Crash prone 'modern roundabouts'

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

Previous topic - Next topic

tradephoric

In 2013, a modern roundabout was constructed at the intersection of Ellsworth & State in Ann Arbor, Michigan.  Three years prior to the roundabout construction, the intersection averaged 16 crashes per year.  The first year after the roundabout was completed, Ellsworth & State experienced the highest number of accidents in the entire state of Michigan (168 total crashes). 

Ann Arbor roundabout has most crashes in Michigan
http://www.mlive.com/news/ann-arbor/index.ssf/2015/03/roundabout_on_ann_arbors_south.html
http://www.semcog.org/Data-and-Maps/Crash-and-Road-Data/Point_Id/81016689/view/RoadIntersectionCrashDetail

Are there other crash prone modern roundabouts that you are aware of? 


froggie

Speaks more to the (in)flexibility of Michigan drivers than it does any inherent safety risk with roundabouts.  New roundabouts elsewhere have seen major drops in the number of crashes.

NE2

pre-1945 Florida route log

I accept and respect your identity as long as it's not dumb shit like "identifying as a vaccinated attack helicopter".

jeffandnicole

Here it is: https://goo.gl/maps/NGJmj .  Is that a 'standard' roundabout?  You almost drive straight out of the roundabout, rather than curve out.  Plus that right turn only bypass comes out very close to the actual roundabout. 

I think there's a legit design issue here.

tradephoric

Quote from: froggie on May 18, 2015, 03:57:00 PM
Speaks more to the (in)flexibility of Michigan drivers than it does any inherent safety risk with roundabouts.  New roundabouts elsewhere have seen major drops in the number of crashes.

The safety statistics most often cited to the public are based on a 2000 study done by the Insurance Institute for Highway Safety.  A few points should be considered.  Of the 24 roundabouts analyzed, only 3 were previously signalized intersections (21 intersections were non-signalized).  Also, no roundabout had an AADT of greater than 31,000 vehicles.  The study observed that modern roundabouts have the following safety benefits:
-   38% reduction in total crashes
-   76% reduction in injury crashes
-   90% reduction in fatal and incapacitating-injury crashes

In 2011, the Wisconsin Department of Transportation funded a study that analyzed 24 roundabouts in their state.  In the Wisconsin study, a roundabout with an AADT of 70,000 was analyzed (over twice the traffic volumes of any roundabout included in the 2000 study).  Of the 24 roundabouts analyzed, 11 saw crashes increase after the roundabout was completed and 6 saw crashes increase by over 40%.   Overall, they observed the following safety benefits:
-   9% reduction in total crashes
-   52% reduction in injury crashes

Here is a summary of the roundabouts analyzed in the Wisconsin study:




Big John

^^ #7 STH 141?
WisDOT should know better than that. :pan:

tradephoric

Quote from: Big John on May 18, 2015, 06:36:15 PM
^^ #7 STH 141?
WisDOT should know better than that. :pan:

Does traffic back up through the roundabout when a train goes by?

https://www.google.com/maps/@44.4588082,-87.950588,141m/data=!3m1!1e3

jakeroot

First year stats are such horse shit. Driver familiarity is a major preventer of collisions at intersections (drivers know what to expect; driver behavior, what each lane does, signal phases, etc). Changing literally anything will throw people off and you're probably gonna have a few additional collisions. I'll admit, that's a big jump, but I guarantee you tradephoric, the collisions will fall. Let's revisit this intersection in a couple years.

tradephoric

Quote from: jakeroot on May 19, 2015, 02:38:26 AM
First year stats are such horse shit. Driver familiarity is a major preventer of collisions at intersections (drivers know what to expect; driver behavior, what each lane does, signal phases, etc). Changing literally anything will throw people off and you're probably gonna have a few additional collisions. I'll admit, that's a big jump, but I guarantee you tradephoric, the collisions will fall. Let's revisit this intersection in a couple years.

Seeing a spike in crashes at SE Michigan roundabouts is not uncommon.  I touched on this in a previous thread.

Quote from: tradephoric on March 30, 2015, 02:32:54 PM
The chart below looks at crash data of 16 major roundabouts built throughout SE Michigan between 2006-2009.  While PDO accidents spike the first year after roundabout construction, injury accidents drop from day one.  By year two people seem to get use to them and the PDO steadily declines from the initial spike.  This has at least been the trend for SE Michigan.


If the crashes at Ellsworth & State jumped from 16 to 32, that would be par for the course.  However, it jumped from 16 to 168 crashes.  There's a problem here.  To put this in perspective, the intersection with the 2nd most crashes in Michigan for 2014 had 87 crashes (12 Mile & Orchard Lake). 

Bickendan

Having glanced at the roundabout in question in Streetview, I don't see anything inherently wrong with the design. My gut says it's driver inflexibility and impatience, which will eventually largely evaporate as drivers get familiar with the design.

tradephoric

Here are a list of roundabouts where total crashes have doubled since the completion of the roundabout (when averaging the 3 years before to 3 years after crash data).   To view the most up to date aerials you may have to input the coordinates into GE.  Do these roundabouts share any attributes that could explain the increase in crashes? 

(950%) Ellsworth Rd & State Rd (Ann Arbor, Michigan):
https://www.google.com/maps/@42.22941,-83.73873,17z/data=!3m1!1e3?dg=brw

(335%) Livernois Rd & Hamlin Rd (Rochester Hills, Michigan):
https://www.google.com/maps/@42.6518911,-83.1526542,137m/data=!3m1!1e3

(297%) Canal & 25th Street (Milwaulkee, Wicsonsin):
https://www.google.com/maps/@43.0311071,-87.9434376,92m/data=!3m1!1e3

(266%) SB US23 & Lee Road (Brighton, Michigan):
https://www.google.com/maps/@42.5059638,-83.759329,117m/data=!3m1!1e3

(135%) Cass Ave & Romeo Plank (Clinton, Michigan):
https://www.google.com/maps/@42.6113704,-82.9322543,117m/data=!3m1!1e3

(127%) STH 32/STH57 & Broadway (De Pere, Wisconsin):
https://www.google.com/maps/@44.4474383,-88.0602622,117m/data=!3m1!1e3

jeffandnicole

Quote from: Bickendan on May 19, 2015, 01:08:08 PM
Having glanced at the roundabout in question in Streetview, I don't see anything inherently wrong with the design. My gut says it's driver inflexibility and impatience, which will eventually largely evaporate as drivers get familiar with the design.

But 3 years?   Usually familiarity takes a week for most drivers; maybe a month for more infrequent drivers.

tradephoric

#12
To jeffandnicole's point, in these studies no crash data is analyzed from the year the roundabout was constructed.  If a roundabout was completed in October, drivers have at least a few months to get use to before any crashes are included in the analysis.

For anyone interested, here's a link to the Wisconsin study:
http://www.topslab.wisc.edu/programs/safety/projects/roundabouts/WI%20Roundabout%20Evaluation%20Volume%202%20Safety.pdf



jakeroot

Quote from: tradephoric on May 19, 2015, 02:27:06 PM
For anyone interested, here's a link to the Wisconsin study:
http://www.topslab.wisc.edu/programs/safety/projects/roundabouts/WI%20Roundabout%20Evaluation%20Volume%202%20Safety.pdf

I'm sure it's in there, but I can't find it. Any idea what the common collisions were? (Like sideswipes, T-bone, etc).

Even if collisions are higher with roundabouts, I'm tempted to say that's preferable to, say, *consistent* fatal collisions. Have these roundabouts had fatal collisions? And were they contributed to driver error or a design fault? I ask that, because, roundabouts should work well -- their design is not the inherent cause of these collisions.

6a

#14
Angle and sideswipe both had gigantic increases (at the Ann Arbor intersection). Like, those two accounted for almost all of the increase. I wonder if jeffandnicole is on to something regarding that right turn bypass bit. Looking at the crash increases it really does make me wonder if it's just people not staying in the correct lane. I'm not terribly familiar with the area - is this intersection near the university?

Closer to home here, I'm really curious to see what happens to this one, currently under construction:



Traffic count sits at 50,000 currently, and even with left turns banned on the E-W route (SR 161) it's the third worst intersection for crashes in the city of Dublin. I believe it will be the first in the area to have a three lane section, but Dublin is the king of roundabouts locally, so like I said, I'm really curious to see the results.

Edit: oh wow, on that Ann Arbor one, you can see right across it. Could that be a cause for distraction?


jakeroot

Quote from: 6a on May 19, 2015, 04:16:01 PM
Edit: oh wow, on that Ann Arbor one, you can see right across it. Could that be a cause for distraction?

That's a very good point. One of the reasons the Carmel, Indiana roundabouts are so successful is because of their decoration.

johndoe

Cool thread, good info.  I wonder if this roundabout has tangents on entry and what the fastest path speeds are. 

tradephoric

Quote from: jakeroot on May 19, 2015, 03:51:46 PM
Quote from: tradephoric on May 19, 2015, 02:27:06 PM
For anyone interested, here's a link to the Wisconsin study:
http://www.topslab.wisc.edu/programs/safety/projects/roundabouts/WI%20Roundabout%20Evaluation%20Volume%202%20Safety.pdf

I'm sure it's in there, but I can't find it. Any idea what the common collisions were? (Like sideswipes, T-bone, etc).

Even if collisions are higher with roundabouts, I'm tempted to say that's preferable to, say, *consistent* fatal collisions. Have these roundabouts had fatal collisions? And were they contributed to driver error or a design fault?  I ask that, because, roundabouts should work well -- their design is not the inherent cause of these collisions.

Check out Table 4 on page 20.  There was big increases in sideswipe-same direction (SSS) crashes in the Wisconsin study (increasing from 10 in the before condition to 75 in the after).   




Brian556

I said it before, I do not like multi-lane roundabouts. They are just too complicated, and can confuse drivers.  Only single-lane roundabouts should be allowed. Even single lane roundabouts can have issues, especially if drivers enter them too fast. Also, if there is too high of a volume of traffic entering from one direction, the other directions can have a hard time even getting a chance to enter the roundabout. The ones in Windermere, FL have this problem, particularly the southern one.
https://www.google.com/maps/place/Windermere,+FL/@28.4949864,-81.5342674,398m/data=!3m1!1e3!4m2!3m1!1s0x88e7635aece2fef1:0xfac6dfbac566065e

There is a new one in Flower Mound, TX that I don't care for due to the rather high traffic volume. To me it's a pain in the ass having to deal with the high traffic volume, and trying to figure out when it is safe to enter the roundabout, especially from a stopped position. I'd personally rather have a signal at this particular location:
https://www.google.com/maps/@33.0664592,-97.0879309,95m/data=!3m1!1e3

jakeroot

Quote from: Brian556 on May 19, 2015, 10:35:24 PM
I do not like multi-lane roundabouts. They are just too complicated, and can confuse drivers.

That is a gross over-generalization. There are plenty of multi-lane roundabouts (near me) that have zero issues.

In your defence, however, there are certainly areas in this country where the "courtesy" behaviour (necessary for a successful roundabout) doesn't exist. Texas springs to mind, along with much of the south and perhaps Midwest. Places like rural New England, the Pacific Northwest, and perhaps pockets of the Southwest are the best candidates for roundabouts, since they have the most courteous (and thus the most observant) drivers (at least in my experience).

renegade

Re:  State and Ellsworth in Ann Arbor:  It's a well-thought-out roundabout  Part of the problem is, northbound on State, there's a traffic signal which favors the side road about 150 yards north of the roundabout.  All the traffic-flow in the world is rendered useless if there's no place for it to go.  The other problem is that no one wants to learn to maneuver properly.  Some enter the roundabout without yielding to traffic already within, while others will gladly stop in the middle to let others in.  Hell, I personally witnessed on driver who said, "Fuck this" and turned left instead of going around.

Don’t ask me how I know.  Just understand that I do.

KEK Inc.

Quote from: jakeroot on May 19, 2015, 11:53:34 PM
Quote from: Brian556 on May 19, 2015, 10:35:24 PM
I do not like multi-lane roundabouts. They are just too complicated, and can confuse drivers.

That is a gross over-generalization. There are plenty of multi-lane roundabouts (near me) that have zero issues.

In your defence, however, there are certainly areas in this country where the "courtesy" behaviour (necessary for a successful roundabout) doesn't exist. Texas springs to mind, along with much of the south and perhaps Midwest. Places like rural New England, the Pacific Northwest, and perhaps pockets of the Southwest are the best candidates for roundabouts, since they have the most courteous (and thus the most observant) drivers (at least in my experience).

I've run into dangerous congestion at these roundabouts when I went kayaking in Olympia last.

https://www.google.com/maps/place/Olympia,+WA/@47.0449015,-122.9119405,18z/data=!4m2!3m1!1s0x5491c9c1ae285569:0x4f146197e2881b83
Take the road less traveled.

jakeroot

Quote from: KEK Inc. on May 20, 2015, 02:45:08 AM
Quote from: jakeroot on May 19, 2015, 11:53:34 PM
Quote from: Brian556 on May 19, 2015, 10:35:24 PM
I do not like multi-lane roundabouts. They are just too complicated, and can confuse drivers.

That is a gross over-generalization. There are plenty of multi-lane roundabouts (near me) that have zero issues.

In your defence, however, there are certainly areas in this country where the "courtesy" behaviour (necessary for a successful roundabout) doesn't exist. Texas springs to mind, along with much of the south and perhaps Midwest. Places like rural New England, the Pacific Northwest, and perhaps pockets of the Southwest are the best candidates for roundabouts, since they have the most courteous (and thus the most observant) drivers (at least in my experience).

I've run into dangerous congestion at these roundabouts when I went kayaking in Olympia last.

https://www.google.com/maps/place/Olympia,+WA/@47.0449015,-122.9119405,18z/data=!4m2!3m1!1s0x5491c9c1ae285569:0x4f146197e2881b83

I've gone through those a couple of times. Never had any issues, but the odd manner in which they facilitate through traffic would more than likely exacerbate traffic flow.

Any idea if Olympia publishes any stats on these roundabouts? The only documents I can find on the so-called "Gateway Project" (as it was called in 2003 -- how original) only seem to be promoting it, if you know what I mean.

KEK Inc.

I think they're a bit too close to each other.  I also went during rush hour.
Take the road less traveled.

jakeroot

Quote from: KEK Inc. on May 20, 2015, 03:26:57 AM
I think they're a bit too close to each other.  I also went during rush hour.

I'd rather have two closely spaced roundabouts than two closely spaced signals. Then again, they don't seem that close?



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.