In defense of neighborhood traffic circles

Neighborhood traffic circle in Berkeley, California

In a post to an e-mail list, I made the statement:

On narrow streets in residential areas…small traffic circles at intersections can slow traffic and reduce the temptation to use those streets for through travel…

…which elicited the following response from Ken O’Brien, whom I consider a friend, but who is also probably the most hard-core supporter of equal treatment for bicyclists and motorists I know. He rejects all special treatments, including ones which I regard as beneficial.

I’m very disappointed to read John Allen supports these foolish structures.

These bogus structures encourage left turning and straight-through traffic to enter intersections swinging right (to avoid and go around the obstacle). They encourage traffic to perform lateral merges (and a lateral merge away from the direction they intend to travel) immediately at the intersection. They break the rule that you want traffic to prepare for lateral merges early and separately from scanning for conditions ahead immediately at the intersection.

Please, please Mr Allen, reconsider your support for these illogical attempts at traffic calming.

I responded:

I can always count on Ken O’Brien to make principled and consistent comments based on traffic theory, and I would agree with most of his comments if such structures were installed on streets with fast or heavy traffic. In fact, one of my complaints about the reconstruction at the rotaries on Concord Avenue in Cambridge [Massachusetts] was the narrowing of the travel lanes, resulting in some of the very problems he describes.

I think that Ken and I can agree that rotaries do not inherently defy the rules of good, simple intersection design. A rotary [the Massachusetts name; also, “traffic circle” and in its modern, improved version, “roundabout”] is a street with several other streets entering at T intersections from the right, and the T intersection is a very ordinary type of intersection.

Drivers must merge with traffic in the rotary and then go around to the desired exit. When Ken says that drivers must make “a lateral merge _away_ from the direction they intend to travel” — well, they intend to travel around the rotary, and so they are merging into the traffic flow of the rotary. As drivers enter the rotary, conflicting traffic comes from only one direction, the left, rather than from three directions as in a “crossroads” intersection. Once in the rotary, it comes from only one direction, the right. To this extent, I do not agree with Ken’s comments about traffic flow. True, traveling straight through or turning left is more complicated and slower in a rotary than at a “crossroads” intersection — if there is no conflicting traffic — but that fact does not imply any inherent disagreement with ordinary traffic rules.

Bicyclists are impeded by the small rotaries much less than motorists are. The only maneuver that is slower for a bicyclist is the left turn: a bicyclist must go around the rotary to make a left turn. Because of the narrowness of a bicycle compared with motor vehicles (other than motorcycles), the small rotary does not slow a bicyclist’s travel at all for through travel or right turns.

Ken says that the small rotaries that “break the rule that you want traffic to prepare for lateral merges early and separately from scanning for conditions ahead immediately at the intersection.”

Actually, as these rotaries are typically on two-lane, two-way or one-lane, one-way streets, they don’t require any merges at all before the intersection, except by operators of slow, narrow vehicles such as bicycles. A bicyclist should merge into the traffic flow before going around the rotary, but then a bicyclist should do this for through travel at any rotary [and at many if not most other intersections]. The merge is particularly easy at the small rotaries because motorists, with their need to steer right around the center island, must slow to bicycle speed before entering the rotary. I’ve been in Berkeley [California] repeatedly and ridden through such rotaries, and I did not get the uneasy feeling I always get with bikeway junctions which defy the rules of traffic flow, the feeling that drivers, including me, need to have eyes in the back of our heads or X-ray vision to see traffic that might conflict with the movements we are preparing.

Now let’s step back a bit and look at the larger issue of reduction of traffic speed and volume in residential neighborhoods.

Wherever residential streets provide useful shortcuts or alternative routes, and especially in cities like Berkeley with a grid traffic pattern, use of residential streets by through motor traffic becomes a nuisance and a hazard. Residents demand a solution. What solutions are available?

Attempting to ban non-resident motorists would be unworkable, and a violation of the basic right of free travel. Location-specific electronic monitoring and control of vehicle speeds is an idea whose time has not yet come. Low speed limits and traffic-law enforcement can work if they have enough political support, but budgets for law enforcement are often inadequate, and while residents want to prevent speeding in their own neighborhoods, they want to avoid speeding tickets in other neighborhoods. Everyone has only one own neighborhood but there are many other neighborhoods, and so the political force to weaken enforcement usually wins out. We’ve seen this happen recently with challenges to photo-monitoring and ticketing of speeding motorists.

Because the “soft” measures aren’t very effective, residents demand, and you will find, some form of traffic calming in almost any residential neighborhood. Then the question becomes: what form of traffic calming?

The “spaghetti” pattern of curved streets in many newer suburban developments? Well, that can’t be retrofitted onto existing neighborhoods, and it’s confusing too — getting lost on the curved streets is very easy. Cul de sac development patterns? They force everyone including bicyclists to take long, roundabout routes on major arterials. Conflicting one-way signs from one block to the next (a favorite approach in the Boston, Massachusetts area)? That’s good at reducing the traffic volume on the residential streets, but works sorely to the disadvantage of bicyclists.

Bicycle-permeable barriers and diverters to break up through routes? Berkeley has some of these and I think they have their place, but they also can pose hazards if they result in nonstandard traffic movements, and they pose an issue of quick access for emergency vehicles. Speed humps and speed tables? I think they also have their place, but they do pose the issue of possible damage to vehicles. Bulb-outs, narrowing of the travel way, chicanes (making the traveled part of the street weave from one side to the other, for example by alternating parking on one side and then the other)? I think these measures have their place too, but they can work sorely to the disadvantage of bicyclists if the travel way is excessively narrowed. Snow clearance also becomes more difficult as the street gets more complicated.

All in all, I think that for bicyclists, the small rotaries are one of the least disadvantageous and most advantageous forms of traffic calming. By preventing motorists from traveling through intersections at high speed, the small rotaries succeed in reducing the speed and volume of traffic on the residential streets very substantially. Through-traveling motorists are discouraged from using these streets and are more likely to use arterial streets instead.

I’m sure that the success of the rotaries depends on design details; in particular, the traffic island must be large enough to accommodate the turning radii of vehicles that use the intersection. Success also depends on location. I have indeed seen one such rotary that failed — installed on Concord Street in Wellesley Lower Falls, Massachusetts and was made of collapsible, reflectorized poles. I wish I had a photo of it, but vehicles damaged it and it had been removed within a few weeks, before I got back to it with a camera.

Concord Street, which becomes Park Street in Weston, is a minor arterial street with a connection at each end to a numbered state highway and another connection in the middle to the Massachusetts Turnpike. As much as residents living along this street might have wanted otherwise, the needs and desires of through-traveling motorists prevailed. The small rotaries in Berkeley, on the other hand, do appear to have succeeded.