Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Research Home
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-RD-97-152

Guidelines for the Use of Raised Pavement Markers




1. Future accident analyses should consider how variables such as road geometry, average daily traffic (ADT), and geographical region will affect the interpretation of results. Previous studies that have averaged over such variables should be re–evaluated.

2. Measurement of driving performance, including analyses of potential accidents because of curve lengthening when RPMs are used, should be investigated further.

3. Studies should report the driving situation that was investigated in terms specific to those set by FHWA guidelines. For example, if spacing specifications are described on the basis of the degree of curvature for horizontal curves, studies should report the road geometry in terms of degree of curvature. The reflectivity of RPMs should be measured in SI units and reported with reference to the minimum reflectivity criteria (yet to be established). These criteria along with required visibility distance should be used to evaluate RPM performance.

4. Guidelines should include specific information as to when and where left edgelines should be supplemented by RPMs, perhaps using lighting and ADT criteria.

5. Good and Baxter indicated that further studies with larger sample sizes should be conducted to confirm their results.(18) It is also imperative that the study include a sample of older drivers to ensure that the benefits of larger, more clear delineation to the older population is not discounted. Studies should also be conservative enough to include potential benefits of wide edgelines to drivers when faced with adverse weather conditions.

6. It would be useful to provide separate guidelines for those States that will invariably install snowplowable RPMs as opposed to non–snowplowable RPMs–especially since the cost–effectiveness may vary between States on the basis of weather conditions.

7. The cost–effectiveness/performance tradeoff should be investigated in order to determine whether RPMs should be used on road sections with broken lines. If RPMs are used to delineate a passing zone, they will be subjected to many impacts, reducing their life span when compared with other uses. However, a recent study by Zwahlen and Schnell (1996) suggested that longer preview distances would be beneficial where passing is allowed on both sides of the road (dashed line sections); their study showed that the driver had the worst and shortest preview conditions (2.14 s) when driving through these sections.(21) More information is required before making a recommendation.

8. Pennell recommended that guidelines for RPMs also include the following: clear markers should supplement white lines, and when two–way markers are used for motorist guidance, the nonreflective side should face wrong–way motorists as opposed to a red reflective side.(19) These factors contradict the current guidelines; further studies are needed to determine which approach is best.

9. Color coding with RPMs is not a current delineation practice, but it was suggested for use by earlier work in areas such as no–passing zones or merge/diverge areas. For example, a no–passing zone might be delineated with white markers at long spacing and yellow markers at short spacing, or by delineating a solid yellow line with markers as a warning where a solid white line is currently used (Taylor et al.).(13) Current standards outlined by the Roadway Delineation Practices Handbook only describe the use of colored RPMs to supplement the color of paint.(2) It might be useful to determine whether the color of RPMs might provide the driver with additional information.

10. The current standards need to be specific for all situations. Currently, the Roadway Delineation Practices Handbook indicates in figure 40 (page 88) that RPMs should be placed inside double–yellow centerlines when used on two–lane, two–way roads, yet figures 44, 45, 47, and 48 all show situations where RPMs are placed outside double–yellow centerlines when used on two–lane, two–way roads.(2) The Handbook needs to clarify which pattern should be used under which circumstances.

11. Although recommendations currently suggest that ramps are delineated with RPMs on the left side only, according to Taylor et al., the "pathway to be driven within" should be defined as opposed to the "line to stay next to." They suggested placing RPMs on both edges for ramps or diverging roadways.(13) According to Kahn (1979) (cited in Migletz et al.) the level of guidance for night driving is increased when RPMs are placed on the inside and outside of curves.(2) These possibilities require further investigation.

12. Those preparing final guidelines for the various regions should consult the recommended guidelines made in New Jersey by Pennell, which support the current standards for: placing RPMs in line with and gap of skip lines, to offset RPMs from a solid edgeline, to offset RPMs to the traffic side of left edgelines, and to place RPMs between double yellow centerlines.(19)

13. The suggestion to use supplemental RPMs on right solid edgelines under hazardous conditions states that the edgeline should have a "reasonable distance of treatment before the hazard." This reasonable distance needs to be quantified, either through empirical research or by analytical means.

14. The current guidelines specify that centerlines are supplemented to indicate a transition from passing in both directions to no passing zone, on a two–lane, two–way road, with RPMs spaced at 2N–24.4 m (80 ft). These are placed between skip lines, and RPMs spaced at N–12.2 m (40 ft)–are placed outside of double solid–yellow lines. However, the guidelines provided by Pennell specify that special treatments should not be used for situations of passing, no passing, or a combination of the two.(19) Although recommendations made by Pennell support certain spacing standards, conflicting recommendations may reflect the need for region–specific guidelines. States that use snowplowable RPMs may require different guidelines for certain circumstances.

15. In figures 41, 42, and 44 in the Roadway Delineation Practices Handbook, alternate examples are given for spacing RPMs with a separation of N instead of 2N.(2) The Handbook needs to be more specific as to when or where this alternate spacing should be used.

16. In figures 41, 42, and 43 in the Roadway Delineation Practices Handbook, examples are given for RPM use without supplementing the painted delineation.(2) The Handbook needs to be more specific as to when or where this alternate system should be used.

17. When referring to segment–to–gap ratios, the MUTCD also states that dimensions other than the 3–1 ratio "may be used as best suits traffic speeds and need for delineation."(1) Since the placement of RPMs is directly dependent on the segment–to–gap ratio, it is suggested that the following two sources be considered to determine if an upper or lower boundary can be specified.

Two sources indicated the possibility that insufficient delineation may induce drivers to speed up in order to increase information flow rate. Allen et al. reported that contrast thresholds reduce under dynamic conditions and reach a maximum at a frequency region of 2.5 Hz.(5) Therefore, in order to receive information at an optimal frequency of 2.5 Hz, the current 12.2–m (40–ft) cycle length would require a driver's speed to be 109.5 km/h (68 mi/h). If the cycle length were reduced to 7.3 m (24 ft), the optimum speed would be 43.5 km/h (27 mi/h). The study found that, when driving with a 12.2–m (40–ft) cycle length, there was an increased driver time delay at slow speeds, which resulted in the driver's "desire" to speed up.(5) Good and Baxter also found that the addition of some, but insufficient, delineation seemed to result in excessive speeds.(18) It was suggested by Allen et al. that if cycle lengths were kept small, perhaps it would avoid inducing inappropriate high speeds under adverse visibility conditions.(5) Recommendations for using a higher segment–to–gap ratio (i.e., shorter cycle length) would change the current 1:3 ratio with a cycle length of 12.2 m (40 ft) (producing a 3.05–m (10–ft) length with a 9.2–m (30–ft) gap) to, for example, a 3:5 ratio with a cycle length of 7.3 m (24 ft) (producing a 2.7–m (9–ft) line length with a 4.6–m (15–ft) gap).

According to these sources, if the ratio is lowered too much, RPMs would be spaced too far apart and would not provide sufficient information. Ultimately, the delineation guidelines need to set strict criteria so that they require little speculation.

18. Pennell suggests that supplemental RPMs be spaced at 6.1 m (20 ft) for painted gores unless the gore exit is shorter than 24.4 m (80 ft), in which case, spacing should be 3.05 m (10 ft).(19) The Roadway Delineation Practices Handbook specifies a spacing of N/4–3.05 m (10 ft)–for the gore area of exit and entrance ramps.(2) This suggestion may have been made because snowplowable raised pavement markers (SRPMs) are used extensively in New Jersey. Consideration should be given as to whether this recommendation should also be implemented in other States where SRPMs are used.

19. Consider the progress of research concerning paint markings, for it might affect the current recommendation for RPMs (use and/or spacing). For instance, Fitzpatrick, Lance, and Lienau (1995) found that drivers moved into or out of the exiting lane further upstream of the lane drop gore after pavement markings were used at freeway lane drop exits. (23) The number of maneuvers also decreased; the largest decrease was in the number of one–lane changes through the gore. If the gore markings are supplemented with RPMs, as recommended in the subsection above, the areas with more frequent lane changes would suffer the greatest damage and loss of markers. Additional research should be done to address the best combination of paint and RPM location/spacing, as required for both day and night delineation.

20. Zwahlen considered the use of RPMs on cloverleaf entrance and exit ramps.(22) He suggested that four RPMs would be required in view to provide adequate preview time for the driver. The optimal spacing would therefore be 7.6 m (25 ft). He concluded that RPMs were not useful as delineation for cloverleaf entrance and exit ramps.(22) However, since he used a sample size of only 11 people, it is recommended that the experiment be redone with a larger sample size. Also, since RPMs were severed to achieve 50 percent of the initial reflectivity and left as long as 2 years, it is recommended that the experiment be redone with frequent measures of RPM reflectivity.

21. Pennell recommends that continuous center left–turn lanes should not be identified by specially spaced RPMs.(19) This suggestion may have been made because SRPMs are used extensively in New Jersey. Consideration should be given as to whether this recommendation should also be implemented in other States where SRPMs are used.

22. Pennell (19) recommends using the same spacing before intersections as is used for through movement; however, previous studies by Taylor et al. (13) and Bali, McGee, and Taylor (1976) (24) support the notion of reduced spacing before intersections. Pennell also made recommendations specific to signalized and unsignalized intersections, a distinction that may be useful for guidelines.(19)

23. One area of research considers the use of coded delineation to provide the driver with advance information about upcoming curves or intersections, but in such a way as to induce the driver to slow down. An initial review of the following research indicates that coded delineation may help the driver maneuver the vehicle both at intersections and through curves.

The use of coded stripes as advance warning for curves was studied by Witt and Hoyos (1976) in a simulation experiment.(25) They found that with coded stripes: (a) the precision of course–following increased, especially in the approach zones to curves; (b) steering wheel turning became more steady and smooth; and (c) the drivers drove slower on difficult road sections and faster on easy road sections. Witt and Hoyos suggested follow–up field research in order for the results to be generalized to the real world.(25) After reviewing literature that investigated the potential benefits of transverse stripes as warnings before curves or at stop approaches, Bali et al. considered them to be effective at reducing the average speed and the variability around that average.(24) Bali et al. suggested the use of transverse stripes painted across approach lane(s) at gradually decreased spacing where a required stop is unexpected or in high accident areas. The use of thermoplastics was also suggested (as paint would wear away too fast), along with a bar width of 609.6 mm (24 in) and an overall pattern length of 152.5 m (500 ft).(24) Taylor et al. suggested that the principle of perceptual modification techniques might be applied to spacing of RPMs or PMDs. They thought that deliberate distortions of relative motion between marker and vehicle might produce the deceleration profiles required.(13) Blaauw also suggested relating the spacing of RPMs to road curvature like that used for stripes and PMDs.(17) In fact, a study by Zwahlen (1993) used an optical illusion in an attempt to reduce driver speeds.(26) He arranged PMDs to provide a perceptual illusion of increased curve sharpness by ascending (or descending) delineation heights. The height increments needed to be relatively large (e.g., 1.07 m (3.5 ft) to 1.65 m (5.41 ft) over a 72–m (236–ft) curve length); the results found a reduction in speed with significant center speed reduction from 49.69 km/h (30.88 mi/h) to 44.09 km/h (27.4 mi/h).(26) A study by Rockwell and Hungerford (1979) (cited in Zwahlen, 1993) also found a decrease in speeds through curves when a novel PMD system was used.(26) Further research is needed to determine whether the effect wears off over time.

There are many coded delineation issues that remain unclear, such as what part of the curve (e.g., radius, degree of curvature) reveals the most essential elements of a turn (e.g., sharpness, width) and whether RPMs present this information early enough to warn the driver. The results of a study by Fildes and Triggs highlight a potentially dangerous illusion in the perception of curvature during the negotiation of bends in the road.(7) They found that, when drivers are asked to make judgments about curvature–similar to a situation commonly experienced as a driver moves along a curved two–lane highway–first, the driver will primarily make that judgment on the basis of the curve's deflection angle, and second, the radius of curvature is likely to be misinterpreted by drivers in their assessment. Subjects in the experiment responded to small–radius, small–angled curves as least curved when in fact they were the most curved.(7) Thus, it seems important to ensure that the driver has as much information about the most important part of the curve as far ahead of time as possible. The curve–assessment process was shown by Cohen and Studach (1977), and McLean and Hoffman (1973) (cited in Fildes and Triggs) (7), and by Shinar et al. (4), to start well in advance of the vehicle entering the curve (at least 100 m (327.9 ft) before); far delineation is therefore essential to provide the necessary information.

The above research should be included in a thorough literature review, and further empirical studies should be conducted to answer questions such as: "Does the effect of coded delineation wear off over time?" and "How are accident rates affected by coded delineation?"

24. The Roadway Delineation Practices Handbook lists the following guidelines in the text and refers to figure 50 for clarification of RPM placement in work zones.(2)

RPMs should supplement paint with a spacing of 12.2 m (40 ft) (a retroreflective RPM is placed midway between each 3.05–m (10–ft) paint stripe).

If RPMs substitute painted for skip lines, a cluster of four nonretroreflective RPMs with a retroreflective RPM every 12.2 m (40 ft) or N is recommended. The nonretroreflective RPMs should be placed 1 m (3 1/3 ft) or N/12 apart to provide the daytime appearance of a skip line.

However, both RPM type and spacing are depicted differently in figure 50 than is written in the text. According to the figure, the following guidelines should be used:

RPMs can be used to substitute paint on two–lane, two–way roads if work zones will last 14 days or less. A series of three RPMs will be intra–spaced at N/12 and inter–spaced at 10N/12 (or N distance from the first RPM in series 1 to the first RPM in series 2).

RPMs can be used to substitute paint on two–lane, two–way roads if work zones will last over 14 days. A series of three RPMs will be intra–spaced at N/12 and inter–spaced at 10N/12 (or N distance from the first RPM in series 1 to the first RPM in series 2), and accompanied by painted edgelines.

RPMs can be used to substitute for paint on two–lane, two–way roads with severe curvature if work zones will last 14 days or less. A series of two RPMs are used, intra–spaced at N/20 and inter–spaced at 18N/20 (or N/2 distance from the first RPM in series 1 to the first RPM in series 2).

This discrepancy needs clarification. It is also important that the guidelines specify under which circumstances paint should be replaced by RPMs rather than be supplemented by them.

25. Pennell recommended that TRPMs should be spaced at 1.5 m (5 ft) or N/8 with retroreflective units at 6.1 m (20 ft) when substituting stripes.(19) This should be considered when determining which source (text or figures) reflects the guidelines of the Roadway Delineation Practices Handbook.

26. The results of the study by Davis were used to make recommendation for the preferred way to delineate at curves, drop–offs, and relocated ramps. (27) The final recommendations for RPMs were listed in the subsection above; there were also a series of suggested study areas that emerged from the survey of professionals who had delineation experience. Some of these were not included in the final experiments, and may warrant further research. For horizontal and vertical curves, consider supplemental retroreflective RPMs spaced 6.1 m (20 ft) apart or retroreflective RPMs spaced 1.8–3.05 m (6–10 ft) apart when used without paint. For relocated entrance ramps, consider supplemental RPMs spaced at 3.05 m (10 ft) and RPMs spaced at 1.5 m (5 ft) when used without paint.

27. More information is required in order to compile a final list of features that should be incorporated for RPM types. A minimum reflectivity criteria for RPM effectiveness has yet to be established and will influence the level of reflectivity required for new RPMs. Cost analyses will determine whether a minimum half–life is required. Research to determine lens types, damage protectors, mountings and other special features will most likely follow production–driven studies. Documentation of acceptable features should be extrapolated from these sources.

28. Zwahlen explains the use of OCARD (ODOT Computer Aided Road Delineation)–a knowledge–based system that computes the delineation layout using consistent and uniform delineation–in his study "Optimal Application and Placement of Roadside Reflective Devices for Curves on Two–Lane Rural Highways."(26) If it is possible to adapt this software system for the use of RPMs, it may aid the traffic engineer in the application of RPMs and make the task more efficient. The potential of this package should be investigated further.

29. As of yet, there are no national guidelines for when to replace a RPM. Since the criteria for determining minimum acceptable reflectivity for RPMs are not yet resolved, further research must be done in order to set a criterion in place. Currently, individual States determine which criteria to use for replacement procedures, based not on minimum reflectivity but on missing markers. Freedman et al. indicate several States that use different criteria to determine "effectiveness" of RPMs and therefore when to replace them: (3)

  • California RPMs are replaced when two or more consecutive markers are missing.
  • Florida RPMs are replaced when eight or more consecutive markers are missing.
  • Texas RPMs are replaced when 50 percent or more markers are missing within 1.6 km (1 mile) of highway.
  • Massachusetts Replaces only reflective lens if casting is intact.
  • Michigan Replaces only reflective lens if casting is intact.
  • New Jersey Replaces only reflective lens if casting is intact.
  • Massachusetts Snowplowable RPMs are replaced when 30 percent or more markers are missing.
  • New Jersey Uses visual inspection.
  • Pennsylvania RPMs are replaced as needed, determined by visual inspection.

Since the effectiveness of RPM delineation requires both adequate reflectivity and adequate numbers, it would be ideal to determine a replacement criterion based on both factors. The Delineation Practices Handbook refers to a study by McNees (1987), which sought to determine a procedure that could be used to evaluate the effectiveness of RPM and raised traffic button systems.(31) The evaluating criteria include both reflectivity and loss of markers as indicators of replacement. However, the proposed procedure still requires further investigation. The experimental procedure and, thus, in turn, the proposed evaluation procedure does not account for additional variability.(31) Further research could determine whether evaluations might be biased by factors such as film processing, comparing scenes taken from different cameras or film sets, or whether color film produces significantly different representations from black and white film. Other factors that might bias the contrast or brightness of the photograph should also be investigated, so that a precise procedure might be provided. Lee, Hostetter, and Leibowitz (1991) include some information about controlling for contrast that is useful.(32)

If the method of evaluating photographs proves insufficient, it is still necessary to apply a criterion to determine whether RPMs should be replaced. One issue to consider however, is how one accurately applies such a criterion to RPMs in the field while they are attached to the road. One study addresses some of the problems in measuring wet night performance and discusses the mobile laser retroreflectometer. This measuring device is described as being able to measure background luminance and it records both the coefficient and the contrast luminance of a line along significant distances of road (DeJaiffe, 1987).(33) The progress of this research should be monitored, along with the possibility of measuring RPMs with a mobile device with similar capabilities.

One alternative to measuring reflectivity in the field is to find a relationship between field and laboratory measurements such that laboratory–controlled studies can accurately reflect what would be found in the field. A study by King and Graham (1989) found that there is a strong relationship between test subject subjective evaluation of field luminance and laboratory evaluation of luminance. (34) This relationship can be expressed mathematically and the resulting equation used to calculate "field factors" relating laboratory–produced evaluations to actual field evaluations. If King and Graham's findings can be verified, future research could then be conducted in the laboratory under controlled and safe conditions with minimum field verification. The possibility of applying this method to RPM research should be considered.

30. Given how the luminosity requirements for delineation are outlined (based on research performed by Freedman et al. (3)) in the Roadway Delineation Practices Handbook (2), if possible, the same basic method should be applied to determine minimum contrast levels for RPMs. If a minimum reflectivity criteria cannot be achieved through this method, information from the study by Blaauw and Padmos should be expanded to determine a criterion on the basis of visibility distances.(6)

31. If contrast can be used to determine a minimum reflectivity criterion, the results of the computerized headlight evaluation model (DETECT) used by Freedman et al. may be useful.(3) They showed that, for centerline visibility, if a single vehicle providing glare is between 100 and 300 m (327.9–983.7 ft) away, the contrast on dry pavement needs to be eight times greater than when there is no glare source in order to preserve a 3–s preview time. On wet tangents, for a single oncoming vehicle at 300 m (983.7 ft) away, the contrast needs to be four to eight times greater than the contrast needed when there is no glare source to preserve 2 s of preview time.(3)

32. Given that Blaauw and Padmos used a required preview time of 5 s, in clear weather the required preview distance for delineation would have had to be 110 m (360.6 ft) when the velocity was 80 km/h (49.7 mi/h) and 140 m (459 ft) when the velocity was 100 km/h (62.1 mi/h).(6) In fog, the required preview distance for delineation would have had to be 100 m (327.9 ft) when the velocity was 70 km/h (43.5 mi/h) and 110 m (360.6 ft) when the velocity was 80 km/h (49.7 mi/h). According to these findings, at a speed of 80 km/h (49.7 mi/h), the driver would have to be able to see delineation at least 110 m (360.6 ft) ahead in clear weather.(6) If these levels were adjusted using a preview time of 4 s (the upper limit for visibility distance to account for older or impaired drivers) and other conditions were taken into account (e.g., glare), it would be a starting point for determining a minimum reflectivity level based on visibility distance as opposed to contrast level.

33. Consider the progress of research concerning the reflectivity of paint and thermoplastic markings. This will ensure that cost–effectiveness estimates for the use of RPMs is accurately portrayed. Some research studies that may give background or updated information are listed below:

  1. Chapman, B. J. (1994). Cost–Effective Marking and Delineation Materials for Highways. Technical Report No. FHWQ–CA–TL–94–06. California State Department of Transportation: Sacramento Division of New Technology, Materials and Research.

  2. Merrit, J. O. and Kerr, S. K. (1977). Driver's Visibility Requirements for Roadway Delineation, Vol. II: Color Identification of Yellow Highway Paint as a Function of Yellow/White Pigment Mixture Ratio. Technical Report No. FHWA–RD–77–166. Washington, D.C.: Federal Highway Administration.

  3. Gatlin, G. R. (1993). Evaluation of Cold Plastic and Hot Spray Thermoplastic on I–20 in Scott County. Technical Report No. MDOT–RD–93–67–20. Jackson, MS: Department of Transportation.

  4. Kidd, S. Q. (1991). Cold Plastic and Hot Thermoplastic, Foil–back Tape, Removable Tape and Paint Pavement Markings. Technical Report No. MSHD–RD–90–67–17. Mississippi State Highway Department, Jackson, MS: Jackson Research and Development Division.

  5. Schrock, M. P. et al. (1993). Developing a Monitoring System of the Dispensing Rate of Glass Traffic–Line Beads. Technical Report No. K–TRAN–KSU–92–1. Kansas State University, Manhattan.

  6. Transportation Research Board (1991). Communications, Traffic Signals, and Traffic Control Devices. Washington, D.C.: National Research Council.

  7. Zwahlen, H. T. and Schnell, T. (1996). Visibility of Yellow Center Line Pavement Markings as a Function of Line Configuration and Line Width. Proceedings of the Human Factors and Ergonomics Society 40th Annual Meeting. Ohio University, Athens, Ohio: Department of Industrial and Manufacturing Systems Engineering, pages 919–921.




Previous | Table of Contents | Next

United States Department of Transportation - Federal Highway Administration