News:

Thanks to everyone for the feedback on what errors you encountered from the forum database changes made in Fall 2023. Let us know if you discover anymore.

Main Menu

Crash prone 'modern roundabouts'

Started by tradephoric, May 18, 2015, 02:51:37 PM

Previous topic - Next topic

tradephoric

Quote from: lordsutch on August 24, 2016, 11:38:33 PM
Quote from: tradephoric on August 24, 2016, 08:44:11 PM
Half of Pennsylvania's Roundabouts Show A Decrease In Crashes
http://wnep.com/2016/08/02/half-of-pennsylvanias-roundabouts-show-a-decrease-in-crashes/

I'll be the first to admit I'm no genius, but if half of the roundabouts in Pennsylvania show a decrease in crashes, that means the other half of roundabouts in Pennsylvania either stayed the same or crashes went up.  Pennsylvania hasn't really built any significant roundabouts to date but thought I'd share this news story anyways.  But yeah, flip a coin to determine if a particular roundabout in PA has seen crashes increase or decrease.
Half of... ten.

Did i say there was more than ten?  I got a hunch if i told you the sun rises in the east you would argue it rises in the west...  that's just the kind of person you are.  Anyways, Pennsylvania really hasn't gotten into the roundabout game and have stuck with constructing little hodunk roundabouts.  I don't even think Pennsylvania has a multi-lane roundabout.  I'm not at all surprised only ten roundabouts were analyzed (apparently you are???... i don't know what to make of your posts anymore lordsutch). 


tradephoric

Quote from: jakeroot on August 24, 2016, 11:42:14 PM
Quote from: tradephoric on August 24, 2016, 10:51:09 PM
Every time i update this chart it's getting more and more yellow (indicating roundabouts that have had circulating lanes removed):

You can highlight the Steptoe Roundabout in Washington. Circulating lanes were removed from two approaches in mid-July.

Thanks for the info Jake.

cjw2001

#702
Quote from: tradephoric on August 24, 2016, 10:51:09 PM
Yet another 2x2 multi-lane roundabout has been downsized and this time it's in Carmel.  The roundabout at 116th & Illinois has historically been a high crash intersection in the city.  The EB right lane on 116th has been re-striped as a right turn only lane. 

Don't confuse temporary changes during US 31 construction with permanent changes.

tradephoric

^You are right (meaning i was wrong lol).  I now see why the 116th & Illinois roundabout was temporary reconfigured during construction at US 31.  If the city were to downsize this roundabout they would likely remove circulating lanes along the N/S leg as opposed to the E/W. 

It will be interesting to see how the 2x2 roundabout at Carmel & Rangeline Road performs once finished.  It could easily become an accident hot spot in the city.  To all my Carmel friends, why is it so difficult to find intersection crash data for Carmel's intersections?  I don't know how many total and injury crashes Carmel & Rangeline Road has had over the past 5 years, but it would be pertinent information moving forward in accessing how effective the new 2x2 roundabout will be. 

cjw2001

#704
Yes as I mentioned earlier in the thread we are just finishing up from 5 years of US 31 construction which had major impacts on local traffic flows (mainly Westfield during the first half of the project and mainly Carmel during the second half, but some overlap).  So I'd be very cautious about making year to year comparisons on accident rates on anything near US 31 during the period of extended disruption.   Quite a bit of local traffic moved from US 31 to Spring Mill, Illinois, and Pennsylvania during the project period.  Also Illinois street south of 111th is a newly constructed connection to 106th, so there is now through traffic on Illinois that didn't previously exist.

The accident rate on the I 465 to US 31 ramps has also been off the charts lately due to the construction conditions.   Now that the ramps are nearing final completion (a little bit of final paving yet to do and two lanes open at the merge onto US 31 instead of one) I would expect those rates to drop off some.   The US 31 to I 465 Eastbound ramp and Meridian North to I 465 Eastbound ramp are now in much better shape as well, with a dedicated lane for each with no merging required in the final configuration.

Also apologies if my 3:30 am response was a bit short.  Couldn't get to sleep last night after all the local tornado activity so I was a bit toasty from being up too late. :spin:


lordsutch

Quote from: tradephoric on August 24, 2016, 11:59:20 PM
Did i say there was more than ten?  I got a hunch if i told you the sun rises in the east you would argue it rises in the west...  that's just the kind of person you are.  Anyways, Pennsylvania really hasn't gotten into the roundabout game and have stuck with constructing little hodunk roundabouts.  I don't even think Pennsylvania has a multi-lane roundabout.  I'm not at all surprised only ten roundabouts were analyzed (apparently you are???... i don't know what to make of your posts anymore lordsutch). 

My point is generalizing (drawing broad conclusions from a set of data) from ten cases is stupid, especially when there's no particular reason to believe Pennsylvania's roundabouts are substantially different from roundabouts in other states (or road users in Pennsylvania are substantially different from those in other states), particularly neighboring ones.

Now if Pennsylvania was doing something that substantially deviated from the FHWA guidance or what other states were doing, it might be worthwhile to isolate what's going on in Pennsylvania, but otherwise what you'd really want to do is add Pennsylvania's data on its roundabouts to the data on roundabouts on other states and use that for a comprehensive analysis. Then you could look at what intersection-specific factors (# of entry lanes per leg, # of circulating lanes, # of lanes, presence/absence of various traffic control devices, diameter of circulating roadway, presence/absence of visual obstructions in center, painted versus substantive dividers, presence/absence of segregated bike and pedestrian facilities, lighting/lack of lighting, overall vehicle volume, truck volume, speed limits on entry roadways, ...) seem to drive crash rates (say with a regression model) and go from there... something you can't really do with 10 data points because the number of meaningful factors that vary between cases exceeds the number of cases being analyzed. And more cases lets you deal with the fact that serious accidents and fatalities are relatively rare events, and thus the law of large numbers doesn't kick in as quickly (i.e. you need more cases to establish that a series of rare events aren't just due to random chance).

kalvado

Quote from: lordsutch on August 25, 2016, 06:54:15 PM
Quote from: tradephoric on August 24, 2016, 11:59:20 PM
Did i say there was more than ten?  I got a hunch if i told you the sun rises in the east you would argue it rises in the west...  that's just the kind of person you are.  Anyways, Pennsylvania really hasn't gotten into the roundabout game and have stuck with constructing little hodunk roundabouts.  I don't even think Pennsylvania has a multi-lane roundabout.  I'm not at all surprised only ten roundabouts were analyzed (apparently you are???... i don't know what to make of your posts anymore lordsutch). 

My point is generalizing (drawing broad conclusions from a set of data) from ten cases is stupid, especially when there's no particular reason to believe Pennsylvania's roundabouts are substantially different from roundabouts in other states (or road users in Pennsylvania are substantially different from those in other states), particularly neighboring ones.
Yet any DOT will happily report how much safer roundabouts are when there are 19 accidents on roundabout compared to 20 on signalized intersection before that.
It is very difficult to obtain reliable data on real life objects. It is easy to have two identical cages with 20 mice each. It is much more difficult to have 2 identical groups of 20 people - someone will move, someone goes on diet, gets another job etc - although clinical trials of new drugs attempt to do that with reasonable success.  It is almost impossible to have 20 identical interchanges in 2 identical towns - because there are no identical towns with identical traffic patterns to begin with.
It is still possible to extract some information from large datasets, but often there is no interest and/or knowledge required to do that, I am not sure if there is even data collection system in place...

tradephoric

For a long time I never questioned the safety benefits of roundabouts.  The national studies said that there was a big decrease in injury crashes and a drop in total crashes.  Worst case was that roundabouts would see about the same number of crashes as the intersection they replaced but the crashes would be less severe resulting in fewer injury accidents.  Even a slight increase in total crashes would be ok in my mind since there would be a reduction in injury crashes.  But then I started reading articles about how some roundabouts were seeing crashes triple and quadruple.  In the Ann Arbor roundabout there was a 10 fold increase in crashes (going from 17 crashes before the roundabout to 170 crashes after).  The crash results of these more complex roundabouts weren't matching the national studies.  I had to find out why.

The main IIHS roundabout study was weighted heavily towards single-lane and simple multi-lane roundabouts (where main-street has multiple circulating lanes but side-street only has one circulating lane... ie. a 2x1 roundabout).  Some of the roundabouts being built today are much more complex than the roundabouts analyzed in the IIHS study.  Most of my posts have focused on a very specific type of modern roundabout — roundabouts with two-circulating lanes throughout the entire roundabout.  I wanted to see if these complex multi-lane roundabouts have the same safety benefits as the national studies suggest.  After querying out a database of over 5,000 roundabouts, I came up with a list of roughly 50 roundabouts that met specific criteria.  From there I found crash rates for roughly half of them.  The average crash rate for these complex multi-lane roundabouts is 4.09.  Just to put that crash rate into perspective, MassDOT reported that their statewide crash rate at signalized intersections is 0.77.  Interestingly, there is an inverse relationship between traffic volumes and the crash rates of signalized intersections.  Some may argue that if I looked at the highest volume signalized intersections in Massachusetts that they too would have high crash rates, but the research suggests the exact opposite would happen.  There is safety in numbers at signalized intersections.

Agencies don't like to see crash rates exceed 2.0 MEV and a crash rate of 4.09 is unacceptable to most road agencies.  If agencies knew in advance that a roundabout they were proposing would have a crash rate of 4.09 after construction they probably wouldn't build it in the first place (and i'm talking about looking at crash data years after construction is complete).  Should these 2x2 and 3x2 roundabouts be built in the first place if they end up having crash rates that are exceedingly high?  The problem is most agencies probably don't recognize there is a problem.  They truly believe that there will be a reduction in crashes since that is what the national studies told them.

tradephoric

Mayor Mike Moore says roundabout in Jeffersonville was a 'bad idea'
http://www.wdrb.com/story/32722447/concerns-about-jeffersonville-roundabout

It's hard to gain public support for roundabouts when the Mayor says the roundabout was a "bad idea" .   Mike has every right to say what he did as the roundabout has been the site of 134 crashes in the first 9 months of operation.  In response, INDOT gave the "learning curve"  argument saying that it can take up to 12 months for people to get used to a new roundabout configuration.  It's nothing more than wishful thinking that crashes will go down.  A complex 3-lane roundabout was built at the terminus of M-5 in Commerce, Michigan in 2011.  Ever since then it's been common for the roundabout to experience over 100 crashes a year.  In 2015 it was the highest crash prone intersection in the entire state of Michigan with 186 total crashes.  Michigan drivers have had 4 years to learn the roundabout, yet there is still a reported crash every other day at this roundabout.  How long does this supposed "learning curve"  take?

tradephoric

Gatineau roundabout logs most crashes of any intersection in Quebec
352 crashes – 213 of them resulting in major damage – occurred there over 5-year period
http://www.cbc.ca/news/canada/ottawa/worst-quebec-intersection-1.3735210

With the Allumettières and Saint-Joseph Blvd. roundabout averaging over 70 crashes a year, this 2x2 multi-lane roundabout has a crash rate approaching 5.0 MEV.

tradephoric

The Lakeville, MN roundabout at CSAH 50 / CSAH60 has been open for a year now.   According to a recent article, there have been 127 crashes in the first 10 months of operations.  In 2011, Dakota County put together an intersection study and came up with the following crash predictions for the roundabout alternative:


https://www.co.dakota.mn.us/Transportation/RoadStudies/Documents/CR50-CR60FullReport.pdf
http://sunthisweek.com/2016/07/15/some-drivers-still-adjusting-to-lakeville-roundabout/

The study predicted the roundabout would experience 7 PDO crashes and a crash rate of 0.44 for the "full planned growth"  period.  Assuming an AADT of 52,000, 127 crashes would equate to a crash rate of 6.69.  Put another way the current crash rate is 1400% higher than the study predicted.  It would take 19 years of 7 PDO crashes (19 x 7 = 133) to catch up to the current number of crashes that took place in the roundabout's first year of operation. 

There was another crash prone 2x2 multi-lane roundabout in Richfield, Minnesota that underwent a lot of study and the lessons learned from that roundabout were incorporated into the Lakeville design.  Even though Minnesota's best practices for roundabout design were included in the Lakeville roundabout, total crashes are still way too high.

tradephoric

If you think about it, the leg of a large traffic circle resembles a standard T-intersection.  Drivers at a T-intersection can safely turn into the right lane even when there is an approaching vehicle in the left lane.  The problem is drivers are treating modern roundabouts the same way they treat large traffic circles.  The driver turning into the right lane just assumes the left-lane driver will continue circulating around the circle, which is wrong.  This picture really highlights the "circular road"  misconception and was captured from a webinar Washington County, Minnesota put together.   It touches on a lot of the issues being discussed in this thread:



Below is the full hour long webinar for anyone interested along with some excerpts I found most interesting.

https://www.youtube.com/watch?v=Ycy0UD4kuw0

Quote5:30  The Radio Drive roundabout... "was averaging 40 crashes per year which was very alarming.  By and large they were certainly all very minor crashes; low speeds, low impact; but certainly from a public relations perspective as far as wanting to implement more roundabouts to help promote traffic safety, this was really a problem and something that we had to address."

7:46  Why is the difference between a modern roundabout and traffic circle important?  ... "it is important because especially as you start to get multiple lanes involved, the legal operations and the proper driver behavior we expect are very different between these two.  A single lane configuration can kind of be safely ignored but in multi-lane conditions it is certainly a problem."

17:08 Prof Hourdos discusses some of the changes made to the 66th & Portland roundabout.  He mentions the turn arrows were changed from "fish-hooks"  to the standard style.  However, if you look at the most recent aerial imagery they have switched back to the "fish-hook"  style. 

21:01 Discusses the results of the 66th & Portland roundabout study.  "The before and after study happened in two stages.  We collected video on the before... 3 months after the changes... and one year after.  There was a slight reduction in yield violations in the immediate 3 month after study.  Unfortunately these were not sustained in the one year after, in which case we actually saw a 30% increase in yield violations."  

23:20  "Because we did a lot of changes [to the 66th & Portland roundabout] at the same time, one clue that led us to believe that the additional signs (probably the change of the "fish-hook"  signs to the straight) and the extension of the solid line at the entrance (was effective) is that at the same time we saw a reduction in violations we also saw a similar reduction to the incorrect lane choice. 

24:25 Conclusion from Richfield is the changes did not have an effect on yield violations but showed improvement on incorrect lane change and turning violations.

26:05 Extend analysis done in Richfield to three additional roundabouts to solidify Minnesota's understanding of what solutions work to reduce crashes at multi-lane roundabouts. 
-University Dr & 5th Ave in St. Cloud
-Hwy 22 at Adams St and Madison Ave in Mankato
-50/60 Roundabout in Lakeville

28:10  "The roundabout in Lakeville implemented some of the guidance provided from Richfield.  They did use the straight arrows instead of the fish-hook arrows and they also extended the solid line 100 feet from the yield line.  This roundabout also has overhead lane designation signs, which is different from the Richfield roundabout (which did not have overhead signs)."
 
31:02  I noticed it says changes in signs and lane markings took place a year after the roundabout in Mankato opened and data collections was repeated.  Agencies are quick to argue that the first year of crash data may have a disproportionately high number of crashes as drivers get use to using the roundabout.  Yet it seems like Minnesota is going to analyze any roundabout changes they make and determine their effectiveness by analyzing the before data (first year of roundabout being opened when drivers don't know how to drive it yet) to after data (roundabout has been opened for a year, and drivers are starting to get used to it).  Agencies shouldn't have it both ways.

36:59  "Do whatever we can to prevent users from having the "circle road"  misconception.

44:15 "We want to minimize crash opportunities.  We have been doing what we can to avoid these 2x2 configurations.  We originally designed for a 20 year volume which really isn't a bad thing in a non-roundabout context but just creates some of those opportunities for mistakes.  It's better to design it for future expansion but don't build it all now if you don't need to. "

45:20 Discusses the roundabout diet (downsizing 2x2 roundabouts to 1x2 roundabouts).

47:12 "The current MUTCD standard sign really implies the concept of a "circle road".  Is there a different type of sign that could convey this more effectively?"  He also discusses yellow delineators, spaced out chevrons and different types of striping.  Overall a very interesting webinar and will love to hear the results of their study.

cjw2001

#712
The roundabout at this ramp in Carmel on the south end of Clay Terrace has signs that say "to both lanes" under the yield sign.  Unfortunately street view isn't available from the front side of the signs.  The square sign says "to both lanes".  Simple but effective.  Also that street view image is a bit compressed -- there are two full lanes entering at that set of yield signs even though it doesn't look big enough in that image.

tradephoric

Quote from: cjw2001 on August 29, 2016, 05:17:34 PM
The roundabout at this ramp in Carmel on the south end of Clay Terrace has signs that say "to both lanes" under the yield sign.  Unfortunately street view isn't available from the front side of the signs.  The square sign says "to both lanes".  Simple but effective.  Also that street view image is a bit compressed -- there are two full lanes entering at that set of yield signs even though it doesn't look big enough in that image.

Here's some signage that is used at the Homer Watson & Block line roundabout:

tradephoric

American's take on the Dutch turbo.  This could help reduce the number of failure to yield crashes if the "circular road" misconception described in the webinar really is a problem.

https://www.google.com/maps/@43.3849407,-87.9229875,97m/data=!3m1!1e3

jakeroot

Quote from: tradephoric on August 30, 2016, 11:20:17 AM
American's take on the Dutch turbo.  This could help reduce the number of failure to yield crashes if the "circular road" misconception described in the webinar really is a problem.

Outside of Victoria International Airport on Vancouver Island, BC, there is a genuine turbo roundabout. You can read more about it in this PDF: http://goo.gl/2gXcGn

Google Maps: http://goo.gl/o7xm5i




DaBigE

Quote from: tradephoric on August 30, 2016, 11:20:17 AM
American's take on the Dutch turbo.  This could help reduce the number of failure to yield crashes if the "circular road" misconception described in the webinar really is a problem.

https://www.google.com/maps/@43.3849407,-87.9229875,97m/data=!3m1!1e3

It may be splitting hairs to some, but I wouldn't classify that one as more than a 2x2 with spirals and right-turn bypass lanes. Turbos tend to feature raised dividers within the circulatory, not just paint.
"We gotta find this road, it's like Bob's road!" - Rabbit, Twister

tradephoric

Take a look at these two pictures below.  In both cases the yellow car assumes the blue car in the inner-most lane of the roundabout will continue circulating through the roundabout ("the circular road misconception" ).  With a 2x2 roundabout the assumption can lead to a crash when the blue car actually takes the red path.    However, with a 1x2 roundabout this assumption by the yellow driver doesn't lead to a crash since both paths the blue car can take is in conflict with the yellow car (so the yellow car is forced to wait regardless.. since it's tough to squeeze two cars into one lane).



This is a fundamental problem I see with 2x2 roundabouts.  How do you design 2x2 roundabouts so the driver in the yellow car knows that the blue car may actually be exiting the roundabout?  In too many cases the yellow car wrongfully assumes the blue car is going to continue circulating around the roundabout and pulls out... leading to a failure to yield crash.

Rothman

Quote from: tradephoric on August 30, 2016, 02:56:01 PM


In too many cases the yellow car wrongfully assumes the blue car is going to continue circulating around the roundabout and pulls out... leading to a failure to yield crash.


Yep.  When cars don't follow the rules of the road, it does sometimes lead to crashes.  Driver error, not design.
Please note: All comments here represent my own personal opinion and do not reflect the official position(s) of NYSDOT.

english si

Quote from: tradephoric on August 30, 2016, 02:56:01 PMThis is a fundamental problem I see with 2x2 roundabouts.  How do you design 2x2 roundabouts so the driver in the yellow car knows that the blue car may actually be exiting the roundabout?  In too many cases the yellow car wrongfully assumes the blue car is going to continue circulating around the roundabout and pulls out... leading to a failure to yield crash.
Design it as an ACTUAL circular road, and so exiting the roundabout requires indication (using your blinkers) to follow the rules of the road (eg 186, which is the British rule governing these matters) as you are turning off it!


Also, in the second example you give the yellow car has either the blue car crossing its path, or seeking the same space on the (at that point one-lane) roundabout. So that's driver error, not a design flaw. But I see what you mean.

DaBigE

Quote from: Rothman on August 30, 2016, 03:15:06 PM
Quote from: tradephoric on August 30, 2016, 02:56:01 PM


In too many cases the yellow car wrongfully assumes the blue car is going to continue circulating around the roundabout and pulls out... leading to a failure to yield crash.


Yep.  When cars don't follow the rules of the road, it does sometimes lead to crashes.  Driver error, not design.

+1.
How many more clues will the entering driver ignore? Traffic within the circulatory has the ROW. Period. End of story. Doesn't matter if you're in the UK or the US. Secondly, if the circulating driver is not signalling left as their circulating, you assume the worst-case, that they're going to exit. Further, many locales post supplementary plaques "To Traffic From Left" or "To Both Lanes" for this very reason. Not much more can be done from a design standpoint other than removing a circulating lane.
"We gotta find this road, it's like Bob's road!" - Rabbit, Twister

tradephoric

Quote from: DaBigE on August 30, 2016, 04:29:31 PM
How many more clues will the entering driver ignore? Traffic within the circulatory has the ROW. Period. End of story. Doesn't matter if you're in the UK or the US. Secondly, if the circulating driver is not signalling left as their circulating, you assume the worst-case, that they're going to exit. Further, many locales post supplementary plaques "To Traffic From Left" or "To Both Lanes" for this very reason. Not much more can be done from a design standpoint other than removing a circulating lane.

BINGO!  Roundabout diets that eliminate circulating lanes at these 2x2 multi-lane roundabouts are becoming more and more common.  Nobody is suggesting that these high crash rate roundabouts be replaced with traffic signals - instead modify them to become low crash rate roundabouts.  I want to reiterate what Joe Gustafson said in the Minnesota webinar:

Quote"We want to minimize crash opportunities.  We have been doing what we can to avoid these 2x2 configurations.  We originally designed for a 20 year volume which really isn't a bad thing in a non-roundabout context but just creates some of those opportunities for mistakes.  It's better to design it for future expansion but don't build it all now if you don't need to."

kalvado

Quote from: tradephoric on August 30, 2016, 06:00:25 PM
Quote from: DaBigE on August 30, 2016, 04:29:31 PM
How many more clues will the entering driver ignore? Traffic within the circulatory has the ROW. Period. End of story. Doesn't matter if you're in the UK or the US. Secondly, if the circulating driver is not signalling left as their circulating, you assume the worst-case, that they're going to exit. Further, many locales post supplementary plaques "To Traffic From Left" or "To Both Lanes" for this very reason. Not much more can be done from a design standpoint other than removing a circulating lane.

BINGO!  Roundabout diets that eliminate circulating lanes at these 2x2 multi-lane roundabouts are becoming more and more common.  Nobody is suggesting that these high crash rate roundabouts be replaced with traffic signals - instead modify them to become low crash rate roundabouts.  I want to reiterate what Joe Gustafson said in the Minnesota webinar:

Quote"We want to minimize crash opportunities.  We have been doing what we can to avoid these 2x2 configurations.  We originally designed for a 20 year volume which really isn't a bad thing in a non-roundabout context but just creates some of those opportunities for mistakes.  It's better to design it for future expansion but don't build it all now if you don't need to."

That way single roundabout can mean job security for the life of the company.
I think that is really a great approach to burning tax dollars providing safe and efficient road environment!

jakeroot

Quote from: tradephoric on August 30, 2016, 06:00:25 PM
Quote from: DaBigE on August 30, 2016, 04:29:31 PM
How many more clues will the entering driver ignore? Traffic within the circulatory has the ROW. Period. End of story. Doesn't matter if you're in the UK or the US. Secondly, if the circulating driver is not signalling left as their circulating, you assume the worst-case, that they're going to exit. Further, many locales post supplementary plaques "To Traffic From Left" or "To Both Lanes" for this very reason. Not much more can be done from a design standpoint other than removing a circulating lane.

BINGO!  Roundabout diets that eliminate circulating lanes at these 2x2 multi-lane roundabouts are becoming more and more common.  Nobody is suggesting that these high crash rate roundabouts be replaced with traffic signals - instead modify them to become low crash rate roundabouts.  I want to reiterate what Joe Gustafson said in the Minnesota webinar:

Quote"We want to minimize crash opportunities.  We have been doing what we can to avoid these 2x2 configurations.  We originally designed for a 20 year volume which really isn't a bad thing in a non-roundabout context but just creates some of those opportunities for mistakes.  It's better to design it for future expansion but don't build it all now if you don't need to."

Well that's fine, but what happens in 20 years when (or if) they meet that volume? If they re-stripe that second lane, will it cause another bunch of headaches? Will drivers be well enough used to the roundabout at that point that collisions are pretty much unheard of? Or, do they replace the junction with a signal?

tradephoric

^^^
Good questions Jake.  If we are OK with seeing crashes on almost a daily basis at some of these complex roundabouts, then keep building them.  If not, then maybe a traffic signal does becomes a better alternative.  In recent years NW Green Bay has constructed 7 triple-lane roundabouts in about a square mile area.  I think that's a mistake as triple lane roundabouts have not been well proven.  I can cite 4 triple-lane roundabouts that have already undergone roundabout diets and another triple-laner where a diet is planned.  The triple-lane roundabouts that haven't had circulating lanes removed are experiencing triple digit crashes each year (like the M-5/Pontiac Trail roundabout in Commerce, Michigan which had 186 crashes in 2015).   It sure seems like there is a limit to how complex a roundabout can be before vehicles start crashing into each other on nearly a daily basis.



Opinions expressed here on belong solely to the poster and do not represent or reflect the opinions or beliefs of AARoads, its creators and/or associates.