Crash prone 'modern roundabouts'

[kalvado]

The Sprinkle Road & I-94 roundabout is crash prone.  In 2019 it was the 5th most crash prone roundabout in Michigan with 71 crashes and 2 injuries and the most crash prone roundabout outside of SE Michigan.  In 2020 total crashes went down to 33 but injuries tripled to 6 (and was still the most crash prone roundabout in the state outside of SE Michigan).

Wasn't the roundabout completed in 2016?  A fuller set of data would tell a better story, especially if crashes have continued to drop off.

Then again, how many times have we called you out on cherry picking data and you've just ignored us?

Is there a way to access full crash data without lengthy work and FOILs, or that is something too specialized? You probably know how that works in NY..

[Rothman]

The Sprinkle Road & I-94 roundabout is crash prone.  In 2019 it was the 5th most crash prone roundabout in Michigan with 71 crashes and 2 injuries and the most crash prone roundabout outside of SE Michigan.  In 2020 total crashes went down to 33 but injuries tripled to 6 (and was still the most crash prone roundabout in the state outside of SE Michigan).

Wasn't the roundabout completed in 2016?  A fuller set of data would tell a better story, especially if crashes have continued to drop off.

Then again, how many times have we called you out on cherry picking data and you've just ignored us?

Is there a way to access full crash data without lengthy work and FOILs, or that is something too specialized? You probably know how that works in NY..

Tradephoric already got the two years' worth.  That means he can get access to the rest.

[jakeroot]

Wasn't the roundabout completed in 2016?  A fuller set of data would tell a better story, especially if crashes have continued to drop off.

Then again, how many times have we called you out on cherry picking data and you've just ignored us?

Every year the 18 1/2 Mile & Van Dyke roundabout in Sterling Heights shows up on the list of intersections with the most crashes in Michigan.  In 2021 there were 168 total crashes and 17 injuries.  That roundabout was completed in 2005.  After nearly two decades, drivers are still having trouble with that roundabout.  You can blame it on driver error, but as the complexity of these roundabouts increase it can become increasingly difficult for drivers to navigate safely.

It seems quite clear to me that crash rates at multi-lane roundabouts do fluctuate, but are likely higher than whatever multi-lane signalized intersection would have otherwise been built. The Sterling Heights roundabout is a classic example of this: it has not once been modified since opening, and drivers still continue to crash at it, at a remarkably high rate.

Driver error.  Again.

Of course, driver error is the cause of almost all crashes. But why do drivers seem to make more errors at multi-lane roundabouts? And why are the crash rates either barely improving, not improving, or even getting worse?

[Rothman]

Wasn't the roundabout completed in 2016?  A fuller set of data would tell a better story, especially if crashes have continued to drop off.

Then again, how many times have we called you out on cherry picking data and you've just ignored us?

Every year the 18 1/2 Mile & Van Dyke roundabout in Sterling Heights shows up on the list of intersections with the most crashes in Michigan.  In 2021 there were 168 total crashes and 17 injuries.  That roundabout was completed in 2005.  After nearly two decades, drivers are still having trouble with that roundabout.  You can blame it on driver error, but as the complexity of these roundabouts increase it can become increasingly difficult for drivers to navigate safely.

It seems quite clear to me that crash rates at multi-lane roundabouts do fluctuate, but are likely higher than whatever multi-lane signalized intersection would have otherwise been built. The Sterling Heights roundabout is a classic example of this: it has not once been modified since opening, and drivers still continue to crash at it, at a remarkably high rate.

Driver error.  Again.

Of course, driver error is the cause of almost all crashes. But why do drivers seem to make more errors at multi-lane roundabouts? And why are the crash rates either barely improving, not improving, or even getting worse?

I'm not so sure this is generally true for two-lane roundabouts.  I can point to various examples in NY where there were low-speed crashes at the beginning of the installation, but those rates did drop off.

But sure, locations where crashes remain frequent should be re-evaluated.

But also, getting back to the one roundabout that caused this discussion, despite tradephoric's opinion on the data he presented, crashes dropped by half.  That's why a broader look time-wise is warranted.

[kalvado]

Wasn't the roundabout completed in 2016?  A fuller set of data would tell a better story, especially if crashes have continued to drop off.

Then again, how many times have we called you out on cherry picking data and you've just ignored us?

Every year the 18 1/2 Mile & Van Dyke roundabout in Sterling Heights shows up on the list of intersections with the most crashes in Michigan.  In 2021 there were 168 total crashes and 17 injuries.  That roundabout was completed in 2005.  After nearly two decades, drivers are still having trouble with that roundabout.  You can blame it on driver error, but as the complexity of these roundabouts increase it can become increasingly difficult for drivers to navigate safely.

It seems quite clear to me that crash rates at multi-lane roundabouts do fluctuate, but are likely higher than whatever multi-lane signalized intersection would have otherwise been built. The Sterling Heights roundabout is a classic example of this: it has not once been modified since opening, and drivers still continue to crash at it, at a remarkably high rate.

Driver error.  Again.

Of course, driver error is the cause of almost all crashes. But why do drivers seem to make more errors at multi-lane roundabouts? And why are the crash rates either barely improving, not improving, or even getting worse?

I'm not so sure this is generally true for two-lane roundabouts.  I can point to various examples in NY where there were low-speed crashes at the beginning of the installation, but those rates did drop off.

But sure, locations where crashes remain frequent should be re-evaluated.

But also, getting back to the one roundabout that caused this discussion, despite tradephoric's opinion on the data he presented, crashes dropped by half.  That's why a broader look time-wise is warranted.

What is the target crash rate for an intersection?  NY average for urban intersections is 0.2-0.5 per million,  looks like 0.2 is a good target number if design safety is a goal.
I am too lazy to look up traffic counts on that thing, but 50k/daily should be on the high side. That is, 1 crash every 100 days, or 3-4 crashes a year is a good target value for intersection evaluation.
We're talking about the drop from 71 to 33 crashes (that drop is in 2020, a covid year), which for me is a drop from "crucify designer in the center of the roundabout" to "firing squad is humane enough"

[tradephoric]

But also, getting back to the one roundabout that caused this discussion, despite tradephoric's opinion on the data he presented, crashes dropped by half.  That's why a broader look time-wise is warranted.

Thirty-three crashes and 6 injury crashes is significant and would be a target in any safety audits performed in that jurisdiction.  Yes, there was a drop in crashes in 2020 which is the same year everyone was being forced to stay in their homes (in Michigan resident's couldn't even travel freely between their vacation homes under Governor Whitmer's stay-at-home order).

[kphoger]

Yes, there was a drop in crashes in 2020 which is the same year everyone was being forced to stay in their homes

Non-fatal crash rates dropped in 2020, but fatal crash rates increased.  The 2020 fatality rate was the highest since 2007, despite a steady decrease over the previous three years.

[tradephoric]

The University of Maryland National Transportation Center published a report in 2016 looking at the safety of roundabouts in Arizona.  This report was done before all the multi-lane roundabouts were constructed along Arizona 260 and 89A that HighwayStar referenced a few posts back.  They found that multi-lane roundabouts increase total crashes in Arizona which also mirrors the findings of other states.  While roundabouts can reduce the severity and number of injuries, there were examples in the report of roundabout locations that saw an increase in injury crashes.  It's not a guarantee a complex multi-lane roundabout will be safer than the signalized intersection it replaces.

EFFECT OF TRAFFIC ROUNDABOUTS ON SAFETY IN ARIZONA
https://mti.umd.edu/sites/mti.umd.edu/files/documents/NTC2015-SU-R-3%20Michael%20Mamlouk.pdf

EXECUTIVE SUMMARY
Seventeen roundabouts in 5 cities in Arizona were used in the study, out of which 11 single-lane and 16 double-lane. Most of the locations of single-lane roundabouts were controlled by 2-way stop signs before the roundabout installation, while most of the locations of double-lane roundabouts were controlled by signals. Accidents data were collected and broken down into 3 levels: damage, injury, and fatality. Equal number of years were used before and after the roundabout conversion at each location. The most recent AADT value at each location was used to backcalculate the AADT value at the time of roundabout conversion, which is the midpoint of the analysis period. The average rates of accidents, damages, and injuries per year and per year per million vehicles were evaluated. It was found that single-lane roundabouts reduced the accident rate of intersections. However, double-lane roundabouts increased the accident rate of intersections. A decision needs to be made as to either remove double-lane roundabouts or find solutions on how to make these roundabouts safe, such as making geometric improvements or educating the public on how to use them. The results also showed that both single- and double-lane roundabout conversions reduced the severity levels of accidents. Considering both accident rate and severity level, warrants needs to be developed for roundabout conversion for both single- and multi-lane roundabout conversion.

[tradephoric]

Non-fatal crash rates dropped in 2020, but fatal crash rates increased.  The 2020 fatality rate was the highest since 2007, despite a steady decrease over the previous three years.

It sounds like we are in agreement then.  The Sprinkle & I-94 roundabout saw a big drop in total crashes in 2020 which follows the national trend of non-fatal crash rates dropping during the pandemic.

[kphoger]

Non-fatal crash rates dropped in 2020, but fatal crash rates increased.  The 2020 fatality rate was the highest since 2007, despite a steady decrease over the previous three years.

It sounds like we are in agreement then.  The Sprinkle & I-94 roundabout saw a big drop in total crashes in 2020 which follows the national trend of non-fatal crash rates dropping during the pandemic.

Are there data points for 2021 at that intersection?  National crash rates rebounded in 2021.

[Rothman]

Non-fatal crash rates dropped in 2020, but fatal crash rates increased.  The 2020 fatality rate was the highest since 2007, despite a steady decrease over the previous three years.

It sounds like we are in agreement then.  The Sprinkle & I-94 roundabout saw a big drop in total crashes in 2020 which follows the national trend of non-fatal crash rates dropping during the pandemic.

Are there data points for 2021 at that intersection?  National crash rates rebounded in 2021.

But also, getting back to the one roundabout that caused this discussion, despite tradephoric's opinion on the data he presented, crashes dropped by half.  That's why a broader look time-wise is warranted.

Thirty-three crashes and 6 injury crashes is significant and would be a target in any safety audits performed in that jurisdiction.  Yes, there was a drop in crashes in 2020 which is the same year everyone was being forced to stay in their homes (in Michigan resident's couldn't even travel freely between their vacation homes under Governor Whitmer's stay-at-home order).

Well, good.  We'll see what the jurisdiction does after conducting their safety audit, then.

[tradephoric]

Are there data points for 2021 at that intersection?  National crash rates rebounded in 2021.

I honestly don't know but the roundabout wasn't in the top 20 list of most dangerous Michigan intersections in 2021, so presumably it had fewer than 71 crashes.  Of the 20 most crash prone intersections in the state, 5 of them were roundabouts. 

What Were Michigan's Most Dangerous Intersections in 2021?
1. 11 MILE RD/I 696 @ VAN DYKE AVE, Warren, 173 Total Crashes, 38 Injuries
2. 10 MILE RD @ I 94, St. Clair Shores, 168 Total Crashes, 42 Injuries
3. 18 1/2 MILE RD @ VAN DYKE AVE, Sterling Heights, 168 Total Crashes, 17 Injuries (Roundabout)
4. MARTIN PKWY @ N PONTIAC TRL, Commerce Township, 140 Total Crashes, 18 Injuries (Roundabout)
5. US 131 @ WEALTHY ST, Grand Rapids, 134 Total Crashes, 26 Injuries
6. ORCHARD LAKE RD @ 14 MILE RD, Farmington Hills, 122 Total Crashes, 15 Injuries (Roundabout)
7. 12 MILE RD @ I 94, St. Clair Shores/Roseville, 121 Total Crashes, 47 Injuries
8. SCHOOLCRAFT RD @ TELEGRAPH RD, Redford Township, 104 Total Crashes, 21 Injuries
9. TELEGRAPH RD @ 12 MILE RD, Southfield, 95 Total Crashes, 15 Injuries
10. 11 MILE RD @ GRATIOT AVE E, Roseville, 92 Total Crashes, 9 Injuries
11. SOUTHFIELD RD @ W 11 MILE RD, Lathrup Village, 91 Total Crashes, 28 Injuries
12. HALL RD @ VAN DYKE AVE, Utica, 86 Total Crashes, 15 Injuries
13. SAGINAW ST @ N HOMER ST, Lansing Township, 85 Total Crashes, 26 Injuries
14. DIX AVE @ M 39, Lincoln Park, 83 Total Crashes, 17 Injuries
15. STATE ST @ ELLSWORTH RD, Ann Arbor, 79 Total Crashes, 1 Injury (Roundabout)
16. MIDDLE BELT RD @ SCHOOLCRAFT RD, Livonia, 78 Total Crashes, 16 Injuries
17. FORD RD @ LILLEY RD, Canton Township, 75 Total Crashes, 32 Injuries
18. FRANKLIN ST SW @ US 131, Grand Rapids, 75 Total Crashes, 10 Injuries
19. 6 MILE RD @ I 96/I 275, Livonia, 73 Total Crashes, 19 Injuries
20. LEE RD @ WHITMORE LAKE RD, Green Oak Township, 71 Total Crashes, 5 Injuries (West Roundabout)

https://www.michiganautolaw.com/legal-resources/most-dangerous-intersections/

[kphoger]

3. 18 1/2 MILE RD @ VAN DYKE AVE, Sterling Heights, 168 Total Crashes, 17 Injuries (Roundabout)
4. MARTIN PKWY @ N PONTIAC TRL, Commerce Township, 140 Total Crashes, 18 Injuries (Roundabout)
6. ORCHARD LAKE RD @ 14 MILE RD, Farmington Hills, 122 Total Crashes, 15 Injuries (Roundabout)
15. STATE ST @ ELLSWORTH RD, Ann Arbor, 79 Total Crashes, 1 Injury (Roundabout)
20. LEE RD @ WHITMORE LAKE RD, Green Oak Township, 71 Total Crashes, 5 Injuries (West Roundabout)

Looking at Google Maps...

3.  Multi-lane roundabout.  2 lanes circulating.
4.  Multi-lane roundabout.  2 lanes circulating.  Median fencing on two approaches.
6.  Multi-lane roundabout.  2 lanes circulating.
15.  Multi-lane roundabout.  2 lanes circulating.
20.  Not quite 2 lanes circulating, SB exit is 1 lane.  Part of a double roundabout (with SB US-23 ramps).

[tradephoric]

^There are 3 roundabouts with triple circulating lanes (at least partial) in the entire state of Michigan. They also happen to be 3 of the top 6 most crash prone intersections in the state.  While these roundabouts do carry a lot of traffic, they don't see nearly the same traffic volumes as a major signalized intersection along the state's many 8-lane boulevards (ie. Telegraph, Woodward, Grand River, Northwester, Hall Road, etc).

3. 18 1/2 MILE RD @ VAN DYKE AVE, Sterling Heights, 168 Total Crashes, 17 Injuries (Roundabout)
4. MARTIN PKWY @ N PONTIAC TRL, Commerce Township, 140 Total Crashes, 18 Injuries (Roundabout)
6. ORCHARD LAKE RD @ 14 MILE RD, Farmington Hills, 122 Total Crashes, 15 Injuries (Roundabout)

[skluth]

I was really excited to see the incorporation of roundabouts into the base build of Cities Skylines 2. It seems they don't share the opinions of some here who don't like roundabouts, especially multi-lane roundabouts. I'm sure I'll be using them all over my future cities.

[kalvado]

^There are 3 roundabouts with triple circulating lanes (at least partial) in the entire state of Michigan. They also happen to be 3 of the top 6 most crash prone intersections in the state.  While these roundabouts do carry a lot of traffic, they don't see nearly the same traffic volumes as a major signalized intersection along the state's many 8-lane boulevards (ie. Telegraph, Woodward, Grand River, Northwester, Hall Road, etc).

3. 18 1/2 MILE RD @ VAN DYKE AVE, Sterling Heights, 168 Total Crashes, 17 Injuries (Roundabout)
4. MARTIN PKWY @ N PONTIAC TRL, Commerce Township, 140 Total Crashes, 18 Injuries (Roundabout)
6. ORCHARD LAKE RD @ 14 MILE RD, Farmington Hills, 122 Total Crashes, 15 Injuries (Roundabout)

It may be too much to ask, but adding traffic counts to the list may really help.
I did commute through a 50k daily intersection. It was tough, but generally not very bad. Roundabouts top at 25-30k daily. FHWA says 40k but I am not sure what is that based upon other than wishful thinking.  I suspect those bad roundabouts handle try to process more than that.

[HighwayStar]

The University of Maryland National Transportation Center published a report in 2016 looking at the safety of roundabouts in Arizona.  This report was done before all the multi-lane roundabouts were constructed along Arizona 260 and 89A that HighwayStar referenced a few posts back.  They found that multi-lane roundabouts increase total crashes in Arizona which also mirrors the findings of other states.  While roundabouts can reduce the severity and number of injuries, there were examples in the report of roundabout locations that saw an increase in injury crashes.  It's not a guarantee a complex multi-lane roundabout will be safer than the signalized intersection it replaces.

EFFECT OF TRAFFIC ROUNDABOUTS ON SAFETY IN ARIZONA
https://mti.umd.edu/sites/mti.umd.edu/files/documents/NTC2015-SU-R-3%20Michael%20Mamlouk.pdf

EXECUTIVE SUMMARY
Seventeen roundabouts in 5 cities in Arizona were used in the study, out of which 11 single-lane and 16 double-lane. Most of the locations of single-lane roundabouts were controlled by 2-way stop signs before the roundabout installation, while most of the locations of double-lane roundabouts were controlled by signals. Accidents data were collected and broken down into 3 levels: damage, injury, and fatality. Equal number of years were used before and after the roundabout conversion at each location. The most recent AADT value at each location was used to backcalculate the AADT value at the time of roundabout conversion, which is the midpoint of the analysis period. The average rates of accidents, damages, and injuries per year and per year per million vehicles were evaluated. It was found that single-lane roundabouts reduced the accident rate of intersections. However, double-lane roundabouts increased the accident rate of intersections. A decision needs to be made as to either remove double-lane roundabouts or find solutions on how to make these roundabouts safe, such as making geometric improvements or educating the public on how to use them. The results also showed that both single- and double-lane roundabout conversions reduced the severity levels of accidents. Considering both accident rate and severity level, warrants needs to be developed for roundabout conversion for both single- and multi-lane roundabout conversion.

Not surprised in the least with those. Although most roundabouts at least seem to replace a regular intersection those were basically barriers on a highway at a junction with some minor road, easy to see how they would be inducing accidents.

[tradephoric]

New Volusia County roundabout has seen over 70 crashes in a year
https://www.clickorlando.com/news/investigators/2023/06/29/newly-constructed-roundabout-in-volusia-county-has-seen-over-70-crashes-in-a-year/

The article referenced the roundabout at West Haven Road & Woodland Blvd constructed by FDOT about 5 miles away from the problem roundabout.  That roundabout only had 17 crashes in a 2 year time period.  The biggest difference is that is a 2X1 roundabout (much simpler to navigate) while the other roundabout is a 2X2 roundabout. 

[jakeroot]

I love that in the story's video, you can see the news vehicle illegally continuing around the roundabout from the outside lane. At least twice (see 3:15). Perhaps an on-screen graphic where the image briefly goes black on white, with a big "NO" text?

It is interesting to me how much confusion can arise from markings with roundabouts (like in the story, dashed lines being confused for "I can change lanes" rather than just guidance markings). Obviously they're designed to reduce confusion, and mostly I'd guess they do. But some drivers seem to totally misinterpret them. Perhaps we should try and make roundabouts more like signalized intersections, where markings are only used for complex situations, like double turns.

tradephoric, have you looked into the Los Alamitos Roundabout in Long Beach, CA? It recently had lane markings added, and entry widths' reduced. Curious to see how the accident rate has changed. Before 2020, it was easily the largest roundabout in the US, or at least the roundabout with the most entry lanes. And curiously, definitely the largest non-rotary (New England-style) circular intersection without any markings.

[tradephoric]

tradephoric, have you looked into the Los Alamitos Roundabout in Long Beach, CA? It recently had lane markings added, and entry widths' reduced. Curious to see how the accident rate has changed. Before 2020, it was easily the largest roundabout in the US, or at least the roundabout with the most entry lanes. And curiously, definitely the largest non-rotary (New England-style) circular intersection without any markings.

There seems to be way too many arrows within the circle and the arrows that do exist are confusing. I believe roundabout specific arrows should only be used as drivers are approaching the roundabout (if used at all) but shouldn't be used inside the circulating lanes.  Also thru arrows in a modern roundabout should direct drivers to exit the roundabout not to continue around it.  I also don't see the need to have lane arrows right after drivers are exiting the roundabout.  Here is my proposed marking/arrow design compared to what is currently out there.  Not saying my proposed design would be any better but it does follow more modern-roundabout design guidelines (which may or may not be a good thing to adhere to at this particularly large circle). 

CURRENT ROUNDABOUT MARKINGS


PROPOSED ROUNDABOUT MARKINGS

[tradephoric]

In the left circle it looks like both lanes are allowed to exit the roundabout at the same time but then in the right circle it looks like you are no longer allowed to from the left most circulating lane.  Can someone explain what is going on here?

[jakeroot]

There seems to be way too many arrows within the circle and the arrows that do exist are confusing. I believe roundabout specific arrows should only be used as drivers are approaching the roundabout (if used at all) but shouldn't be used inside the circulating lanes.  Also thru arrows in a modern roundabout should direct drivers to exit the roundabout not to continue around it.  I also don't see the need to have lane arrows right after drivers are exiting the roundabout.  Here is my proposed marking/arrow design compared to what is currently out there.  Not saying my proposed design would be any better but it does follow more modern-roundabout design guidelines (which may or may not be a good thing to adhere to at this particularly large circle). 

....

Really appreciate the analysis, tradephoric. I also like your redesign.

In the left circle it looks like both lanes are allowed to exit the roundabout at the same time but then in the right circle it looks like you are no longer allowed to from the left most circulating lane.  Can someone explain what is going on here?
https://i.imgur.com/qwrJqOx.png

Nice spot, that is confusing!

This is a great example of how markings, if not done perfectly (and we have not yet found perfect markings, IMO), seem to create a more dangerous situation as drivers can fail to correctly interpret the markings. Eg, "I can change lanes here because the markings are dashed" even though they are dashed to indicate to entering drivers that they can cross over that line, not to allow drivers to stay left across it (in fairness, just writing that makes it seem confusing). The other is the type of driver who sees the roundabout as a huge one-way road, thinking you can only exit from the outside lane...so they just use the outside lane the whole time, even if "turning left".

The old circle, while perhaps not exactly 2009-MUTCD in its design (no surprise having been designed in part by British consultants in the 1990s), did not really hold your hand the way modern roundabouts do. I really do think marking-less roundabouts are something that needs to be tried again. I am growing tired of drivers who are misinterpreting lane lines and creating dangerous situations as a result, and then blaming (rightly or wrongly) the municipality.

[Occidental Tourist]

Oh, it's a shitshow, all right.  The markings for PCH north coming into the circle are also confusing.  I've used the inside lane to enter the circle only to have someone from the outside lane cut over in front of me to try to get into the circle as well.

[jamess]

New Volusia County roundabout has seen over 70 crashes in a year
https://www.clickorlando.com/news/investigators/2023/06/29/newly-constructed-roundabout-in-volusia-county-has-seen-over-70-crashes-in-a-year/

The article referenced the roundabout at West Haven Road & Woodland Blvd constructed by FDOT about 5 miles away from the problem roundabout.  That roundabout only had 17 crashes in a 2 year time period.  The biggest difference is that is a 2X1 roundabout (much simpler to navigate) while the other roundabout is a 2X2 roundabout.

Horrendous bike lane design. They direct people onto the sidewalk path but didnt build ramps

On
https://goo.gl/maps/HvqxerJPBRYYhmBe8

Off
https://goo.gl/maps/BMbASEeeTsKSLsAC8


In the on link, you can see bikes are forced to make a right turn


They got one corner right.

https://goo.gl/maps/WfX6thrF1oKF2v6b6

https://goo.gl/maps/a7cuMcbCY6eU7UL87

A licensed professional stamped this?

[JoePCool14]

It really does boggle the mind sometimes how stuff makes it past QAQC. No one notices these things throughout the process? Or maybe they do notice them and can't be bothered to fix them. Lazy.