An expert symposium cosponsored by FHWA looked at the use of color in conveying safety messages, particularly in highway applications.

For thousands of years, humans have used color as a tool to communicate with each other. Today, color permeates daily life. Shoppers rely on color when selecting produce at the supermarket, sports teams use colors to differentiate themselves from opposing teams on the field, and marketers use color to convey information, differentiate products, and move merchandise. In the context of highway transportation, colors shoulder even greater responsibilities, where human lives are at stake. To wit, red, yellow, and green traffic signals let motorists know when to stop and go at intersections, while orange and black road signs alert drivers to use caution when driving through work zones, among other colors used in various signs to convey information, warnings, and other purposes.

"Color provides a backdrop against which one's imagination can soar," says Michael Halladay, former acting associate administrator for the Federal Highway Administration's (FHWA) Office of Safety. "Or it can serve to bring specific items to your attention."

Halladay delivered the keynote address—"How Safe Are Your Colors?"—at a 1-day expert symposium following the 77th annual meeting of the Inter-Society Color Council (ISCC). At the annual meeting, held in September 2008 in Baltimore, MD, color and its many uses in the modern world took center stage. Founded in 1931, ISCC (www.iscc.org) is a nonprofit professional society that works to disseminate knowledge related to color description and specification. The breadth of the council's interests was evident at the 2008 annual meeting, which featured sessions titled "The reds of love and rage: A note on the risk of eliciting negative emotions" and "What color is that cheese doodle, really? A day in the life of a color specifier."

Following the annual meeting, ISCC and FHWA cohosted a 1-day expert symposium for transportation professionals: "Perception, Measurement, and Application of Safety Colors." Presentations at the symposium shined a light, so to speak, on the connections between color and highway applications—specifically, looking at the role color plays in highway safety. In addition to Halladay's keynote address, three other presenters shared updates on research related to the use of color in the highway community.

Roxane Mukai, traffic manager at the Maryland Transportation Authority, addressed the process of developing the distinctive purple color used for electronic toll collection in the Northeastern United States. Tom Hicks, director of the Office of Traffic and Safety at the Maryland State Highway Administration (SHA), discussed the installation of fluorescent yellow traffic signs at critical locations. And John Molino, Ph.D., senior research psychologist at SAIC, who supports visibility research at FHWA's Turner-Fairbank Highway Research Center (TFHRC), shared initial findings from a study of colored retroreflective sign sheeting under daylight conditions.

Color and Highway Safety

During the keynote address, FHWA's Halladay underscored the inherent power of color to bring things—such as roadway signs, traffic signals, and lane delineation—to the attention of motorists. Color has the capacity to make safety-related messages stand out against the visual clutter of today's roadway environments. FHWA and State and local highway agencies need to understand this power in order to use color effectively, he said.

A critical use of color is to warn of danger or signal the need for caution, often at distances that preclude other types of signals. An example is the use of red signal lights by maritime, rail, and road authorities to indicate that the observer must yield right-of-way to another party. When combined with an icon such as Mr. YukTM, the grimace-faced symbol created by the Pittsburgh Poison Center to educate the public about poison prevention, color provides a means of communication to even those who cannot read.

In the world of transportation, where information needs to be communicated successfully over a variety of distances and understood quickly even in the midst of information overload, safety colors need to be unambiguous. Colors selected for use in the highway environment need to be based on a high degree of probability for positive identification by drivers, Halladay said, and restrictions might need to be placed on other uses of specified colors in a given milieu, such as the use of fluorescent yellow-green for advertising along a roadway.

Halladay also discussed the careful balance required between establishing a desired outcome, such as specifying the color requirements for traffic signs, and accounting for the practicality of meeting the requirement. For example, when specifying a safety color for highway traffic signs, standards should allow for a reasonable service life in the field, while providing a high probability of correct identification and appropriate response among motorists. In addition, the requirement should be relatively easy to verify through established measurement procedures, such as ASTM International's standard test method E1349 (the consensus standard for measurement of surface color), so agencies can ensure that critical signs continue to meet the needs of the driving public.

Halladay closed his presentation by laying out three challenges facing color producers, instrument manufacturers, and highway agencies in terms of establishing and maintaining viable standards for safety colors. First, does the context in which the safety color is used, such as the use of fluorescent orange in a work zone for cones and workers' vests, result in a higher degree of proper identification? Even if it does, is it practical or even desirable to modify existing color regions or boxes (the allowable range of a given color used for a specific application) to further minimize the potential for improper color identification? Second, is it possible to develop a measurement procedure that closely matches the human perception of the apparent color of retroreflective materials under various daytime observations? Is such a procedure needed, or would it be more reasonable to work on a simpler measurement procedure that provides more reproducible values? And, finally, is there a point where the use of additional colors in highway traffic signs will result in confusion rather than simplifying the process of information transfer to road users?

These questions, among others, could represent future avenues of research for those studying safety color in general and the use of color in the highway community in particular.

Electronic Tolling Goes Purple

Roxane Mukai's presentation, "Color Me Purple: Or How the E-ZPassSM Interagency Group (IAG) Selected the PMS 259 Color Purple," outlined the history of the selection of purple as the designated color for electronic toll collection. The effort began in 1991 when seven toll collection agencies from New Jersey, New York, and Pennsylvania formed an alliance that became known as the E-ZPass IAG. The IAG selected the name E-ZPass from thousands of potential names and began establishing a distinctive identity for the program, including the design and color of a logo.

In 1991, purple was one of three colors on the Manual on Uniform Traffic Control Devices' (MUTCD) reserved color list. Of those three, purple was the least used color and offered the highest contrast level, making it ideal for a new application. A review of available purple color samples from vendors led to selection of Pantone® Matching System® (PMS) 259 Purple for signs identifying electronic toll lanes.

"A key consideration was that all IAG members needed to agree to use the selected color for E-ZPass lanes," Mukai said, "providing a high degree of uniformity for signing of electronic toll collection." Today, 24 member agencies in 13 States have joined the IAG and adopted PMS 259 Purple for their electronic tolling signage.

Mukai noted that the IAG has been studying the use of purple pavement markings to further delineate lanes for electronic toll collection. The group will strive to ensure that the appearance of the colored pavement markings closely matches the accepted sign color. She also reported that efforts to produce a purple E-ZPass signal light have not been as successful, due to difficulty in obtaining a true purple light and because of small-field tritanopia, a phenomenon that results in a loss of sensitivity to blue light when the signal appears very small. As a result, a purple light might appear red with a blue haze surrounding the signal, potentially causing confusion that could result in erratic behavior as drivers approach toll booths.

On January 2, 2008, FHWA published a Notice of Proposed Amendments to the MUTCD supporting designation of the color purple for use in identifying electronic toll collection. FHWA is analyzing comments on the proposal for the next edition of the MUTCD. Rulemaking could be completed by late 2009. For more information, visit http://mutcd.fhwa.dot.gov/res-notices.htm.

Fluorescent Yellow On Traffic Signs

During a presentation titled "Implementing Fluorescent Yellow on Traffic Signs at Critical Locations," Tom Hicks of Maryland SHA discussed the regulation of traffic control devices (TCDs), which the MUTCD defines as "all signs, signals, markings, and other devices used to regulate, warn, or guide traffic, placed on, over, or adjacent to a street, highway, pedestrian facility, or bikeway by authority of a public agency having jurisdiction." He stressed that TCDs should fill a need, command attention, convey a clear and simple meaning, command respect from road users, and provide adequate time for a proper response. With these issues in mind, SHA conducted evaluations on the use of fluorescent yellow traffic warning signs.

Previous human factors testing indicated that, compared to standard yellow, fluorescent yellow signs would improve detection and recognition under a variety of environmental conditions. As reported in a 2001 study by researchers at North Carolina State University, The Effect of Fluorescent Yellow Warning Signs at Hazardous Locations (PB2001-107693), not only did recognition improve during daytime (when fluorescence assists in detection against high levels of visual clutter in the roadside environment), but also at dawn and dusk, and during fog and rain. The reason for this improvement in performance is that fluorescent signs provide higher luminance than their standard counterparts, while retaining highly saturated colors. Hicks explained that the luminance of the fluorescent yellow signs is as much as twice that of standard yellow signs during daytime and anywhere from two to three times as great at night.

Based on this earlier study, SHA researchers in December 2001 decided to evaluate fluorescent yellow signs on Maryland highways. As part of the evaluation, the researchers evaluated stopping behavior at a site with limited sight distance to the STOP sign along one approach. The researchers conducted an evaluation of free-flowing traffic before and after installation of fluorescent yellow STOP AHEAD signs and found that nonstopping behavior decreased from approximately 10 percent of traffic (on the occluded, or obstructed, leg) to approximately 3 percent. In addition, the average distance at which drivers initiated braking increased significantly.

Subsequent to the success of the experimental applications, the researchers determined that fluorescent yellow signs provide equivalent durability in retention of both color and retroreflectivity. Over time, SHA amended its policy on the use of fluorescent yellow signs, ultimately resulting in a final policy decision in 2004. The current policy states that fluorescent yellow will be used on all road systems on all warning signs, with the exception of school/pedestrian warning signs and freeway incident traffic management signs, which will use the reserved fluorescent yellow-green and fluorescent pink colors, respectively.

The 2003 MUTCD does not specifically address fluorescent yellow, but in 1999 FHWA made an official interpretation allowing its use in highway signs. The addition of fluorescent yellow to the section on sign colors is among the proposed amendments for the next edition of the MUTCD.

Testing Sign Retroreflectivity

John Molino's presentation, "A Comparison of Colored Retroreflective Sign Sheeting Under Daylight Conditions: Perception Versus Measurement," briefed symposium attendees on initial results from an ongoing FHWA study. He explained that daylight measurements of the color of retroreflective materials used for roadway signs are generally more variable than measurements of nonretroreflective materials. Retroreflective surfaces have spherical or prismatic reflectors that direct light back in nonuniform ways. These materials are designed to reflect a maximum amount of light from the headlights of a vehicle back to the eyes of a driver. The chromaticity (quality of color) of the material under daylight viewing conditions is generally a secondary concern, Molino said.

The reproducibility of color measurements of retroreflective materials during daylight hours in the field shows considerable variability. Slight variations between colorimeters (instruments used to measure color) and field spectral photometers (instruments used to measure luminance or brightness) can result in significant differences in measured color. Furthermore, previous FHWA research has shown that field measurements of chromaticity do not correspond to perceived color judgments made by human observers.

"Understanding and reducing these inconsistencies is important to FHWA for defining the size and shape of the color boxes [the coordinates in color space that define the boundaries of acceptable colors] used to specify colors for roadway signs," Molino said, especially as these color boxes are incorporated into Federal regulations (23 CFR Part 665, 2002, Color Specifications for Retroreflective Sign and Pavement Marking Materials).

In this FHWA study, researchers measured the color and luminance of retroreflective sign materials under laboratory and field conditions. Then they compared the instrument measurements with the perceptual judgments of color and brightness made by human observers in the field. Next, the researchers employed a percentage estimation technique for the perceptual rating of hue and saturation, along with a similar perceptual rating procedure for brightness. The researchers determined the hue, saturation, and brightness for four types of retroreflective sheeting materials and one diffuse plaque, four quadrants of the color box (as described in the Federal regulations), and six colors specified for use in roadway signage. Altogether they evaluated 120 color and material combinations (five types of sign materials x four quadrants x six colors).

Research participants sat outdoors under a shade tent on a closed road and viewed 7.5- by 7.5-inch (19- by 19-centimeter) test samples at a distance of about 100 feet (30 meters). After viewing a sample for 10 seconds, participants were asked to respond with hue, saturation, and brightness ratings, using a rating scale ranging 0 to 100 percent for all three judgments. A total of 17 participants rated all 120 samples twice a day for 4 days. To supplement the test participants' subjective brightness ratings, the research team obtained separate subjective brightness rankings for one quadrant of the yellow and red colors of the five materials.

Then the researchers plotted the hue scaling data on uniform appearance diagrams (UADs). These diagrams represent a two-dimensional perceptual color space based on red-green and blue-yellow dimensions. The UADs are based upon the opponent process theory of human vision, where red and green operate as an antagonistic color pair, designated as the a dimension, and blue and yellow operate as a second antagonistic color pair, designated as the b dimension. These two dimensions are used to establish the coordinate system for the UADs for perceived color. A third dimension, designated the L dimension, is composed of black and white and is achromatic (devoid of color). This third dimension is associated with the luminance, or brightness, of the stimulus. For objective instrument measurements, the corresponding three dimensions (L, a, and b) can be combined into a three-dimensional color space, referred to as a "Lab," or "LAB," space. A widely used version of this formulation, which scientifically describes how the average human eye sees color, is known as CIELAB, where the CIE stands for Commission Internationale de l'Éclairage.

After plotting the UADs, the researchers compared these perceptual hue and saturation data to CIELAB plots produced from the instrument chromaticity measurements made in the laboratory and in the field to determine how well the instrument color measurements corresponded to perceived color appearance. The researchers also compared the instrument measurements to determine how well the laboratory measurements corresponded to the field measurements. Similarly, the researchers compared the instrument-measured luminance determinations from the various sample stimuli in the laboratory and in the field with each other, and with the brightness ratings reported by the study participants.

The study used sheeting samples specified in the ASTM standard specification D4956-07. On average, the research team found that participants rated sheeting types VIII (super-high-intensity retroreflective sheeting having highest retroreflectivity characteristics at long and medium road distances) and IX (very-high-intensity retroreflective sheeting having highest retroreflectivity characteristics at short road distances), and proposed type XI (which uses a new, highly efficient optical design for reflecting light) closer to the laboratory measurements than sheeting type III (high-intensity retroreflective sheeting that is typically encapsulated glass-bead retroreflective material) and diffuse plaque (no retroreflective properties). In the supplemental brightness determinations, participants ranked sheeting types VIII, IX, and proposed XI higher than sheeting type III for both red and yellow colors. Diffuse plaque received the highest brightness rankings of all five materials for both colors.

Overall, the participants' responses aligned well with the instrument measurements. The responses actually indicated a greater perceptual separation between hues, especially those for the red and orange color boxes, than the measurements revealed. That is, the human observers seemed more sensitive to hue differences than the instruments. However, study participants tended to perceive the colors to be less saturated, on average, when compared to instrument measurements, and detected little difference in the saturation of the four quadrants for yellow and orange. According to Molino, this implies that, within certain limits, the fading of yellow and orange signs, although not desirable, might not pose as significant a problem as physical measurements might indicate.

A comparison of the luminance measured in the field with the participants' ratings of brightness indicated that people are less sensitive to differences in the luminance of signs that are relatively bright (as is typical for white and yellow backgrounds) than to differences in the luminance of relatively dark signs (especially blue). The results indicate that for people with normal color vision, the probability of improper color recognition of traffic signs, at least for the colors evaluated, appears to be very low. However, the subtle differences between instrument measurements and human perception, and the relative lack of human sensitivity to differences in color saturation and sign luminance, indicate a possibility that existing measurement procedures can fail to properly account for how the driving public views individual signs.

Ultimately, the researchers found that the red versus orange sign materials and, to some extent, the orange versus yellow sign materials were more difficult for observers to discriminate. Although not likely to be a source of confusion with the present definition of acceptable colors, Molino says, care needs to be taken when considering modifications to color boxes in these color regions. The final report will be available by fall 2009 at www.fhwa.dot.gov/research/tfhrc/programs/safety/.

Ongoing Color Studies

Colored materials used in highway traffic signs and pavement markings need to provide reasonable service life under difficult environmental conditions, and, increasingly, contain limited or no toxic chemicals. As directed by Congress in Section 1907 of the Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users, FHWA is conducting demonstration projects in Alaska and Tennessee to study the safety and environmental impacts and cost effectiveness of different pavement marking systems and the effect of State bidding and procurement processes on the quality of pavement marking materials employed in highway projects. The projects will include an evaluation of the impacts and effectiveness of advanced acrylic water-borne pavement markings and lead-free thermoplastic yellow paint, which could become environmental friendly alternatives to existing paint systems that contain toxic chemicals.

The presentations shared during the ISCC expert symposium underscore the importance of the highway safety community having an understanding of perceptual responses to color and the process of specifying and measuring color. When it comes to color in the roadway environment, one might go so far as to say, "Safety is in the eye of the beholder."

Carl K. Andersen is the roadway team leader in the Office of Safety Research and Development at TFHRC and also manages the Arens Photometric and Visibility Laboratory. He holds a master's degree in physics from the Naval Postgraduate School, Monterey.

For more information, visit www.fhwa.dot.gov/research/tfhrc/labs/visibility/ or contact Carl Andersen at 202-493-3366 or carl.andersen@fhwa.dot.gov.