From Periphery to Center: Does Bike Redistribution Work?
D.C.'s Capital Bikeshare program. Photo via Park View, D.C.

How can bikesharing programs, like D.C.'s Capital Bikeshare, ensure that stations have an even distribution of bicycles? Photo via Park View, D.C.

The Transport Politic‘s Yonah Freemark has been writing recently about the efficiency of bikesharing models that major cities around the world have been adopting. He focuses on the issue of redistribution.

Bikesharing systems have opened in cities, such as Denver, Co.; Minneappolis, Minn.; London, England; Montreal, Canada, Melbourne, Australia, Shanghai, and Washington, D.C.   Boston, Mass.New York City, and Budapest, Hungary are also planning, seeking funding or studying the potential for bikeshare systems. (For a more comprehensive list go to Wikipedia or check out this blog based in D.C.)  As bikeshare systems gain in popularity, so do logistical concerns. Bike availability at stations and open spots for depositing bikes has become an issue for a number of cities. The New York Times documented the occurrence in Barcelona’s bike program, called Bicing, in 2008.

Most bike programs account – at least in part – for the discrepancies in bike movement by  manually recirculating bikes.

Too Many or Too Few in London

Still, cities with bikeshare schemes are experiencing too many or too few bikes in certain areas. Freemark says London’s system, Barclays Cycle Hire, is not working in specific locales due to the directional flow of commuters. Job centers and residential areas are isolated from one another. According to Freemark, “this may put a strain on bike sharing, since to work, the concept requires a relatively even pattern of bike pick-ups and drop-offs at every station.”  Stations like King’s Cross and Waterloo Stations experience peak usage when dozens of extra bicycles are left undocked by users.  To anticipate high usage the system is also leaving dozens of extra bikes available for users.  This video from the BBC shows workers of the bike system simply presiding over “dumped bikes” at the King’s Cross station.

Looking at the  real-time map system of Barclays Cycle Hire in London, which we recently wrote about, you can see some stations have no bikes and others are totally full. (Again look in “grid view” to most easily see bike availability.) For another view, Freemark highlights a visualization of bike availability by station in Oliver O’Brien’s spatial analysis blog, Suprageography.

Another Issue: Bike Station Density

For Washington, D.C., the bikesharing system Capital Bikeshare has been in place for only a few weeks. The bikes span the city and Arlington County – 114 stations and more than 1,000 bikes. There are plans to expand it.  Below is an image from Transport Politic that shows, in the top row, the densest areas of bike station placement in one square-mile areas of Washington D.C., Montreal and Paris.  In the second row are the least dense areas for bike usage in the three cities, which include the neighborhoods of Anacostia in Southeast D.C., southeast of Parc Maisonneuve in Montréal, and Montreuil, east of Paris.

Photo by The Transport Politic

Photo by The Transport Politic

The Transport Politic sums up his findings:

The charts demonstrate the fundamental difference between Washington’s proposed system and those in Montréal and Paris. In the center-cities, the French-speaking cities have roughly three times the densities of bike stations as the District proposes; in areas far from downtown, the difference is even more pronounced. Indeed, the minimum density of stations anywhere in the Paris or Montréal bike-sharing zones is higher than the maximum density promoted for Washington.

Low density for bike stations  is a problem for a number of reasons, according to Freemark:

  • short neighborhood commutes become more difficult which reduces chances to attract occasional riders
  • insufficient density can increases the likelihood that stations run out of bicycles run or out of dock spaces
  • Availability in Real-Time

    D.C. has opted to spread out its stations, unlike Paris where almost every station is within 200 meters.

    A check of bikeshare programs in D.C. using its bikeshare availability app, on a sunny but chilly afternoon, revealed that stations appear to be either mostly full (seven to 10 bikes) or mostly empty (two or less.) A few bike share stations in the neighborhoods of Columbia Heights and U Street had very low bike availability, as did two stations near the Crystal City Metro stop in Arlington, Va., and a station in Anacostia (an area that Freemark highlights in his map above.) One station on Independence Avenue and 12th Street in downtown D.C., near the USDA office, was totally empty for a number of hours according to the application. Still no stations were totally full, i.e. unaccepting of bikes.

    Downtown D.C. bike availability on October 5, 2010. As you can see a few stations are empty or close to being empty. Photo via Rhys Thom's i phone.

    Downtown D.C. bike availability on October 5, 2010. A few stations are empty or close to being empty. Image via iPhone screenshot.

    Chris Holben, bikesharing project manager for the District Department of Transportation (DDOT), said that the contracted company by DDOT and Arlington County, Alta Bicycle Share, responsible for installation, maintenance and redistribution of bikes, “found out right away the system could not run itself.”

    The company’s contract states explicitly that a station cannot be full or empty for more than three hours at a time between the hours of 6:00 a.m. and midnight. Obviously, the company’s goal is to try to beat that. They have laptops in their trucks that map the bike distribution in real-time.

    Complex Commuting Patterns

    Even with mobile apps and online data like those in London, a full or empty station leaves a rider without a nearby option when parking or picking up a bike, which is especially inconvenient in D.C. where bike spacing is more spread out.  Plus, lots of people don’t have the convenience of checking station availability on their smart phones.

    Freemark’s analysis of Montreal and Paris suggests “you either have to put a lot of stations in a community, or not serve it at all. It’s the low station density middle ground that causes problems.”

    But we also have to remember D.C. is far smaller than Paris and Montreal. D.C.’s daytime population pulses to nearly a million – 172% of its nighttime population. By c0ntrast the central city of Paris has about 2 million inhabitants within city limits, and the metropolitan area has about 11 million. In Montreal proper, the population is close to two million and the metropolitan area nears four million.

    Freemark concludes that the cities most amenable to bike sharing are those in which urban districts are not separated by usage, but rather those that support integrated, mixed urban use.

    But D.C. officials are more optimistic than Freemark. DDOT’s Holben says he thinks the majority of  the District’s users will employ the bikes for commuting, which means downtown stations will be full and the suburbs empty. Despite the one-way strain, he says, “We will have to redistribute – there’s going to be movement we don’t necessarily know about.”

    Plus, all of these bike sharing schemes are relatively new, and it is safe to assume the early adopters are commuters. As the bike sharing programs gain in popularity, tourists, students and other casual users may more evenly redistribute the bikes. In Paris the distribution problem for Velib was largely centered around hills. Communities and neighborhoods at the tops of hills were empty during the first year of the program, but Velib added a simple behavioral fix: “Return your bike to one of 100 stations perched over 60 meters above the rest of the city, and 15 minutes of free riding is added to your account.” So far it’s been successful.

    It also seems more data on commuting patterns, such as the rise of reverse commuting, or usage in specific neighborhoods would benefit bikesharing programs and alternative transport options, in general. Ultimately these systems are designed to benefit the user. Holben says, “we’ve been lenient thus far. We don’t want you to abandon the bike.” Or be without one. So he says, “you can call us.”

    Holben cited one commuter who had a meeting to attend but no space for her bike, so Capital Bikeshare stored her rented bike and she called when she was done with her meeting. Obviously it’s not the norm, but hopefully these innovative systems will work out their kinks.

    Print Friendly

    • Pingback: The Future of Sustainable Urban Mobility: Integrate, Integrate, Integrate | TheCityFix.com

    • Pingback: The Future of Sustainable Urban Mobility: Integrate, Integrate, Integrate | thecityfix.com

    • Pingback: WANTED: Bike Share in New York City! | TheCityFix.com

    • Pingback: WANTED: Bike Share in New York City! | thecityfix.com

    • Merin

      People should not overlook the lessons learned from SmartBike. I communted to and from work every day over a two-year period using SmartBike: U Street to Metro Center and vice versa. In that period, there were maybe four times when the bike rack at U Street was empty in the morning, and not one time when Metro Center was empty in the evening.

      Managing that flow just depends upon two things: tracking user demand/habits and anticipating user flow. Obviously the latter is dependent upon the former. It seems to me that SmartBike got that formula down pretty quickly. I think Capital Bikeshare will too once they’ve got enough data to track and make assumptions of user behavior.

    • Adam L

      What does it matter how many people commute into D.C. from Maryland and Virginia? The Capital Bikeshare system is meant to be used by people who live and work in either the District or Arlington, not for the 400,000+ people who commute from the suburbs. The actual comparison when trying to determine how to manage a bikesharing system should be the number of current and potential bicycle commuters who both live and work in either D.C. or Arlington.

    • Adriana

      The whole reason why Montreal’s Bixi works so well is simple convenience.

      1) Stations are everywhere – if one is full the one located a block over probably isn’t.

      2) It feels safe to ride – not only does central Montreal have a great lane network, much of it separated from cars in some way, but drivers are incredibly aware of cyclists and respect those who share the roads that have no separated lanes.

      Number (2) evolves, but (1) must be met – using Bike share has to make people’s lives easier – otherwhile they just won’t use it.

    • http://thecityfix.com/members/paytonc/ paytonc

      @Mark: That wasn’t an oversight, but just a matter of funding. The initial bike stations were paid for by DDOT and Crystal City BID, which is why all of the bikes are in the District and CC. Phase 2 will ask USDOT to double the size of the system, adding many more stations throughout Arlington — but in the meantime, you can ask your local authorities (Arlington County, the BIDs in Ballston, Clarendon, Rosslyn, and Columbia Pike, local universities) to chip in to add more stations. The capital cost is around $10K per station, IIRC.

    • http://thecityfix.com/members/paytonc/ paytonc

      Great piece. Bike sharing functions as a walk extender. As such, having a high density of stations is analogous to having high-frequency transit service in terms of usability. Having stations placed every 3-5 minutes along someone’s walk route is much more useful than having stations every 10 minutes — particularly when some stations are full/empty.

      As for urban form, I’ll be curious to see how well CaBi works when the downtown stations are fully up (they seem to be avoiding the area for now). Walking and biking work best in mid/high-density, high-mixed-use areas: trips are short, space is at a premium, and origins/destinations are highly dispersed. DC’s mid-rise downtown probably works better for this than other American cities, but only time will tell.

    • http://blog.robpitingolo.org Rob

      Jonna, really good post.

      This is something that I’ve really been wondering about since they started installing CaBi stations around DC. Last weekend I rode past 3 stations. Two of them (one in Dupont, one in Columbia Heights) were completely empty. Another station in Columbia Heights was about half-and-half.

      I seriously question if bike sharing can be a reliable means of commuting to work. Since more people work in the central business district than anywhere else that’s served by CaBi, it seems like the risk is that all the downtown stations get jammed up every morning and it becomes difficult for people to find a place to drop the bike near their destination. Or, in the afternoon, that downtown stations are empty as people take CaBi bikes out to the neighborhoods where they live.

    • http://thecityfix.com/members/ericaschlaikjer/ Erica Schlaikjer

      Thanks for catching our error! We have corrected the statement to read, “D.C.’s daytime population pulses to nearly a million — 172% of its nighttime population.”

    • Mark

      I still think it’s crazy that the Rosslyn/Courthouse/Clarendon Neighborhoods were left out of the launch — especially given the proximity to Georgetown and other parts of DC.

      Once again way to much time is spent with commuters (if you were commuting wouldn’t you be 10x more likely to buy your own bike) and not enough time spent on the casual biker who a well designed program makes a ton of sense for.

    • http://landsharkblog.com Tyler

      Very interesting thoughts… I go by the one at the Navy Yard metro and just last night/this morning, I was checking out the usage and positioning of the bikes. There were 5 there yesterday afternoon, with 2 rows of space. But when I went by this morning, there were no bikes! That’s a lot of empty spots.

      I wonder how many bikes they put at the stations when they open them. They must leave a few empty spaces, or are there bikes in the system for every space at a station?

    • math nerd

      I have to object to the statement that DC’s daytime population is “up by 172% from its residential population.” I think you mean “up by 72%” – if the day and night populations were the same, you wouldn’t say “up by 100%.”

    • Pingback: Morning Notes | ARLnow.com

    • http://thecityfix.com/members/jmckone/ Jonna McKone

      Hi Kelly,

      Thanks for posting! I guess I should have pointed out the station at Columbia Heights or the one in Crystal City – definitely make sense that the stations were empty then!

    • kelly

      “One station on Independence Avenue and 12th Street in downtown D.C., near the USDA office, was totally empty for a number of hours according to the application.”

      Of course, this station was just installed today and so it isn’t too surprising that it is empty… Based on the pictures of how the stations are installed, it looks like they don’t bring bikes with the station, but rather the bike crew comes around after the station install crew.

      (The other noticably very empty station in Georgetown was also installed today!)