San Francisco, California
Prepared by Michael Sallaberry, Transportation Engineer, San Francisco Department of Parking and Traffic; 2014 updates by Jill Mead, UNC Highway Safety Research Center
Bicycle lanes and wide curb lanes are common on-street facilities for accommodating and attracting bicyclists. As it is a goal of the city and county of San Francisco to encourage cycling as a viable transportation option, efforts are constantly made to find and create opportunities for the installation of bicycle facilities. However, with a population of about 780,000 people in a 47-square-mile space, San Francisco is a very dense and congested city where a variety of mode users compete for limited street space. While this reality is one reason that bicycling is a popular way to travel through the city, it also complicates the installation of bicycle facilities.
In order to implement the city’s bicycle route network, motor vehicle lanes must often be removed to create space for bicycle facilities (often referred to as a “road diet”). San Francisco is a walkable city where mass transit is heavily used and elevated freeways are being torn down rather than constructed. The effects of such road diets on all road users must, however, be considered and sufficiently studied before final approval and implementation.
Although road diets have been implemented to create room for bicycle facilities on at least 16 streets throughout the city, this case study will focus primarily on the experience with Valencia Street, with passing reference to another road diet on Polk Street. Additionally, experiences with proposing and studying road diet projects in general will be shared as appropriate.
Valencia Street is a 19.1 m (62 ft 6 in)–wide street through a shared-use area of mostly two- to three-story buildings with commercial at street level and residential units above, and metered on-street parking on both sides. The street lies in a grid pattern and is paralleled by four other north-south arterials. Before the project, the arterial was a four-lane street with an Average Daily Traffic (ADT) of approximately 22,000 vehicles per day. A motor coach transit line with a headway of 15 to 20 minutes travels along the street. There is a heavy pedestrian presence because the street is a popular area with restaurants, nightclubs, and a variety of shops. All intersections have signals. A photo of Valencia Street with four lanes before the road diet is shown below.
Figure 1. Valencia Street before road diet.
Though the bicycle community wanted a road diet performed along Valencia Street, the local department of transportation was not willing to reduce capacity along this important north-south corridor. Valencia Street can be used as a surface street alternative to the Central Freeway, which was damaged by the 1989 Loma Prieta earthquake. Eventually, after a series of community meetings and public hearings, the city Board of Supervisors voted on a resolution in November 1998 calling for the removal of two travel lanes and the installation of bicycle lanes and a median lane for left turns on a one-year trial basis. In March of 1999, work was completed on Valencia Street with the road diet performed from Market Street at its north to Tiffany Avenue to the south, a length of approximately 1.8 miles.
Please see figure 2 below for a picture of Valencia Street after the road diet.
Figure 2. Valencia Street after road diet.
To minimize the loss of capacity along Valencia Street and reduce the impacts to parallel streets, changes were made to the signal timing along Valencia Street and Guerrero Street one block to the west. On Valencia Street, the green time was maximized for the Valencia Street split while still maintaining time for pedestrians crossing Valencia Street. On Guerrero Street, the signal offsets were modified to promote a smoother progression at 25 mph, as the speed limit was lowered from 30 mph to address citizen concerns along the primarily residential street. The speed limit change and signal timing modifications were intended to address speeding concerns and help mitigate the likely increase of traffic along Guerrero Street.
Before the work was started, baseline data were collected for use in a before-after report evaluating the road diet. As the project was done temporarily for a one-year trial period, the results of the report would be presented at various public hearings with the project to be voted on by the Board of Supervisors. If the project were rejected, the street would be returned to its previous four-lane configuration.
Traffic volumes were recorded along Valencia Street and the four parallel arterials surrounding it to determine if there was spillover traffic from Valencia Street and where it went. The counts were taken using pneumatic devices laid across the roadway that automatically counted vehicles. The counters were installed at the same location on all five streets.
After determining the green times for Valencia Street, it was predicted that 10 percent of Valencia Street traffic would divert to parallel streets after the road diet was performed. Following is a table showing before and after ADTs for the five roadways along the corridor. As expected, Valencia Street traffic volumes dropped by 10 percent.
Collision data were also collected to determine if safety was improved with the new design. As the trial was for one year, it was difficult to come to any statistically significant conclusion for the before-after report. However, as it has now been a few years since the installation of the bike lanes, the collision data analyzed include a larger sample size.
The table below summarizes the collision data results. The values in the table are average monthly collision totals for each respective collision type, and not rates.
|Before (1/95-12/98)||After (3/99-12/01)||Percent Change|
* Collisions per month
** Bicycle collisions not included during 1996 and 1997 due to lack of reporting so the before period reflects only 1995 and 1998 data.
Total collisions declined by 20 percent, though the overall drop was less dramatic when one considers that the ADT along Valencia Street dropped by approximately 10 percent. Also, a signal upgrade project was completed along Valencia Street in 1997 that increased signal visibility and helped reduce the overall collision rate. Thus, it is difficult to come to any definite conclusions regarding the effect of this road diet on overall collision patterns along Valencia Street.
Although bicycle collisions increased by approximately 50 percent, the increase was outpaced by the 140 percent rise in ridership along the street. This net decrease in collision rate for cyclists mirrors the increased comfort cyclists report feeling along the street.
Collisions involving pedestrians dropped by 36 percent. This could be viewed as a byproduct of the traffic calming effect people along the street have anecdotally observed. With lower speeds and fewer lanes, motorists are able to avoid collisions with pedestrians more easily. According to anecdotal accounts, pedestrian volumes on Valencia Street have increased the past few years as the street has thrived commercially and attracted even more foot traffic.
Bicycle counts were taken along Valencia Street before and after. Ideally, counts also would have been taken on parallel streets to determine how much of the rise in cyclists along Valencia Street was attributed to new cyclists or to cyclists transferring from parallel routes. Also, a number of counts should have been taken to come up with an average that better accounts for fluctuations in cycling volumes that occur with time of year, weather conditions, etc.
A bicycle count taken on Valencia Street prior to the road diet showed 88 bicyclists per afternoon peak hour. After the road diet, a count yielded 215 bicyclists per hour, a 140 percent increase. As no counts were taken on parallel streets before the road diet, it is difficult to know what percentage of these cyclists were new cyclists or cyclists from parallel streets. Speaking with cyclists, however, it is clear that many were new cyclists willing to try bicycling once they saw the bike lanes installed.
Public response was recorded using a hotline voicemail system that was advertised on two signs installed prominently along the roadway. The number of e-mails and letters submitted were also considered. Care must be taken to ensure that the source of public input is considered. For instance, do 200 form letters sent as part of a mail in campaign outweigh 20 individually written letters? Regardless, the ability to directly hear from the public was instrumental in understanding how various people responded to the changes and what successes or problems were associated with the changes.
Public response to the road diet project was supportive. A hotline was advertised along Valencia Street on two prominent signs directly after the road diet. From the 286 recorded calls, 259 were supportive of the project while 27 were opposed. Of letters and e-mails received, 39 supported the project while three did not. A postcard campaign led by the local bike coalition yielded 484 supportive post cards and four not supportive.
As this was the first road diet studied in San Francisco, there were some data that could have been collected for a more complete study but were not. They include: transit travel time and delay data, travel time and delay data for motorists, double parking observations, and spot speed surveys. Other data that could have been collected for a very thorough before-after study could include: noise levels, cyclist compliance with laws, and surveys of residents, merchants, cyclists, motorists, and pedestrians.
Although the project was initially controversial within the local department of transportation and some members of the community, the general consensus is that the project is a success. Bicycling along the street has increased dramatically and has made the street the second most heavily used bicycle route in the city. Collision rates for cycling have dropped on the street. The merchants association has shown support for the road diet that has made the street seem like more of a destination rather than through arterial. Although some traffic has spilled over to adjacent streets, it is likely that much of that traffic is through traffic with no intention of stopping along the street anyway. Thus, merchants’ fears that less traffic meant less business were not substantiated, in general.
With public outreach initiated by the bicycle coalition and mandated by the nature of a one-year trial, giving stakeholders plentiful opportunities to be involved in the process was an important aspect of the project’s success. Also, the use of a trial allowed everyone to see how the project operated in real life, especially useful for skeptics. It is important to have a trial of sufficient length to allow any changes in traffic patterns to come to an equilibrium. One year is a good length, with six months as a possibly sufficient length of time. With any trial, the process should be made clear to thecommunity so that there are no misguided expectations.
As this was the first trial road diet in the city, some data was not collected that would have been helpful. The effect on transit was not sufficiently studied. Travel time and delay studies for both transit and motor vehicles would have been helpful. Also, bicycle counts on parallel streets would have provided a better picture of where the increase of cyclists originated. While speed data would be helpful on road diet projects in general, the nature of Valencia Street is such that speeds are so variable given the short blocks, the changing traffic levels, the presence of double parking, etc. that collecting consistent before and after data would have been difficult.
Although the road diet created significantly more work when it was designated a trial, it was worthwhile to study and thoroughly discuss the project. Since the Valencia Street project, the city government and public has been generally more receptive to the idea of road diets. One example of a road diet whose approval was made more likely by Valencia’s success was Polk Street, a similarly controversial project.
Polk Street is a 13.6 to 15.1 m (44 ft, 9 in to 49 ft, 9 in)-wide street with metered on-street parking on both sides. Like Valencia Street it travels through a shared-use area and lies in a grid pattern with one and two-way parallel arterials. Before the project, the street was a three-lane street with two lanes serving the heavier southbound direction. Depending on which section of Polk Street, the ADT ranged from 11,000 to 16,000 vehicles per day. A motor coach transit line with a headway of 10 to 20 minutes travels along the street and pedestrian presence is significant. Nearly all intersections have signals. Polk Street was installed as a six-month trial and also underwent a review of a before-after report. As withValencia Street, the road diet on Polk Street was also eventually approved as a permanent installation.
For paint and sign work, and labor spent writing the report, the road diet cost $130,000.
Michael Sallaberry, P.E.
Assistant Transportation Engineer
San Francisco Department of Parking and Traffic
25 Van Ness Avenue, Suite 345
San Francisco, CA 94102
(415) 554-2352 (fax)
Bicycle Hotline (415) 585-BIKE