Crash prone 'modern roundabouts'

[tradephoric]

The larger agencies are looking at the numbers and trying to figure out what went wrong. But in most cases, since the serious injury crashes are down and traffic is generally flowing better than before, they're going to move on to one of their many other fires they have to put out. Unfortunately, real-world engineering becomes a balancing act - there are always going to be trade-offs and compromises.

Agencies aren't just moving on when they see a big spike in crashes at these complex roundabouts.  The reality is many of these 3x2 complex roundabouts have been downsized to 2x2 or even 2x1 roundabouts (Superior Street & 14th St roundabout in Lincoln).  Drastically reducing the capacity of the roundabout just years after it was built doesn't sound like nothing.    The roundabouts that haven't been downsized end up near the top of crash prone intersection lists (last year 3 of the top 5 most crash prone intersections in Michigan were at 3x2 roundabouts) and agencies are left defending what is almost indefensibly high crash rates.
   
It's true, serious injury crashes and fatalities are down at the complex roundabouts analyzed in the Minnesota study, but total injury crashes rose by 6%.  Not to mention there was a 212% increase in PDO crashes.  Doing a before/after crash cost analysis, the social impacts of the roundabouts are worse than the intersections they replaced.  Similar case if you look at the social impacts of the complex roundabouts in the Region of Waterloo in the 2018 crash report that was just released.  You seem to be underestimating the impacts these problematic complex roundabouts are having.  Look at this chart of crash rates that was included in the Minnesota study.  Full dual roundabouts far and away have worse crash rates than other traffic control devices in Minnesota.

[jeffandnicole]

Questions for discussion: what makes multi-lane roundabouts broadly acceptable in the United Kingdom but not in the United States? Is the crash data substantially the same between the two countries? If not, why not?

I tried doing searching for "roundabout crash analysis in France", "roundabout crash analysis in Paris France" and the same replacing "analysis" with "data" (and a few other similar searches).  Very few relevant results appear, and of the few, the analysis just talks mostly about the data collection rather than the results.  One search led me to roundabouts in Korea, but again, no real conclusion about their safety was mentioned.

You need to know where to look for.
https://link.springer.com/book/10.1007/978-3-642-84537-6#toc - is there anything remotely comparable for US? Published in 1991, mind it..

Yeah, I didn't see that in my searching.  But even then, most results appear to be studies and predications. Other than the first result, we're still not giving an analysis of how they function. And even then, we would need to pay for the reports listed.  So after everything that has been claimed regarding how the US doesn't analyze recently built rounds, it's becoming clearer that nearly no other country, on a regular basis, analyzes them either!

[tradephoric]

A year after the 161-Riverside roundabout was modified to reduce a circulating lane of traffic, crashes have dropped from 2.5/week to 2/week.  That's still over 100 crashes a year and much higher than the 15 annual crashes the intersection averaged before the roundabout.  After modifications to the roundabout, injury accidents dropped from 37 to 15 per year.  That's a significant drop, but the pre-roundabout intersection only averaged 15 crashes total.  Of course the article makes no mention of the pre-roundabout crash rates, but here is the link to that.  Back in 2014 the city was citing a crash rate of 0.98 at the intersection to justify the construction of the roundabout.  Now even with the "modified" roundabout, the crash rate is 5.7. 

Dublin officials say 161-Riverside roundabout fixes lessen crashes
https://www.thisweeknews.com/news/20191022/dublin-officials-say-161-riverside-roundabout-fixes-lessen-crashes

[kalvado]

Questions for discussion: what makes multi-lane roundabouts broadly acceptable in the United Kingdom but not in the United States? Is the crash data substantially the same between the two countries? If not, why not?

I tried doing searching for "roundabout crash analysis in France", "roundabout crash analysis in Paris France" and the same replacing "analysis" with "data" (and a few other similar searches).  Very few relevant results appear, and of the few, the analysis just talks mostly about the data collection rather than the results.  One search led me to roundabouts in Korea, but again, no real conclusion about their safety was mentioned.

You need to know where to look for.
https://link.springer.com/book/10.1007/978-3-642-84537-6#toc - is there anything remotely comparable for US? Published in 1991, mind it..

Yeah, I didn't see that in my searching.  But even then, most results appear to be studies and predications. Other than the first result, we're still not giving an analysis of how they function. And even then, we would need to pay for the reports listed.  So after everything that has been claimed regarding how the US doesn't analyze recently built rounds, it's becoming clearer that nearly no other country, on a regular basis, analyzes them either!

Pay... this is academic research; unfortunately a lot of publications are not free; fortunately many universities have access (including access from university library by general public); even more fortunately there are other ways.... Bottom line - PM if you want to read those

As for analysis - a pretty broad one from UK:  https://journals.sagepub.com/doi/abs/10.3141/2585-02
And people are looking beyond today's needs: https://ieeexplore.ieee.org/abstract/document/7504449/authors#authors

[tradephoric]

A injury accident involving three-vehicles at the Morse Blvd and Stillwater Trail roundabout occurred in the Villages on Friday.  This comes a few weeks after a serious injury accident at a Village roundabout that left a motorcyclist in the ICU:

Driver taken by ambulance from three-vehicle crash in roundabout in The Villages

[DaBigE]

A injury accident involving three-vehicles at the Morse Blvd and Stillwater Trail roundabout occurred in the Villages on Friday.  This comes a few weeks after a serious injury accident at a Village roundabout that left a motorcyclist in the ICU:

Driver taken by ambulance from three-vehicle crash in roundabout in The Villages
~IMAGE CENSORED~

Posting the link wasn't enough? What purpose did spreading a photo of someone being wheeled away on a stretcher add to this? It's bad enough that the newspaper published it.

Further, it was a rear-end chain-reaction crash...one that could have happened just as easily at any other intersection type:

The driver of a maroon minivan at about 10:30 a.m. Friday had been northbound on Morse Boulevard at the roundabout at Stillwater Trail when it crashed into the back of a RoMac Building Supply Co. window installation truck hauling glass, according to the Sumter County Sheriff's Office. The RoMac truck was pushed into the rear of another vehicle.

[kalvado]

A injury accident involving three-vehicles at the Morse Blvd and Stillwater Trail roundabout occurred in the Villages on Friday.  This comes a few weeks after a serious injury accident at a Village roundabout that left a motorcyclist in the ICU:

Driver taken by ambulance from three-vehicle crash in roundabout in The Villages
~IMAGE CENSORED~

Posting the link wasn't enough? What purpose did spreading a photo of someone being wheeled away on a stretcher add to this? It's bad enough that the newspaper published it.

Further, it was a rear-end chain-reaction crash...one that could have happened just as easily at any other intersection type:

The driver of a maroon minivan at about 10:30 a.m. Friday had been northbound on Morse Boulevard at the roundabout at Stillwater Trail when it crashed into the back of a RoMac Building Supply Co. window installation truck hauling glass, according to the Sumter County Sheriff's Office. The RoMac truck was pushed into the rear of another vehicle.

But it did happen at roundabout, contributing to roundabout crash rate.
While I agree posting individual crash reports is meaningless, message of roundabouts safety underperforming by a huge margin is clear

[tradephoric]

Every year the roundabouts in Augusta, Maine are the most crash prone intersections in the state.  You can literally query out any year in Maine's crash database going back to 2003 and the Cony Circle and Memorial Circle will top the list (MaineDOT's crash database is cited in the article).  Also if you look at pre-2010, the Cony Circle had an insanely high number of injury crashes (highest average in the state each and every year).  The circle was reconfigured in 2008 (to conform to more modern roundabout standards) and the number of injury accidents have dropped but still higher than you would like to see at a roundabout.

8 Investigates Maine's high-crash locations for 2019
https://www.wmtw.com/article/8-investigates-maines-high-crash-locations-for-2019/30379657

[tradephoric]

^I won't post a picture of any of the many injury crashes that have occurred in the Cony circle so not to be censored.  Posting a picture involving an injury accident is outrageous apparently.

[tradephoric]

And here's how to properly navigate the Cony Circle.  Simple right?

[kphoger]

Yes, it should be straightforward.  And yet, by my count, there are a total of nine conflict points that are not merely glancing blows:

[jakeroot]

I've been through those Augusta roundabouts before (way back in 2011); Cony Circle is impressive to me, even if it's #2 most crash-prone in the state. It's really hard to configure a five-way intersection, so credit where credit's due to Augusta's engineers of yore.

On the other hand, Memorial Circle is arguably simpler, yet features more crashes (at least for 2019). That's not good for a regular four-leg roundabout.

[DaBigE]

^I won't post a picture of any of the many injury crashes that have occurred in the Cony circle so not to be censored.  Posting a picture involving an injury accident is outrageous apparently.

Get over yourself. Apparently you're new to social media. News outlets (at least around here) get crucified when they post that kind of picture on social media. You never know who may not know about their loved one being involved in an accident. Finding out via social media is not usually a way loved ones prefer to find out. Your decision to continue one newspaper's bad decision doesn't make it any better.

What's arguably more troubling, your sharing of it had zero to do with furthering the cause of this topic other than over-sensationalizing the situation. It provides zero context to any possible cause of the crash or what role the roundabout may or may not have played. There were two other photos you could have posted, but you chose the one to spin the crash as bad as it possibly could be. It's one thing if the photo would have been included with the attachment on its own, but you had to do extra work and knowingly post the extra photo. You could have just left it with a link to the article.

[kalvado]

Yes, it should be straightforward.  And yet, by my count, there are a total of nine conflict points that are not merely glancing blows:

Which is the least relevant number for analysis.

[kphoger]

Those conflict points are what I think drive the hidden danger of multi-lane roundabouts.  In the literature, you can better your sweet bippy you'll always find an illustration touting the reduced number of conflict points a roundabout has, and also an explanation of how all the conflict points that remain are merely glancing blows unlikely to result in serious damage or injury/death.  Then that's used to help justify building a mutli-lane roundabout that has more conflict points than illustrated and also does indeed have non-glancing conflict points.

Yield-only approaches are fine and dandy when all trajectories gradually sweep into and out of each other.  But, when half the trajectories actually cross each other, mere yield signs are insufficient to prevent accidents.

[jakeroot]

^I won't post a picture of any of the many injury crashes that have occurred in the Cony circle so not to be censored.  Posting a picture involving an injury accident is outrageous apparently.

Get over yourself. Apparently you're new to social media. News outlets (at least around here) get crucified when they post that kind of picture on social media. You never know who may not know about their loved one being involved in an accident. Finding out via social media is not usually a way loved ones prefer to find out. Your decision to continue one newspaper's bad decision doesn't make it any better.

What's arguably more troubling, your sharing of it had zero to do with furthering the cause of this topic other than over-sensationalizing the situation. It provides zero context to any possible cause of the crash or what role the roundabout may or may not have played. There were two other photos you could have posted, but you chose the one to spin the crash as bad as it possibly could be. It's one thing if the photo would have been included with the attachment on its own, but you had to do extra work and knowingly post the extra photo. You could have just left it with a link to the article.

This seems to be arguing for argument's sake.

The man is on a gurney with no visible injuries, and there is a dented car. I do not understand why this would be problematic. The man cannot be identified in the photo (his face is not visible), and there is no front licence plate.

He could have chosen to not share the image, but from my perspective, it's just a picture of the crash. No forum rules disallow this.

[kphoger]

If that article were published in Mexico, you might have seen blood on the pavement and organs hanging out of the abdomen on the front page of the newspaper.

[jakeroot]

If that article were published in Mexico, you might have seen blood on the pavement and organs hanging out of the abdomen on the front page of the newspaper.

The last place I would expect censorship would be this forum. Mexican journalism? Guess I can't say I'm familiar, but I appreciate their straightforwardness.

[kalvado]

I've been through those Augusta roundabouts before (way back in 2011); Cony Circle is impressive to me, even if it's #2 most crash-prone in the state. It's really hard to configure a five-way intersection, so credit where credit's due to Augusta's engineers of yore.

On the other hand, Memorial Circle is arguably simpler, yet features more crashes (at least for 2019). That's not good for a regular four-leg roundabout.

The other factor to consider is total traffic through the intersection. I cannot lookup numbers at the moment, but I suspect we're dealing with circle operating over capacity, which was actually designed without understanding of those limitations. While traffic light on 5-leg would have issues as well, other approaches to distribution of traffic may be considered to unload the sour point - instead of trying to manage unmanageable.

[DaBigE]

^I won't post a picture of any of the many injury crashes that have occurred in the Cony circle so not to be censored.  Posting a picture involving an injury accident is outrageous apparently.

Get over yourself. Apparently you're new to social media. News outlets (at least around here) get crucified when they post that kind of picture on social media. You never know who may not know about their loved one being involved in an accident. Finding out via social media is not usually a way loved ones prefer to find out. Your decision to continue one newspaper's bad decision doesn't make it any better.

What's arguably more troubling, your sharing of it had zero to do with furthering the cause of this topic other than over-sensationalizing the situation. It provides zero context to any possible cause of the crash or what role the roundabout may or may not have played. There were two other photos you could have posted, but you chose the one to spin the crash as bad as it possibly could be. It's one thing if the photo would have been included with the attachment on its own, but you had to do extra work and knowingly post the extra photo. You could have just left it with a link to the article.

This seems to be arguing for argument's sake.

The man is on a gurney with no visible injuries, and there is a dented car. I do not understand why this would be problematic. The man cannot be identified in the photo (his face is not visible), and there is no front licence plate.

He could have chosen to not share the image, but from my perspective, it's just a picture of the crash. No forum rules disallow this.

I never said it was against the forum rules. Is it in poor taste? I think many would agree that it is. You don't know if the injured person is known by a member here. Does it help further the investigation into roundabout performance? No, which is the whole point of this 90+ page thread. I asked why the photo was posted, and all I got was a smartass remark. All it does is help sensationalize a crash report and waste bandwidth; it added nothing to help determine why certain roundabouts have poorer performance than others. 

[jakeroot]

I've been through those Augusta roundabouts before (way back in 2011); Cony Circle is impressive to me, even if it's #2 most crash-prone in the state. It's really hard to configure a five-way intersection, so credit where credit's due to Augusta's engineers of yore.

On the other hand, Memorial Circle is arguably simpler, yet features more crashes (at least for 2019). That's not good for a regular four-leg roundabout.

The other factor to consider is total traffic through the intersection. I cannot lookup numbers at the moment, but I suspect we're dealing with circle operating over capacity, which was actually designed without understanding of those limitations. While traffic light on 5-leg would have issues as well, other approaches to distribution of traffic may be considered to unload the sour point - instead of trying to manage unmanageable.

Worst-case, they could always signalize the roundabouts. But I cannot even begin to imagine how bad the traffic would be.

At the Memorial Circle, I'm picturing a giant "squareabout" going in front of the CVS and then behind the triangle building. There would be a left turn from Cony St to northbound Bangor St. If I can find the time, I'll mock something up.

[jakeroot]

I never said it was against the forum rules. Is it in poor taste? I think many would agree that it is. You don't know if the injured person is known by a member here. Does it help further the investigation into roundabout performance? No, which is the whole point of this 90+ page thread. I asked why the photo was posted, and all I got was a smartass remark. All it does is help sensationalize a crash report and waste bandwidth; it added nothing to help determine why certain roundabouts have poorer performance than others.

I know you didn't say that, but you're making a big deal out of something which isn't inherently wrong.

I think it's a bit ridiculous to presume that the only images that should be posted to this thread, should be those that "further the investigation into roundabout performance". I'll agree that the image from the article did absolutely nothing for this thread (it could have easily been omitted), but you could make the same argument about many of the photos in this thread; many are, at best, tangentially related to the actual topic at hand. But virtually none of the photos actively detract from the conversation. That, in my opinion, is the key point here: his photo may have been in poor taste to some eyes, but generally speaking, I don't think it detracted from the conversation.

[jakeroot]

I was reading this document, prepared by Carl Clayton, P.Eng. P.E., Chuan Kua, P.Eng, and Brice Stephenson, P.Eng, which talks about how Edmonton, Alberta's traffic circles compare to so-called "modern roundabouts".

The document was prepared for the Innovative Intersection and Interchange Designs Session of the 2004 Annual Conference of the Transportation Association of Canada in Quebec City.

I quite like the final paragraph:

Collision rates for the Edmonton roundabouts are relatively high in comparison to other intersections in Edmonton. In instances where roundabouts were converted to typical signalized intersections, the collision rates appear to have decreased rather than increased. This is in conflict with the findings outlined in literature and by practitioners that well designed roundabouts typically have lower collision rates than other forms of at-grade intersections. A possible explanation for this might be that the geometry of Edmonton roundabouts, while conducive to maximizing traffic flows, degraded the superior level of safety typically attributed to roundabouts. Alternatively, another possibility might be that as much of the safety reviews prepared for roundabouts are on lower volume single lane roundabouts, perhaps the safety potential of a roundabout decreases as traffic volumes approach the upper capacity limit. In either case, it would appear to be prudent to avoid trying to accommodate traffic volumes at the higher end of the range recommended in the various design guidelines by using geometry beyond that recommended in the current state-of-the-art guidelines or oval shaped roundabouts.

Why do I like this paragraph? Well, it's 2004, and they're already noticing a trend that larger, heavy-volume roundabouts tend to have higher crash rates than signalized intersections. Edmonton has been lucky enough to have installed, left in place for some time, and then removed some of these roundabouts; evidently, they're a bit safer operating as regular signalized intersections.

Tangentially-related: the pentagon-shaped Bonnie Doon roundabout in Edmonton has been replaced by a traffic light due to an LRT line. This is problematic for locals, as the roundabout served all directions of the five intersection legs, unlike the new signal. Glad I got to drive it a few years ago.

[tradephoric]

Collision rates for the Edmonton roundabouts are relatively high in comparison to other intersections in Edmonton. In instances where roundabouts were converted to typical signalized intersections, the collision rates appear to have decreased rather than increased. This is in conflict with the findings outlined in literature and by practitioners that well designed roundabouts typically have lower collision rates than other forms of at-grade intersections. A possible explanation for this might be that the geometry of Edmonton roundabouts, while conducive to maximizing traffic flows, degraded the superior level of safety typically attributed to roundabouts. Alternatively, another possibility might be that as much of the safety reviews prepared for roundabouts are on lower volume single lane roundabouts, perhaps the safety potential of a roundabout decreases as traffic volumes approach the upper capacity limit. In either case, it would appear to be prudent to avoid trying to accommodate traffic volumes at the higher end of the range recommended in the various design guidelines by using geometry beyond that recommended in the current state-of-the-art guidelines or oval shaped roundabouts.

Edmonton really is an interesting case study.  According to Edmonton's 2018 crash report, the unsignalized multi-lane roundabout at 107 Avenue & 142 Street was the most crash prone intersection in the city with 93 crashes.  Yet the 118th Avenue & St Albert Trail signalized roundabout in Edmonton didn't even crack the top 10, and it's a triple-lane roundabout no less.


https://www.edmonton.ca/transportation/PDF/2018MVC_AnnualReport.pdf

[tradephoric]

I'd love to simply ignore this thread entirely, but I keep making the same mistake of coming back when there's a new post, in the [insane?] hopes there is going to be some new information that I could maybe use. But it looks like I am better-off sticking with my Google News feed and industry-supported sources. If I want crash reports, I'll just go straight to the source and cut out the snarky middleman.

Can you go to your industry-supported sources and track down crash data for the Taylor Street and Shawano Avenue roundabout in Green Bay?  I believe that roundabout was constructed in 2011 and still haven't been able to track down crash data for it.  Let's figure out how well (or poorly) it's doing.