#22 – Blue Bike Lanes at Intersection Weaving Areas

Portland, Oregon

William W. Hunter and Libby Thomas, UNC Highway Safety Research Center


Intersection and intersection-related locations account for 50 to 70 percent of bicycle–motor vehicle crashes (Hunter, Stutts, Pein, and Cox, 1996). In Portland, OR, both motorists and bicyclists had expressed concern about a number of locations where bicycles and motor vehicles came into conflict when motor vehicles turned, changed lanes, or merged across bike lanes at or near intersections. Colored pavement, raised crossing paths, and other measures have the potential to alert motorists and cyclists to these intersection conflict zones, thereby increasing yielding behaviors and reducing conflicts and crashes. Such treatments have been found to be effective in several European and Canadian cities (Pronovost and Lusginan, 1996; Jensen, 1977; Leden, 1977; Leden, Gårder, and Pulkkinen, 1998).

The 10 Portland sites selected for treatment and study were all sites with a high level of bicyclist-motorist interaction and a history of complaints. All were in areas with existing bicycle lanes. Prior to treatment, all of the bike lanes were outlined with dashed lines at the conflict areas. All except one of the sites also had in place traditional regulatory signs to alert motorists to “YIELD TO BIKES.” The signs had been in place for some time and were in good repair. At one location, Hawthorne Bridge, where there was no yield sign for motorists, bicycles had been yielding to motor vehicles before the blue pavement and signs were added.


Each of the 10 sites were locations where the bicyclist travels through (straight ahead) and the motorist crosses the bicycle lane to: exit a roadway (group 1), move into a right turn lane (group 2), or merge onto the bicyclist’s street from a ramp (group 3). (See figures 1–3 for examples.)

At all 10 sites, the conflict areas of the bicycle lanes were marked with light blue paint or with blue thermoplastic intended to highlight the conflict zone. The intent was to increase awareness and safe behaviors by both cyclists and motorists and yielding behaviors by motorists. Light blue was chosen because it doesn’t have another meaning to motorists (as do red and green, sometimes used in other countries), can be detected by color-blind individuals, and usually is relatively visible in low-light or wet conditions. Additionally, blue was overwhelmingly favored by participants in a number of public presentations, as well as by bicycling professionals, and prior studies suggested that it would be an effective color.

The first sites were painted blue with glass beads applied to the wet paint at a total cost of $900. Unfortunately, within two to three months, the paint was worn away at some of the locations with higher traffic volumes. Therefore, at eight of the sites, a more expensive, thermoplastic, skid-resistant material was applied.

At each location, one of several innovative “YIELD TO BIKES” signs was installed with a design appropriate for the particular motorist maneuver and configuration at that site (Fig. 4).

Figure 1. Hawthorne bridge conflict area. Motorists exit right to an off-ramp, crossing the through bike lane that weaves left to cross the bridge. Example of a group 1 configuration.

Figure 2. Motorists approaching Grand Avenue weave across a bicycle through lane to enter a right-turn- only lane on Madison Street. (Group 2 site)

Figure 3. Bicyclists approaching the Broadway Bridge travel straight, while motorists from Interstate Avenue entering Broadway cross the bike lane conflict area (outlined by the dashed striping). (Group 3 site)

Group 1 - Exit right across bicycle lane

Group 2 - Lane change across bicycle lane

Group 3 - Entering roadway/merge across bicycle lane

Evaluation and Results

Videotape analysis was used to compare before and after behaviors of both motorists and bicyclists in the conflict areas. Twenty hours of “before” treatment video data (two hours per site) and 30 hours of “after” data (two or four hours per site) were collected. Videotaping was performed at peak-hour ride times on days with good weather. Video data were compared with observations conducted before videotaping, and there was no evidence that the presence of the camera affected rider or motorist behavior. Each bicyclist traveling through a site was an observation, while each vehicle traveling through a site in the presence of a bicycle was also an observation. Videotapes were analyzed to code signaling, slowing and stopping, and yielding behaviors for both bicyclists and motorists, as well as head-turning or scanning behavior for bicyclists only.

Videotapes were also analyzed to code conflicts “before” and “after” treatment. Conflicts were defined as an interaction between motorist and bicyclist where at least one of the parties had to make a sudden change in speed or direction to avoid the other (a stringent definition).

Bicyclists’ opinions on the treatment were solicited through an in-the-field, oral survey of 200 riders who had just traveled through one of the sites. A survey was also mailed to about 1,200 owners of vehicles who had been spotted driving through the same site as determined from license plate numbers. Responses were received from 222 of the vehicle owners. Additionally, city staff members performed test rides on wet treated surfaces to evaluate slipperiness. The sites were also informally evaluated for durability and wear of the markings.

As mentioned above, the painted markings did not last more than two months at high traffic locations. Almost a year after the thermoplastic treatments were applied, six of those eight locations showed little wear. One was in fair condition, and one was in poor condition because it may have been installed incorrectly. Thus, the higher cost for thermoplastic application may be offset by greater durability and lower maintenance costs. Neither the paint nor the thermoplastic was slippery, but neither material was as visible at night as had been expected.




Motorist behaviors changed significantly in one or more ways at most sites. From the data pooled across sites, significantly more motorists slowed or stopped at the conflict area in the “after” period than in the “before” period (87 percent after compared to 71 percent before). Fewer motorists signaled their intentions after the blue pavement was installed (63 percent after compared with 84 percent before), but this result could partially be because the motorists yielded more frequently.


Most observable bicyclist characteristics (age group, helmet use, passengers carried) remained the same for the before and after periods, with the exception that there were 29 percent females before and 21 percent in the after period over all the sites. The percentage of bicyclists following the marked path through the conflict areas significantly increased over all sites from 85 percent before to 93 percent after the blue markings were added. Bicyclists slowing or stopping on approach to the conflict areas decreased from 11 percent to 4 percent after the treatment. Reduced slowing is interpreted to signify bicyclists’ increased comfort in approaching the conflict areas.

Some desirable bicyclist behaviors decreased, however, after the treatment. Considerably fewer bicyclists turned their heads to check for motor vehicle traffic after the treatment than before (43 percent before, 26 percent after). Additionally, as with motorists, fewer bicyclists (4 percent) used hand signals to indicate their intended movement after the blue pavement was installed, although few bicyclists (11 percent) used hand signals in the before period either. It also should be noted that bicyclists would not be expected to signal at sites where they were riding straight ahead (all but two of the sites).

Motorist and bicyclist interactions

A significantly higher percentage of motorists over all sites yielded to bicyclists after the blue pavement was installed — 92 percent in the after phase compared with 72 percent in the before period. Conflicts, as defined in this study, were infrequent in both periods, with eight coded in the before period and six coded in the after period. Conflict rates were therefore quite small — 0.95 per 100 entering bicyclists in the before period. This rate decreased to 0.59 per 100 after the blue pavement was installed.

There were differences by site and by type of site (group) in some of the outcomes noted above (for full report and analyses, see Hunter, et al. 2000). For example, after blue pavement was installed for the group 1 and group 3 sites described above, the percentage of bicyclists using the marked pathway increased significantly and the percentage of bicyclists slowing or stopping decreased significantly. Also, the percentage of motorists yielding to bicyclists increased significantly. Unfortunately, bicyclists turned to check for traffic less frequently at those groups of sites. In the group 2 sites, where motorists were shifting into a right-turn lane across a through bicycle lane (as opposed to entering or exiting the roadway), cyclists actually increased their scanning behavior and motorist signaling also increased significantly. The percentage of bicyclists using the painted area at the group 2 sites decreased after treatment, and motorist yielding did not change significantly at the group 2 sites.

Survey results

The majority of bicyclists indicated the following:

A majority of surveyed motorists noticed the blue markings and the signs. More motorists who noticed the signs also correctly interpreted that the blue pavement meant they should yield to cyclists. Nearly 50 percent of the motorists who responded said the treatment helped increase awareness of the conflict areas, while others expressed concern about creating a false sense of security for bicyclists.

Conclusions and Recommendations

These results suggest that colored bike lanes and accompanying signs may be one way to heighten both motorist and bicyclist awareness of some types of intersection and merge conflict areas, thereby creating a safer riding environment. Motorist yielding behavior increased overall and at six of 10 individual sites. Slowing by bicyclists approaching the conflict areas also decreased, signaling an increased comfort level among cyclists. Some of the treated areas still are in good condition, even five years after the thermoplastic markings were installed. Some are somewhat worn, but still functional. Others are greatly worn where traffic is heavy. The thermoplastic coloring seems to last two to three years in places with heavy traffic. Five years following installation, Portland’s bicycle coordinator still has a high opinion of the value of the blue pavement markings. He has more sites identified for implementing this treatment when funds become available to install and maintain them.

More evaluations are needed of the use of this treatment as well as when and where such applications are appropriate, the effects and use of signs with markings, and the types of materials and colors that should be used. Additionally, bicyclists should be encouraged to continue their vigilance and scanning behavior after colored pavement markings are installed in conflict areas.

Costs and Funding

Painted sites materials and labor $900/10 sites
Blue thermoplastic sites materials $9,700
Blue thermoplastic sites labor $6,300
Total $16,000/8 sites (1998)



Mia Birk, Principal
Alta Planning + Design
144 NE 28th Ave
Portland OR 97232
(503) 230-9862

Roger Geller
Bicycle Coordinator
City of Portland Office of Transportation
1120 SW 5th Avenue, Room 800
Portland, OR 97204
(503) 823-7671 (voice)
(503) 823-7609 (fax)
(503) 823-6868 (TDD)

The modification (blue bike lanes) that is the subject of this case study is not compliant with the Manual on Uniform Traffic Control Devices, nor is it currently being considered for inclusion. Accordingly, it is imperative that any jurisdiction wishing to utilize blue bike lanes (or any other non-approved traffic control device) should seek experimental approval from the Federal Highway Administration. For information on how to do so, please visit this Web site: http://mutcd.fhwa.dot.gov/kno-amend.htm.