Crash prone 'modern roundabouts'

[jakeroot]

Had the IIHS said that the tight circle of a roundabout only "encourages" drivers to slow down, then roundabouts would be at the same level as traffic signals - where if you ignore the encouragement a deadly crash can occur.

So is your problem the roundabouts or the FHWA's grammar?

[kalvado]

Had the IIHS said that the tight circle of a roundabout only "encourages" drivers to slow down, then roundabouts would be at the same level as traffic signals - where if you ignore the encouragement a deadly crash can occur.

So is your problem the roundabouts or the FHWA's grammar?

Let me try to interpret:
Looks like problem is with state/local authorities taking that statement at face value, and failure to address underlying conditions. Which is actually a problem.
We talked about a roundabout at the end of tunnel. Clearly that is not safest possible design, and clearly roundabout itself is not the cure or problem, it is just part of a failure to create safe situation.
Just another anecdotal situation:
we had another roundabout built nearby. An intersection before that one was built had a hill blocking the view when approaching from one direction. Making things worse, traffic light was visible from far away - and often people sped up to catch the light. That was a relatively slow light, giving priority to higher traffic cross street.  At the same time, another car could be making a turn while someone flies over the hill at them.. I had some close calls personally, both as bullet and target.
Now that hill was smoothed quite a bit during roundabout construction. That alone could make light controlled intersection much better - but was funded only as part of a much bigger project including roundabout. This is what I call "addressing underlying condition". And roundabout should not take whole - only  small part, if any - credit for making that intersection safer (although I didn't see any data or crashes)

[kphoger]

...

Let me try to interpret:
Looks like problem is with state/local authorities taking that statement at face value

I have trouble believing that any DOT has in fact interpreted the word "forces" in the IIHS statement ("the tight circle of a roundabout forces drivers to slow down," for those who don't want to dig back several pages to find the original statement) to mean that a roundabout literally and physically prevents a driver from going above a certain speed when entering the intersection–and has then installed a roundabout based on the expectation that zero drivers will ever enter at faster than x mph.

To suggest that the reason roundabouts are being constructed is that the word "forces" in that statement was taken as absolutely literal......is silly.

[kalvado]

...

Let me try to interpret:
Looks like problem is with state/local authorities taking that statement at face value

I have trouble believing that any DOT has in fact interpreted the word "forces" in the IIHS statement ("the tight circle of a roundabout forces drivers to slow down," for those who don't want to dig back several pages to find the original statement) to mean that a roundabout literally and physically prevents a driver from going above a certain speed when entering the intersection–and has then installed a roundabout based on the expectation that zero drivers will ever enter at faster than x mph.

To suggest that the reason roundabouts are being constructed is that the word "forces" in that statement was taken as absolutely literal......is silly.

If it looks like a duck...

[tradephoric]

We know the rate of deadly crashes at signalized intersections versus roundabouts: there's empirical data on the point, from jurisdictions all over the world. There's no need to base the safety claim on whether or not drivers actually are forced to slow down or whether or not they obey traffic signals; it can be derived from simple observation of fatality incidence over time, without a causal argument.

It's difficult to determine how effective roundabouts are at reducing fatal crashes - assuming they reduce them at all - because fatal crashes are such a rare event.   In addition, signalized intersections that have been the site of a fatal crash are often the intersections specifically targeted for roundabout conversion.  Comparing the safety of the worst performing signalized intersections to roundabouts isn't the same as comparing roundabouts to a standard signalized intersection.  Interestingly, when a roundabout is the site of a fatal crash they don't become a target to be converted back to a signalized intersection. 

On September 21, 2014 a 23 year old motorcyclist slid into the roundabout island on Fischer-Hallman Road in Kitchener and was pronounced dead at hospital.  Investigator Staff Sgt. James Strand concluded that "speed and poor rod conditions were the reason for this collision happening and not the roundabout" .   Since this crash wasn't blamed on the roundabout, the Region of Waterloo can keep claiming that there has never been a fatality in a local roundabout.

Roundabout death? Motorcycle speeding, lost control on wet road, before Kitchener crash
http://www.therecord.com/news-story/6225063-roundabout-death-motorcycle-speeding-lost-control-on-wet-road-before-kitchener-crash/

[intelati49]

We know the rate of deadly crashes at signalized intersections versus roundabouts: there's empirical data on the point, from jurisdictions all over the world. There's no need to base the safety claim on whether or not drivers actually are forced to slow down or whether or not they obey traffic signals; it can be derived from simple observation of fatality incidence over time, without a causal argument.

Since this crash wasn't blamed on the roundabout, the Region of Waterloo can keep claiming that there has never been a fatality in a local roundabout.

{{citation-needed}}

Just because the roundabout isn't at fault doesn't mean that information disappears. (Or at least isn't in Missouri)

And if they're claiming that, they're the idiots.



Article with corresponding crash

[jakeroot]

If it looks like a duck...

It doesn't look like a duck.

[lordsutch]

Yes, it's difficult to determine in any single location due to fatalities being a rare event. That's why the data is aggregated. Here are two meta-analysis studies that aggregate local/regional studies into a more complete picture:

Road safety effects of roundabouts: A meta-analysis
Effects on Road Safety of Converting Intersections to Roundabouts: Review of Evidence from Non-U.S. Studies

[kalvado]

Yes, it's difficult to determine in any single location due to fatalities being a rare event. That's why the data is aggregated. Here are two meta-analysis studies that aggregate local/regional studies into a more complete picture:

Road safety effects of roundabouts: A meta-analysis
Effects on Road Safety of Converting Intersections to Roundabouts: Review of Evidence from Non-U.S. Studies

In short: decrease of number of fatal accidents, no effect to possibly increase of property-only crashes; no breakdown on roundabout type - small side roads vs major roads, 1-lane vs 2-lane vs 3-lane; 2 US papers in past 2 years.

[lordsutch]

In short: decrease of number of fatal accidents, no effect to possibly increase of property-only crashes; no breakdown on roundabout type - small side roads vs major roads, 1-lane vs 2-lane vs 3-lane; 2 US papers in past 2 years.

To be precise (I believe the first article includes all the studies that were looked at in the second, since it's newer): ~65% decrease fatality accidents, ~35% decrease in injury accidents, a statistical wash on PDO.

There are extant studies that break down effectiveness by roundabout type but they are based on smaller samples or case studies, and there you run into the rare events problem (somewhat resolved by lumping injury and fatality accidents into a single category).

The general evidence is that injury and fatality accidents are reduced regardless of roundabout size, but the effects are smaller at higher volume, higher capacity roundabouts. PDO tends to go up, as you have more vehicles getting into low angle-of-incidence, low net-momentum collisions due to failure to yield and/or failure to stay in lane. That PDO goes up is kind of what you'd expect since vehicles generally don't pass next to each other on a single-lane roundabout but can and do on a multi-lane roundabout, and staying in lane on a curve takes more driver attention than staying in lane in a straight line across a typical signalized intersection. Hence the recommended use of turbo markings etc. rather than the old--school UK approach of "you're supposed to know where the lanes are even though there are no markings."

[tradephoric]

Yes, it's difficult to determine in any single location due to fatalities being a rare event. That's why the data is aggregated. Here are two meta-analysis studies that aggregate local/regional studies into a more complete picture:

Road safety effects of roundabouts: A meta-analysis
Effects on Road Safety of Converting Intersections to Roundabouts: Review of Evidence from Non-U.S. Studies

Thank you for posting those aggregate studies.  First off the safety findings in those studies aren't as impressive as the 2000 IIHS study.  This is important as a lot of complex roundabouts have been constructed in this country after agencies have highlighted the safety statistics found in the IIHS study.  Do you realize the only 2x2 roundabout analyzed in the IIHS study was a roundabout on Avon Road in Vale, Colorado?  Avon road has a speed limit of 25 mph.  Conveniently road agencies in America have promoted the safety statistics of double-lane roundabouts for the next 20-years based on a roundabout in the middle of a ski-resort.  But i digress...

The real issue is not all roundabouts are created equal.  Of the roughly 5,000 roundabouts in America only a few hundred have complex 2x2 or 3x2 geometries.  That's less than 5% of the total roundabouts in this country.  If these complex roundabouts have poor safety records, they would be masked in these aggregate studies since single-lane roundabouts would be weighted heavily in the analysis.  When looking into just these complex roundabouts a troubling safety trend appears.  The average crash rate of these complex roundabouts is about 4.0 MEV; roughly 4X higher than the crash rate of a typical signalized intersection.  It's harder to tie down the injury crash rates, but if PDO crashes are 4X higher it's not a stretch to assume that they won't see injury crashes reduced by 35%.  Put another way, the aggregate studies say that there is a ~35% reduction in injury crashes, but that's assuming total crashes are roughly the same.

The FHWA and IIHS need to analyze the safety performance of complex roundabouts; ones with 2x2 and 3x2 geometries.  Only then will we get a good understanding if these roundabouts are even worth building from a safety perspective.  Imagine if they found complex roundabouts see a 300% increase PDO crashes, a 22% increase in injury crashes, and a 0% increase in fatalities.  Would communities be so eager to build THESE roundabouts then?  I don't have a problem with agencies promoting roundabouts, but we should at least know which types of roundabouts will reduce injury crashes and which ones won't.

[lordsutch]

Well, then you can either accept evidence from other countries that have a lot more multi-lane roundabouts already (which do show improvements in both injury and fatality rates), or you can wait until sufficient data is accumulated on multi-lane roundabouts in the U.S. alone to draw those conclusions from a large sample over time and space.

What we do know, based on the laws of physics, is that injury and (particularly) fatality accidents are highly associated with two factors in collisions: high speed and high net momentum. Putting aside the semantics of the word "force" for a moment, the evidence is that most drivers traverse roundabouts at lower speed than they would traverse a signalized intersection and that most drivers follow lane markings and curbs so the angle of incidence in the event of a collision is low with both vehicles moving on roughly the same vector, hence low net momentum.

This design is effective to the extent that fatality accidents at roundabouts are almost exclusively the result of driver impairment or criminal activity in some form (DUI, driving while senile, heart attack behind the wheel, evading police in a high-speed pursuit, etc.), as virtually all of your examples to date have been. By contrast a momentary lapse in concentration in a roundabout that results in a collision will, at worst, give you and/or the other driver whiplash and a very expensive repair bill, rather than a guaranteed trip to the hospital or the morgue from a side impact or head-on crash when you try to beat a red or don't see it in time.

[kphoger]

lordsuch & tradephoric:  Thank you both for quickly bringing this thread back into an intelligent, reasonable, and interesting conversation.

[kalvado]

Well, then you can either accept evidence from other countries that have a lot more multi-lane roundabouts already (which do show improvements in both injury and fatality rates), or you can wait until sufficient data is accumulated on multi-lane roundabouts in the U.S. alone to draw those conclusions from a large sample over time and space.

What we do know, based on the laws of physics, is that injury and (particularly) fatality accidents are highly associated with two factors in collisions: high speed and high net momentum. Putting aside the semantics of the word "force" for a moment, the evidence is that most drivers traverse roundabouts at lower speed than they would traverse a signalized intersection and that most drivers follow lane markings and curbs so the angle of incidence in the event of a collision is low with both vehicles moving on roughly the same vector, hence low net momentum.

This design is effective to the extent that fatality accidents at roundabouts are almost exclusively the result of driver impairment or criminal activity in some form (DUI, driving while senile, heart attack behind the wheel, evading police in a high-speed pursuit, etc.), as virtually all of your examples to date have been. By contrast a momentary lapse in concentration in a roundabout that results in a collision will, at worst, give you and/or the other driver whiplash and a very expensive repair bill, rather than a guaranteed trip to the hospital or the morgue from a side impact or head-on crash when you try to beat a red or don't see it in time.

That still doesn't address issues of throughput or tradeoffs between property and injury accidents.
Would roundabout be still better compared to intersection with wavy approaches, for example?

[UCFKnights]

Well, then you can either accept evidence from other countries that have a lot more multi-lane roundabouts already (which do show improvements in both injury and fatality rates), or you can wait until sufficient data is accumulated on multi-lane roundabouts in the U.S. alone to draw those conclusions from a large sample over time and space.

Do other countries have the amount of improper lane usage in their roundabouts as compared to the US? Its still my core concern with multilane roundabouts, and why they fail so quickly when you add lanes... many Americans simply cannot understand that a left cannot be made from the right lane, as if another car fails to stop for you, you will crash. When I go through multilane roundabouts, I see someone use the wrong lane for their turn just about every single time. And from talking to my cop friends, they seem unaware that using the wrong lane in the roundabout causes an accident and file their accident reports incorrectly quite frequently.

[lordsutch]

In my experience driving in the UK and observing in Ireland, where lane marking is rare except at signalized roundabouts, lane discipline is nonetheless quite good.

But they have 50 years of experience with them, while on the continent where they're more recently introduced, turbo markings and the like seem to have been settled on as the best approach.

Even in the UK signalized or complex roundabouts will a typically have a "Get in lane" lane assignment sign with destinations in advance of the roundabout, in addition to the standard diagrammatic guide sign, something that really should be introduced here, like here: https://goo.gl/images/FyOjMu

[tradephoric]

A presentation that evaluates the safety of Wisconsin roundabouts was given during the 5th International Conference on Roundabouts held on May 8-10, 2017 in Green Bay.   All the roundabouts evaluated in this presentation were built in 2009 and before.   This means that the explosion of triple-lane roundabouts built in Wisconsin after 2009 were not evaluated in this analysis - and these are the types of roundabouts that seem to have the highest crash rates. 



http://teachamerica.com/RAB17/RAB174A_Bill/index.htm

Of the 29 multi-lane roundabouts evaluated, there was a 35% increase in total crashes and a 26% decrease in injury and fatal crashes (KABC crashes).   If you look at the 39 urban roundabouts evaluated, there was a 48% increase in total crashes and just an 18% decrease in injury and fatal crashes (KABC crashes).   It's great that all this data exists but it would be nice if WTOP broke it down further.  What would happen if they evaluated the safety performance of urban multi-lane roundabouts that have two-approach lanes at all legs of the roundabout (IE 2x2 and 3x2 roundabouts)?   It's quite possible you would see an increase in both injury crashes and total crashes if just those types of multi-lane roundabouts were evaluated.

[tradephoric]

By the City's [of Lincoln] own admission the chain link fencing was a temporary measure.  They never had any intention to keep the chain link fencing in place.  If it was shown to be ineffective at reducing crashes they would have ripped it out.  If it was shown to reduce crashes (as they believe it did) they would replace it with permanent fencing.  But there's a bigger issue here then what type of fence the city decided to use.  How does the city know the fencing was effective at reducing crashes at the roundabout in the first place?  The problem is at the same time they installed the fencing they removed a circulating lane of traffic inside the roundabout.  Those two changes combined saw a reduction in crashes at the roundabout, but how do we know if the reduction was due to the fencing or the reduction of a circulating lane? 

Luckily, they have added fencing at the most crash prone roundabout in Michigan (186 crashes in 2015).  No other changes were made other than the fencing so we will see if it actually helps reduce the total number of crashes. 

http://www.wxyz.com/news/region/oakland-county/drivers-have-mixed-feelings-about-new-fences-at-m-5-roundabout

We now have crash data to see if the screen fencing added to the M-5 & Pontiac Trail roundabout in Michigan was effective at reducing total crashes.  When comparing the 6 months before the screen fence was added to the 6 months after, there was no reduction in the total number of crashes.  From an operational standpoint, they do believe the screen fence provided greater gaps for drivers to enter the roundabout.  This data was presented by Tom Blust at the 5th International Conference on Roundabouts:


http://teachamerica.com/RAB17/RAB176A_Blust/index.htm

[tradephoric]

Samuel Bobko gave a presentation at the 5th International Conference on Roundabout that focused on a multi-lane roundabout in Cleveland Ohio at I-71 and Quigley Road.   It was constructed in 2007 and he acknowledged that there was an elevated crash issue at the roundabout.  The presenter thinks that the increase in crashes starting around 2011 was due to I-71 being under construction and more drivers were getting off and going through the roundabout. He then goes through low cost improvements made to the roundabout that reduced the collisions at the roundabout.  While the presenter hits on the usual changes —lane arrows, pavement markings, signage changes — one of the biggest changes that likely reduced the crashes was the removal of a circulating lane inside the roundabout. 

The presenter claims that the roundabout was never constructed for safety.   But isn't it always about safety when it comes to modern-roundabouts?   We have been conditioned to believe that roundabouts are safe and this roundabout was no exception.  Here is the official press release from ODOT back in 2006:

August 18, 2006, Garfield Heights, OH...
This is the first modern roundabout to be built in Northeast Ohio. The only other roundabout in Ohio is in the City of Dublin. A roundabout was chosen for this location because it will increase capacity and lower delays, calm traffic, improve aesthetics, improve safety, and be able to handle U-turn movements.
https://www.dot.state.oh.us/projects/ClevelandUrbanCoreProjects/Innerbelt/News%20Releases/QuigleyRdPressRelease.pdf

Here's the before and after crashes at the I-71 and Quigley Road roundabout.  Even after the improvements were made to the roundabout the crashes in 2015 and 2016 were still much higher than the pre-roundabout crashes.  And there doesn't appear to be a big drop in injury crashes either.  I'm not seeing the safety improvements ODOT were advertising back in 2006 play out at this roundabout.  This is yet another example of a multi-lane roundabout that failed to improve safety.

[kalvado]

The presenter claims that the roundabout was never constructed for safety.   But isn't it always about safety when it comes to modern-roundabouts?   We have been conditioned to believe that roundabouts are safe and this roundabout was no exception.  Here is the official press release from ODOT back in 2006:

August 18, 2006, Garfield Heights, OH...
This is the first modern roundabout to be built in Northeast Ohio. The only other roundabout in Ohio is in the City of Dublin. A roundabout was chosen for this location because it will increase capacity and lower delays, calm traffic, improve aesthetics, improve safety, and be able to handle U-turn movements.
https://www.dot.state.oh.us/projects/ClevelandUrbanCoreProjects/Innerbelt/News%20Releases/QuigleyRdPressRelease.pdf

Well, you cannot have a cake and eat it too. At some point I expect rhetoric shift toward "it is about capacity, and accident rate is increased only by that much.. "

[tradephoric]

Well, you cannot have a cake and eat it too. At some point I expect rhetoric shift toward "it is about capacity, and accident rate is increased only by that much.. "

There is a growing consensus that there is a crash problem at these multi-lane roundabouts.  But you still have roundabout experts saying that the injury crashes have been taken care of at these high crash roundabouts.  The problem is they have convinced themselves that roundabouts are safer and won't even entertain the idea that certain types of roundabouts may be increasing injury and fatal crashes.

[kalvado]

Well, you cannot have a cake and eat it too. At some point I expect rhetoric shift toward "it is about capacity, and accident rate is increased only by that much.. "

There is a growing consensus that there is a crash problem at these multi-lane roundabouts.  But you still have roundabout experts saying that the injury crashes have been taken care of at these high crash roundabouts.  The problem is they have convinced themselves that roundabouts are safer and won't even entertain the idea that certain types of roundabouts may be increasing injury and fatal crashes.

Thing is, I don't understand how that happened. Why would engineers jump on the idea?
My only explanation is that roundabout is the only big thing traffic engineers in many areas can design, especially in northeast-midwest. Few big projects going on; even big rebuilds seem to be done exactly and carefully by old blueprints...

[tradephoric]

Armdale Roundabout safety probed
http://thechronicleherald.ca/metro/1195561-armdale-roundabout-safety-probed

The Armdale rotary in Halifax, Nova Scotia was completed in 1956 and was converted to modern roundabout standards in 2007 (although I'll note there is no truck apron at the roundabout).  A few years after the conversion, municipal staff noted a "sharp and sustained"  increase in accidents (jumping from the low 60s to the high 90s).  Here are a few recent news reports of injury accidents that have occurred at the roundabout:

http://www.cbc.ca/news/canada/nova-scotia/armdale-roundabout-crash-sends-3-to-hospital-with-serious-injuries-1.3116629
http://www.cbc.ca/news/canada/nova-scotia/armdale-roundabout-motorcyclist-crash-1.3300385

Major crashes still occurring at the Armdale roundabout.  A crash this week saw a small compact car take out a steel light pole and a wooden pole all in one shot!  Car must have been flying to do that much damage...

Roundabout crash knocks down poles, slows traffic in Halifax
http://thechronicleherald.ca/metro/1473156-roundabout-crash-knocks-down-poles-slows-traffic-in-halifax

Armdale Roundabout car-pedestrian crash sends 72-year-old man to hospital
http://www.cbc.ca/news/canada/nova-scotia/armdale-roundabout-car-pedestrian-crash-sends-72-year-old-man-to-hospital-1.3177515

[tradephoric]

The IIHS remade their roundabout safety video that reflects the reality of how people are driving though them.  Just imagine if the IIHS made a video that promoted the safety of traffic signals by making definitive statements about how red lights force drivers to stop.  They would lose credibility because everyone knows red lights don't force drivers to stop.  But it's acceptable for them to argue that roundabouts are safe simply because roundabout have tight curves.  That's justification that roundabouts are safer?

https://www.youtube.com/watch?v=1mZODRwyKZk

[jakeroot]

Okay that was actually pretty funny.

But, you still seem to be taking this whole "force drivers to slow down" thing a little too seriously.