Crash prone 'modern roundabouts'

[kphoger]

I'm curious to know when the roundabout in question was constructed.  The engineers should have known better than to build that large of an ICD and that straight of approaches if it was constructed in the last decade.

You seem to be implying that a smaller ICD would have been more effective at reducing the number of crashes.  To your original question, the M-5/Pontiac Trail roundabout was constructed in 2011.  While there was an increase in crashes after the roundabout was constructed, the crash rate wasn't as high as other multi-lane roundabouts cited in this thread.  It wasn't until 2015 when the roundabout had a big spike in crashes (likely due to it being part of a major detour route).  Now you are throwing the roundabout design under the bus because of an abnormally high number of crashes in 2015.  Yes, a roundabout that experiences 100 crashes in a year is a problem, but you are trying to convince me a smaller ICD would have done better.  I'm not buying it. 

My assertion is that the roundabout should be designed with a smaller ICD and more-flared approaches in order to reduce driver speed, which is what the fencing is supposed to accomplish.

The approaches do have some flare, but not enough. If a driver is physically able to enter the circulating roadway at such speeds, then the geometry of the roundabout is not doing its job. Approaches need to be flared enough that a driver cannot exceed the target speed upon entry. A smaller ICD does the same for vehicles already in the roadway; this, IMO, is a lesser issue but does contribute to a driver's perception as to whether or not it's OK to push the envelope.

[tradephoric]

My assertion is that the roundabout should be designed with a smaller ICD and more-flared approaches in order to reduce driver speed, which is what the fencing is supposed to accomplish.

It seems like you can get more flare at larger diameter roundabouts.  If a roundabout is too small you are limited to how much you can flare the approach out (before you overshoot the circle so to speak).  I have a few other points why I believe larger ICD roundabouts may be more effective at reducing crashes:

#1. You don't want the fastest path of a roundabout to resemble a straight line if your goal is to keep speeds down.  Which red line resembles a drag strip? 

Small ICD roundabout:


Large ICD roundabout:


#2. Drivers have more reaction time when the approach legs of a roundabout are spread out.  Imagine CAR A enters the roundabout mistakenly believing CAR B is about to exit.  Would u want CAR B to take 1.7 seconds or 3.4 seconds to travel the blue path?  Now it's true that the average circulating speeds might be SLIGHTLY higher at the larger ICD roundabout but the speeds wouldn't be drastically different.  Instead of 3.4 seconds reaction time at the large ICD roundabout, it might really be 3 seconds because the car is circulating at 23 mph as opposed to 20 mph.  Still, would you want 1.7 second reaction time or 3 second reaction time when taking evasive action?

Small ICD roundabout:


Large ICD roundabout:


#3.  Look at the empirical evidence.  The 2x2 multi-lane roundabouts with smaller central island diameters have a greater increase in crashes than the roundabouts with larger central island diameters:

[tradephoric]

One other point about the fencing idea.  Sure, you restrict the visibility of vehicles inside the roundabout but you also restrict the visibility of pedestrians inside the crosswalk (who are crossing the exit leg of the roundabout).  Is it really a good idea to restrict the visibility of pedestrians?

[jakeroot]

One other point about the fencing idea.  Sure, you restrict the visibility of vehicles inside the roundabout but you also restrict the visibility of pedestrians inside the crosswalk (who are crossing the exit leg of the roundabout).  Is it really a good idea to restrict the visibility of pedestrians?

I think it's better to hide the pedestrians until they are about to cross each carriageway. That way, overly-courteous drivers in the other carriageway don't stop too early. Use those yellow flashing beacons if visibility is truly an issue.

[jeffandnicole]

Fencing also keeps pedestrians crossing where they should be crossing, rather than jumping out into the street wherever they feel like.  Good for motorists too, as they shouldn't come upon a ped in a random part of the road. As long as the fence is 3 feet high or so, visibility isn't reduced for either peds or motorists.

[kalvado]

One other point about the fencing idea.  Sure, you restrict the visibility of vehicles inside the roundabout but you also restrict the visibility of pedestrians inside the crosswalk (who are crossing the exit leg of the roundabout).  Is it really a good idea to restrict the visibility of pedestrians?

I think it's better to hide the pedestrians until they are about to cross each carriageway. That way, overly-courteous drivers in the other carriageway don't stop too early. Use those yellow flashing beacons if visibility is truly an issue.

It's even better to hide pedestrians until they are past windshield.

[tradephoric]

The Lee Road roundabout in Livingston County has been previously discussed on this thread.  It had the second highest number of INJURY crashes in Livingston County over the past 5 years and by far the highest number of total crashes.

http://www.livingstondaily.com/story/news/local/community/livingston-county/2015/11/23/livingston-countys-worst-intersections-crashes/75170418/

The Livingston County Road Commission is now considering a proposal which would eliminate a circulating lane within the roundabout. 

http://www.livingstondaily.com/story/news/local/community/green-oak-township/2016/06/10/lee-road-roundabout-redo-considered/85645684/

Here seems to be the trend when it comes to complex 2x2 and 2x3 multi-lane roundabouts in the United States:

1.   Agencies propose a complex multi-lane roundabout, citing safety studies that are heavily skewed towards single-lane and/or simplified 2x1 multi-lane roundabouts (ie. the 2000 Insurance Institute for Highway Safety study).
2.   The multi-lane roundabout is built and there is a big spike in total crashes.
3.   Agencies "tweak"  the roundabout to combat the high crash rate, making minor pavement marking/signage changes.  While these tweaks can reduce the crash rate, they often leave a lot to be desired (ie. there is still too many crashes even after the tweaks).
4.   Agencies decide to remove a circulating lane inside the roundabout to simplify the design (converting 2x2 roundabouts to 2x1 or even single lane roundabouts).   Eliminating a circulating lane can often be done without affecting operation too much since the roundabout was originally designed for 2035 (or whatever) traffic volumes.  Simplifying the roundabout often solves the crash problem. 

Of the 2x2 and 2x3 roundabouts cited in this thread, many are seeing circulating lanes removes due to the high crash rates.  At what point do agencies question whether these complex roundabouts are acceptable in terms of safety?  Should these complex roundabouts be built in the first place if they are just going to be downsized within years of being constructed? 

[tradephoric]

The 10 Most Dangerous Roundabouts in Metro Detroit:

1. M-5/Martin Parkway at Pontiac Trail — Commerce Township Total crashes: 186

2. Van Dyke Ave. (M-53) at 18 1/2 Mile Road — Sterling Heights Total crashes: 124

3. State Street at Ellsworth Road — Ann Arbor/Pittsfield Township Total crashes: 114

4. Maple Road at Farmington Road — West Bloomfield Township Total crashes: 84

5. 14 Mile Road at Farmington Road — Farmington Hills/West Bloomfield Total crashes: 79

6. Maple Road at Drake Road — West Bloomfield Township Total crashes: 63

7. Lee Road at Whitmore Lake Road — Brighton/Green Oak Township Total crashes: 60

8. Martin Parkway at Oakley Park Road — Commerce Township Total crashes: 53

9. Hamlin Road at Livernois Road — Rochester Hills Total crashes: 41

10. Romeo Plank Road at 19 Mile Road — Clinton Township Total crashes: 37

http://womc.cbslocal.com/2016/05/11/the-10-most-dangerous-roundabouts-in-metro-detroit/

[tradephoric]

The roundabout at Van Dyke & 18 1/2 Mile in Metro Detroit was built over a decade ago.  In 2015 it had 124 crashes.  It may be an example of a roundabout that was overbuilt:

[tradephoric]

New 2015 crash data released for SE Michigan.  The Ellsworth roundabout in Ann Arbor (which averaged 15.5 crashes before the roundabout) has averaged 140 crashes the first 2 years of operation.    If you dig deeper with the numbers, there was 2 injury crashes at the intersection from 2011-2012 (before data) and 9 injury crashes from 2014-2015 (after data). 


http://semcog.org/Data-and-Maps/High-Frequency-Crash-Locations

So to recap...
Average total crashes before roundabout = 15.5
Average total crashes after roundabout = 140
That's a 803% increase    in total crashes....

Average injury crashes before roundabout = 1
Average injury crashes after roundabout = 4.5
That's a 350% increase    in injury crashes... 

[7/8]

New 2015 crash data released for SE Michigan.  The Ellsworth roundabout in Ann Arbor (which averaged 15.5 crashes before the roundabout) has averaged 140 crashes the first 2 years of operation.    If you dig deeper with the numbers, there was 2 injury crashes at the intersection from 2011-2012 (before data) and 9 injury crashes from 2014-2015 (after data). 


http://semcog.org/Data-and-Maps/High-Frequency-Crash-Locations

So to recap...
Average total crashes before roundabout = 15.5
Average total crashes after roundabout = 140
That's a 803% increase    in total crashes....

Average injury crashes before roundabout = 1
Average injury crashes after roundabout = 4.5
That's a 350% increase    in injury crashes...

One problem I see with this roundabout is the lack of any obstruction in the centre. Even though it's counter-intuitive, it's better to block people from being able to see across the middle, since it's an unnecessary distraction. Putting some tall plants in the middle would force people to pay attention to cars on their left, like they should.

I still agree though that those crash statistics are still crazy, especially the increase in injury crashes!
https://www.google.ca/maps/@42.2292761,-83.7391719,3a,75y,54.44h,74.65t/data=!3m7!1e1!3m5!1srAaeg2wFciLs9vpdpmQH9A!2e0!6s%2F%2Fgeo2.ggpht.com%2Fcbk%3Fpanoid%3DrAaeg2wFciLs9vpdpmQH9A%26output%3Dthumbnail%26cb_client%3Dmaps_sv.tactile.gps%26thumb%3D2%26w%3D203%26h%3D100%26yaw%3D250.00459%26pitch%3D0!7i13312!8i6656

[kalvado]

New 2015 crash data released for SE Michigan.  The Ellsworth roundabout in Ann Arbor (which averaged 15.5 crashes before the roundabout) has averaged 140 crashes the first 2 years of operation.    If you dig deeper with the numbers, there was 2 injury crashes at the intersection from 2011-2012 (before data) and 9 injury crashes from 2014-2015 (after data). 


http://semcog.org/Data-and-Maps/High-Frequency-Crash-Locations

So to recap...
Average total crashes before roundabout = 15.5
Average total crashes after roundabout = 140
That's a 803% increase    in total crashes....

Average injury crashes before roundabout = 1
Average injury crashes after roundabout = 4.5
That's a 350% increase    in injury crashes...

One problem I see with this roundabout is the lack of any obstruction in the centre. Even though it's counter-intuitive, it's better to block people from being able to see across the middle, since it's an unnecessary distraction. Putting some tall plants in the middle would force people to pay attention to cars on their left, like they should.

I still agree though that those crash statistics are still crazy, especially the increase in injury crashes!
https://www.google.ca/maps/@42.2292761,-83.7391719,3a,75y,54.44h,74.65t/data=!3m7!1e1!3m5!1srAaeg2wFciLs9vpdpmQH9A!2e0!6s%2F%2Fgeo2.ggpht.com%2Fcbk%3Fpanoid%3DrAaeg2wFciLs9vpdpmQH9A%26output%3Dthumbnail%26cb_client%3Dmaps_sv.tactile.gps%26thumb%3D2%26w%3D203%26h%3D100%26yaw%3D250.00459%26pitch%3D0!7i13312!8i6656

For a small roundabout like that, car across the circle is a potential conflict if making 270 turn. 

[7/8]

New 2015 crash data released for SE Michigan.  The Ellsworth roundabout in Ann Arbor (which averaged 15.5 crashes before the roundabout) has averaged 140 crashes the first 2 years of operation.    If you dig deeper with the numbers, there was 2 injury crashes at the intersection from 2011-2012 (before data) and 9 injury crashes from 2014-2015 (after data). 


http://semcog.org/Data-and-Maps/High-Frequency-Crash-Locations

So to recap…
Average total crashes before roundabout = 15.5
Average total crashes after roundabout = 140
That’s a 803% increase    in total crashes….

Average injury crashes before roundabout = 1
Average injury crashes after roundabout = 4.5
That’s a 350% increase    in injury crashes…

One problem I see with this roundabout is the lack of any obstruction in the centre. Even though it's counter-intuitive, it's better to block people from being able to see across the middle, since it's an unnecessary distraction. Putting some tall plants in the middle would force people to pay attention to cars on their left, like they should.

I still agree though that those crash statistics are still crazy, especially the increase in injury crashes!
https://www.google.ca/maps/@42.2292761,-83.7391719,3a,75y,54.44h,74.65t/data=!3m7!1e1!3m5!1srAaeg2wFciLs9vpdpmQH9A!2e0!6s%2F%2Fgeo2.ggpht.com%2Fcbk%3Fpanoid%3DrAaeg2wFciLs9vpdpmQH9A%26output%3Dthumbnail%26cb_client%3Dmaps_sv.tactile.gps%26thumb%3D2%26w%3D203%26h%3D100%26yaw%3D250.00459%26pitch%3D0!7i13312!8i6656

For a small roundabout like that, car across the circle is a potential conflict if making 270 turn. 

Here's a similar-sized (maybe even smaller?) roundabout in Kitchener, ON (Lancaster St and Bridge St) that has trees and bushes in the middle. Maybe this policy differs by region?

[jakeroot]

Here's a similar-sized (maybe even smaller?) roundabout in Waterloo, ON (Lancaster St and Bridge St) that has trees and bushes in the middle. Maybe this policy differs by region?

http://i.imgur.com/Dwtc7dI.png

I'm not sure if it's official policy to decorate the center of the circle, but, it's highly recommended. Several guides that I've seen in the past recommend center decorations to prevent drivers from seeing more than just what's to the left of them.

[kalvado]

For a small roundabout like that, car across the circle is a potential conflict if making 270 turn. 

Here's asimilar-sized (maybe even smaller?) roundabout in Waterloo, ON (Lancaster St and Bridge St)[/url] that has trees and bushes in the middle. Maybe this policy differs by region?

Which tells nothing about safety of the design, any statistical data on this one?
With 30 meters diameter a car across has about 50 meters to travel before conflict. That is about 5 seconds at recommended speed, and 3 seconds since seeing it with decoration on the center. Is that enough to accelerate and clear the path? Most drivers don't really utilize full 0-60 acceleration rating  at each start... 

[tradephoric]

I'll bring this up here since it was mentioned in another thread.  Another complex crash prone roundabout is being simplified.  The Steptoe St. & Columbia Park roundabout in Tri-Cities, Washington is having a circulating lane removed inside the roundabout.  Yet another example of combating a high crash rate roundabout by removing a circulating lane:

http://goo.gl/kWNe0i

[kalvado]

I'll bring this up here since it was mentioned in another thread.  Another complex crash prone roundabout is being simplified.  The Steptoe St. & Columbia Park roundabout in Tri-Cities, Washington is having a circulating lane removed inside the roundabout.  Yet another example of combating a high crash rate roundabout by removing a circulating lane:

http://goo.gl/kWNe0i

Next step should be removal of central island and adding color coded signal light...

[7/8]

For a small roundabout like that, car across the circle is a potential conflict if making 270 turn. 

Here's a similar-sized (maybe even smaller?) roundabout in Kitchener, ON (Lancaster St and Bridge St) that has trees and bushes in the middle. Maybe this policy differs by region?

Which tells nothing about safety of the design, any statistical data on this one?
With 30 meters diameter a car across has about 50 meters to travel before conflict. That is about 5 seconds at recommended speed, and 3 seconds since seeing it with decoration on the center. Is that enough to accelerate and clear the path? Most drivers don't really utilize full 0-60 acceleration rating  at each start...

It's not in the top 100 most crash-prone intersections in the Region of Waterloo.
http://www.regionofwaterloo.ca/en/gettingAround/resources/2014-Collision-Ranking---By-Rank.pdf

I can't find any statistics for this roundabout though

[tradephoric]

It's not in the top 100 most crash-prone intersections in the Region of Waterloo.
http://www.regionofwaterloo.ca/en/gettingAround/resources/2014-Collision-Ranking---By-Rank.pdf

I can't find any statistics for this roundabout though

I don't see the Homer Watson Blvd and Block Road roundabout on that list.  Did they just leave out roundabouts in their analysis or is that intersection not part of the Waterloo region?

It's somewhat surprising they are going ahead with building a massive 3x2 roundabout at Ottawa Street & Homer Watson Blvd after all the issues with the Homer Watson Blvd & Block Road roundabout.  Based on your link, Ottawa Street & Homer Watson Blvd is the most crash prone intersection in the entire region.  Time will tell if the crash problem gets better or worse after the roundabout is completed. 


http://www.waterlooregionconnected.com/printthread.php?tid=339&page=1

[7/8]

It's not in the top 100 most crash-prone intersections in the Region of Waterloo.
http://www.regionofwaterloo.ca/en/gettingAround/resources/2014-Collision-Ranking---By-Rank.pdf

I can't find any statistics for this roundabout though

I don't see the Homer Watson Blvd and Block Road roundabout on that list.  Did they just leave out roundabouts in their analysis or is that intersection not part of the Waterloo region?

I decided to read the full report this time, and I found this on page 35



So, there are 13 roundabouts in the "Top 100" chart I posted earlier, but the other 7 roundabouts (as of 2014) were there for less than five years, so they were included in the table above. Unfortunately, there isn't much data there. I also still don't see the Bridgeport and Lancaster roundabout I mentioned earlier, maybe it's newer than I thought.

It's somewhat surprising they are going ahead with building a massive 3x2 roundabout at Ottawa Street & Homer Watson Blvd after all the issues with the Homer Watson Blvd & Block Road roundabout.  Based on your link, Ottawa Street & Homer Watson Blvd is the most crash prone intersection in the entire region.  Time will tell if the crash problem gets better or worse after the roundabout is completed. 


http://www.waterlooregionconnected.com/printthread.php?tid=339&page=1

I know what you mean, and I'm sure this will be controversial. But the transportation departments in the Region and the Cities here seem to like roundabouts . They're also putting in two roundabouts on Erb St west of the current roundabout with Ira Needles/Erbsville; they're building a new Costco and they believe it'll help with traffic flow.

[kalvado]

I know what you mean, and I'm sure this will be controversial. But the transportation departments in the Region and the Cities here seem to like roundabouts . They're also putting in two roundabouts on Erb St west of the current roundabout with Ira Needles/Erbsville; they're building a new Costco and they believe it'll help with traffic flow.

Lets call it religious attitude. After all, it is for everyone's safety! And don't let stupid numbers to obscure the dogma 

[lordsutch]

Lets call it religious attitude. After all, it is for everyone's safety! And don't let stupid numbers to obscure the dogma 

Funny, that's what a lot of us would say about tradephoric's repeated goalpost-shifting in this thread.

[kalvado]

Lets call it religious attitude. After all, it is for everyone's safety! And don't let stupid numbers to obscure the dogma 

Funny, that's what a lot of us would say about tradephoric's repeated goalpost-shifting in this thread.

You see, I tried real hard - but looks like there is no solid justification for roundabouts published in English. There is some anecdotal evidence that those sometimes cause problems, and equally anecdotal evidence that is may sometimes work. But I didn't find any understanding of WHY and WHEN it goes one way or the other. But everyone is sure it is best thing since sliced bread.
There are some quite interesting works on lane capacity, for example. I didn't see anything of remotely similar quality about roundabouts... 

[cjw2001]

Living in an area (near Carmel IN) with well over 100 roundabouts nearby, I can definitely state that the roundabouts provide a huge advantage for traffic flow.  It is so much better now for traveling the area without having to stop and wait at so many lights and stop signs.  Traffic used to back up significantly before the roundabouts were added.   Is it perfect?  No.  But it is far and away better than it was before the roundabouts were added.  Both my travel times and my gas mileage have noticeably improved.

I have no safety concerns driving the many roundabouts in the area.  A little bit of defensive driving (anticipate that there are other drivers on the road that don't know what they are doing) goes a long way toward avoiding any issues. 

The roundabouts are very popular with the local residents, and another 32 are going to be added within the next several years.  Can't wait for the further improvements in local traffic.

[tradephoric]

I have no safety concerns driving the many roundabouts in the area.  A little bit of defensive driving (anticipate that there are other drivers on the road that don't know what they are doing) goes a long way toward avoiding any issues. 

Here is the citywide crash data for Carmel dating back to 2010.


There has been a 39% increase in total crashes (2442 crashes in 2015 / 1754 crashes in 2010) and a 56% increase in injury crashes (241 injury crashes in 2015 / 154 injury crashes in 2010).  The population of Carmel has increased 12% in the same time-frame (88,713 in 2015 / 79,191 in 2010).  Population data: http://www.carmel.in.gov/index.aspx?page=45

Ironically, both of Carmel's fatal crashes in 2014 occurred at roundabouts.  Carmel still has its fair share of "dangerous"  traffic signals; yet the "safe"  roundabouts were the site of the fatal crashes.  We've all been told by the "experts"  that roundabouts reduce fatal crashes by 90%.... perhaps they fabricated that safety statistic. 

http://fox59.com/2014/08/06/one-dead-in-carmel-accident-passenger-taken-to-hospital/
http://fox59.com/2014/09/26/person-dead-after-motorcycle-accident-in-carmel/