Why Not Fix the Flatbush/Atlantic/Fourth Hellscape With a Traffic Circle?

The Flatbush/Atlantic/Fourth traffic circle concept from Perkins Eastman, with bike lane in blue.
The Flatbush/Atlantic/Fourth traffic circle concept from Perkins Eastman, with bike lane in blue.

Could a traffic circle tame cars and trucks at the chaotic intersection of Flatbush, Atlantic, and Fourth avenues in Brooklyn? A team of architects thinks so.

Earlier this week DOT held a public workshop about improving pedestrian safety in this area, where heavy traffic on wide, two-way streets meeting at irregular angles creates exceptional hazards. Motorists injured 78 people walking and biking there between 2010 and 2014, according to DOT, and have killed four pedestrians and one cyclist in the project area since 2008. Over half of pedestrian injuries occurred while the victim was crossing with the signal.

DOT’s draft plan would add pedestrian islands, curb extensions, and crosswalks, while leaving the basic geometry of the streets intact.

Perkins Eastman transportation project designer Jonathan Cohn, who lives in Park Slope, presented a different idea at the DOT workshop — a traffic circle.

It’s a rough concept that has yet to be fleshed out in detail, but one that merits strong consideration. With a traffic circle, left turns across multiple lanes of traffic would be eliminated, which should provide a major safety boost. The traffic pattern would be simpler for pedestrians to negotiate. Cohn’s concept also calls for a two-way bikeway around the edge of the circle.

“The flow of vehicle traffic becomes predictable,” said Perkins Eastman’s Justin Wolf. “It’s steady, it’s predictable, and that creates a safer environment.”

DOT staff at the workshop said there’s not enough space at Flatbush/Atlantic/Fourth for a traffic circle, but Cohn thinks that obstacle is surmountable, the Brooklyn Paper reports. He suggested the city could “take a bit” off the lot currently occupied by a PC Richards, which developer Forest City already plans to replace with an office tower. A safer, more welcoming pedestrian environment at the doorstep of the tower might be worth whatever tradeoff Forest City might have to make.

It’s worth a thorough study to see if a traffic circle could work at this location, even if it requires some negotiations with property owners.

For a very rough idea of how the Flatbush/Atlantic/Fourth site compares to two other large traffic circles in NYC, take a look at these Google satellite views at the same scale.

Above is the Flatbush/Atlantic/Fourth convergence…
…and here is Park Circle, at the southwest corner of Prospect Park, which has a two-way bike lane around most of its perimeter.
This is Columbus Circle at the southwest corner of Central Park. Note the central area is smaller than the grassy area in Park Circle. It still manages to be a well-used public space despite the overbuilt roads encircling it…
The public space in the middle of Columbus Circle. Via Google Street View
  • bggb


  • Bernard Finucane

    Excellent idea, but the building on the triangle ends up a little off center, which is why I hesitated to suggest it.

    Also it would still require bending Flatbush to meet Atlantic at a perpendicular. How else would that intersection work?

  • Jonathan R

    How do you keep vehicles exiting onto eastbound Atlantic Ave from crashing into entering vehicles from westbound Flatbush Ave? Traffic lights like at Columbus Circle? Seems like that would back up traffic a great deal.

  • Matthew

    this appears to mean traffic heading south on Flatbush could no longer make right turns to head south on 4th avenue. which is a wonderful idea, although MTA buses (B103-Ltd and B37, I believe) make that turn to serve 4th ave. they could probably just go south on 3rd ave, then turn east on Dean to catch up to 4th Ave.

  • J

    I’m certainly willing to entertain the idea of a circle, but Columbus Circle is not exactly a model of high-quality urban space. Park Circle has gotten a lot better but the inside is kind of wasted space. It would be interesting to see how the % of usable space, and the crossing distances, ped detours, and opportunities compare between the proposals. RIght now there isn’t enough information for me to say one is better than the other.

  • In a comment on another thread, I mentioned my discomfort with the two traffic circles at Prospect Park, with one on Kings Highway, and with the many in Washington. I totally forgot about Columbus Circle.

    Even with the signals, dealing with that circle on the bike is not very easy. If you’re coming north on Eighth Avenue, you’re probably on the left side because that’s where the bike lane (which stops a few blocks short of the circle) is. This puts you flush against the centre of the circle, where there is no room; so you’re forced to take the lane. But then, whether you want to go to Broadway or to Central Park West, you have to cut to the right across several other lanes. Unpleasant.

    If you approach the circle from the right edge of Eighth Avenue, then forget about Broadway; you’re going to Central Park West.

    It’s just not clear to me why people think that traffic circles are so good at calming traffic and at creating a nicer place for bicyclists. I think that it’s better to just let these streets remain straight lines, and to control the access to the intersection with traffic signals that give traffic from every street a straight phase and a turning phase.

  • ItsEasyBeingGreen

    Columbus Circle (and Frederick Douglass Circle) should just get a protected bike lane already, considering that 8th Avenue has one and CPW, the continuation of 8th Ave, has a fairly OK functioning Class II lane.

    I’m not sure what makes you uncomfortable about the protected lane at Park Circle considering its completely phase-separated. It’s inconvenient but not uncomfortable. It’s not comfortable only as a potential jaybiker (which I know you are not) and jaywalker since the exits from Park Circle are still at quite oblique angles rather than right angles.

  • Danny G

    If the multiple subway lines running underneath those streets don’t make a traffic circle unfeasible from the get-go…

    You might be better off moving whatever future high-rise buildings that are planned to replace Modell’s and PC Richards inside the circle, shaving a little off of the mediocre Target mall building, and stretch the circle east into a large oval so that Flatbush and Atlantic can intersect this oval separately.

    I know that’s a lot to digest without a diagram.

  • ItsEasyBeingGreen

    The wonderful thing about this Flatbush/Atlantic/4th proposal is the right angles at all arms off the circle. The existing circles of NYC in the other pictures still don’t get this right and have the arms at oblique highway offramp angles which encourage people to drive through the turns at full speed.

  • Yeah, I guess Park Circle is not so bad. It’s definitely the best of the bunch that I mentioned.

    Maybe the problem there is in my head. I’m just not used to using traffic circles; so even a relatively benign one just feels a bit weird.

  • van_vlissingen

    Wouldnt it be easier to close Flatbush to everything but busses and people from here to GAP? (Truck route would need to be remapped)

  • Larry Littlefield

    There aren’t a lot of truck routes in Brooklyn. This is two of them. And that has to be considered.

  • Bernard Finucane

    This is how the intersection could be made more viable. The trick is to curve streets when they intersect at an oblique angle to make them perpendicular.

  • kevd

    Since it has flatbush and Atlantic merging SE of the traffic circle, it won’t really work.

  • kevd

    Jonathon, they just drew a circle on the map and sent out press release. They didn’t actually figure out the details!

  • Jonathan Cohn

    Heading NW from Altantic and Flatbush, entering the circle would be controlled by lights before the merge.

  • Hilda

    Traffic circles; It has worked in other places in NYC, but it really works globally. Also, please note that Jonathon Cohn is also the architect that proposed closing Broadway, in Manhattan, to create much needed green space. Truly someone thinking outside the box if you look locally, but his ideas are very well supported if you broaden your outlook.
    Unfortunately I was not able to attend the meeting, but any good designers job is not to denounce an idea, but to determine the obstacles, concerns and feasibility of making that idea work. Sometimes it is found that it is not worth it. But sometimes it’s possible to figure out a way to get it done, and the work is in selling the idea, not simply figuring out how to make it work. Protected bike lanes, prior to 2009, had not been done before in the U.S., but were prevalent all across many European countries as typical, if not essential.
    As advocates, we shouldn’t let the city dismiss a good idea, smoky because it might not be easy to push through. We should insist that it is studied and, if proved to have merit, should be supported.
    Otherwise, what the hell are these workshops for?

  • kevd

    So it isn’t really a traffic circle and lacks all a traffic circle’s advantages. It’s just a signalized intersection that is round (-ish)

  • Miles Bader

    Also, my personal experience is that they really suck for pedestrians…

    I lived for years in the UK, where roundabouts are de rigueur, and while I found them kind of fun on a bike (zoom right through, nice swooping action), I hated them as a pedestrian, because cars never seemed very careful entering or exiting, and there were no guaranteed times when traffic was actually stopped for you to cross.

    The other problem as a pedestrian was that the sight-lines were all screwed up, because of the circle you were often in a place where you couldn’t see along both axes of the road, and there it was often very confusing where you needed to be looking to anticipate traffic.

    Together with the amount of space required (which doesn’t work well in an urban context, and the space required often comes at the expense of pedestrian use), and I’ve always been a bit mystified as to why so many walkdable-streets/bicycle advocates seem to love them… I’ve always wondered if they’ve actually ever used one…

  • Traffic circles can be signalized. You’re thinking of a roundabout.

  • Joe R.

    Which is why roundabouts make more sense. They’re also smaller in diameter, and so have a greater traffic calming ability as a result.

  • Joe R.

    Honestly, if you’re going to start sticking in traffic lights, you might be better off with a small-scale version of something like this:


    Besides being a signature item, it’ll let pedestrians and cyclists cross this clusterf*ck of an intersection without being delayed waiting for cars.

  • Joe R.

    It’s sort of like what people say about democracy, namely that it’s not a good form of government until you’re tried all the others. Traffic signals absolute suck for pedestrians and especially cyclists. They increase trip times by two to five times, they increase the energy a cyclist expends, they make walking or cycling unnatural and unpleasant with the repeated starting and stopping. Even worse, statistically they don’t even improve safety despite perceptions to the contrary.

    There are several nice things about roundabouts:

    1) Little or no delay compared to traffic signals.
    2) The geometry enforces safe, slow speeds at junctions.
    3) Left turns are just as safe as any other type of turn.
    4) There are only three potential points of conflict-the crosswalk before entering, the entry to the roundabout itself, and the crosswalk when exiting.

    In NYC roundabouts seem more appropriate for arterial-arterial intersections. The room exists for them there, plus these intersections often have very long light cycles to allow pedestrians to cross. For arterial-minor street intersections putting a yield or stop sign on the minor street makes more sense. Judicious use of roundabouts at major intersections, plus elimination of traffic signals in between these intersections, would make walking and especially cycling much faster and more pleasant.

    That said, sure, roundabouts probably fail when you have large numbers of pedestrians, motorists, and cyclists but then so do signalized intersections. Point of fact there really isn’t any good way to deal with large numbers of street users with disparate needs short of grade separating them from each other. As much as I’d love the concept of pedestrian and cycling sky bridges, NYC probably would never spend the money on it.

  • Syd Chan

    What about getting a citizen’s group to pay for a traffic engineering study from a private engineering firm to see if a traffic circle is feasible at that location, with the volume of traffic that passes through? It’s easier to convince NYCDOT to make a change with a traffic study in hand that they can easily digest and understand.

  • Bernard Finucane

    See my comment in the thread for a solution to this problem.

  • Miles Bader

    I disagree, I’ve tried both, and found roundabouts significantly worse for me as a pedestrian. Not just in crowded areas either, even isolated little roundabouts were annoying.

    I did like them as a bicyclist however.

    Remember that a pedestrian has essentially no penalty from a short stop.

  • Joe R.

    In cases like the intersection we’re talking about it’s not a short stop. In case you’re unfamiliar, NYC uses ridiculously complex phasing at these 5 or 6 way intersections. That often means red light cycles of 2 or 3 minutes. I’ve approached intersections like this on my bike, noticed a light which remained red for well over a minute as I was approaching, and then went through it on the assumption the traffic signal was broken. Pedestrians aren’t going to wait 2 or 3 minutes to cross the street. That’s why a roundabout is a better answer in this case. They just need to cross one or two lanes at time.

  • kevd

    they can be. but then what is the advantage over any other intersection?
    aside from inconveniencing pedestrians?

  • kevd

    your diagram seems much better.

  • Miles Bader

    I may well be that in this particular case, a roundabout would be a better solution; one of the big problems with roundabouts is a greater amount of “perceptual confusion” compared to a conventional intersection, so in the case of an already confusing intersection (unusual numbers of streets, weird angles), it may not be any worse.

    My complaint is really against what seems to be a generally uninformed faith in the merits of roundabouts I see on forums like this.

    [For this reason, one of the worst types of roundabouts I’ve encountered are very small ones which would typically be simple unsignalled intersections if they were conventional. These seem to be super popular in Britain.]

  • Joe R.

    I’m not sure I even see the point of using a roundabout in what would otherwise be an unsignalized intersection. It may be that Britain overdoes roundabouts the way NYC overdoes traffic signals. For the vast majority of NYC intersections, roundabouts really aren’t the best answer but I think judicious use at arterial-arterial intersections would be a big win all around.

  • ahwr

    It can be nice to force people to slow down enough to accommodate poor sight lines. You seem like the sort to prefer concrete to stop signs.


    It sounds nice. right? It is a mess when you don’t know if the person turning towards you while you are crossing the street is about to hit you so you need to jump out of the way or if they are going straight and are just forced to get close to you by the road design.

  • ahwr

    In cases like the intersection we’re talking about it’s not a short stop. In case you’re unfamiliar, NYC uses ridiculously complex phasing at these 5 or 6 way intersections.

    Sometimes. And sometimes, like Atlantic and Flatbush, they just ban left turns. Probably because traffic volumes are too high for complex signal phasing to work. But it has the effect of reducing pedestrian delay from the three minutes you’re imagining. It’s not in another country, head over there with a stop watch one day and see how long you really have to wait if you miss a light. Compare it to the traffic circles and see which is actually a longer delay. The answer might surprise you. The traffic circles shown above, as well as the high volume ones in other countries, are larger than the one shown in the drawing. You’d have to take a lot more building to fit one in here. And this is not an appropriate place to just ask drivers to yield. Pedestrians would gridlock auto traffic. Maybe you think who cares about people in cars or buses that get stuck? People will only tolerate so much delay, stressed drivers would bully their way through for the same reason you’re convinced pedestrians would be crossing against the light. It’s not something anyone would try here. Which means signals. So pedestrians will still have to wait for the light.

  • Joe R.

    Whatever the red light cycle might be in that location, it’s likely unacceptably long for pedestrians or cyclists (unacceptable by my definition being more than about 10 seconds). You could enforce failure to yield with cameras. Also, it’s always the same excuse why we can’t use any other solution but signals—namely traffic volumes are too high. Why don’t we take steps to reduce these volumes given that they cause a whole bunch of other problems like aggressive driving? It’s frankly insulting in a place like NYC for either pedestrians or cyclists to have to wait for cars to cross intersections, ever.

    Note also I’m suggesting using a roundabout in this location, not a traffic circle. It’ll easily fit. If it can’t handle traffic volumes during parts of the day eventually drivers will get the message and choose alternate routes.

  • Joe R.

    But that’s a particularly poor place for a roundabout. The best way to handle low volume intersections like that is stop or yield signs on the minor street.

  • ahwr

    Both are minor streets.

  • Joe R.

    So pick whichever one is longer and give it priority by putting stops or yields on the other street. Or just have uncontrolled intersections. Those look like very low volume streets where nearly anything would work reasonably well. I suspect the mini-roundabout was just used to get drivers to slow down.

  • ahwr

    >Also, it’s always the same excuse why we can’t use any other solution but signals—namely traffic volumes are too high.

    Also different culture.

  • ahwr

    >Why don’t we take steps to reduce these volumes given that they cause a whole bunch of other problems like aggressive driving?

    > If it can’t handle traffic volumes during parts of the day eventually drivers will get the message and choose alternate routes.

    You know this is a great way to produce more aggressive driving, right?

  • ahwr

    Changes the character of the street when cars don’t have to stop or slow down.

  • chandru

    Traffic circles without signals (as exist all over Europe) are absolutely horrible for peds. since you’re dependent on driver’s courtesy stopping (even if it’s the law) to cross. Works in France (quite courteous), and here in CA, where drivers have been conditioned since the 70’s to STOP for peds, crosswalk or no. Compare to Boston or Chicago. We might be in-between.

    So traffic signals are necessary. but you can see how it messes up traffic at PPW & 17th…too many lights & stops, many backups.

    What might work is signals only to enter the circle (which would solve the ped crossing issue) and the normal yield-to-left for cars (which most stupid Brooklyn drivers would either 1) ignore or 2) be terrified of.)

    No simple solution, though having a greenspace inside the circle is intriguing.

  • Joe R.

    Like I said, use cameras to enforce the yields. Or have a hover ring for peds and cyclists to cross. Signals on a roundabout or traffic sort kind of defeat the purpose of it.

  • Joe R.

    Then do what is done with minor streets in Europe. Bollard them off at intervals so the only people driving there are the ones who live there or are visiting someone. No reason drivers should be speeding and using streets like those as through streets. Bollarded off periodically though from motor vehicles they would function as great cycling routes.

  • Joe R.

    There really is no advantage to a traffic circle or roundabout if you signalize it. In fact, it’s probably worse than regular signalized intersection.

    This being NYC I’m not surprised so many people are suggesting signalizing a traffic circle. We would probably put traffic signals in grocery stores too if the hardware were less costly. NYC never met a traffic signal it didn’t like. The rest of the world is moving away from traffic signals while in NYC they’re still springing up like mushrooms. Just two weeks ago they installed yet another traffic signal on 73rd Avenue, a block away from two existing traffic signals which were installed a few years ago (none of the three are really needed given that the cross streets are crappy little minor streets with hardly any traffic).

  • neroden

    This sort of location — many streets coming in at random angles — is what roundabouts (traffic circles) are great for. Do it.

  • neroden

    Oh god, the micro-roundabouts, where they really just need a stop sign. England loves them. They’re horrible.

    The mega-roundabouts on expressways with two, three or even four lanes going through the circle are horrible too.

    One-lane roundabouts where multiple roads come together at weird angles are pretty damn good though.

  • neroden

    Looks good. Because there are truck routes involved here (sigh) you may actually have to provide a rounded corner for the southerly Atlantic/Flatbush intersection….

    …but I have a better idea. See that zig-zag in the lower corner of your picture, with Pacific, Flatbush, Atlantic, and Fort Greene Place? Make a second roundabout. Then combine the two roundabouts into a “bow tie” in the middle. The center would have a sort of “bow tie” shaped plaza, and structurally it would be one big roundabout — you would just be driving around a bow-tie shape instead of a circle-shape.

    Think you could come up with a picture for that? Pretty please?

    I know it would make for a convoluted drive for someone turning left from 4th onto Flatbush, but I think it would be worth it.

    Another way to look at this: straight-line northbound Atlantic into northbound Flatbush, straight-line southbound Atlantic into southbound Flatbush, eliminate the contrary direction of traffic, make the mini-block of 4th between Atlantic and Flatbush southbound only, and built a street in front of Barclays Center which is northbound only. But get rid of the lights and point all the roads into a roundabout.

    If PC Richard & Son and Modell’s Sporting Goods can be demolished, you could build a much more elegant-looking roundabout. I might suggest a much enlarged Community Garden. There would have to be a safe means to enter the center, because it would be quite large.

  • Bernard Finucane

    Well if you say pretty please 🙂 I had no idea my Paint skills were in such demand…

    I made the circles the same size.

    This is without bus stops, so a lot of the white stuff on the right would actually be grey. This setup saves a lot of spce on turn lanes. No need to demolish anything, as far as I can see.

    Also no changes at Pacific and Flatbush, but I would suggest pedestrianizing Pacific between Flatbush and 4th.

  • Tyler

    Now, if only we can get people to stop exiting the existing traffic circles from the extreme center lane… hmm….

  • GiuseppeG

    Apologies for my terrible drawing, but it feels like what you mean is something along the lines of this?


DOT Reveals a Flatbush Ave Pedestrian Safety Plan By Atlantic and Fourth

Last night DOT presented its initial concept for pedestrian safety improvements near the convergence of Flatbush, Atlantic, and Fourth avenues in Brooklyn [PDF]. The intersection is located at the center of Brooklyn’s largest transit hub, where the Long Island Railroad meets eight subway lines and four MTA bus routes. The Brooklyn Academy of Music and the Barclays […]

DOT Unveils Short-Term Ped Fixes Near Brooklyn Traffic Hub

A sidewalk addition will keep traffic from turning onto Hanson Place from Flatbush and Fourth Avenue. Streets near the Brooklyn Academy of Music and the LIRR’s Atlantic Terminal are set to receive a basket of pedestrian improvements that may get underway as soon as November. Speaking last night to the CB2 transportation committee and about […]

Eyes on the Street: Pedestrian Haven on Hanson

Courtesy of Brownstoner, here’s a recent shot of sidewalk construction between Flatbush Avenue and Hanson Place, near the Brooklyn Academy of Music. The short passage — or slip ramp, as DOT calls it — allowing cars to turn onto Hanson from Flatbush and Fourth Avenue is now closed to car traffic. This is especially good […]