Poles and posts accounted for a little less than 5 percent of all fatal crashes (see Table 2). Roadside Safety: Current Practices The literature on roadside safety establishes three roadside crash strategies that can be considered when seeking to im- prove run-off-road crash statistics (1). First, the ideal scenario is to prevent vehicles from leaving the travelway, thereby eliminating the crash entirely. Preventing the conditions that lead to run-off-road crashes (conditions such as driving while impaired or fatigued) and alerting a driver that he or she is leaving the travelway would be potential countermeasures included in this category. The second strategy, which is based on the idea that run- off-road events are impossible to prevent entirely, is to design a roadside that is “forgiving.” In other words, a roadside should be designed to minimize the consequences of a run- off-road event. Under current practice, the ideal roadside C H A P T E R 2 State-of-the-Art Summary

allows errant vehicles to come to a controlled stop before encountering an object located along the roadside by includ- ing a clear zone adjacent to the travelway. In many situations, however, such as urban roadways located in narrow rights- of-way, a clear zone may be impractical. Thus, in many cases, in which the provision of a clear zone and/or wider right-of- way may be desirable from a safety perspective, achieving this clear zone may be infeasible. Under these circumstances, design agencies should strive to minimize the severity of an impact with a fixed object should such a crash occur. This literature review summarizes known roadside design safety guidance for roadways in urban areas. There is often little substantive knowledge on the safety impacts of various design treatments, leaving the definition of what constitutes a “safe” facility open to question. The ability to clearly and reasonably evaluate and demonstrate the safety impacts will go a long way toward resolving many of the contentious issues that relate to the design of urban roadways and toward satis- fying the needs and interests of project stakeholders. This re- view, therefore, summarizes focused research on the safety of roadside treatments in urban areas with particular attention to the high-speed (thus more severe) crash locations such as suburban-to-urban arterial transitions where land use is less dense, on-street parking is rarely permitted, and the presence of driveways/intersections is considerably less frequent than it is in more congested urban business corridors. 6 Crash Type On Roadway Run-Off- Road Shoulder Median Other/ Unknown % Run- Off-Road Total Fatal Crashes Single Vehicle 6,507 12,340 2,431 1,022 353 54.5 22,653 Multiple Vehicle 15,647 297 302 198 92 1.8 16,536 Total 22,154 12,637 2,733 1,220 445 32.2 39,189 Injury Crashes Single Vehicle 154,000 320,000 14,000 48,000 28,000 56.7 564,000 Multiple Vehicle 1,235,000 7,000 1,000 7,000 2,000 0.6 1,252,000 Total 1,390,000 327,000 16,000 54,000 30,000 18.0 1,816,000 Property-Damage-Only Crashes Single Vehicle 328,000 598,000 31,000 81,000 277,000 45.5 1,314,000 Multiple Vehicle 2,957,000 11,000 3,000 14,000 5,000 0.4 2,990,000 Total 3,284,000 609,000 34,000 94,000 282,000 14.1 4,304,000 All Crashes Single Vehicle 488,000 930,000 48,000 129,000 306,000 48.9 1,900,653 Multiple Vehicle 4,208,000 18,000 5,000 21,000 7,000 0.4 4,258,536 Total 4,697,000 948,000 53,000 150,000 313,000 15.4 6,159,189 Source: Adapted from Traffic Safety Facts 2005 (2). Table 1. Crashes by number of vehicles and relation to roadway (2005). First Harmful Event Fatal Crashes % of All Fatalities Collision with Motor Vehicle in Transport 15,357 39.2 Pedestrian 4,520 11.5 Bicycle 776 2.0 Object Not Fixed Other Object Not Fixed 1,209 3.1 Overturn (Rollover) 4,266 10.9 Tree/Shrub 3,215 8.2 Pole or Post 1,852 4.7 Culvert/Ditch/Curb 2,591 6.6 Embankment 1,444 3.7 Guardrail 1,189 3.0 Bridge 336 0.9 Fixed Object Other Fixed Object 1,812 4.6 Other Unknown First Harmful Events 622 1.6 Total 39,189 100.0 Source: Adapted from Traffic Safety Facts 2005 (2). Table 2. Fatal crashes by most harmful event (2005).

In the first volume (and subsequent volumes) of NCHRP Report 500, prospective engineering countermeasures and their associated effectiveness are classified as “Tried,” “Exper- imental,” or “Proven” (3, pp. V-2 through V-3). This classi- fication permits readers to understand the level of testing performed on a specific countermeasure perceived to be effective for a safety improvement program. Summarized versions of the definitions given in the first volume of the NCHRP Report 500 series for “Tried,” “Experimental,” and “Proven” are given below (3, pp. V-2 through V-3): • Tried (T)—Strategies that have been implemented at a number of locations but for which valid safety evaluations have not been identified. As a result, these strategies should be used with caution until information about their effec- tiveness can be accumulated and they can be reclassified as Proven strategies. • Experimental (E)—Strategies that appear sufficiently promising so that application and testing appear feasible for a small-scale evaluation. These strategies do not have any valid safety evaluations or large-scale applications and warrant pilot studies to help elevate them to the category of Proven strategies. • Proven (P)—Strategies that have been used in more than one location and for which properly designed evaluations were conducted to show their level of effectiveness. A user can apply a proven strategy with some level of confidence, but is also aware of appropriate applications as a result of these previous studies. The roadside-object literature review included in this re- port focuses on proven safety strategies for urban roadside safety; however, Tried and Experimental strategies are also included in an effort to provide a comprehensive listing of known or perceived applications. In addition, this chapter re- views the following: • Roadside crash statistics, in an effort to identify the specific nature of roadside crashes in urban areas; • The various strategies currently in use in urban environ- ments to keep vehicles from leaving the travelway; and • General information, safety research, and proposed safety strategies for a variety of potential roadside objects com- mon to the urban environment. Although this review targets the design of roadsides in urban areas, much of the literature on roadside design has been based on studies of rural environments. As a result, the literature on rural roadside safety is included when it is applicable. Finally, this review focuses specifically on those roadways classified as urban arterials, collectors, and local streets because urban stakeholders are most vocal about wanting roadside treatments that balance the demands of agencies, stakeholders, and users on these roadways. While highways, freeways, and other high-speed, limited-access roadways may have impor- tant roadside safety issues, the design of such roadways is out- side the scope of this study. Examining Roadside Safety in Urban Environments The majority of travel undertaken in the United States occurs on urban roadways. Of the 2.9 trillion miles of travel in 2003, roughly 1.8 trillion—62 percent—occurred in urban areas (4). Urban roadways experience higher levels of traffic congestion, particularly during morning and evening peak periods, and are much more likely to incorporate multi- modal travel, including transit, bicycling, and walking. Trip characteristics differ as well. While rural roadways experience more freight and long-distance, inter-regional trips, most urban travel is characterized by intra-regional travel, partic- ularly household-related travel, such as work or shopping trips. Thus, it is not surprising that the nature of urban road- side crashes may also differ from the nature of rural roadside crashes. When one compares fatal crash frequency in rural areas with crashes in urban areas, it is clear that fatal rural roadway crashes occur more often than fatal urban roadway crashes. While on-roadway crashes are slightly more frequent in rural areas, off-roadway crashes, which include rollover as well as fixed-object crashes, are considerably more frequent in rural areas. In the categories of fixed-object and rollover crashes only, fatal crashes are still more frequent in rural environ- ments. Approximately 60 percent of fatal fixed-object crashes and 77 percent of fatal rollover crashes occurred in rural environments. Although focusing solely on fatal crashes risks underesti- mating the likelihood of a roadside crash for urban areas, fatality crash information is dependably reported and does provide some indication regarding crash trends. Table 3 shows fatal crash conditions for common roadway classes for urban and rural environments. Fatal fixed-object crashes for the roadway classes shown are more pronounced in rural areas. Pedestrian and bicycle fatalities, on the other hand, are much more of an urban problem. For the road classes under con- sideration, fatal pedestrian crashes are almost twice as likely to occur in urban environments as in rural ones. Pedestrian and bicycle activity is more common in urban environments, and the increased presence of pedestrians and bicyclists in- creases the possibility that such a crash will occur. In general, fatal crashes with most types of fixed objects occur more often in rural environments than in urban ones (see Table 4). When one considers specific roadway classes, however, several exceptions emerge, particularly on roadways 7

classified as minor arterials. As depicted in Table 4, principal arterials in urban areas have fatal crashes involving utility poles, light poles, and sign poles more often than their rural counterparts. Preventing Vehicles from Leaving the Travelway The logic behind keeping vehicles on the travelway is simple: if a vehicle does not leave the travelway, it cannot be involved in a roadside crash. The difficulty posed by strate- gies aimed at keeping vehicles on the roadway is that, unlike providing appropriate clear zones or effective impact attenu- ators, these strategies are oriented toward the driver rather than the vehicle. The literature on roadside safety shows that, of the strate- gies identified in this report, knowledge on the design factors that may help prevent vehicles from leaving the roadway is the least developed. Typically, research has focused on the geo- metric characteristics of locations where vehicles leave the travelway and has found that a disproportionate share of these crashes is associated with shifts in the horizontal curvature of the roadway, particularly an isolated, sharp, horizontal curve. Approximately 45 percent of all fixed-object crashes (5) and up to 77 percent of tree-related crashes (6, 7) are due to vehicles traveling off the roadway on the outside of a horizontal curve. Because of such findings, the principal design strategy for keeping vehicles on the roadway is to address horizontal shifts in the roadway. While eliminating the isolated sharp curve is perhaps the most effective treatment, such applications are often prohibitively costly for mitigating safety problems on existing roads. As an alternative, designers have adopted a secondary strategy, which is to delineate potentially hazardous environments, such as curves, using traffic control devices such as posted advisory speeds, chevrons, and other markings or strategies to more clearly indicate the edge of the travelway. Another secondary strategy that is commonly employed to alert drivers before their vehicle leaves the travelway is to place rumble strips on the shoulder of the roadway. Delineate Potentially Hazardous Roadside Environments A common practice in minimizing run-off-road crashes is to use signs to delineate potentially hazardous roadside con- ditions. Signs and other indicators are used to increase the driver’s awareness of changes to the operating characteristics of the roadway. Under conventional practice, the delineation of hazardous roadside conditions is limited almost exclusively to the use of signs to denote shifts in the horizontal curvature of the roadway or other oncoming hazards. Other features, such as the physical characteristics of the surrounding roadway, 8 Rural Urban Crash Condition Principal Arterial Minor Arterial Collector Local Total Rural Principal Arterial Minor Arterial Collector Local Total Urban Motor Vehicle Collision 2,256 1,912 2,211 891 7,270 2,262 1,442 447 877 5,028 Ped/Bike 269 221 377 322 1,189 1,353 865 257 748 3,223 Overturn (Rollover) 472 420 905 560 2,357 153 116 57 184 510 Fixed Object 821 1,088 2,740 1,811 6,460 805 851 442 1,058 3,156 Other Causes 119 154 271 270 814 145 148 66 266 625 Total 3,937 3,795 6,504 3,854 18,090 4,718 3,422 1,269 3,133 12,542 Source: Fatality Analysis Reporting System (118). Table 3. Fatal crash conditions on arterial, collector, and local roads (2005). Rural Urban Fixed Object Principal Arterial Minor Arterial Collector Local Total Rural Principal Arterial Minor Arterial Collector Local Total Urban Tree/Shrub 184 247 835 650 1,916 121 185 113 334 753 Utility Pole 39 73 181 138 431 103 116 59 119 397 Culvert/Ditch/Curb 143 237 583 401 1,364 240 239 111 242 832 Embankment 151 182 491 191 1,015 30 45 27 54 156 Guardrail 123 116 147 56 442 53 54 26 28 161 Building/Fence/Wall 30 50 160 110 350 40 46 26 82 194 Light/Sign Poles 86 84 139 95 404 108 81 36 72 297 Bridge 13 23 65 39 140 21 28 6 15 70 Other Fixed Object 52 76 139 131 398 89 57 38 112 296 Total 821 1,088 2,740 1,811 6,460 805 851 442 1,058 3,156 Source: Fatality Analysis Reporting System (118). Table 4. Fatal fixed-object crashes on arterial, collector, and local roads (2005).

also may give drivers cues as to safe operating behavior. Many governing jurisdictions consider the design of the street and the surrounding environment collectively, thus taking advan- tage of environmental characteristics to alert the driver of safe operating behavior. Identifying Hazardous Conditions Using Signage A common practice aimed at keeping vehicles on the road- way is to post advisory speeds or other signage applications to denote potentially hazardous conditions. While such a prac- tice makes sense, the inconsistency of the practice of posting advisory speeds (8) and the variability of posted speed limit practices (9, 10, 11, 12) have led drivers to regularly disregard speed signs. Further limiting the effectiveness of the signage practice is the fact that drivers are not merely disregarding signs; they are failing to notice them. Studies have found that drivers typi- cally comprehend only 56 percent of the signs posted along the roadway (13). Further, even when drivers are conscientious in their attempts to adhere to factors such as posted speeds, they naturally increase speeds toward the roadway’s design speed when they begin to shift their concentration away from monitoring the speedometer (14). This has implications for both geometric design and broader design practice, as it in- dicates that violations of driver expectations may, to some extent, be directly associated with the design speed of the spe- cific road. Overall, these findings suggest that signage and other similar applications may have only a moderate effect on preventing run-off-road events. Enhancing Lane Delineation Run-off-road crashes often occur during reduced visibility conditions (e.g., at dusk, dawn, and night, and during rain). Thus, enhanced lane delineation may help keep an alert driver from departing the road unexpectedly. There are several methods for improving lane delineation in urban areas. These may include curb lines, edge striping, street lighting, reflec- tive pavement markers, and pavement texture and/or color treatments. The location of a concrete curb adjacent to an asphalt road is common for many urban regions. The contrasting light color of the curb helps define the edge of the travelway. In re- gions where both roads and curbs are made of concrete, the face of the curb may sometimes be painted to create a con- trast. One disadvantage to using the curb line as the sole method for lane edge delineation is that frequent curb cuts or disruptions (at driveways or pedestrian crossing locations) may misdirect drivers who are fatigued or impaired. Infor- mation about the curb condition as a common urban road- side feature is included later in this chapter. In rural and select urban environments where roadways do not have curb lines, a common method for lane edge delin- eation is edge striping (using reflective paint for low-volume roads and thermoplastic stripes for more densely traveled facilities). The use of edge striping in the urban environment varies. Many jurisdictions elect to use edge striping for major facilities only and allow the standard center striping combined with the curb line to delineate lane edges for lower speed local roads. A common strategy for enhancing roadway visibility on urban streets is the use of street lighting. This not only illu- minates the travelway, but also provides safety and security for adjacent pedestrian facilities. Lighting is further discussed later in this chapter. The use of reflective pavement markers (raised or snow- plowable) can also help delineate the vehicle travelway. These pavement markers often require regular maintenance (for lens replacement or replacement of missing markers), so extensive use of reflective pavement markers is generally reserved for high-volume locations or for locations that are perceived to be high risk. Finally, many jurisdictions are experimenting with alter- native pavement treatments. The use of skid-resistant pavement surfaces has been a common recommendation for minimizing run-off-road crashes during inclement weather (1); however, an additional strategy is to change actual pavement color or pavement type at critical locations such as pedestrian cross- walks (for transverse delineation) and pavement edge (for lon- gitudinal delineation). The City of Charleston, South Carolina, has maintained several cobblestone roads in their historic peninsula district. These roads serve to clearly identify the motor vehicle space from the adjacent pedestrian space and also provide the added benefit of dramatically slowing motor vehicle operating speeds on these roads. The rough cobble- stone pavement treatment is not conducive, however, to safe bicycle activity. In Denmark, the road surface treatments help road users to clearly define who is to use a specific area of the road system. These road surface treatments also help to de- fine transitions from public to private space (15). Variations in road surface can be achieved by using patterns, textures, and similar treatments. Taking Advantage of Characteristics of the Surrounding Environment While signage is most typically used to delineate hazardous conditions, the FHWA scan of European practice suggests that signage is only one means of informing the driver of changes in appropriate driving behavior (16). Drivers are monitoring both traffic signs and the physical environment as part of the driving task. While adequate signage is important for encour- aging safe driver behavior under changing environmental 9

conditions, environmental hazards are signaled by more than just the signs posted adjacent to the travelway. The geometric design of the roadway and the characteristics of the surround- ing environment provide the driver with cues regarding safe operating behavior. One observation from the previously mentioned scanning tour was that Europeans try to make the entire roadway send a clear and consistent message regarding safe operating be- havior. Thus, design speeds are related to the physical envi- ronments in which the roadways are located, and the posted speed is meaningfully related to both. Typically, European design guidance specifies tight design ranges for each road- way class, with a range of typically not more than 20 km/h (approximately 12 mph) for any single road type in the urban environment. By narrowly specifying an appropriate design speed range, designers are able to minimize the instances—such as an isolated sharp curve—that may be a potential hazard. An important aspect of this European practice is that it is adopted for the purposes of enhancing the safety of the road- way. Agencies adopting such practices typically aim to achieve, at a minimum, a 40-percent reduction in crashes over a 5-year period, and, in many cases, agencies aim to have zero fatalities over a 10-year period (16). In a study of how people conceptualize urban environments, Kevin Lynch found that features such as architecturally unique buildings, key viewsheds, and other environmental stimuli serve as central reference points by which individuals orient themselves and cognitively map their travel progress (17). The observation that such features figure prominently in the way individuals visualize their travel activity suggests that environ- mental features provide drivers with important cues regard- ing appropriate driving behavior. The use of environmental factors to help inform drivers of safe operating conditions has received little attention in the literature (18), although the field of traffic psychology has begun to strongly encourage the use of environmental features as a key strategy for enhancing transportation system safety (19). Transit New Zealand’s Guidelines for Highway Landscaping encourages agencies to use highway planting to help drivers understand the road ahead (20). Plantings are recommended to help with curve delineation, headlight glare reduction, visual containment, and speed awareness and stimulation. In 2001, the City of Las Vegas, Nevada, developed a guide for neighborhood traffic management. For this effort, they performed a community survey in which respondents rated pictures of various street cross sections (21). The most popular images were tree-lined streets in residential areas and com- mercial buildings placed close to the road in business districts. Both the trees and buildings provide a sense of enclosure that frames the street and narrows the driver’s field of vision. The Las Vegas guide further suggests that when the buildings are set farther back from the street, the roadway appears to be wide and conducive to excessive speeds. The enclosed environment helps to mitigate speeding. In New Zealand, this enclosed environment is captured using a vertical elements technique in which the heights of vertical features are designed to be greater than the width of the street to provide the optical ap- pearance of a narrow street (22). These vertical elements can include trees, light poles, and other elements as long as the human-made objects are frangible, and trees or shrubs have narrower trunks and do not interfere with sight lines. Rumble Strips Physical rumble strips are grooves placed into the roadway or paved shoulder and are aimed at alerting the driver of potentially hazardous conditions (see Figure 1). A similar alert can also be achieved using thermoplastic rumble strips which are generally placed on the surface of the road in con- junction with lane edge delineation. While transverse rumble strips are often used for purposes such as alerting the driver of a downstream stop condition such as a toll booth or a stop- controlled intersection, longitudinal rumble strips are also effective for alerting the driver that he or she is leaving the travelway. Although rumble strips do not have speed-reducing capabilities (23), they cause a vehicle to vibrate and make noise when it crosses over them, thereby signaling to the driver that greater attention to the traveling environment is warranted. The sound made by a vehicle crossing rumble strips typically does not exceed that of the ambient sound experienced by the driver (24); thus, the ability of rumble strips to alert drivers 10 Photo reprinted from “New Focus for Highway Safety.” (119) Figure 1. Rumble strips.

to hazardous conditions is largely restricted to the vibration the rumble strips produce. This vibration is nevertheless a substantial cue for increasing the driver’s awareness of the roadway environment. Several studies of the effectiveness of rumble strips have determined that placement of rumble strips can decrease the number of run-off-road crashes be- tween 30 and 85 percent (25, 26). Applicability of Shoulder Rumble Strips to Low-Speed Urban Roadways While shoulder-based rumble strips have proven effective in reducing run-off-road crashes on interstates and freeways (particularly in rural environments), their applicability to lower-speed roadways may be limited. The use of physical (grooved) shoulder rumble strips assumes the existence of a level, paved shoulder. In many urban environments, raised curb is used in lieu of shoulders. This prevents the possibility of introducing physical rumble strips as a potential treatment for eliminating run-off-road urban crashes; however, ther- moplastic rumble strips may be used in the urban setting to achieve a similar result. Another issue affecting use of shoulder rumble strips in urban areas is that when a shoulder is available in urban areas, in many cases it serves as a travelway for bicyclists (27). Beyond the physical unpleasantness that rumble strips may pose for the bicyclist, the application of rumble strips can potentially result in the loss of control of the bicycle (28). Given the potentially negative influence on bicycle use in urban areas, as well as the frequent use of raised curb, the use of shoulder rumble strips is typically not appropriate on low-speed urban roadways. Finally, a common complaint about rumble strips in urban environments is that the noise they generate disrupts the peaceful environment of the adjacent land and the residents of the area (particularly during the quieter night hours). This perceived adverse affect on adjacent property owners result- ing from the use of rumble strips also limits their use in urban areas. Mid-Lane Rumble Strips A potential rumble strip treatment that may be more appli- cable to urban roadways—particularly urban arterials—is the use of mid-lane rumble strips. In this treatment, rather than applying rumble strips to the shoulder of the road, rumble strips are placed in the center of the vehicle travel lane. In this application, as vehicles leave their travel lane, their tires cross over the rumble strips, thereby producing the sound and vibration associated with shoulder rumble strips, without re- quiring a shoulder-based treatment (25). While mid-lane rumble strips are largely untested, they nevertheless present possibilities for improving roadside safety in urban areas. Placing a mid-lane rumble strip in the outside travel lane can be used to produce the same effect on the vehicle as a shoulder- based treatment (i.e., sound and vibration) without necessi- tating a roadside treatment. The use of a mid-lane treatment raises two potentially im- portant questions. First, what is the appropriate location of such a treatment for a curbed urban roadway? Second, what are the impacts of such a treatment on motorcyclists? In the case of roadways where the shoulders are curbed, or where there is a limited operational offset, the mid-lane rumble strip can be oriented to correspond to the expected location of the left tire of the roadway’s design vehicle. In these cases, the narrowest vehicle—a passenger vehicle—is the appropriate design vehicle for the treatment. Thus, the left tire of passen- ger vehicles will be used to delineate the appropriate position of mid-lane rumble strip treatments (or the right tire, assum- ing a treatment oriented toward preventing a crash into the median). While such an application will do little to address the safety needs of larger design vehicles, it should, nevertheless, have an effect on decreasing the rates of passenger-vehicle run-off-road crashes. Addressing the needs of motorcyclists is more difficult. While it has been demonstrated that motorcyclists can safely navigate rumble strips (24), the vibration associated with rumble strips can create discomfort when the rider is forced to travel over them for prolonged periods. An assumed min- imum motorcycle tire width of approximately 13 cm (5 in.) can reasonably be accommodated with a left- or right-tire offset on a 3-m (10-ft) travel lane. Nevertheless, such an appli- cation should be further researched before being employed in practice. While the use of mid-lane rumble strips seems promising, it is important to consider the longer-term behavioral impacts that may result from a widespread use of mid-lane rumble strips. In urban areas, travel is often characterized by frequent lane changing. Where mid-lane rumble strips are common, drivers may become acclimated to the sound and vibration they produce and cease to treat them as special events that re- quire increased attention to the driving task. In addition, the placement of mid-lane rumble strips should not occur at locations of heavy pedestrian activity, such as mid-block pedestrian crossings. Both the raised and grooved rumble strips create a potential tripping hazard for pedestrians by in- troducing an uneven walking surface. Safety of Urban Roadside Elements An urban environment is characterized by many potential roadside hazards. To improve roadside safety, many of these objects can be removed or relocated; however, it is probable that numerous prospective roadside hazards must be retained 11

to facilitate the needs of the community or the road users. As a result, this chapter reviews known roadside objects and strategies that may help improve the safety of their place- ment. Table 5 provides an overview of common urban road- side features and features often sought by local stakeholders to increase the aesthetic quality of urban roadsides. Each of these items is reviewed in greater detail in this chapter. Removal/Relocation/Placement of Roadside Objects Engineers are encouraged to identify potentially danger- ous objects adjacent to the travelway and remove them, ideally through the use of a clear zone. The recommended standard practice for higher-speed roadways is the provi- sion of a lateral clear zone that will enable at least 80 per- cent of errant vehicles to stop or return to their travel lane safely. The appropriate width of clear zones is ultimately based on the slope of the roadside, daily traffic volumes, and speed (1). The opportunities for providing a clear zone in urban areas are often limited due to the restricted width of the existing right-of-way and the density of adjacent roadside develop- ment. Use of the available right-of-way includes many com- peting demands. Further, many communities seek to provide a physical buffer zone adjacent to the travelway to encour- age pedestrian activity or to enhance the aesthetic quality of the roadway. Often, this involves the planting of trees or in- clusion of landscaping in a buffer area between the sidewalk and the vehicle travelway. Placement of mature street trees in close proximity to the road can present a hazard to the motorist. Minor departures from the travelway under these conditions can result in a potentially serious fixed-object crash, particularly at high speeds. Often, a configuration with rigid objects located immediately adjacent to the travelway is a result of a road-widening project where the only way to accommodate increasing vehicle capacity demands within the constraints of the current transportation infrastructure was by further encroaching on the existing roadside (29). Under current urban roadside design guidelines, engineers are provided with a special designation, the operational offset, which effectively permits the location of fixed objects 0.5-m (1.5-ft) from the curb face (1, 30). This offset value is a min- imum suggested distance associated with avoiding such op- erational issues as car-door and vehicle mirror conflicts with roadside objects and minimizing the impact to traffic opera- tions; it is not provided for safety purposes (31). The opera- tional offset should not be considered as an acceptable clear zone, but simply as a minimum value to ensure elimination of traffic operational conflicts. Where a clear zone cannot be achieved, the individual road should be tailored for the con- ditions at a specific site. The influence of supplemental factors such as crash history, future traffic, and heavy vehicle pres- ence should be included in the decision process. For evaluation of changes to the roadside such as removal of potential hazards, an engineer must determine whether the benefits associated with relocating a hazardous object outweigh the cost of doing so. The “cost” may take many forms, such as societal impacts or actual removal dollars, so elaborate cost-benefit methodologies have been developed to estimate the relative benefits of removing these objects (29, 32, 33, 34, 35). If a potentially hazardous object must be located adja- cent to the travelway, the principal means of addressing run-off-road crashes where adequate clear zones cannot be provided is to ensure that any object placed in the clear zone is “crashworthy,” that is, any object located in the clear zone is designed to minimize the severity of a potential crash. NCHRP Report 350 (36) provides specific standards and test conditions, such as soil and vehicle specifications, that are used for evaluating the crashworthiness of roadside fixtures such as guardrails, utility poles, and light supports. The reader is referred to NCHRP Report 350 for a full con- sideration of the test specifications used in the evaluation of crashworthiness. Two strategies exist, both of which are subject to the test conditions contained in NCHRP Report 350. The first is to incorporate frangible roadside objects and hardware into the 12 Features Immediately Adjacent to the Travelway Safety Barriers Curbs Barriers and Guardrails Shoulders Bridge Railings Channelization Crash Cushions and End Terminals Medians Roadside Grading Static Roadside Objects Dynamic Roadside Features Mailboxes Bicycle Facilities Landscaping, Trees, and Shrubs Parking Street Furniture Sidewalks and Pedestrian Facilities Utility Poles, Luminaires, and Sign Posts/Hardware Table 5. Common urban roadside features.

design of the roadside environment, and the second is to shield or cushion potentially hazardous objects and environments. A more detailed description of known safety strategies for the various urban roadside elements previously identified in Table 5 is presented in the following sections. Features Immediately Adjacent to the Travelway Physical features immediately adjacent to the travelway are the first objects encountered when an errant vehicle exits the road. These features can include curbs, shoulders, channelized islands, medians, and roadside grading. This section reviews each feature and known safety issues regarding each item. Curbs General information. Much of the rural research re- garding pavement edge treatments evaluates the influence of graded or paved shoulders on safety performance at the time a vehicle enters the roadside environment. In an urban envi- ronment, very few road edge treatments include roadway shoulders as a transition from the travel lanes to the adjacent roadside environment. Instead, curb is commonly used in urban environments as it can help direct storm drainage (thereby reducing the need for roadside ditches and wider right-of-ways) and provides visual channelization to help delineate the pavement edge. There is, therefore, a need to understand the safety of curbs in the urban environment. An important issue of concern for addressing roadside safety at curbed locations is the influence that various curb types may have in causing vehicles to trip or launch during a run-off-road event where the “first harmful” object the vehi- cle encounters is the roadside curb. The vertical curb has an almost vertical face and is generally between 150 to 225 mm (6 to 9 in.) in height above the driving surface of the adjacent pavement. The vertical curb is used as a means for discour- aging motorists from intentionally leaving the roadway. A sloping curb has an angled surface, a height of 150 mm (6 in.) or less, and is designed for use on higher-speed roadways (greater than 70 km/h [approximately 45 mph]) or at loca- tions where a vehicle may need to leave the roadway for emer- gency purposes. The sloping curb is designed so that a vehicle can traverse the curb without damaging the vehicle (30). Curbs “A” and “B” in Figure 2 depict example vertical curbs, while Curbs “C” through “E” represent various sloping curb configurations. The AASHTO publications A Policy on Geometric Design of Highways and Streets (referred to hereafter as the Green Book, [37]) and the Roadside Design Guide (1) both indicate that a vertical curb, struck at higher speeds, may cause an errant vehicle to mount and/or launch. The Roadside Design Guide currently provides the following recommendations for the design of vertical curbs (1, p. 10–7): At speeds over 40 km/h (25 mph), a vehicle can mount the curb at relatively flat angles. Consequently, when sidewalks or bicycle paths are adjacent to the traveled way of high-speed facilities, some provision other than curbing may need to be made for the safety of pedestrians and bicyclists. The Roadside Design Guide further suggests provision of a minimum horizontal clearance of 0.5 m (1.5 ft) beyond the face of curbs to any obstructions. This distance is the opera- tional offset previously discussed. The Green Book recommends the use of curbing on road- ways with speeds of approximately 73 km/h (45 mph) or less (37). The Green Book further notes that when vertical curbs are used on these lower-speed roadways, placement of verti- cal curb will preferably be offset 0.3 to 0.6 m (1 to 2 ft) from the edge of the travelway. The Green Book recommends against the use of curbs along high-speed arterials such as freeways, but indicates that when used on these facilities, a curb “should be of the sloping type and should not be located closer to the travelway than the outer edge of the shoulder.” (37, p. 322) A Guide for Achieving Flexibility in Highway Design (31) also indicates that vertical face curbs at low-speed (40 km/h [25 mph] or less) locations have limited redirectional capa- bilities for errant vehicles. For speeds above 40 km/h (25 mph), the curb can influence driver behavior, but does not provide a vehicle redirection function. Safety research. The research supporting the statements summarized above and found in common design guidelines spans curb crash testing and computer modeling over a period of many years. However, crash testing standards, computer modeling capabilities, and typical study vehicles have changed during this period. In general, researchers have performed test- ing on vertical curbs, sloping curbs, and curbs with adjacent 13 Graphic adapted from A Policy on Geometric Design of Highways and Streets, 4th ed. (37). Sample Sloping Curbs Sample Vertical Curbs (C) (E)(D) (B)(A) Figure 2. AASHTO example curbs.

barriers such as guard rails. In 1972, Dunlap et al. (38) per- formed several roadside curb evaluations including tests for five standard curbs and eight curb/guardrail combinations. In 1974, Olson et al. (39) evaluated curbs using crash testing combined with computer simulation. These two research studies were among the first to suggest the following commonly accepted concepts regarding curb safety: • Curbs 150 mm (6 in.) tall or less do not redirect vehicles at speeds above 73 km/h (45 mph) and should therefore not be used for high-speed roads, • Impacting curbs 150 mm (6 in.) tall or less will generally result in either no injury or minor injuries only, and • Combinations of lower speeds and small approach angles produce the greatest effect on vehicle path correction. A study performed in the 1970s at the Texas Transportation Institute evaluated curb placement in conjunction with traffic barriers and sloped medians (40). The researchers concluded that the traffic barriers should not be immediately adjacent to curbs as vehicles may vault or underride the barrier. They also concluded that grading the median or roadside level with the top of the curb will help reduce problems with barriers and guardrail interactions near curbs. An evaluation performed for the Nebraska Department of Roads (NDOR) included crash tests as well as simulations of sloping curbs and curb-guardrail combinations (41). The re- searchers’ evaluation included various degrees of impact and vehicle trajectory. They concluded that the three sloping curbs tested (two NDOR standard curbs and one AASHTO standard curb) were traversable for a wide range of impact conditions and had very little likelihood of causing vehicle rollovers. The researchers further determined that the chance that a vehicle could underride a guardrail was slight, and the chance that a vehicle would be vaulted by the curb-guardrail combination was greatest when the barrier was located any- where from 0.45 to 3.7 m (1.5 to 12.1 ft) behind the curb. This range of offset values applied to both a small and a large test vehicle. A report commissioned by the Florida Department of Trans- portation (42) simulated the trajectories of three design vehi- cles hitting sloped (125 mm [5 in.] tall) and vertical (150 mm [6 in.] tall) curbs at approach speeds of approximately 57, 73, and 90 km/h (35, 45 and 55 mph) and impact angles rang- ing from 3 to 15 deg. The model results found that the verti- cal curbs would deflect the Ford Festiva test vehicle for all approach speeds at angles of impact up to 12 deg. For a Chevy C2500 pickup, the vertical curb would deflect vehicles oper- ating at 90 km/h (55 mph), but only when the angle of impact was 3 deg or less. The sloping curb was not shown to redirect the vehicle under any combination of approach speed or angle. None of the impacts were shown to result in a rollover or substantial vertical displacement of the vehicle, nor were the events shown to result in more than minor damage to the vehicle. The AASHTO Highway Safety Design and Operations Guide (30) indicates that the potential for vehicle vaulting or rollover for curbs higher than 100 mm (4 in.) is a factor of the vehicle’s weight, speed, suspension system, angle of impact, and vehicle lane tracking. As a result, small cars are generally overrepresented in serious curb-related crashes. The poten- tial for a vehicle to vault precludes the exclusive use of a curb as sufficient protection for pedestrian facilities or roadside elements. In 2005, Plaxico and colleagues published NCHRP Report 537: Recommended Guidelines for Curb and Curb-Barrier In- stallations in which they evaluated roads with operating speeds of 60 km/h (40 mph) or greater and the potential influence of curb or curb-barrier combinations at these locations (43). They determined that the most significant factor influencing vehicle trajectory is curb height. As a result, shorter curbs with flatter sloping faces should be used at higher speed locations. They also determined that a lateral distance of approximately 2.5 m (8.2 ft) is needed for a traversing vehicle to return to its predeparture vehicle suspension state. As a result, guard- rails should not be placed closer than 2.5 m (8.2 ft) behind curbs on roads where vehicle speeds are greater than 60 km/h (40 mph). As the research performed by Plaxico and col- leagues did not focus on low-speed roads, the placement of guardrails behind curbs for speeds lower than 60 km/h (40 mph) is not known. In summary, curbs can provide positive (visual) guidance for drivers, but curbs do not have the ability to redirect errant vehicles upon impact (unless the vehicle speed is quite low and the vehicle impact angle is extremely small). If an errant vehi- cle approaches the curb at a small deflection angle, the impact of the curb is unlikely to be the cause of serious injury to the vehicle occupants; however, the curb may affect a vehicle’s tra- jectory, resulting in impact with a second, more substantial roadside hazard. A barrier or guardrail must be placed behind the curb in such a way as to avoid vaulting the errant vehicle. Strategy summary. A variety of strategies have been pro- posed, applied, and/or tested for safe application of curb treatments. Common strategies are as follows: Purpose Strategy Prevent curb from vault- • Use appropriate curb heights with ing vehicles known influences on vehicle trajectories (P) • Locate barriers behind curbs an appropriate distance to improve curb-barrier interactions (P) • Grade adjacent terrain flush with the top of the curb (P) 14

Shoulders General information. The common edge treatment for urban roads is a curb or curb with gutter; however, many roads exist in urban environments with a graded or paved shoulder instead of a curb located immediately adjacent to the travelway. The purpose of a shoulder is to provide a smooth transition from the travelway to the adjacent roadside while facilitating drainage and promoting various other shoulder functions (as listed in Table 6). The shoulder width is included as part of the clear zone width; therefore, the values shown in Table 6 should not be confused with clear zone requirements. There are many recommendations regarding appropriate shoulder widths for lower speed roads. These widths vary depending on the function of the shoulders as well as the available right-of-way. Table 6 shows suggested shoulder widths from the AASHTO publication A Guide for Achieving Flexibility in Highway Design (31). This information was first compiled for a 1982 NCHRP study (44). These widths are recommended for shoulder functional use and do not reflect identified widths for safety purposes. Because right-of-way costs are high in urban environments, the use of paved or graded shoulders in these environments often is the result of previously rural roads being incorpo- rated into urbanized land use without the companion road- way improvements. Often the road with a shoulder will have a drainage ditch located parallel to the road, so care must be taken to maintain traversable conditions in the event that an errant vehicle exits the road, travels across the shoulder, and then encounters the roadside grading. There are many research studies that have evaluated the safety benefits of shoulders and companion shoulder widths. Several of these studies are included in the safety research sec- tion that follows. Safety research. The research regarding shoulder safety has been generally divided into three categories—safe shoulder width, pavement edge treatments, and safety of paved versus graded shoulders. The research regarding these three areas of shoulder safety is summarized in the following: • Safe shoulder width. Much of the research into the appro- priate width of shoulders focuses on the high-speed rural condition. Early research indicated that crash frequency tended to increase with shoulder width. For example, Belmont published a paper on rural shoulder widths in 1954 and a subsequent paper in 1956 (extending the study to lower volume rural roads) and suggested that wider shoul- ders for higher speed, high-volume rural roads resulted in increased crash rates, while the trend appeared reversed for lower volume, high-speed roads (45, 46). Subsequent to these early studies, numerous researchers have studied the shoulder width question. In an unpublished critical review of research in this area (47), Hauer re-evaluated many of the original shoulder width studies using the original data and concluded that shoulder width safety is a sum of several opposing tendencies. These can be summarized as follows: – The shoulder is even and obstacle free and available for drivers of errant vehicles to use to regain control of their vehicles, correct for their error, and resume normal travel; 15 Functional Classification Shoulder Function Arterial m (ft) Collector and Local m (ft) Drainage of Roadway and Shoulder 0.3 (1) 0.3 (1) Lateral Support of Pavement 0.45 (1.5) 0.3 (1) Encroachment of Wide Vehicles 0.6 (2) 0.6 (2) Off-tracking of Wide Vehicles 0.6 (2) 0.6 (2) Errant Vehicles (Run-off-road) 0.9 (3) 0.6 (2) Bicycles 1.2 (4) 1.2 (4) Pedestrians 1.2 (4) 1.2 (4) Emergency Stopping 1.8 (6) 1.8 (6) Emergency Vehicle Travel 1.8 (6) 1.8 (6) Garbage Pickup 1.8 (6) 1.8 (6) Mail and Other Deliveries 1.8 (6) 0.6 (2) Emergency Call Box Services 2.4 (8) 1.8 (6) Law Enforcement 2.4 (8) 1.8 (6) Parking, Residential 2.4 (8) 2.1 (7) Routine Maintenance 2.4 (8) 1.8 (6) Major Reconstruction and Maintenance 2.7 (9) 2.7 (9) Parking, Commercial 3.0 (10) 2.4 (8) Parking, Trucks 3.0 (10) N/A Slow-Moving Vehicles 3.0 (10) 2.7 (9) Turning and Passing at Intersections 3.0 (10) 2.7 (9) Sources: Adapted from A Guide for Achieving Flexibility in Highway Design (31) and NCHRP Report 254: Shoulder Geometrics and Use Guidelines (44). Table 6. Acceptable shoulder widths for shoulder functions.

– Wide shoulders may induce voluntary stopping and therefore place a hazard immediately adjacent to the travelway; – Wide shoulders may entice drivers to use them as addi- tional lanes or for passing maneuvers on the right; and – Wider shoulders may encourage higher operating speeds. Evaluation of crash data without comprehensively consider- ing these four contrasting tendencies may permit researchers to arrive at a variety of conclusions regarding shoulder width safety. In general, on roads with wider shoulders, travel speeds are higher and crashes are more severe. How- ever, wider shoulders result in fewer run-off-road crashes, and therefore this benefit must be included. • Pavement edge treatments. A common problem with roadway shoulders is that they may not be flush either with the travelway pavement surface (for the case of graded shoulders) or with the adjacent roadside grading (for the case of paved shoulders). There are many reasons that pavement drop-offs may develop in the shoulder region. Erosion of the soil next to the pavement, rutting by frequent tire wear, and pavement overlay maintenance are examples of how, over time, a pavement drop-off may develop. When a drop-off is encountered by an errant vehicle, the vehicle’s tires may have difficulty mounting the extra pave- ment lip, causing the vehicle to further lose control. In the late 1970s and early 1980s, researchers at Texas A&M University performed a series of evaluations on pave- ment edge drop-offs (48, 49). They determined that ver- tical drop-offs as small as 7.6 cm (3 in.) could result in a severe crash if encountered by an errant vehicle. The Texas A&M researchers developed pavement edge shapes to pro- vide a more beveled edge and determined that for speeds up to 90 km/h (55 mph), a 45-deg angle could be applied to the drop-off. This sloped edge would then enable errant vehicles to regain access to the travelway safely. Currently, the Federal Highway Administration promotes a pavement edge treatment called the safety edge that uses a similar 45-deg angle with construction standards that permit com- paction to provide pavement edge stability. • Safety of paved versus graded shoulder. The safety of paved versus graded shoulders is less controversial than the pavement width consideration. Several studies have indi- cated that the addition of any paved shoulder will help with crash reduction. Zegeer, Deen, and Mayes concluded that increasing the width of a paved shoulder for rural roads by 0.3 m (1 ft) would reduce crashes by approximately 6 per- cent (50). They also concluded that paving at least 0.3 m (1 ft) of a shoulder would reduce crashes by 2 percent. More recently, McLean found that for Australian roads the application of sealed shoulders with widths from 1.5 to 2 m (5 to 6.6 ft) would result in a decrease in crash rates and, therefore, be a cost-effective treatment (124). Strategy summary. Common shoulder treatment strate- gies are as follows: Purpose Strategy Discourage run-off-road Provide wider shoulders suitable crashes for shoulder function (P) Provide traversable • Eliminate pavement drop-offs (P) transition for errant • Add a pavement safety edge (T) vehicles • Provide a paved or sealed shoulder (P) Channelization/Medians General information. The separation of traffic move- ments by the use of a raised median or turning island is often referred to as channelization. For the purposes of this review, a flush or traversable median or island is considered part of the roadway, while a raised median and raised turn island are considered part of the roadside. Channelized islands are generally used to reduce the area of pavement at an intersection while providing positive guid- ance to turning vehicles. Channelized islands can be used for pedestrian refuge and traffic control device placement, and they can also be planted with landscaping treatments that contribute to an improved visual environment (15). For a raised island to be visible, it should have a minimum size of 5 m2 (50 ft2) for urban conditions (37). The orientation of the curb on a raised island should be slightly skewed to the adja- cent travel lane to give an illusion of directing vehicles into the travel lane. Other cross-sectional characteristics of raised islands are similar to those of raised medians. The raised median provides the primary function of sep- arating opposing directions of vehicle travel. This physical separation has the added benefit of improving access man- agement (restricting frequent left turns into driveways), pro- viding a location for pedestrian refuge (assuming the median has adequate width), and providing road edge delineation during inclement weather conditions (particularly snow). A median may simply be raised using a vertical or sloped curb. In urban regions, median width can vary dramatically de- pending on the proposed function of the median. As sug- gested in the Maryland publication, When Main Street Is a State Highway: Blending Function, Beauty and Identity, the use of a median can dramatically improve the visual quality of a facility (51). Safety research. Many of the recent research studies about raised median safety focus on the influence of the median on access management and the resulting reduction of crashes due to restricted left-turn movements. Although this crash reduction strategy falls outside the scope of this litera- ture review, it is worth noting that the median condition has added safety benefits that should be considered in a 16

comprehensive crash evaluation. In this review, however, the evaluation will be focused on median crashworthiness for the purposes of roadside safety. In general, median research (excluding access management studies) has focused on the crash condition with specific atten- tion to the following questions: • Do medians prevent pedestrian cross-median crashes? What is the influence of a median barrier that completely prohibits pedestrian crossing? • Do medians reduce the number of crashes and crash severity? • Can landscaping and trees be safely located in medians? • Should median barriers be used to improve median safety? A median barrier review is included later in this chap- ter in the section on safety issues, so this section focuses on the influence of a median on crashes and landscaping treat- ments. Several researchers have weighed the merits of a raised median (divided highway) versus no median or a flush median. Unfortunately, in most of the before-after re- search studies, a jurisdiction was implementing a median improvement in conjunction with other improvements, such as road widening, lane narrowing, and so forth. As a result, the influence of divided versus undivided has re- sulted in a wide variety of crash observations. Harwood (52) studied several median conversion configurations from un- divided to divided operations. After controlling for a vari- ety of variables, he concluded that the influence of the median on safety was small. Many studies have resulted in similar observations, that is, that raised medians have a neg- ligible effect on crash frequency. Crash severity varies de- pending upon the median width (wider medians reduce the chance for head-on collisions), the use of median barrier (to be discussed later), and the placement of rigid objects in the median area. The issue of landscaping and the specific evaluation of tree placement are further discussed in the landscaping section; however, a recent three-phase study performed at California Polytechnic State University (53) specifically evaluated the placement of large trees in raised medians on urban and suburban highways. They evaluated sites with and without large trees and determined that at a 95-percent level of statistical confidence, an increased number of fatal or injury crashes were associated with the presence of median trees. The association between median tree crashes and left-side-only crashes, however, was only marginally significant. The three-phase study also indicated that me- dian trees on urban and suburban highways were associated with an increase in collision frequency. Study researchers were not able to identify any systematic relationships be- tween the left-side crash rates and median widths or tree setbacks. They also found that with the increase in fixed- object collisions came a decrease in head-on and broadside collisions. Finally, the researchers found non-intersection locations with median trees were positively associated with hit-pedestrian collision. Another common application for raised medians is as one element in a gateway treatment at transition locations between rural and urban areas. The raised median studies for the purpose of use as gateway strategies are reviewed later in this chapter in the section discussing traffic calm- ing applications. Research into appropriate median widths for safety pur- poses are focused primarily on the high-speed rural condi- tion. Similar median width studies for urban environments are not available. Strategy summary. Common channelized island and median strategies are as follows: Purpose Strategy Reduce likelihood of Widen median (T) run-off-road collision Reduce crash severity • Place only frangible items in channelized island or median (P) • Shield rigid objects in median (P) Roadside Grading General information. The terrain adjacent to an urban road should be relatively flat and traversable. In general, the placement of common urban roadside features such as side- walks and utilities tends to create a flatter urban roadside. The primary risk for irregular terrain adjacent to the travelway is that an errant vehicle will either impact a rigid obstacle or that the terrain will cause the vehicle to roll over. Rollovers were responsible for 20 percent of the fatal crashes in 2002, and the largest number of rollovers occurred after a vehicle impacted an embankment or a ditch (25, 54). The principal cause of rollovers is a vehicle “tripping” on an element of the roadside environment, such as a ditch or an embankment; neverthe- less, sharp pavement drop-off on the shoulder may also lead to vehicle tripping for roads without a curb. To prevent vehi- cle tripping, the grade of ditches, slopes, and embankments should be minimized as much as possible, and pavement drop-offs must be kept to a minimum. These strategies are potentially more relevant to rural and suburban environments than to urban ones, however. In urban areas, the roadside is typically characterized not by shoulders and embankments, but by curb and gutter applications and by adjacent roadside development. This is evidenced when one compares the absolute number of rollover crashes in urban environments with the number of rollover crashes in rural 17

environments. In 2002, there were roughly 1,800 rollover crashes in urban areas, compared with over 6,200 for rural re- gions. Accounting for exposure, roughly one rollover per billion miles of travel occurs in urban areas whereas almost six rollovers per billion miles of travel occur in rural environments. While the conditions that lead to rollover crashes are not clear, crash data analyses indicate that these crash types are generally associated with high-speed travel. Of the roughly 8,000 rollovers that occurred in 2002, only about 600 occurred on roadways classified as urban minor arterials, collectors, and locals. The sideslope of an urban road should, in general, slope from the edge of the right-of-way toward the curb of the road. This slope will prevent any road drainage from encroaching on adjacent property and enables the drainage to be contained within a closed drainage system. As a result, the slope is often quite flat (1V:6H typically) for curbed urban roads. For roads without a curb, the design guidelines for rural roadside con- ditions should be applied. That is, the terrain, including drainage channels, should be safely traversable by a motor vehicle, and the placement of obstacles such as headwalls must be flush with the ground surface and designed to be navigated by an errant vehicle. Safety research. The research team was not able to locate research specific to the urban roadside slope and safety implications associated with this terrain. Most of the studies applicable to the urban condition focused on the presence of roadside obstacles rather than the companion roadside slope. Strategy summary. Common grading strategies are as follows: Purpose Strategy Minimize crash likeli- Maintain traversable grades that are hood free of rigid obstacles (P) Minimize crash severity • Flatten grades to reduce chance of vehicle rollover (P) • Create an object setback policy (T) Static Roadside Treatments Mailboxes General information. The Roadside Design Guide (1) de- tails the preferred specifications for the design and installa- tion of mailboxes. In general, AASHTO recommends the use of a 100-mm by 100-mm (4-in. by 4-in.) wooden post or a 38-mm (1.5-in.) light-gauge pipe for mounting mailboxes, with these posts embedded no deeper than 600 mm (24 in.) in the ground. Mailboxes should further be mounted to their supports to prevent the mailbox from separating from the post during a crash event. Also of concern is the potential hazard associated with larger mail collection boxes, a common feature in urban environments, as well as neighborhood de- livery units, which are associated with apartment complexes. Crash tests of these features have shown them to fail safety requirements, and the Roadside Design Guide recommends placing them outside of clear recovery areas. While making mailboxes crashworthy will satisfy safety associated with mailbox-related crashes, it is important to recognize that the placement of mailboxes may have an important impact on the overall safety of the roadway. Mail- boxes should not obstruct intersection sight distance, nor should they be located directly on higher-speed roadways, where stopping associated with mail delivery and collection can lead to substantial speed differentials between vehicles on the travelway, thereby increasing the possibility of a rear-end collision. Where such conditions exist, the Roadside Design Guide recommends the use of a 2.4-m (8-ft) mailbox turnout lane adjacent to the travelway to allow vehicles to leave the travelway for mail collection and delivery purposes. This turnout concept does not apply to urban curbed streets. At curbed residential locations, the Roadside Design Guide rec- ommends that the minimum distance from the roadside face of the mailbox to the face of the curb should be 150 mm (6 in.), with a preferred offset ranging from 200 to 300 mm (8 to 12 in.). A common issue regarding the placement of mailboxes in an urban environment is that the governing jurisdiction (often a city or county) may not adopt the guidelines com- monly accepted by state departments of transportation. Many urban jurisdictions allow home owners to construct a mailbox of their choosing. In areas in which mailbox vandalism is common, home owners have begun to erect increasingly rigid (less forgiving) mailbox units. A rigid brick mailbox is a com- mon site along many urban residential roadways. The problem of rigid mailbox units is compounded by the general place- ment of such mailboxes adjacent to a driveway (to make it easy for the home owner to retrieve mail). Since the curb has a secondary function of delineating the edge of the roadway, a mailbox placed on the departure side of a driveway (where a curb cut interrupts the roadway delineation) is particularly vulnerable to errant vehicles that exit the road to the right. Safety research. NCHRP Report 350 provides recom- mended procedures to ensure roadside features such as mail- boxes are crashworthy (36). Since mailboxes are a common fixed object adjacent to urban streets (particularly residential), they warrant particular attention when reviewing urban road- side safety. Many urban jurisdictions do not require crash- worthy mailboxes. There are several yielding mailbox designs approved for the National Highway System (NHS) that could be incorporated in an urban setting. Chapter 11 of the Road- side Design Guide (1) provides a comprehensive summary of 18

the safe placement of mailboxes. The use of yielding mail- boxes is promoted in the Roadside Design Guide. This permits convenient mailbox placement adjacent to the road. Mailbox placement for urban commercial locations is not included in the chapter and is a less common problem. In addition to yielding mailbox support design, some jurisdictions promote the placement of reflective object markers on the mailbox or post to improve nighttime visibility (55). Strategy summary. Common mailbox safety strategies are as follows: Purpose Strategy Minimize crash likeli- • Remove or relocate mailboxes to hood safe locations (P) • Add reflective object markers to improve nighttime visibility (T) Minimize crash severity • Develop policies to require crashworthy mailboxes in urban environments (P) • Shield rigid mailboxes where practical (P) Landscaping, Trees, and Shrubs General information. Several types of roadside land- scaping are commonly employed to enhance the aesthetics of roadside environments. These treatments may include the placement of shrubs, street trees, or alternative treat- ments such as landscape berms. In addition to the concern of traversability in the event that an errant vehicle encoun- ters roadside landscaping, a common safety issue of adjacent landscape treatments is sight distance and the impact land- scape treatments may have for intersection, driveway, and stopping sight distance considerations. Regional jurisdic- tions often have landscaping design guidelines, landscaping policies, and street tree master plans. These documents address a variety of landscaping issues including plant type, maintenance, and plant placement. Since trees, in particu- lar, can vary from small, flexible species up to more rigid varieties, the careful selection of tree species is critical. In addition, different tree species can have substantially differ- ent root systems. Species selection should also focus on the potential for the tree system to adversely impact road surface and pedestrian facilities due to pavement heave and cracking. Placement criteria, in some cases, is based on the func- tional purpose or posted speed limits of adjacent roads. Com- mon landscape placement issues addressed in jurisdiction plans include the following: • Proximity to intersections, • Proximity to driveways, • Maintaining a clear vision space, • Lateral offset placement of trees and landscaping, • Longitudinal placement of trees and landscaping, • Median planting strategies, and • Strategic placement strategies for visual perception. These specific placement strategies are further described in the following: • Proximity to intersections. Sight distance should be main- tained in the proximity of intersections. As a result, many landscape guidelines restrict tree placement in the imme- diate vicinity of intersections. The North Carolina Tradi- tional Neighborhood Development (TND) Guidelines (56) and the City of Seattle Street Tree Planting Procedures (57), for example, recommend that trees should be located no closer than 9 m (30 ft) from intersection corners. Land- scape Design Guidelines for the City of Simi Valley (58) requires a clearance distance of 10.7 m (35 ft) from the ex- tended curb at the near side of the cross street curb. Another approach to intersection clearance is based on street type and intersection configuration. For example, the Montgomery, Alabama, Street Tree Master Plan (59) uses street type and traffic control to determine tree offsets from intersections. Example minimum tree placement guidelines at intersections for Montgomery are depicted in Table 7. Figure 3 is based on an FHWA publication and demon- strates the sight triangle that is required to be free of trees at an example intersection (60). Higher-speed vehicles that do not stop at the intersection require more sight distance than stopped vehicles, as shown. Guidelines for Tree Planting and Maintenance on Urban Roads, published by the Traffic Authority of New South Wales, further recommends that skewed intersections, 19 Intersection Control Major Street Neighborhood Street Traffic Light 9.1 m (30 ft) -- Four-Way Stop 9.1 m (30 ft) 4.6 m (15 ft) Major Street Two-Way Stop 12.2 m (40 ft) -- Neighborhood Street Two- Way Stop 9.1 m (30 ft) 4.6 m (15 ft)–Stops 9.1 m (30 ft)—Does Not Stop Source: Adapted from Montgomery Street Tree Master Plan (59). Table 7. Guidelines for minimum tree placement offsets at intersections for Montgomery, AL.

locations with high turning speeds, or locations where fast- approaching vehicles veer into the left lane to avoid impacting right-turning vehicles are all locations where cluster run-off-road crashes can be expected and where additional space free of roadside objects such as trees should be provided (61). • Proximity to driveways. The placement of trees near drive- ways poses similar sight distance issues as those identified for intersections. As an example, the City of Simi Valley guidelines requires a 1.5-m (5-ft) clearance between trees and driveway edges (58). By contrast, the Montgomery recommendations indicate that trees should not be placed within 4.6 m (15 ft) of driveways (59). The City of Seattle, in Street Tree Planting Procedures, requires maintaining a minimum distance between trees and driveways of 2.3 m (7.5 ft) with a recommended distance of 3.0 m (10 ft) (57). Many landscape policies do not directly stipulate tree placement near driveways, but use an approach similar to the New South Wales guidelines, which simply state that drivers exiting driveways should be able to see approaching traffic and pedestrians (61). As an example, the attractive landscaping depicted in Figure 4 seems reasonable upon initial inspection; however, at this location the intersecting roadway is characterized by a horizontal curve. With the landscaping placement close to the driveway, the driver of an exiting vehicle cannot detect approaching vehicles with- out edging into the active travel lane; therefore, these road- side treatments encroach on the required sight distance. As can be viewed in the photo, the adjacent property owner also positioned large “ornamental” rocks at the corner, thereby adding a rigid obstacle in the immediate vicinity of the roadway. • Maintaining a clear vision space. Traditional Neighbor- hood Development (TND) Guidelines recommends that vertical space ranging from 0.6 to 2.1 m (2 to 7 ft) above ground be maintained to preserve lines of sight (56). The AASHTO publication Highway Safety Design and Opera- tions Guide (30) recommends that the vertical “clear vision space” range from 1 to 3 m (3.3 to 10 ft) to ensure clear sight distance for drivers in low-riding sports cars as well as drivers in high trucks and buses. This vertical clear space is common to many regional landscaping plans. The “clear vision space” is essentially the space above shrub growth and below tree overhang. A low tree overhang can also create an obstacle for pedestrian access, as shown in Figure 5. Figure 6 depicts another type of encroachment into the vertical clear vision space. Often landscape berms are used to screen adjacent parking from the roadway. The photo on the left shows a longitudinal landscape berm and the effect it has on horizontal sight distance. At the location shown, the road has a horizontal curve to the right (in the direction the vehicle shown is traveling) and has numer- ous driveways that are not easily visible due to the berm height. The photo on the right depicts the same location but more clearly shows that the height of the berm exceeds the height of a typical passenger car. This type of landscape treatment is a common roadside treatment in many urban areas. • Lateral offset placement of trees and landscaping. Tra- ditional Neighborhood Development (TND) Guidelines 20 Photo by Karen Dixon. Figure 4. Landscaping in sight triangle for driveway. Graphic adapted from Trees in Hazardous Locations (60). Figure 3. Intersection sight triangles.

recommends that planting strips located between the curb and sidewalk should be at least 1.8 m (6 ft) wide (56). This resource further suggests that for streets with design speeds at or below 32 km/h (20 mph) or for streets that permit on-street parking, small street trees can be planted within 0.9 m (3 ft) of the back of curb or along the approximate centerline of the planting strip. The Seattle planting proce- dures permit tree planting 1.1 m (3.5 ft) from the face of curb (57). The Montgomery, Alabama, plan recommends that at neighborhood street locations trees should be in- stalled at a point equidistant between the pavement edge and the right-of-way limits or, for residential neighborhoods, equidistant between the pavement edge and the edge of the sidewalk (59). For major street locations, trees should not be located closer to the edge of pavement than two-thirds of the distance from the pavement edge to the right-of-way limits. The Georgia Department of Transportation Online Policy and Procedure System (62) recommends that in an urban environment, trees with diameters less than 100 mm (4 in.) should be laterally positioned 1.2 m (4 ft) for posted or design speeds of 56 km/h (35 mph) or less, 2.4 m (8 ft) for posted or design speeds of 64 to 72 km/h (40 to 45 mph), and outside the clear zone for speeds greater than 72 km/h (45 mph). For larger trees, the minimum lateral 21 Photos by Karen Dixon. Figure 6. Landscape berm that blocks horizontal sight distance. Graphic reprinted fromVegetation Control for Safety. A Guide for Street and Highway Maintenance Personnel, FHWA-RT-90-003 (120). Figure 5. Overhead object hazard.

placement should be 2.4 m (8 ft) from the curb or 1.2 m (4 ft) from the outside shoulder in central business districts or commercial locations for posted or design speeds of 56 km/h (35 mph) or less. Similarly, for speeds of 64 km/h (40 mph) the lateral clearance of 3.0 m (10 ft) should be maintained. For speeds of 72 km/h (45 mph), a lateral off- set of 4.3 m (14 ft) is recommended. The Georgia Depart- ment of Transportation further requires that large trees be placed beyond the clear zone limits for speeds greater than 72 m/h (45 mph). AASHTO (30) suggests that landscape design include consideration of the mature size of trees and shrubs and how they will influence safety, visibility, and maintenance costs as they mature. In addition, if on-street parking is permitted, the landscape border area needs to be wide enough to accommodate the planned landscaping and still permit access to parked vehicles. New Zealand’s Guidelines for Highway Landscaping (20) recommends plant layering, an approach in which plants are grouped according to height, as depicted in Figure 7. This plant layering approach permits the use of roadside landscaping and, as indicated in the guide, will do the following: – Allow wider clear zones for rigid objects, – Permit the inclusion of large trees in the roadside design, – Allow appropriate sight distance, and – Permit visually appealing plant compositions. Finally, at horizontal curve locations the lateral offset of landscaping features must not obstruct sight distance, as shown in Figure 8. • Longitudinal placement of trees and landscaping. In ad- dition to lateral offset placement standards and constrain- ing the placement of trees near intersections or driveways, longitudinal placement strategies include placement to help develop tree canopies and placement to help avoid conflicting obstacles. One source recommends that trees be spaced so that mature tree canopies grow within 3 m (10 ft) of each other to help provide shade (63). This placement results in tree spacing from 7.6 to 15.2 m (25 to 50 ft) de- pending on the tree type. The City of Montgomery plan suggests tree placement approximately 9.1 m (30 ft) on center, but also emphasizes that canopy trees should not be positioned under service wires (59). Other placement constraints in the Seattle procedures include space separating trees from underground utility lines of 1.5 m (5 ft), a minimum space of 3.0 m (10 ft) sep- arating trees from power poles (4.6 m [15 ft] recom- mended), and a space of 6.1 m (20 ft) separating trees from street lights or other existing trees (57). The City of Simi Valley requires a separation of 4.6 m (15 ft) between trees and street light standards, 3.0 m (10 ft) between trees and fire hydrants and alleys, and 1.5 m (5 ft) between trees and water meters or utility vaults (58). • Median planting strategies. Many jurisdictions maintain similar lateral offset guidelines as for the right side of the roadside (in the direction of travel). The New South Wales guidelines note that where planting is required in a median, trees should be located a minimum lateral distance from the nearest travelway of 2.5 m (8.2 ft) (61). The City of Simi Valley guidelines suggest that median tree spacing can vary based on tree type (58). Shrubs should have a mature 22 Graphic reprinted from Trees in Hazardous Locations (60) Figure 8. Sight distance around a horizontal curve. Reprinted from Guidelines of Highway Landscaping (20) Figure 7. Example of plant layering.

height of 0.8 m (2.5 ft), and ground cover should be set back from the curb edge a minimum of 0.5 m (1.5 ft). • Strategic placement strategies for visual perception. One report from Denmark suggests that the traffic-related feature of roadside plantings may be due to the visual narrowing of the driver’s field of view, which results in speed reductions (15). This speed reduction hypothesis is echoed in other lit- erature, but it has not yet been empirically substantiated. Safety research. There is considerable anecdotal infor- mation in the literature that supports the potential benefits of roadside landscaping placement for health and driver well- being. Similarly, impact with a rigid object, such as a large tree, is a known hazard, and this danger has been consistently reinforced in rural roadside research. While shrubs are often classified with trees for the purposes of analyzing crash data, it is important to consider the safe placement of shrubs sep- arately from that of trees and other landscaping elements, as many types of landscaping elements are considered frangible. The majority of roadside landscape safety literature has focused on the safety condition of street trees. The placement or removal of street trees is often one of the most contentious elements with respect to the design of roadsides in urban areas. Urban stakeholders often seek to incorporate street trees in the design of urban roadsides; however, when trees are placed ad- jacent to the travelway, they can become rigid, fixed-object hazards. Current practice discourages the placement of trees with mature-tree caliper widths greater than 100 mm (4 in.) along the roadside (1, 37). This maximum tree size is based on the crash tests of 100 by 100 mm (4 by 4 in.) wooden signposts. A tree, unlike a wood signpost, has a root system; however, the wood post used for sign supports has long been the reference for tree size on the assumption that if the wooden signpost is safe, then a tree of a similar size should also be safe (64, 65). In 1986, the FHWA commissioned a study to reduce haz- ards due to trees (66). Although the study focused generally on the rural environment, the researchers found that for fatal tree crashes, the median tree diameter at breast height was 508 mm (20 in.), whereas the median tree diameter for nonfatal tree crashes was 381 mm (15 in.). FHWA’s Roadside Improvements for Local Roads and Streets further notes that trees with multi- ple trunks, groups of small trees, or a combination of a small tree and another fixed object can act as a potential hazard and should be considered collectively (67). For combined effects, the cross section should not exceed 83.87 sq cm (13 sq in.). A 1990 study performed by Turner and Mansfield evaluated urban tree safety in Huntsville, Alabama, based on a study of tree crashes (7). The study presents aggregate information on the characteristics of urban run-off-road crashes into trees, but did not include information on the specific road characteris- tics of the environments in which these crashes occurred. The authors concluded that mature trees with diameters larger than 10 cm (4 in.) should not be permitted within a roadside clear zone region. The authors further suggested that if the trees are needed for aesthetics or environmental reasons, the tree loca- tion should be behind a barrier, ditch, or retaining wall. A 1999 study conducted in Washington State examined both rural and urban environments and developed models for both conditions as well as models combining the urban and rural data (68). The researchers determined that the vari- able representing the number of isolated trees in a section had a negative sign (indicating a decrease in accident frequency) for urban areas. This same variable had a positive sign (indicating an increase in accident frequency) for rural loca- tions. The authors also evaluated crash severity. The models they developed predicted that in an urban environment isolated trees can be expected to result in possible injury while the presence of tree groups can be expected to result in disabling injuries or fatalities. The authors had a similar finding for tree groups in rural environments. The authors attributed the reduction of accident frequency due to isolated trees in urban environments to the fact that there are fewer trees in urban environments than in rural ones. In 1999, Kloeden and colleagues published a study of crashes that occurred between 1985 and 1996 in Southern Australia (69). The researchers did not separate the crashes into urban and rural categories; however, they did perform evaluations based on the speed zones associated with the crash locations. Table 8 depicts the distance to roadside haz- ards for fatalities that occurred during the study period for speed zones of 80 km/h (50 mph) or less. The offset values shown in this table are rounded to the closest meter. This Australian study found that 58.6 percent of roadside hazard fatalities were due to vehicle impacts into trees. A study published in 2003 evaluated five arterial roadways in downtown Toronto and sought to understand the safety impacts of placing landscape elements, such as mature trees, adjacent to the travelway. The sites were selected because all five sites were undergoing various environmental and aesthetic improvements due to community concerns associ- ated with major road reconstruction projects. The sites were tracked from 1992 to 1995 as they underwent these improve- ments. This study found a statistically significant reduction in mid-block crashes from the pre- to post-test conditions, although the authors did not elaborate on the nature of the crashes that were investigated as part of the study (70). These results, however, cannot be uniquely attributed to trees since the projects involved major reconstruction. Researchers from Monash University Accident Research Centre in Australia published a study in 2003 that evaluated roadside safety issues in Victoria from 1996 to 2000 (71). They determined that 4.1 percent of collisions with roadside trees resulted in a fatality, compared with only 2.3 percent of other roadside object crashes. They also noted that the likelihood of 23

a fatality is greater for collisions at higher speeds and that the most frequently impacted roadside hazards were trees, poles, fences, and embankments. A study published in 2005 by Bratton and Wolf performed an analysis using national crash data and concluded that crash frequency is generally higher and injury level more severe in higher speed rural areas (72). They also noted that crashes involving trees are more injurious than all crashes in general. Bratton and Wolf were not able to identify a signifi- cant difference between tree collision rates in urban and rural areas. Although they noted that trees, as fixed objects, statis- tically increase the likelihood of injury in accidents, trees are involved in a small overall percentage of these crash events. The authors went on to note that since the clear zone concept does not appear to be a feasible notion for the urban envi- ronment, designers should develop a way of safely integrat- ing trees into the urban roadside environment. Researchers at California Polytechnic State University performed a three-phase study completed in 2004 (with in- terim reports in 2002 and 2003) in which they evaluated the street tree application specifically for the urban median con- dition (53). In the initial stages of the research, the re- searchers performed a literature review and noted the wide variety of anecdotal evidence and conflicting empirical evi- dence produced for previous studies into the roadside safety of the urban street tree. They noted that there are a variety of clearance standards used throughout the United States for recommended offset values to roadside hazards such as large trees (defined as trees with diameters greater than 100 mm [4 in.] for their study) and very little direction regard- ing appropriate placement of trees in medians. As a result, Phases 2 and 3 of the study were focused on evaluating urban street trees in curbed urban and suburban highway medians with a variety of median widths, including narrow medians. The researchers concluded that large trees located in medians are associated with more total crashes as well as more fatal and injury crashes. The presence of median trees was statistically significant for the severity model developed for the study. The researchers also found a positive rela- tionship between median trees and hit-pedestrian crashes at non-intersection locations. The median width results were inconclusive. Strategy summary. A variety of strategies have been pro- posed, applied, and/or tested for safe application of landscap- ing and tree placement adjacent to the roadside. The safety of these strategies in some cases is well known; in other cases, strategies hold promise but their exact influence on safety is unknown. Common strategies are as follows: Purpose Strategy Prevent large trees from • Restrict/Refine planting growing in hazardous guidelines regarding tree and locations landscaping placement (T) • Implement plant layering strategies (T) Eliminate hazardous • Remove or shield isolated large tree conditions trees (diameter of 100 mm [4 in] or more) (P) • Shield tree groups (P) • Establish urban lateral offset guidelines for large trees (T) • Delineate hazardous trees to improve visibility (E) Minimize level of Reduce travel speed on adjacent severity road (P) 24 Distance of Roadside Hazard from Road (m)(ft) Number of Crashes Percentage (%) Cumulative Percentage (%) 0 (0) 34 22.2 22.2 1 (3) 38 24.8 47.1 2 (7) 30 19.6 66.7 3 (10) 18 11.8 78.4 4 (13) 12 7.8 86.3 5 (16) 5 3.3 89.5 6 (20) 3 2.0 91.5 7 (23) 1 0.7 92.2 8 (26) 3 2.0 94.1 9 (30) 1 0.7 94.8 10 (33) 3 2.0 96.7 14 (46) 2 1.3 98.0 15 (49) 2 1.3 99.3 16 (52) 1 0.7 100.0 Total 153 100.0 Source: Adapted from Severe and Fatal Car Crashes Due to Roadside Hazards (69). Note: Crashes involving multiple fatalities are only counted once in this table. Table 8. Offsets to roadside hazards in fatalities.

Street Furniture General information. In many urban areas, the use of street furniture is a common approach to improving the aes- thetic quality of a street. Street furniture includes items placed adjacent to the road that are there to improve the adjacent land use or to improve the transportation operations. In some jurisdictions, street lights and signs are included in the cate- gory of street furniture; however, for the purposes of this review, street furniture is considered to be supplemental items such as benches, public art, trash receptacles, phone booths, planters, bollards, fountains, kiosks, transit shelters, bicycle stands, and so forth. Often the placement of these de- vices can obscure sight distance, so their location should not occur in the sight triangles of intersections or driveways. Many street furniture items are placed along the right-of-way by property owners, as in the case of the placement of a side- walk cafe in front of a restaurant, and are thus largely outside the engineer’s control. Transit shelters are provided to protect transit riders from inclement weather and must be located close to the curb to facilitate short bus dwell times. One interesting way that some jurisdictions manage the placement of street furniture by land owners is by a permit- ting process that requires vendors or property owners to acquire liability insurance for street furniture located adja- cent to the road. Seattle is one city with this insurance requirement. Other jurisdictions restrict the placement of private street furniture by a permitting process comple- mented by a tax for each square meter or square foot of pub- lic right-of-way used. Dublin, Ireland, is one city that uses this approach. Safety research. The research team for this project was not able to locate any research evaluating the relative hazard that may be posed by street furniture. Nevertheless, street furniture can potentially create sight distance obstructions when located near an intersection, particularly when large numbers of people congregate as a result of the street furni- ture. It is also important that the sight distance of pedestrians be maintained when placing street furniture proximate to the roadway. Strategy summary. Common street furniture safety strate- gies are as follows: Purpose Strategy Minimize likelihood of • Locate street furniture as far crash from street as possible (P) • Restrict street furniture placement to avoid sight distance issues for road user (P) Minimize crash severity Develop street furniture that meets basic crashworthy standards (E) Utility Poles, Luminaires, and Signposts/Hardware General information. In both the national and interna- tional literature, the placement of utility poles, light poles, and similar vertical roadside treatments and companion hard- ware is frequently cited as an urban roadside hazard. For example, Haworth and Bowland have noted that while im- pacts with trees were more common outside the Melbourne, Australia, metropolitan area, single-vehicle crashes with poles or posts were more common within the metropolitan region (73). A 1998 study by the European Transport Safety Council identified collisions with utility poles or posts as one of the top two roadside hazards for Finland, Germany, Great Britain, and Sweden (74). Safety research. This section discusses the research litera- ture for utility poles, lighting supports, and signposts/roadside hardware. • Utility poles. Utility poles are prevalent in urban environ- ments and can pose a substantial hazard to errant vehicles and motorists. A study of utility pole crashes, for example, found that crash frequency increases with daily traffic volume and the number of poles adjacent to the travel- way (33). Utility poles are more prevalent adjacent to urban roadways than rural highways, and demands for opera- tional improvements coupled with limited street right-of- way often lead to placing these poles proximate to the roadway edge (see Figure 9). The absolute number of fatalities related to utility poles has remained around 1,200 since the mid 1980s (see Figure 10) (75). Utility poles are involved in the sec- ond highest number of fixed-object fatalities (trees are involved in the highest number of fixed-object fatalities). In a statistical study performed by Washington State Department of Transportation researchers, utility poles were identified as one of the roadside features that can significantly affect the injury level or severity of run-off- road crashes (68). The literature regarding crashes related to utility poles has identified utility poles as being principally an urban hazard, with urban areas experiencing 36.9 pole crashes per 161 km (100 mi) of roadway, compared with 5.2 pole crashes per 161 km (100 mi) of roadway for rural areas (76). Zegeer and Parker found that the variable that had the greatest ability to explain crashes related to utility poles was the average daily traffic (ADT) along the roadway (33). The significance of ADT as the critical variable explains the importance of vehicle exposure to understanding run-off- road crashes with utility poles. A common recommendation for addressing the utility pole safety issue is to place utilities underground and thereby remove the hazardous poles. The removal of all 25

poles in the urban roadside environment is not practical as these poles often function as the supports for street lights and other shared utilities. There are, however, several known utility pole hazardous locations that should be avoided when feasible. In general, utility poles should be placed in the following locations: – As far as possible from the active travel lanes, – Away from access points where the pole may restrict sight distance, – Inside a sharp horizontal curve (as errant vehicles tend to continue straight toward the outside of curves), and – On only one side of the road (66, 77). State of the Art Report 9: Utilities and Roadside Safety summarizes categories for utility pole safety solutions as the following: changing the pole position; using safety de- vices (crash cushions, safety poles, guardrail, and barriers); or warning motorists of the obstacles (78). The report in- cludes several initiatives in which pole relocation or re- moval is currently targeted as a safety strategy. Additional ways to minimize utility pole crashes are plac- ing utilities underground (where feasible), using shared poles to reduce pole density, and relocating poles to less vul- nerable locations. The delineation of poles using reflective tape or buttons may also help an alert driver identify a util- ity pole and avoid it; however, this delineation treatment may also act as an attractor for impaired drivers who are at- tempting to guide their vehicles by road edge delineation. The Land Transport Safety Authority in New Zealand rec- ommends that utility poles and large trees be highlighted using a uniform method that cannot be removed, such as reflectorized markers or paint markings (79). Increasing the lateral distance of utility poles from the travel lanes appears to be a promising improvement strategy. Many jurisdictions maintain an operational offset of 0.5 m (1.5 ft), but several agencies are seeking to increase the pole placement offset in urban regions. Haworth and colleagues (80) observed that in metropolitan Melbourne, Australia, poles involved in fatal crashes were most often less than 2 m (6.6 ft) from the edge of the road. The Clear Roadside Committee established by the Georgia Utilities Coordinat- ing Council suggests that for curbed sections, poles should be placed as far as is practical from the face of the outer curbs, with the following goals: – Lateral clearance of 3.6 m (12 ft) from the face of the curb to the face of the pole. – For speed limits greater than 56 km/h (35 mph) but not exceeding 72 km/h (45 mph), a lateral clearance of 2.4 m (8 ft). 26 Photos by NCHRP Project 16-04 research team. Figure 9. Road widening resulting in utility pole hazards. Graphic reprinted from State of the Art Report 9: Utilities and Roadside Safety (75). Figure 10. Fatalities related to crashes with utility poles (1980–2000).

– For roads with posted speed limits less than or equal to 56 km/h (35 mph), a lateral clearance of 1.8 (6 ft) (81). Similar to the Georgia policy, the Maine Utility Pole Lo- cation Policy suggests that offsets should be greater than 2.4 m (8 ft) for roadways with posted speed limits of 40 to 55 km/h (25 to 35 mph) and that offsets should be greater than 4.3 m (14 ft) on roadways with posted speed limits of 65 to 70 km/h (40 to 45 mph) (82). In Sweden, emphasis is placed on system level improve- ments. Based on the idea that the transportation system itself is unsafe, Sweden is redeveloping the system to re- duce user errors that lead to injury or death. One of the strategies to achieve this goal is to modify the system to ensure that users are not exposed to impact forces that can kill or severely injure the users (71). Finally, some utility pole literature suggests the use of breakaway poles. In the event that an errant vehicle impacts a breakaway pole, the pole will swing upward and then back down (thereby permitting the impacting vehicle to travel safely under the pole). One concern with breakaway utility poles has been whether they pose a threat to pedestrians when they swing back down after swinging up to avoid a crash. A 1970s series of case studies performed in Australia by McLean, Offler, and Sandow evaluated crashes in the proximity of Stobie poles (utility poles with two rolled-steel joists separated by concrete) (83). In evaluating the risk to pedestrians, the researchers determined that there were no cases in their studies where a pedestrian was in the imme- diate vicinity of a collision between a car and a Stobie pole. • Lighting and visibility. The design of luminaire posts is di- rected by NCHRP Report 350, and substantial research has been devoted to designing these light poles to be yielding upon impact (using breakaway bolts) (36). Multiple designs for these posts are included in the current edition of the Roadside Design Guide (1), and specifications for evaluat- ing these features are contained in AASHTO’s Standard Specifications for Structural Supports for Highway Signs, Luminaires, and Traffic Signals (84). An important issue in addressing roadside safety is the role of lighting in making potentially hazardous roadside environments visible to the road users (motor vehicle driv- ers, bicyclists, and pedestrians), particularly during night- time hours. Other issues with roadside lighting, as cited in the literature, include frequency and spacing of lights (56) and lighting color and associated visibility (85). These is- sues are beyond the scope of this study. • Signposts and roadside hardware. The design of signposts is directed by NCHRP Report 350, and there has been substan- tial research devoted to designing these features to be tra- versable (36). Multiple designs for these features are included in the current edition of the Roadside Design Guide (1), and specifications for evaluating these features are contained in AASHTO’s Standard Specifications for Structural Supports for Highway Signs, Luminaires, and Traffic Signals (84). Crash severity can be minimized through the use of tra- versable hardware, such as break-away light posts, mail- boxes, and utility poles. The current standard for break- away hardware, as contained in the Roadside Design Guide and NCHRP Report 350, is that breakaway features func- tion omni-directionally to ensure that the features do not constitute a hazard from any impact direction (1, 36). To prevent vehicle snags, the stub height after breakaway should not exceed 100 mm (4 in.) (see Figure 11). While an unobstructed and traversable roadside is pre- ferred, it may be necessary at some locations to use break- away features, which will minimize the severity of the initial impact by an errant vehicle. Breakaway poles and similar features must be designed to prevent intrusion on the pas- senger compartment of the vehicle, either by minimizing the weight and load of such features, or by providing a sec- ondary hinge, at least 2.1 m (7 ft) above the ground, that permits the vehicle to pass safely beneath the post upon im- pact. The current edition of the Roadside Design Guide pro- vides specifications for these devices and suggests that the concern for pedestrians in urban areas has led to a trend of using fixed supports for some urban locations (1). 27 Graphic reprinted from Roadside Design Guide (1) with permission. Figure 11. Stub of a breakaway support.

Strategy summary. Common strategies for utility poles, lighting supports, and signposts/roadside hardware are as follows: Purpose Strategy Treat individual high • Remove or relocate poles (P) risk pole locations • Place poles on inside of horizontal curves and avoid placement on outside of roudabouts or too close to intersection corners (P) • Use breakaway or yielding poles (T) • Shield poles (P) • Improve pole visibility (E) Treat multiple poles in • Establish urban clear zone offset high risk locations guidelines for pole setback distances from curb (P) • Place utilities underground while maintaining appropriate nighttime visibility (P) • Combine utilities/signs onto shared poles (reduce number of poles) (P) • Replace poles with building- mounted suspended lighting (where suitable) (E) (86) Minimize level of Reduce travel speed on adjacent severity road (P) Safety Barriers Roadside barriers are subject to NCHRP Report 350 testing criteria (36). There are several types of safety barriers that may be present in an urban environment. These include the following: • Barriers (flexible, semi-rigid, and rigid); • Bridge railings; and • End treatments (crash cushions and end terminals). Generally, most of the research on safety barriers has been oriented toward the design of barriers and their placement to shield vehicles from hazardous roadside conditions. The Roadside Design Guide and NCHRP Report 350 provide con- siderable information on placement and design of safe bar- rier systems (1, 36). FHWA maintains a roadside hardware website that provides information about specific roadside hardware that has been tested (see http://safety.fhwa.dot.gov/ roadway_dept/road_hardware/index.htm). In the urban environment, many of the safety barriers common to rural environments may not be suitable due to constraints regarding space available for flared end treatments, the constraining influence of safety barriers on pedestrian activity, and the potential obstruction of sight distance at the many intersections and driveways in the urban environment. Also, in locations with bicycle activity, safety barriers located immediately adjacent to the road may expose cyclists to un- necessary risks because the barriers may give a sensation of “squashing” the cyclist between the barrier and an adjacent motor vehicle (79). Safety research. Considerable research has been per- formed on a variety of traffic barriers. NCHRP Report 490: In- Service Performance of Traffic Barriers includes an extensive literature review that discusses the evolution of traffic barrier crashworthiness (87). The following summaries briefly re- view the application of these tested barriers (barriers, bridge rails, and end treatments) in an urban environment. • Barriers. Barriers can be categorized as flexible (cable barri- ers and W-beam guardrail with weak post), semi-rigid (thrie beam and W-beam guardrail with strong post), and rigid (concrete barrier system such as the New Jersey barrier). Because guardrails are most typically associated with rural and higher speed environments, the use of guardrails in urban environments is often restricted to protection of bridge approaches and departures. In fact, the conventional use of guardrail is to shield roadside objects from impact that pose a greater threat than impact to the guardrail it- self. Since the placement of guardrails at locations with frequent driveways is problematic due to the numerous breaks in the barrier treatment and the adverse effect of the guardrail on driveway and intersection sight distance, the use of conventional guardrail is minimal in urban low-speed corridors. AASHTO indicates that for very low traffic volume locations, traffic barriers are not generally cost-effective (32). This recommendation is confirmed by research performed by Stephens (88) and Wolford and Sicking (89). In the event that an engineer does endorse the use of a barrier in an urban environment, factors in addi- tion to the lateral offset, deflection distance, terrain effects, flare rate, and length of need (common to rural placement design) must be supplemented by consideration of corner sight distance, pedestrian activity (with particular atten- tion to the needs of persons with disabilities), and bicycle activity (1). When the use of a protective barrier is warranted in the urban environment, the application of aesthetic barrier treat- ments may be considered. These treatments perform the same general function as a guardrail (shield hazardous envi- ronments) while enhancing the aesthetics of a roadway. The use of barriers in an urban environment can be to shield roadside obstacles (such as rigid utility poles), sepa- rate motorized and nonmotorized traffic, and provide a physical separation between the active travel lanes and pe- destrian activity. For barriers with a shielding objective, sev- eral different barriers may be considered. For example, the 28

California Department of Transportation published a report in 2002 with a focus on suitable aesthetic barriers (90). This report includes the status of crash testing as well as the advantages and disadvantages of each treatment. Candidate barriers for urban environments include concrete barriers with textured and patterned surfaces, timber guardrail, pre- cast concrete guardwall, and stone masonry guardwall. The placement of barrier in the vicinity of a median is a common strategy in rural environments to help prevent head-on collisions by errant vehicles. Washington State, for example, performed an evaluation of median treatments for multi-lane, divided state highways with full access con- trol (91). The researchers determined that the placement of a barrier for medians up to a width of 15 m (50 ft) is cost- effective. In an urban environment, the median width is nar- row and often serves the combined functions of separating opposing directions of travel and acting as pedestrian refuge at certain locations. Currently, concrete Jersey barriers are used most commonly for medians in urban locations. The placement of barriers adjacent to the road intro- duces a new roadside hazard. For example, use of rigid bar- riers tends to result in a greater number of minor crashes, but dramatically reduces the number of serious or fatal head-on and run-off-road crashes (92). Lee and Manner- ing (68) determined that for urban environments, guard- rails are significantly associated with an increase in crash frequency, but the severity of these crashes is likely to re- sult in possible injury only. In locations where aesthetics are important and a bar- rier is required, jurisdictions may develop crash-tested options such as the Vermont-approved stone masonry system shown in Figure 12 (in addition to the various aes- thetic barriers identified in the California report previ- ously indicated [90]). • Bridge rails. In both urban and rural environments, bridges should be equipped with rails that do not permit vehicles to penetrate the space beyond the rail (i.e., structural ade- quacy). Transitions to guardrails must be located at both the approach and departure end of all bridge rails. Bridge rails must be designed to retain a large passenger car at the legal driving speed for local streets and roads (94). The bridge rails must, therefore, be structurally designed and maintain their structural integrity after impact. • End treatments. For locations where the end of the barri- ers cannot be adequately flared or protected, it is necessary to use an end treatment such as a barrier terminal or crash cushion. Aesthetic enhancements to end treatments have not received much attention, so conventional treatments are necessary in the urban environment. These treatments should not allow a vehicle to penetrate, vault, or roll upon impact. They should have the strength and redirectional qualities of a standard barrier. Dynamic Roadside Conditions Bicycle Facilities General information. Bicycle facilities consist of road and roadside features intended for bicycle operation. These facilities may include standard lanes, wide outside lanes, bicycle lanes, and off-road bicycle paths. Accompanying bicy- cle facilities may be bicycle hardware located along the road- side, such as bicycle racks. In general, the literature regarding the relationship between bicycle facilities and roadside safety is limited. Wide shoulders and bicycle lanes provide an addi- tional “clear” area adjacent to the travelway, so these features could potentially provide a secondary safety benefit for motorists, provided bicycle volumes are low. These bicycle facilities will also further separate the motor vehicle from any roadside obstructions and improve the resulting sight distance for motor vehicle drivers at intersecting driveways and streets. A second area of consideration is the placement of bicycle- supportive hardware, such as bicycle racks, adjacent to the travelway. Bicycle racks are commonly made of steel or other metals and are typically bolted to the ground to secure locked bicycles from potential theft. These features are not designed to be yielding should a run-off-road event occur. To date, there has been little evaluation of the potential roadside hazard posed by such treatments, although they can clearly present a poten- tial fixed-object hazard. Making such features yielding would potentially minimize the core function of these features— providing a secure location for locking up bicycles. Thus, a potentially more desirable alternative may be to encourage the placement of these features outside of the clear zone. Safety research. Most of the bicycle research focuses on specific bicycle safety issues such as safety helmets and train- ing. Several studies have developed and reviewed previous bicycle suitability or compatibility criteria (95, 96). The crite- ria for determining bicycle suitability include available lane width, traffic volume, and vehicle speeds. One best practices 29 Photo reprinted from Guardrail Study (93). Figure 12. Stone masonry barrier.

review has suggested that road safety for cyclists could be en- hanced by increased law enforcement to ensure that cyclists do not ride in the wrong direction (against traffic) or at night without adequate lighting (97). This best practices review also noted that the use of extruded curbs to separate a bike lane from traffic should be avoided. Since the focus of this research effort is roadside safety for the urban environment, it is helpful to understand the mag- nitude of the safety risk to cyclists as they encounter roadside environments. One FHWA report using hospital emergency department data noted that 70 percent of reported bicycle injury events did not involve a motor vehicle and 31 percent occurred in non-roadway locations. For bicycle-only crashes, a total of 23.3 percent of the recorded crashes occurred at sidewalk, driveway, yard, or parking lot locations (98). Stutts and Hunter (99) evaluated bicycle–motor vehicle crashes and determined that some factors associated with the crash were variables such as age, gender, impairment, and time of day. Roadside variables (sidewalk, parking lot, and driveways) were not statistically significant in their model. Strategy summary. Common strategies to improve bi- cycle safety as well as bicycle–motor vehicle interactions are as follows: Purpose Strategy Reduce likelihood • Use wider curb lanes (P) of crash • Increase bicycle enforcement (T) • Increase operational offsets (P) Reduce severity of crash Locate bicycle racks as far away from road as possible (T) Parking General Information. In many urban environments, lim- ited off-street parking often necessitates the use of on-street parking to address the needs of local businesses and stakehold- ers. As noted in the Green Book (37), cars typically park 150 to 305 mm (6 to 12 in.) from the curb and have a normal width of roughly 2.1 m (7 ft). Thus, approximately 2.4 m (8 ft) are needed to comfortably accommodate on-street parking. One common strategy in larger cities is to design wider outside park- ing lanes, such as 3 m (10 ft), and convert them to travel lanes during peak periods and anticipated high-volume conditions. On-street parking can potentially have mixed results on a roadway’s safety performance. On the one hand, these fea- tures narrow the effective width of the roadway and may result in speed reductions, thereby leading to a reduction in crash severity. Conversely, on-street parking may also lead to an increase in collisions associated with vehicles attempting to pull in or out of an on-street parking space. In addition to vehicle conflicts, on-street parking serves as a physical buffer between the motor vehicle path and pedestrian facilities. The added safety buffer provided to the pedestrian, however, can be countered by the pedestrian who elects to cross the street mid-block in areas not designated for pedestrian crossing. The parked vehicles may act as a shield to prevent proper sight distance for the drivers of adjacent motor vehicles, often resulting in new conflicts between motor vehicles and pedestrians stepping between parked cars. Similarly, there is an inherent conflict between the motor ve- hicle and drivers exiting or entering their parked vehicles on the traffic side of the roadway. An additional concern often cited regarding on-street parking is the effect it may have on reducing emergency serv- ices’ response rate. Often the narrowing effect of on-street parking can be compounded by illegal parking too close to critical locations such as intersections. The Local Govern- ment Commission released a publication called Emergency Response Traffic Calming and Traditional Neighborhood Streets (100) that suggests that the adverse effects of on-street parking can be mitigated by implementing the following strategies: • Placing a double set of driveways periodically (per fire department recommendation) to enable local access; • Placing alleys on short blocks across from each other; • Placing mailbox clusters, curb extensions, or similar treat- ments where residents will find it inappropriate to park; and • Enforcing parking criteria to minimize illegal parking. Finally, the severity of a roadside hazard constituted by a collision between a parked vehicle and a moving vehicle is minimal. Since on-street parking is generally parallel to the moving vehicles, the impact by a moving vehicle is likely to be a sideswipe crash. This is one of the less severe crash types. For locations with head-in or reverse-in parking, the crash severity likelihood is increased as the moving vehicle may impact a ve- hicle in reverse. Proper sight distance and separation of parked vehicles from the active travel lane (often by the use of a bulb- out at the intersection) will help minimize safety risks between moving and parked or parking vehicles. As indicated in the pre- vious section, curb extensions or bulb-outs can pose a hazard to bicyclists by forcing them into the active travel lane. In the event that the parking lanes are not occupied, the extension could also create a hazard for drivers unfamiliar with it. On-street parking is generally not considered appropriate for higher speed roads such as suburban to urban transitional arterials. Strategy summary. Common on-street parking strate- gies are as follows: Purpose Strategy Reduce likelihood of Restrict on-street parking to low- crash speed roads (P) Reduce crash severity Where parking is appropriate, use parallel parking rather than angular parking (P) 30

Sidewalks and Pedestrian Facilities General information. Sidewalks and pedestrian facili- ties, in general, do not pose a particular hazard to motorists. The safety concern about locating these facilities adjacent to the road is the risk to the pedestrians using the facilities. Pro- viding safe facilities for pedestrians is an obvious strategy for increasing pedestrian safety. While shared streets may be appropriate if vehicle speeds and volumes are kept extremely low (see Figure 13), for most roads in urban areas, vehicle speeds warrant the use of sidewalks (21, 101, 102). The Green Book (37) recommends the use of sidewalks on urban streets, with sidewalk widths ranging between 1.2 and 2.4 m (4 and 8 ft) depending on the roadway classification and nearby land use characteristics (see Table 9). Of the roughly 75,000 pedestrian-related crashes that occur each year, almost half occur while the pedestrian was at a non-intersection, on-roadway location (see Table 10). The conventional approach to examining pedestrian and bicycle safety is to examine crash records to determine the non-motorized user’s action prior to a crash event. Complete and accurate motor vehicle crash data for pedestrian crashes can be difficult to find. Stutts and Hunter evaluated police- reported pedestrian crashes and compared these to hospital emergency room records (99). They found that crash events that occurred in parking lots, driveways, and other off-road locations were reported less frequently than those occurring in the roadway. A 1999 FHWA study (98) also reviewed hos- pital emergency room records and determined that 11 percent of the pedestrian–motor vehicle events recorded occurred at roadside locations (sidewalk, parking lot, and driveway). Of the data for which pre-crash action is known, improper crossings were the largest single crash category, accounting for 20 percent of the total crashes. The categories that have the most direct relationship to the roadside fell into the four cate- gories most strongly related to the design of roadsides: jogging, walking with and against traffic, and miscellaneous activity. As shown in Table 11, approximately 13,600 crashes were classified as “darting into road” crashes, which is where a pedestrian rushes into the street and is struck by a motor vehicle. The majority of individuals involved in “darting- into-road” crashes were children between the ages of 5 and 9, who may have been using the street for play activities (103). As Whyte has indicated, the street is often the preferred play location for children, even when parks and other recreational amenities may be available (104). An additional feature of the roadside environment is a pedestrian buffer area. The pedestrian buffer is a physical distance separating the sidewalk and the vehicle travelway. Buffer areas typically serve a host of secondary purposes as well—they provide locations for on-street parking, transit stops, street lighting, and planting areas for landscape mate- rials, as well as a location for a host of street appurtenances, such as seating and trash receptacles. Buffer strips may be either planted or paved. The Green Book supports the use of buffer strips on urban arterials, collectors, and local streets (37). 31 Photo reprinted from walkinginfo.org (121). Figure 13. Low-speed shared street. Road Class Side of Street Specification Urban Arterials Both Border area (buffer plus sidewalk) should be a mi ni mu m of 2.4 m (8 ft) and preferably 3.6 m (12 ft) or mo re. Collector Both sides of street for access to schools, parks, and shopping. Both sides of streets desirable in residential areas. 1.2 m (4 ft) mi ni mu m in residential areas. 1.2 to 2.4 m (4 to 8 ft) in co mme rcial areas. Local Both sides of street for access to schools, parks, and shopping. Both sides of streets desirable in residential areas. 1.2 m (4 ft) mi ni mu m in residential areas. 1.2 to 2.4 m (4 to 8 ft) in co mme rcial areas, although additional width may be desirable if roadside appurtenances are present. Source: Developed from A Policy on Geometric Design of Highways and Streets, 4th ed. (37). Table 9. Green Book sidewalk specifications (37).

Safety research. Several studies have evaluated potential countermeasures to improve pedestrian safety. Generally, these studies were based on case studies, statistical models, or subjec- tive evaluation. Landis and colleagues (105) modeled several roadside walking environment variables to evaluate the pedes- trian’s perception of risk versus actual risk. The researchers expected that as the number of driveways increased they would observe a decrease in pedestrian safety, but this hypothesis was determined not to be statistically significant. They did find that motor vehicle volume and vehicle speeds were significant fac- tors in pedestrian safety. Corben and Duarte evaluated the high number of pedestrian injuries along Melbourne’s arterial roads and recommended the adoption of three practices: • Reduce traffic volumes, • Reduce road widths, and • Reduce vehicle speeds (106). Corben and Duarte further suggested that strategies for reducing the vehicle speed can include public awareness and enforcement campaigns; gateway treatments (such as road narrowing, changing pavement texture, and implementing roundabouts); and streetscape improvements (106). Gateway treatments are addressed in the traffic calming section of this document, which follows this section. Cottrell reviewed a variety of options for improving pedestrian safety in the State of Utah (107). Many of his recommendations are similar to those already reviewed; however, he also included improving sidewalk security and visibility with street lights as an impor- tant issue. Cottrell further noted that the failure to remove snow from sidewalks during winter conditions may result in pedestrians entering the street either to cross it or to walk along the cleared road. The document that to date most ex- haustively summarizes strategies for improving pedestrian safety is NCHRP Report 500: Guidance for Implementation of the AASHTO Strategic Highway Safety Plan—Volume 10: A Guide for Reducing Collisions Involving Pedestrians (123). This document includes methods for enhancing pedestrian safety in the road as well as adjacent to the road. The recom- mendations in the document are consistent with those of the research studies summarized in this review. Strategy summary. Common strategies for eliminating or minimizing motor vehicle–pedestrian crashes at roadside locations are as follows: Purpose Strategy Reduce motor vehicle- • Provide continuous pedestrian pedestrian crash facilities (P) likelihood at roadside • Install pedestrian refuge medians locations or channelized islands (see previous section on medians and islands) (P) • Offset pedestrian locations away from travelway with pedestrian buffers (P) • Physically separate pedestrians from travelway at high-risk locations (P) • Improve sight distance by removing objects that obscure driver or pedestrian visibility (T) • Maintain pedestrian facilities free of leaves, snow, or tree roots (T) • Improve visibility by installing illumination for nighttime conditions (T) • Enforcement and public awareness campaigns (T) Reduce severity of Reduce roadway design speed/ motor vehicle- operating speed in high pedestrian pedestrian crashes at volume locations (T) roadside locations Traffic Calming Applications—Gateway Treatments Although traffic calming applications have been prevalent throughout Europe for several decades, traffic calming is relatively new in the United States, first emerging in the late 1990s as a strategy for addressing community livability 32 Pedestrian Crash Type No. of Crashes Percent No Action 23,502 31.6 Darting into Road 13,594 18.3 Improper Crossing 15,344 20.6 Inattentive 521 0.7 Jogging 211 0.3 Pushing Vehicle 103 0.1 Walking with Traffic 1,746 2.3 Walking against Traffic 1,184 1.6 Playing, Working, etc. in Roadway 8,074 10.8 Other 5,964 8.0 Unknown 4,249 5.7 Total 74,492 100.0 Source: General Estimates System. (122). Table 11. Pedestrian crashes by type (2002). Pedestrian Location Number Percent Intersection - In Crosswalk 14,674 19.7 Intersection - On Roadway 15,319 20.6 Intersection - Other 1,391 1.9 Intersection - Unknown Location 810 1.1 Non-intersection - In Crosswalk 381 0.5 Non-intersection - On Roadway 35,785 48.0 Non-intersection - Other 3,518 4.7 Non-intersection - Unknown Location 173 0.2 In Crosswalk - Unknown if Intersection 16 0.0 Other Location 1,830 2.5 Unknown Location 595 0.8 Total 74,492 100.0 Source: General Estimates System (122). Table 10. Pedestrian location during a crash (2002).

concerns associated with high vehicle traffic volumes and speeds. The practice of traffic calming applications has sparked a substantial debate regarding the safety and appro- priateness of these applications, particularly when applied to roadways intended for higher volumes and/or higher operat- ing speeds, such as minor arterial roadways (23, 108, 109). Traffic calming applications come in a variety of forms, including partial and full road closures and alterations in the de- sign of intersections and curb lines. This report is specifically fo- cused on traffic calming strategies deployed at arterial transitions from higher speed rural conditions to locations with lower speed arterial characteristics. This transitional traffic calming strategy is known as a gateway and is reviewed in the following summary. General information. The traffic calming strategy known as a gateway is defined by Burden as “a physical or geometric landmark on an arterial street which indicates a change in environment from a major road to a lower speed residential or commercial district (110).” Burden goes on to suggest that gateways can be a combination of street narrowings, medians, signs, arches, roundabouts, or other features. The objective of a gateway treatment is to make it clear to a motorist that he or she is entering a different road environment that requires a reduction in speed. Drivers need a certain transitional speed zone with the explicit guidance and roadway features to inform and encour- age them to gradually slow down before they reach the urban residential area for a safe entry. A transitional speed zone can also help drivers to speed up within a certain timeframe when leaving an urban area. This transition area is extremely im- portant for drivers who are not familiar with the urban area. They rely on the roadway features to indicate changes in sur- roundings that require an adjustment in their driving speed and behavior. The gateway concept was presented in a 1998 paper by Greg Pates in which he depicted the region between a rural area and a fringe area (transitioning into urban use) as the gateway location where travel speed should be reduced and motorists should become more alert (111). Safety research. Little information exists on the safety performance of gateway treatments. At present, such treat- ments have not been subject to extensive crash testing, undoubtedly because of the large degree of variation in the design and materials used in the construction of such features. Nevertheless, as noted in Skene, such features are often used by Canadian and British transportation profes- sionals in speed transition zones to provide the driver with visual cues of a forthcoming change in safe operating conditions (108). In the United States, these gateway treat- ments are typically aimed at delineating the boundaries of specific communities. Most of the research regarding gateway treatments focuses on their influence on operating speed or road users’ percep- tions of gateway treatments and their understanding that they are, in fact, transitioning into a different and slower speed environment. In a 1997 study in the United Kingdom, researchers at the Transport Research Laboratory performed a before-after evaluation of a series of traffic calming strategies on major roads (112). The traffic volume on candidate roads was greater than 8,000 vehicles per day, and at least 10 percent of the traf- fic was composed of heavy vehicles. Speeds at inbound gate- ways were reduced at eight out of nine locations tested. Mean speed reductions ranged from 5 to 21 km/hr (3 to 13 mph). The study evaluated a variety of treatments, including speed reduction signage, narrowings, dragon teeth marking, speed cushions, colored pavement, and advanced signing. The re- searchers determined that the signing provided a high visual impact and resulted in large speed reductions. Physical de- vices such as the speed cushions resulted in a greater level of speed reduction than signing alone. The use of colored bands placed laterally across the road and placed in a series seemed to result in some speed reduction, but did not result in large decreases in speed. The researchers did not test the speed re- duction signs and so could not comment on the effectiveness of these devices. Dragon-teeth marking, identified as one of the strategies, is depicted as a schematic and a photograph in Figure 14. 33 Photo and graphic reprinted from Traffic Calming on Major Roads (112) under the terms of the Click-Use License. Figure 14. Dragon-teeth marking.

34 Berger and Linauer describe the operating speed influence of five raised island configurations used as gateway treatments from high-speed rural locations in Austrian villages (113). The five gateway island configurations are shown in Figure 15. The most dramatic influence on speeds occurred for Island Number 5, where the path of the approach lane was shifted dramatically. Table 12 demonstrates the range of speed re- ductions observed in this Austrian study, and Figure 16 shows the speed profile for Island Number 5. A study performed for traffic calming strategies deployed from 1993 to 1996 in Ireland evaluated the transition zone as the area between a high-speed and low-speed road (114). The researchers evaluated the gateway transition from rural to urban environments in two phases: • From the “Traffic Calming Ahead” sign to the “Do Not Pass” sign and • From the “Do Not Pass” sign to the gateway treatments in the form of raised islands. In the first phase, the researchers observed that the “Traf- fic Calming Ahead” sign at the beginning of the transition zone reduced inbound traffic speed. They determined this by comparing the speed reduction results in the transition zone with and without a “Traffic Calming Ahead” sign. With the “Traffic Calming Ahead” sign present, the 85th percentile speeds ranged between 90 and 100 km/hr (56 and 62 mph) at the start of the transition zone. The 85th percentile speeds at the “Do Not Pass” signs were reduced by 6 to 8 km/hr (4 to 5 mph). At locations without the traffic calming signs, the 85th percentile speeds were observed to be reduced by only 2 to 3 km/hr (1.2 to 1.9 mph) at the same approach location. In the second phase, the speed reduction analysis results indicated that the gateway with raised traffic islands was an effective traffic calming treatment. They found that speed reductions of approximately 14 km/hr (9 mph) relative to the speed recorded at “Do Not Pass” signs were achieved at gate- ways with raised islands compared with a reduction of only 10 km/hr (6 mph) at gateways without raised islands. A case study performed in Canada evaluated traffic calm- ing strategies on an arterial road connecting two residential areas (115). Mohawk Road is a two-lane road with a 50-km/hr (31-mph) speed limit and was originally designed to service rural conditions. Speeding is very common on this road, and the test data showed that about 67 percent of the vehicles were exceeding the speed limit at the test location. After evaluating several alternatives, the jurisdiction finally elected to implement a series of landscaped speed control medians of various dimensions to help slow traffic down. Based on a before-after site evaluation, the speed reduction at control sections (where the strategies were not deployed) ranged from 85 percent to 88 percent for initially observed speeds, while the speeding percentage reduction in test sections ranged from 47 percent to 67 percent of the original speeds. This 20-percent speeding reduction was determined to be statistically significant at a 99-percent confidence level. The researchers were not able to compare corresponding crash data for the site. Ewing reviewed a highway reconstruction project deployed in Saratoga Springs, New York (116). The case road transi- tioned from a four-lane, semi-rural highway with a flush, painted median and a speed limit of 88 km/hr (55 mph) to a three-lane urban road with a raised median and a posted speed limit of 48 km/hr (30 mph). The length of road avail- able for this transition was 550 m (1,800 ft). Because the road passes the Saratoga Spa State Park, the Lincoln Baths, and the Museum of Dance, local representatives wanted a gateway for the transition of the road. The three photos shown in Fig- ure 17 depict the gateway transition ultimately constructed for this facility. Roundabouts are another commonly recommended gate- way treatment. Pates has discussed how Norwegian trial projects using roundabouts experienced average speed re- ductions of 10 km/hr (6 mph) (111). Ewing (23) and Zein and Montufar (92) identify roundabouts as safe traffic calm- ing alternatives to conventional intersections that can serve as both psychological and physical indicators of a transition from a rural high-speed environment to the lower speed urban street. Ewing also indicates that the center islands of the roundabouts can be landscaped and possibly include sculptures or monuments. Although the research team was unable to locate published research regarding the use of street art in roundabout medians, they did speak with re- searchers from both the United Kingdom and Australia. Representatives from both countries suggested that the ap- plication of street art in roundabouts is generally hazardous if these items are placed in the center of the first roundabout encountered by the driver on a rural road. The use of a series of roundabouts as a transition, with the street art located in the subsequent roundabouts, however, is a common prac- tice and appears to be a safe strategy for these transitional regions. The evaluation of gateway treatments is a new area of re- search for transportation and, as a result, very little is known about crashworthiness issues. As reviewed in this summary, the focus in using gateway treatments has been on the re- sulting speed reduction (thereby reducing ultimate crash severity).

35 Graphic reprinted from “Raised Traffic Islands at City Limits—Their Effect on Speed” (113). Figure 15. Gateway islands used in an Austrian study. Island Number Speed 1 2 3 4 5 Previous 54.0 58.0 60.0 65.0 65.0 Vmean (km/h) Subsequent 54.1 48.4 44.1 47.2 40.1 Previous 62.0 67.0 70.0 76.0 77.0 V85 (km/h) Subsequent 61.0 54.5 50.5 55.2 44.6 Previous 70.0 88.0 86.0 95.0 97.0 Vmax (km/h) Subsequent 76.2 59.3 56.1 65.8 46.9 Source: Adapted from “Raised Traffic Islands at City Limits—Their Effect on Speed” (113). Table 12. Gateway median speed results from Austrian study (113). Graphic reprinted from “Raised Traffic Islands at City Limits—Their Effect on Speed” (113). Figure 16. Island Number 5 speed profile.

Strategy summary. Common traffic calming gateway strategies are as follows: Purpose Strategy Reduce likelihood of Apply speed reduction signs, run-off-road crash pavement markings, and other gateway treatments (T) Reduce severity of run- • Construct gateway raised median off-road crash treatments (T) • Construct roundabouts with traversable island centers in initial islands (T) Literature Review Conclusion This review has examined the knowledge and practice of roadside safety as they relate to the design of roadsides in urban areas. A brief review of roadside crash statistics demon- strates that roads in the urban environment, although gen- erally lower speed facilities than their rural counterparts, nevertheless suffer from run-off-road collisions with roadside hazards. The operational offset concept for curbed urban roadways should, therefore, not be mistaken as a safety stan- dard, and urban setback policies should be considered for future adoption. Although they are not the focus of this research, the strate- gies used in urban environments to help prevent errant vehicles from leaving the travelway and encountering a fixed roadside object are briefly reviewed in this report. Finally, this review has summarized common urban road- side objects and known safety issues associated with these roadside objects. A collection of safety strategies (experi- mental, tried, and proven) for each set of roadside hazards is included. In general, this review makes it clear that the safety impli- cations of many roadside features (such as curbs, signs, and utility poles) are well understood; however, there are signifi- cant gaps in understanding how many urban roadside features (such as landscape buffers, trees, and lighting) should be addressed for safe urban roadside development. 36 Photos courtesy of Reid Ewing © (116) Figure 17. Example of gateway transition in Saratoga Springs, NY.

Next: Chapter 3 - Findings and Applications »
Safe and Aesthetic Design of Urban Roadside Treatments Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB's National Cooperative Highway Research Program (NCHRP) Report 612: Safe and Aesthetic Design of Urban Roadside Treatments explores recommended design guidelines for safe and aesthetic roadside treatments in urban areas. The report also examines a toolbox of roadside treatments designed to balance pedestrian, bicyclist, and motorist safety and mobility.

NCHRP Report 612 includes four appendices, three of which are available online. The fourth, Appendix C, is included with the report.

Appendix A includes detailed information about the urban control zone corridor sites.

Appendix B includes the summary statistics for the report's case study sites.

Appendix C includes an urban roadside design toolbox, and

Appendix D provides draft language for the urban chapter in the AASHTO Roadside Design Guide.

Tables 19 and 20 on p. 43 of NCHRP Report 612 include incorrect information. The corrected tables are available online.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!