Guidelines for Nighttime Visibility of Overhead Signs (2016)

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.

5 Background Traffic control devices (TCDs) must be visible at an appro­ priate distance for drivers to respond to them properly. With a lack of natural light at night, other light sources must be used to provide the luminance necessary for drivers to see TCDs. Guide and street name signs placed overhead or mounted on the shoulder tend to be manufactured with retroreflective sheeting, which, upon illumination from a vehicle’s head­ lamps, return some of the light back to the driver. When retro­ reflective sheeting is not used, the signs must be internally or externally illuminated, as directed in the MUTCD (4). There are several factors that influence the visibility of guide signs. This chapter summarizes previous research and other sources of information regarding issues and complexities of guide sign visibility at night. Included are results of a survey of state departments of transportation (DOTs) that specifically inves­ tigated sign lighting policies and practices. Luminance Requirements for Sign Visibility Light within the visible spectrum is necessary for the human eye to perceive objects. Objects that do not independently emit light must be illuminated in order to be seen, and illu- minance describes the intensity of the light incident on the surface. Luminance describes the intensity of light reflected at the surface in the direction of a viewer. As luminance is the light from the perspective of a driver, several studies have investigated the ability of drivers to see objects or road fea­ tures based on luminance. Early laboratory research employed practices similar to a common eye exam. In 1977, Richards used a static vision test­ ing method by seating subjects in front of an eye chart (5). By applying four lighting levels from a projector calibrated to simulate a vehicle’s headlamp, Richards provided luminance at levels from 0.03 to 34 cd/m2. Not only did acuity decrease with age and luminance, but the acuities at each luminance value decreased with letter contrast. Schnell et al. presented subjects with an image of a 2­in. sym­ bolic sign, instructing them to walk toward the screen until the symbol was identifiable (6). Luminance was measured from the front of the screen, and the researchers concluded that 82 cd/m2 was the maximum background luminance beyond which there was no improvement in detecting the symbol. Interactions between color and luminance and their influ­ ence on sign recognition have also been studied (7, 8). Forbes determined that signs with greater luminance require shorter subject glances (7). Padmos found that color recognition occurs at lower luminance levels than legibility (8). Early color rec­ ognition helps drivers detect and comprehend the message of a traffic sign earlier because the color is associated with the sign’s meaning. Carlson and Hawkins studied the effects of luminance on the legibility distance of overhead guide signs (1). They varied the luminous intensity of a test vehicle’s headlamps at 32 differ­ ent levels while study participants read the signs at distances corresponding to specific legibility indices. Figure 1 shows the cumulative distributions of correct readings by legibility index and luminance. The findings were used to develop sign retro­ reflectivity requirements for overhead signs based on providing a minimum amount of luminance so that 50 percent of elderly drivers have a 40­ft/in. legibility index. The corresponding luminance is 2.3 cd/m2. Figure 1 illustrates a great amount of diversity in the visual abilities of drivers. In Carlson and Hawkins’s study, the 10 per­ cent of elderly drivers with the best vision needed less than 1 cd/m2 to correctly read the overhead sign at a 40­ft/in. leg­ ibility index (1). The 10 percent of elderly drivers with the poorest vision needed more than 10 cd/m2 of luminance to correctly read the sign at the same location. Luminance of approximately 30 cd/m2 was needed to reach 100 percent correct responses at a distance corresponding to a 40­ft/in. legibility index. The amount of luminance necessary to cor­ rectly read the overhead sign decreased at closer distances (i.e., a legibility index of 20 or 30 ft/in. instead of 40 ft/in.). C H A P T E R 2

6The 2.3 cd/m2 luminance value that met the needs of 50 per­ cent of older drivers at a 40­ft/in. index was adequate for approximately 80 percent of the drivers at a 30­ft/in. index and 100 percent of the drivers at a 20­ft/in. index. The luminance and legibility data presented in Figure 1 were collected in a rural environment with no distracting objects or glare sources. Follow­up research revisited the issue of luminance required to correctly read overhead guide signs but increased the complexity of the visual background by including roadway lighting and glare sources (9). The study used signs of the following color combinations: white on green, white on blue, and white on brown. By including additional light sources, the research identified the lumi­ nance needed for nighttime legibility under four different environments: rural/dark, rural/dark with roadway lighting, rural/dark with glare, and rural/dark with roadway lighting and glare. When glare was added to the rural/dark condi­ tions, the amount of luminance needed to correctly read the signs nearly doubled. Roadway lighting added to the glare condition countered the impact of the glare, and only 15 percent more luminance was needed to achieve the same legibility. Findings were mixed when roadway lighting was added without the glare sources. The previous findings indicate that increased luminance results in increased distance at which drivers can read signs. The results are limited, however, because they primarily rep­ resent the experience of drivers in dark and rural conditions and under low workload. The research by Holick and Carlson suggested that drivers require more luminance to view signs as more light sources are added to a scene, but there is a lack of information about how light and driving scenarios that are more complex than rural conditions interact to affect the detectability and legibility of signs (9). Lighting Sources The luminance to read traffic signs at night can come from lights added to the signs or from vehicle headlamps. As men­ tioned, the MUTCD requires that signs without retro reflective sheeting be illuminated by additional sign lighting, and some agencies light their signs even if the signs are retroreflec tive. This section discusses policies, guidelines, and practices related to sign lighting and vehicle headlamps. Sign Lighting The consistent illumination provided by permanent sign lighting (whether external or internal) facilitates the rapid and accurate recognition and understanding of a sign’s message at night. This is especially helpful in situations with high traf­ fic volume, complex design, adverse weather, and increased ambient luminance. The additional lighting may even be nec­ essary if the retroreflective sheeting is not efficient enough for sign legibility or when recognition and legibility distances need to be increased. This section presents some guidelines associated with providing sign lighting and trends in sign light­ ing found among transportation agencies. Guidelines for Sign Lighting The MUTCD contains several statements concerning the illumination and visibility of signs, especially for nighttime conditions. Section 2A.07 states the following: Regulatory, warning, and guide signs and object markers shall be retroreflective or illuminated to show the same shape and similar color by both day and night . . . and The uniformity of the sign design shall be maintained without any decrease in visibility, legibility, or driver comprehension during either daytime or nighttime conditions. While the MUTCD includes minimum maintained sign retroreflectivity levels (Section 2A.08), it contains no specifi­ cations for the amount of lighting needed when signs are not retroreflective. Additionally, the retroreflectivity levels in the MUTCD are considered minimums, which were established based on driver visibility needs in dark/rural conditions. There is no information about when retroreflective sheeting alone and illuminated by headlamps does not provide a high level of visibility and legibility for other conditions. It is possible that areas of high complexity or greater ambient luminance would reduce sign visibility and legibility. MUTCD Section 2E, which discusses guide signs on free­ ways and expressways, states that the legends should be retro­ Figure 1. Cumulative percent of correct responses for three legibility indices (LIs) when varying sign luminance (1).

7 reflective and the backgrounds that are not independently illuminated should be retroreflective. Additionally, it states that the illumination provided by low­beam headlamps is relatively small. Such information supports the following guidance in Section 2E.06: Overhead sign installations should be illuminated unless an engi- neering study shows that retroreflectorization alone will perform effectively. The type of illumination chosen should provide effective and reasonably uniform illumination of the sign face and message. Though not directly related to sign illumination, the MUTCD covers situations of inadequate sight distance and unique roadway geometries by including guidance for advance street name signs for conventional roadways (placed at a distance appropriate for a driver to properly decelerate and turn) and pull­through signs for freeways and expressways. On free­ ways, there are requirements for guide signs to be repeated on approaches to interchanges (they must be placed 1 mi in advance of and at the theoretical gore and are recommended at 2­mi and 0.5­mi locations). The redundancy of these types of signs is one way to address the limitations of complex situa­ tions and undesirable geometrics or line of sight obstructions. Despite the amount of detail in the MUTCD indicating where signs are to be placed, what information they must con­ tain, and how that information is to be relayed to the driver, there are few specifics that describe how to achieve high vis- ibility and high legibility for guide signs other than stating that the legends must be retroreflective and the whole sign should be illuminated unless an engineering study indicates that illu­ mination is unnecessary. There is no information about how to determine whether lighting is needed and the amount of lighting to provide. Some gaps in the lighting guidelines of the MUTCD are filled by the AASHTO Roadway Lighting Design Guide of 2005 (3). Section 10.2 states that retroreflective signing materials by themselves (without sign lighting) may perform adequately if “the sign is in an area that contains a low­to­intermediate ambient light level, and there is at least 1,200 ft (366 m) or more of tangent sight distance in advance of the overhead sign.” Additionally, the design guide states that “high levels of ambient luminance may make sign lighting warranted regard­ less of the retroreflective properties of the sign face material.” Background ambient luminance is divided into three classifi­ cations and described qualitatively, as shown in Table 1. The AASHTO guidelines also specify the amount of lighting, in terms of both illuminance and luminance, to provide based on the level of background ambient luminance. These levels are shown in Table 2. IESNA is another group that provides sign lighting guide­ lines and recommendations. IESNA guidelines identify five factors that should be considered when evaluating the legi­ bility of guide signs (2): 1. Ambient luminance 2. Sign luminance above ambient luminance 3. Retroreflectivity of sign legend and background materials 4. Contrast between sign legend and background 5. Uniform ratio of sign lighting Level of Ambient Luminance Description Low Low levels of ambient luminance exist in rural areas without roadway and/or intersection lighting. Objects at night are visible only in bright moonlight. There is very little or no other lighting in the area. Medium Medium levels of ambient luminance exist in intermediate areas with some roadway and/or intersection lighting. May contain small areas of commercial lighting. High High levels of ambient luminance exist in urban areas with high levels of roadway lighting. May contain brightly lighted commercial advertising signs, building facades, and/or highly illuminated parking facilities. Table 1. Ambient luminance descriptions (3). Ambient Light Level Sign Illuminance Sign Luminancea fc lx cd/m2 cd/ft2 Low 10–20 100–200 22–44 2.2–4.4 Medium 20–40 200–400 44–89 4.4–8.9 High 40–80 400–800 89–178 8.9-17.8 Note: Adapted from Table 10-1 in Roadway Lighting Design Guide, 2005, by the American Association of State Highway and Transportation Officials. a Based on a maintained reflectance of 70 percent for white sign letters. Table 2. AASHTO recommended sign lighting levels (3).

8Table 3 shows the recommended IESNA lighting levels, based on the same three ambient lighting classifications. Based on the MUTCD guidelines and those in Tables 2 and 3, there appears to be a general consensus that sign light­ ing is not needed in rural areas as long as the retroreflective sheeting meets the MUTCD minimum standards. There is ambiguity, however, regarding the use of sign lighting in urban areas with medium or high levels of ambient lumi­ nance or for unique conditions, such as unusual geometrics or areas of frequent dew, fog, or frost. The MUTCD suggests that lighting may be appropriate for some conditions, but practitioners may have difficulty determining whether or not lighting should be used because of the subjectivity in speci­ fying the level of ambient luminance or other appropriate conditions. An example of research that has provided specific recommendations was produced for the Florida Department of Transportation and suggests that sign lighting be used for overhead signs on curves in urban areas when the curve radius is shorter than 2,500 ft (10). Beyond the question of whether or not sign lighting is needed, there is also an interesting conflict between the guidelines of the MUTCD and those adopted by AASHTO and IESNA. For areas with low ambient luminance, AASHTO and IESNA recommend sign luminance levels in the range of 20 to 44 cd/m2, as shown in Tables 2 and 3. The mini­ mum maintained retro reflectivity levels in the MUTCD were derived from human factor studies performed in a dark, rural setting (low ambient luminance) (1). It was found that luminance of 2.3 cd/m2 was sufficient for half of older drivers to correctly read overhead guide signs at an index of 40 ft for each inch of legend letter height. Beyond 20 cd/m2 would have met the needs for nearly all of the older drivers, suggesting that a guideline of 20–44 cd/m2 is too conservative for a rural setting. Additional lighting and glare sources were added in a follow­up study to represent conditions closer to those of the medium level of ambient light shown in Tables 2 and 3 (9). Under these conditions, the required luminance was near 10 cd/m2—still much lower than the AASHTO and IESNA recommended range of 40 to 89 cd/m2. Another ambiguity comes from the use of both illuminance (the measure of light reaching the sign) and luminance (the measure of light reflected back to the driver) in Tables 2 and 3. Each table has luminance calculated from a constant propor­ tion of illuminance. The reflective efficiencies used are 70 and 45 percent. It is not clear whether these values are representa­ tive of modern sign sheeting products, where retroreflectivity varies by the angles of light incident on and reflecting from the surface. Additionally, there are several factors that affect the luminance of a sign. As luminance is also the measure that best represents the perspective of a driver viewing a sign, the guidelines may be most applicable by providing luminance levels alone and not values of illuminance. As sign sheeting becomes more efficient (which has regularly happened since the publication of the AASHTO and IESNA guidelines), less illuminance is needed to provide drivers with a comparable amount of luminance. A final difficulty in applying the AASHTO and IESNA guidelines is the distance from the sign at which the luminance should be measured. While illuminance does not change at the sign face, the luminance will change with both distance and the angle made from the light source and the location where the light is measured. The research conducted for developing the MUTCD minimum retroreflectivity levels used luminance measured at a distance corresponding to an index of 40 ft for each inch of legend letter height. Sign lumi­ nance should be measured at the distance at which a driver is expected to read a sign. For overhead guide signs, that dis­ tance may be several hundred feet because the legends are often 16 in. (uppercase letters) or taller. The MUTCD, AASHTO, and IESNA guidelines attempt to provide practitioners information to help determine when sign lighting is appropriate, and, to some extent, the amount of light­ ing that should be used. There are some apparent inadequacies and inconsistencies in the guidelines, however. This research was intended to produce information that would resolve these limitations. Lighting Trends The cost of lighting overhead guide signs and the evolution of retroreflective sign sheeting products has led to a growing Ambient Light Level Sign Illuminance Sign Luminancea fc lx cd/m2 cd/ft2 Low 13 140 20 1.9 Medium 26 280 40 3.7 High 52 560 80 7.4 a Sign luminance is based on maintained reflectance of 45 percent for white sign letters, assumed to be diffuse. Source: Recommended Sign Lighting Levels (IES RP-19-01 Deprecated—please check www.ies.org/bookstore for updates), published with permission by the Illuminating Engineering Society of North America. Table 3. IESNA recommended sign lighting levels (2).

9 interest among transportation agencies to determine when sign lighting is appropriate. As retroreflective sign sheeting materials have become more efficient in terms of returning headlamp illumination back to the driver, there has been a trend to turn off or remove most overhead guide sign lighting. Additionally, new fonts have been designed to perform best with newer sheeting materials, thus adding more legibility to overhead guide signs and further pushing the issue of whether lighting is needed. Surveys indicate that many transportation agencies have systematically adopted policies of using highly efficient sign sheeting for overhead guide signs to replace the use of sign lighting. In a 2008 survey by the Wisconsin DOT, only six of the responding 30 states still used sign lighting for overhead guide signs (11). The general consensus of the six states still lighting signs was that lighting was used on a case­by­case basis. The primary concern of the agencies using lighting was maintaining adequate visibility during dew, frost, fog, snow, or when unusual roadway geometrics limit the amount of headlamp illumination reaching the sign. A similar survey conducted by AASHTO indicated that 21 out of 36 state DOTs (62 percent) have deactivated sign lighting due to the cost sav­ ings from improved retroreflective sheeting (12). The 15 states still lighting overhead signs use lighting for urban areas, free­ ways, and exit signs. Only five of the responding states indicated that sign lighting is used in the design of new projects. Findings in a survey for the Kansas DOT concur with the conclusions of the other surveys, indicating that most states are moving away from overhead sign lighting, especially outside city limits (13). Half of the respondents indicated that sign lighting is being eliminated in all places. An interesting finding by the Wisconsin DOT is that the states that no longer use sign lighting have received little or no complaints regarding the change (11). While it is clear that headlamp luminance reflected from modern sign sheeting is sufficient for legibility in rural and dark areas, it seems there has been no study (by a transportation agency or otherwise) confirming that the headlamp luminance is unconditionally sufficient in all areas. The hesitation of some transporta­ tion agencies to remove sign lighting in urban areas (based on the survey results) suggests that there may be conditions for which highly efficient retroreflective sheeting alone is inadequate. Survey of State Transportation Departments One of the tasks of this research project involved a survey of 11 state transportation departments about lighting over­ head guide signs and street name signs in their jurisdiction. The 11 agencies were selected based on responses to previous (Washington State DOT and AASHTO) studies suggesting that they have policies for lighting signs. The survey was con­ ducted to gather information about the agencies’ decisions to light signs. While the agencies were known to light signs at one point based on the previous surveys, Table 4 indicates the basic response for each agency’s current policy. Six of the 11 surveyed states no longer light overhead guide signs. The primary reason for discontinuing the policy to pro­ vide overhead guide sign lighting was cost. Three of the six states that no longer provide lighting stated that newer and brighter retroreflective materials had improved nighttime vis­ ibility such that lighting overhead signs was no longer neces­ sary, even though they had not formally researched the issue. In multiple cases, the procedure for phasing out sign lighting involved replacing the sheeting and turning off and removing the lighting equipment. The four states that light overhead signs on a case­by­case basis provided a variety of scenarios for which sign lighting is used in their jurisdiction. Limited sight distance or unusual geometry, frequent fog, and continuous roadway lighting use are some of the criteria used to determine whether signs should be lit. These reasons are consistent with the 2008 Wisconsin DOT survey (11). Florida was the only state to still have a policy for lighting all overhead guide signs. The only regular exception is if there is no access to nearby electricity. The states that use lighting were asked how the agency determines the amount of illumination provided for over­ head guide signs. The most common answer was that the level of illumination is determined based on AASHTO or IESNA guidelines, such as those in Tables 2 and 3. Two of the states indicated that they have reduced the level of illumination as the retroreflective sign sheeting has improved. The same 11 states were asked about their policies for and experiences using internally illuminated street name signs at signalized intersections. Texas and Florida were the only states whose transportation departments illuminate street name signs: Texas on a case­by­case basis that is being phased out, Table 4. Summarized policies for lighting overhead guide signs. Survey Response State Agency No longer light overhead guide signs Delaware, Illinois, Mississippi, Ohio, Oregon, Texas Light overhead guide signs on case-by-case basis Maryland, New Jersey, Pennsylvania, Virginia Light all overhead guide signs Florida

10 and Florida with a policy for lit street name signs when pos­ sible. The other agencies do not have internally illuminated street name signs within their jurisdiction, though there is no prohibition against local municipalities lighting street name signs as long as they are responsible for the costs and maintenance. The surveys of state agencies indicate a consistent reduc­ tion in the use of sign lighting as agencies either eliminate it entirely or require it less frequently. Retroreflective Sign Sheeting Retroreflectivity is an optical property of a material that enables incoming light to be reflected back to its source. It is measured as the ratio of light reflected back to the recep­ tor compared with the amount that is emitted by the source. The ratio fluctuates based on the reflective elements in the sign sheeting and the viewing angle formed between the light source (headlight), the viewing surface (sign face), and the receptor (driver’s eyes). Both ASTM and AASHTO have created specifications for retroreflective sheeting. Most of the previous research on retro­ reflectivity has been based on the ASTM standard; there­ fore, the ASTM standard is referenced in this report rather than the AASHTO standard. The latest ASTM D4956 speci­ fication contains seven classifications for rigid sign sheeting (Types I, II, III, IV, VIII, IX, and XI). Initially, the classifica­ tions increased sequentially based on retroreflective perfor­ mance, but new materials have been added in chronological order of development since 1989. As a result, the current clas­ sification system does not necessarily indicate relative per­ formance. Rather, each type of sheeting material has unique specifications for performance at various angles of retroreflec­ tion. Retroreflectivity tends to decrease as the angles of retro­ reflection widen. The result is that each type of material has a different level of performance depending on the entrance and observation angles determined by the heights of the head­ lamps and driver and the position of the vehicle with respect to a sign. Type IX sheeting, for example, tends to be less bright at long distances than Type VIII sheeting, but brighter at short distances. Sheeting Types I through III are beaded, and Types IV through XI are microprismatic. Premium types of micro­ prismatic sheeting such as Type XI can be more expensive than other microprismatic materials, so some states, such as Missouri and Texas, now specify that sign legends be micro­ prismatic (Type VIII for Missouri and Type XI for Texas) and the background be composed of either a Type III or Type IV sheeting (14). Such combinations can achieve high legend luminance and high contrast at less expense than purchasing a sign made entirely with Type XI sheeting. Standards for mini­ mum levels of retroreflectivity are detailed in the MUTCD. Visual Performance of Retroreflective Signs Research from the 1960s investigated the legibility of differ­ ent combinations of materials used on overhead guide signs. It was concluded that many material types might provide satisfactory legibility without the use of sign lighting, though the conclusion was drawn based on the results from young drivers (15). Research from the 1970s suggested that sign lighting on overhead guide signs could be eliminated when using Type III sheeting if there is a straight approach to the sign (16, 17). It was also suggested that sign lighting be used on curves or where only the low beams of vehicle headlamps were allowed. In 1984, Gordon evaluated the request by the Cali­ fornia Department of Transportation (Caltrans) to use non­ illuminated overhead signs (18). The Caltrans review team concluded that button copy signs with opaque backgrounds functioned satisfactorily without external sign lighting. There were, however, recommendations to maintain sign lighting for freeway off­ramp and lane­assignment signs that call for immediate lane changes. Additionally, sign lighting is to be used where fog and dew occur frequently. Zwahlen et al. investigated the feasibility of removing light­ ing from overhead guide signs when retroreflective sheeting is used (19). They evaluated four different sheeting combina­ tions with and without exterior sign lighting. Based on the field and photometric evaluations, they concluded that either white Type VII or Type IX legends on green beaded Type III backgrounds could provide adequate appearance, conspi­ cuity, and legibility without the use of additional sign light­ ing. The same researchers later performed a more thorough investigation with only older drivers using six material and lighting combinations (20). The researchers found that unlit signs (illuminated only by headlamps) composed of Type IX on Type IX sheeting or Type VII on beaded Type III sheeting performed better than the lighted signs composed of Type III on Type III sheeting. Although it seems the study was executed only on rural roads, the researchers recommended that the Ohio DOT discontinue its practice of lighting signs. Multiple studies indicate that signs made with micro prismatic sheeting tend to perform better than signs of older sheet­ ing types, even when the older signs are lit (21–24). Bullough et al. compared the performance of unlit signs made from new sheeting (Types VII, VIII, and IX) with lit signs of older sheet­ ing (Types I and III) installed along an expressway in an urban area (23). Based on the photometric measurements of the signs (values of luminance and contrast) and the resulting relative visual performance and response times, the researchers calcu­ lated that visibility of the unlit signs with high­performance sheeting was similar to that of the older signs equipped with external sign lighting. The Indiana DOT evaluated the fea­ sibility of discontinuing lighting overhead guide signs on freeways based on comparisons of the conspicuity, legibility,

11 and appearance of various combinations of sign sheeting types used on legends and backgrounds (24). The DOT deter­ mined that it should start using microprismatic sign sheeting (Types IV through XI) on its overhead guide sign legends and backgrounds, and discontinue lighting such signs. Factors Limiting Retroreflection There are several factors that limit the amount of lumi­ nance a retroreflective sign can direct back to drivers. One is the physical degradation of the sign sheeting, which slowly occurs over time. Research simulating the long­term degra­ dation of sign sheeting indicates that white prismatic sheet­ ing will satisfy MUTCD minimum requirements for at least 20 years (25). In the same study, the green sheeting samples used for sign backgrounds also did not degrade to an unac­ ceptable level. Even though the products are likely to meet the MUTCD minimum criteria for a long time, it is important to recognize that those guidelines represent minimums and that the retroreflective performance regularly declines through­ out the time the sign is in service. Another physical factor that affects retroreflectivity is the presence of dew, frost, and dirt that can accumulate on signs. While dew and frost are pres­ ent only during certain times of the day, the effects of dirt are not restricted to these periods. It is not uncommon for some agencies to have sign cleaning programs, at least for signs that are reasonably accessible. The second factor limiting the luminance reflected back to drivers, which has already been mentioned, is the angle of retro reflection formed between the headlamps, the surface of the sign, and the driver’s eyes. This angle is different for each combination of driver, vehicle, and sign location, and con­ tinually changes as a driver approaches a sign. Each classifica­ tion of sign sheeting has different performance specifications for certain angles of retroreflection, so some products per­ form better than others depending on the geometric condi­ tions and location of the vehicles with respect to a target sign. Research conducted for Florida DOT (that was later adopted into policy) recommended that Type XI sheeting be used for overhead guide signs as long as the sign is not on a curve with a radius shorter than 2,500 ft in urban areas or 800 ft in rural areas (10). Sign lighting should be used in those particular instances because the geometric conditions result in wide angles of retroreflection. Implementing Newer Retroreflective Sign Technology As agencies replace lighting with efficient sheeting, there appears to be a consistent thought that as retroreflective sign sheeting becomes more efficient, the need for sign lighting decreases. Since the recommended values of illuminance and luminance in the IESNA and AASHTO guidelines change based on ambient luminance, it would seem sensible that signs in rural areas with little visual clutter do not need to be lit. However, there is still the question of whether headlamps are sufficient as the only source of nighttime illumination (which occurs when signs are not lit). The following section discusses characteristics of vehicle headlamps that affect the amount of illuminance on a sign. Vehicle Headlamps Headlamps have regularly evolved since their first use in the 1880s. While this evolution has led to overall improvements in performance, there have been significant changes to the light distribution, impacting the light available for retroreflection from traffic signs. Only after headlamp specifications were introduced in the 1990s were standards for retroreflective signs addressed (26). Before 1997, the Federal Motor Vehicle Safety Standard (FMVSS) 108 included headlamp intensity and dis­ tribution requirements for all vehicles sold in the United States (27). It allowed more light to be emitted by headlamps above the horizontal plane than used in European and Japanese vehicles. Light above the horizontal plane is helpful for illu­ minating overhead guide signs. In 1997, the FMVSS Standard was revised to form a global compromise of the specifications for the United States, European, and Japanese headlamps (27). The most significant change was that the headlamps projected less light above the horizontal plane, reducing the amount of light available for overhead signs. Chrysler et al. showed that the evolution of headlamps has resulted in less and less light illuminating traffic signs (28). Sivak et al. showed how recent changes in headlamp design affect the illumination of traffic signs, including overhead signs (29). They compared the differences between 1997 tungsten­ halogen headlamps and 2004 high­intensity discharge (HID) low­beam headlamps for U.S. vehicles. Figure 2 shows the dif­ ference between the median 2004 HID luminous intensities and the median 1997 tungsten­halogen luminous intensities for the central part of the beam pattern (2004 HID minus 1997 tungsten­halogen). Figure 2 shows that the newer headlamps provide more illumination on the pavement in front of the vehicle, but less illumination above the horizontal plane. Sivak et al. indicated that traffic signs, including overhead signs, are less visible using the newer low­beam headlamps. The reduction of light on traffic signs reached up to 69 percent for overhead signs, 64 percent for right­shoulder­mounted signs, and 67 percent for left­shoulder­mounted signs. Headlamp degradation, which became more common as replaceable bulbs became the standard over sealed beam lamps, also negatively affects the illumination. Modern head­ lamps with replaceable bulbs suffer from yellowing and fog­ ging caused by factors like acid rain, condensation, dirt, and

12 heat that did not have as great of an effect on sealed beam headlamps (30).The evolution of headlamps has resulted in less illumination reaching traffic signs, whether a result of changes in the distribution of light or the construction that allows for degradation from the natural elements. Because overhead signs without sign lighting must be retroreflec­ tive and rely on headlamp illumination, these changes affect the sign’s visibility. Roadway Lighting Although not directed at signs, roadway lighting provides illumination that may affect the visibility of the sign. Road­ way lighting is intended to enhance the ability of road users to identify and respond to unexpected hazards. There have been numerous studies on roadway lighting, but none appear to identify the effects of roadway lighting on the luminance and visibility of traffic signs, especially overhead guide signs. Roadway lighting is not intended to light overhead retrore­ flective signs, though it does provide some useful illumination if sign lighting is out of service (2). There have been several studies of the operational effects of roadway lighting, with findings of increased speeds and capacity (31–33) and reduced crash frequencies (34–45). Elvik and Vaa (46) and Monsere and Fischer (47) found that reduc­ tions in roadway lighting resulted in increases in crashes. These positive effects of roadway lighting often overshadow the nota­ ble costs and negative effects. The initial installation, regular maintenance, and energy consumption are clear capital costs of lighting. In addition, roadway lighting causes light pollu­ tion, disability glare, and discomfort glare. Each of these nega­ tively affects sign visibility. It has been estimated that about 35 to 50 percent of light pollution is caused by roadway lighting, which can come in the form of sky glow, light trespass, and glare (48). While sky glow and light trespass (light entering a property or building from an outside source) negatively affect the well­being of res­ idents, bright lighting and glare can reduce contrast sensitivity and color perception and negatively affect older drivers whose eyes cannot quickly adjust to different levels of lighting. Sev­ eral state and local governments have addressed the negative environmental effects of roadway lighting through ordinances to evaluate (and mitigate when appropriate) light trespass or sky glow (12). Though there is inconsistency in lighting ordinances from one agency to another, some of the mitiga­ tion measures include shielding luminaires and dimming or turning off the lights during curfew times. These inconsisten­ cies have resulted in the development of the Model Lighting Ordinance (MLO) by IESNA and the International Dark­Sky Association to standardize ordinances and eliminate confu­ sion that may arise as engineers and developers work in differ­ ent regions. The MLO recommends methods for controlling light pollution while maintaining necessary light for areas that need it through listing specific levels of lighting, types of lumi­ naires to use, and methods for shielding light from unintended targets. Disability glare has a direct link to the physiology of the eye and is caused by light scatter from the ocular media in the eye (49). Light entering the eye collides with components of the ocular media such as the cornea, lens, and vitreous humor. At each collision, photons scatter and cast a veil of light across the retina. Up to 30 percent of the stray light is from the cornea, approximately the same amount is from the lens, and the rest is scattered in the retina itself (50, 51). Note: The solid lines below the horizontal represent the edges of a straight and level roadway with two 3.7-m wide lanes. The dashed line below the horizontal represents the road midline. The red lines above the horizontal represent the positions of three types of signs (right, left, and overhead) from the perspective of the left and right headlamps. Figure 2. Difference between the 2004 high-intensity discharge and 1997 tungsten-halogen headlamps (29).

13 Measurements presented by Adrian and Bhanji showed that much of the scattering also occurs in the vitreous humor (52). The veil of light has the effect of reducing the contrast of an object, which would have the same effect as increasing the background luminance of the object. Discomfort glare is from a light source that causes the observer to feel uncomfortable. Van Bommel and deBoer stated that discomfort glare is based primarily on the observer’s light adaptation level, and the size, number, luminance, and location of the light sources in the scene (53). The definition of discomfort, however, is not precise, and some research has shown that a person’s response to a glare source is based more on his or her emotional state than on the light source itself. Disability and discomfort glare are difficult to identify and quantify in order to comprehensively evaluate their effects on traffic sign visibility, but both types of glare represent condi­ tions that drivers can encounter in urban areas with multiple, bright light sources. Effects of Complex Environments on Sign Visibility The previous sections introduced several topics associated with lighting that affect sign visibility, specifically the needs of drivers, guidelines and trends for sign lighting, retroreflec­ tive sheeting, changes in headlamp illumination, and road­ way lighting. This section discusses how the complexity of the visual field may influence the visibility of traffic signs. Visually Complex Backgrounds Lerner et al. indicated that irrelevant visual information leads to information overload in drivers, disrupting the detec­ tion and processing of information relevant to the driving task (54). It should not be surprising that increases in the com­ plexity of a roadway and its background environment have adverse operational effects, such as increased crash rates and reduced traffic flow (55–57). Since visual clutter competes for a driver’s attention, potentially affecting the conspicuity, vis­ ibility, and legibility of a sign, the complexity of the visual field should be accounted for in an assessment of a sign’s visual performance. Sign conspicuity is a measure of how noticeable a sign is in its surrounding environment and how well it attracts a driv­ er’s attention. Signs in rural areas at night tend to have a high level of conspicuity because there are few objects competing for a driver’s attention. In urban areas, where ambient lumi­ nance increases the visibility of other objects, the visual field is much more complex and a sign’s conspicuity is reduced. It has been shown that visual complexity negatively affects an object’s conspicuity (58–63). Complexity also has been shown to affect sign visibility from a perspective of conspicuity and legibility (64–66). Schieber and Goodspeed found that a sign’s visibility improves most when increasing a sign’s brightness if the sign is in a visually complex environment, such as an urban setting (67). Little, if any, improvement to a sign’s visibility can be expected when increasing sign brightness in rural areas with low complexity. Other research indicates that increasing sign brightness can mitigate the adverse effects of visual complexity (61, 68). A final element of the visual performance of signs is legibil­ ity. It is accepted that a sign has already been detected when it is read. While there has been a substantial amount of research on sign legibility, few studies (if any) have included the effects of background complexity. Mace et al. found that visual com­ plexity had no effects on the legibility of warning signs (69), but Holick and Carlson found that background complexity influenced how much luminance was necessary to read over­ head guide signs (9). With increased background complexity, there was an increase in the minimum luminance at which a sign was legible. Measuring Background Complexity Despite the amount of research investigating how complex environments affect the visual performance of signs, there has been no systematic or quantitative method to classify the level of background complexity. One proposed method is the use of image processing technology, which has been utilized in target recognition, traffic surveillance, pavement crack estimation, remote sensing, and some medical applications. Methods have included evaluating color distribution and variance (70), analy­ zing edge level percentages (71), and determining entropy of an image (72, 73). Although one of the benefits of image pro­ cessing is the ability to control the amount of subjectivity in the analysis, it is not uncommon for subjectivity to be intro­ duced in the form of ratings or rankings. This is done so ele­ ments derived from the image processing can be identified as contributing to a human’s perspective of the desired measure (such as complexity). Okawa (70) and Cardaci et al. (72) are two examples where test participants were used to rate images that were also processed with a specific algorithm. The research described in this report developed a technique to measure the visual complexity of an environment as one part of evaluat­ ing a sign’s visual performance. The technique was used in the analysis presented in Chapter 4. Measuring Sign Visual Performance: Recognition and Legibility Reading traffic sign messages during the driving task involves components of both recognition and legibility. Legibility is the ability to read the sign message without first having an expec­ tation for what the sign says. Since drivers have specific desti­ nations, they tend to use search tactics that reflect top­down processing for recognition on guide and street name signs based

14 on the expected message of the sign. In some circumstances, however, unfamiliar drivers may read a sign without having initial clues regarding its message. These instances involve bottom­up processing, indicating a reliance on legibility over recognition since the driver has no prior expectation for the sign’s message. The expectation for a sign’s message increases the distance at which a driver would otherwise be able to read a sign. Recognition and legibility distances are indicative of a sign’s visual performance. Studies measuring a sign’s recognition or legibility tend to evaluate the distance from a sign at which a driver can read or identify the legend. Recognition distance tends to be longer than legibility distance because of the clues provided by the expectation for a particular message. Several researchers have measured legibility or recognition distances, or both, for signs under different study conditions. A short review of some of the study findings associated with guide signs is presented in Table 5. Average legibility index or recognition index for each study is provided. The average legibility index ranges from 30–44 ft/in.; recognition index ranges from 48–75 ft/in. Although only general findings are reported in Table 5, the studies primarily analyzed the factors (such as driver age and Road Conditions Recognition Index Legibility Index General Description Cl os ed C ou rs e 75 ft/in. 40 ft/in. Garvey et al. (74)—Older study participants read shoulder- mounted signs from the passenger seat to evaluate different fonts on ASTM Type III and Type IX sheeting. Test words were unfamiliar location names. The indices combined both daytime and nighttime viewing. 48.5 ft/in. 40.5 ft/in. Hawkins et al. (75)—Older and younger study participants (more older than younger) read shoulder-mounted and overhead signs at night from the passenger seat of a sedan. The signs used three different fonts and Type III sheeting. Test words were not driving related, and the recognition and legibility tasks may have been affected by complicated tasks assigned to the participants. There were small differences between the indices for overhead- and shoulder-mounted signs. n/a 38.4 ft/in. Carlson (76)—Younger, middle-aged, and older study participants read shoulder-mounted and overhead guide signs at night while slowly driving. Two different fonts were tested with Types III, VIII, and IX sheeting. Test words were not driving related. n/a 29.8 ft/in. Chrysler et al. (77)—Older (55–74 years) participants read sign messages at night while slowly driving. Two fonts and three sheeting types were tested. Four sign colors were tested, but only the average legibility index for white on green is shown here. Test words were all four letters in length and driving related. n/a 34.4 ft/in. Carlson et al. (78)—Younger and older study participants read an internally illuminated overhead guide sign while driving on a closed course at night with different levels of illumination. Test words were driving related. The reported value is an average of the legibility indices at different luminance levels. n/a 41.2 ft/in. Miles et al. (79)—Younger and older study participants read overhead and shoulder-mounted guide signs with different fonts while driving at night. The legends were not driving- related words and were constructed of Type XI sheeting. O pe n Ro ad n/a 32.5 ft/in. Carlson et al. (78)—Younger and older drivers read shoulder-mounted guide signs with legends of different sheeting types representing three illumination levels and two environments (rural and suburban). n/a 31.5 ft/in. Chrysler et al. (80)—Older drivers at night read street name signs on the right side of the road while approaching intersections of different levels of complexity. The average value represents signs with Types III, VII, and IX sheeting. n/a 43.8 ft/in. Funkhouser et al. (81)—Study drivers of mixed ages read overhead guide signs on a toll facility at night. The guide signs were of microprismatic sheeting. Both purple and green backgrounds were tested, but only the average for green signs with 16-in. legends is reported here. Table 5. Recognition and legibility indices from guide sign studies.

15 sheeting type) that influence visibility, whether measured as recognition or legibility. Recognition and legibility distances tend to be higher on closed courses and when the tasks given to the participants are not complex. Conclusion There are several factors that affect the visibility of traffic signs at night. Multiple organizations have established guide­ lines to regulate the physical properties that are under an agency’s control to ensure visibility from a perspective of light and appearance. However, there are several other con­ siderations related to the driver that limit the ability to truly control the driving experience and, subsequently, a sign’s visibility. Minimum luminance levels for sign visibility have been studied multiple times. Those studies, however, have been conducted almost exclusively in rural and dark condi­ tions. Other research and anecdotal evidence suggest that increased sign luminance is necessary in urban areas or locations where drivers may be affected by lights or other objects that are unrelated to the driving task and reduce the overall visibility of a sign. These may be sources of glare or distractions that increase the workload of the driving task. While lighting guidelines (such as those in Tables 2 and 3) have attempted to address the diverse conditions that can be encountered on the road, there are some apparent defi­ ciencies in the established lighting levels, and perhaps in the subjectivity used to determine the in­situ conditions. This research was conducted in an attempt to amend those deficiencies.