Crash prone 'modern roundabouts'

[jakeroot]

I have absolutely no citations to back this up, only word of mouth, but it was widely believed that when these roundabouts were designed, school bus traffic wasn't taken into consideration.

Are there a lot of school buses going through that area? Even more than usual should be outweighed by the sheer number of cars. Unless the school buses are stopping immediately after exiting the roundabout. In that case, roundabouts are pretty annoying, as stopped cars tend to lock up the roundabout (as drivers seem to be less attentive to box-blocking at roundabouts compared to signals).

[kalvado]

I have absolutely no citations to back this up, only word of mouth, but it was widely believed that when these roundabouts were designed, school bus traffic wasn't taken into consideration.

Are there a lot of school buses going through that area? Even more than usual should be outweighed by the sheer number of cars. Unless the school buses are stopping immediately after exiting the roundabout. In that case, roundabouts are pretty annoying, as stopped cars tend to lock up the roundabout (as drivers seem to be less attentive to box-blocking at roundabouts compared to signals).

Pretty often, drivers are not supposed to see what is going on on the other side - and that is done on purpose. It is fairly easy to find yourself in a situation when you don't see a problem before it is too late.
Another fine print issue is what "yield to traffic already in a circle" means. A  long and/or heavy vehicle which cannot accelerate like a car can easily end up blocking traffic even with ample space upstream. Not to mention that they also need to catch the smallest workable gap, meaning they are not too shy about blocking upstream traffic.

[tradephoric]

And the firm didn't even come up with a design that eliminates circulating lanes inside the roundabouts, which has historically been the only long term fix to reducing the crash rates at these problem roundabouts.

If that has indeed been established by your own research in this thread, does the firm even know that's the case?

Here are some of the changes planned at the Hilliard roundabouts that are mentioned in the article. 

-Reshaping the central traffic island by installing temporary curbs.

-The traffic "islands"  in the roundabout will be modified.

-More changes are planned at Main and Cemetery, including resurfacing, revised markings, overhead lane-control signs and "Yield to both lanes"  signs.

-The flashing LED-bordered yield signs will be placed in the center island

Even if these changes help reduce the crash rate, it's hard to pinpoint what change actually made the difference.  Take the 14th Street and Superior roundabout in Lincoln as an example.  The 3x2 roundabout had a very high crash rate so the city downsized the roundabout to a 2x1 roundabout and the site distance was limited by constructing chain link fence running along the splitter islands with slats in the medians.  The result was a dramatic drop in crashes with the crash rate dropping below pre-roundabout levels.  Now a lot of the success was attributed to the fencing installed along the splitter islands to limit drivers views. 

IMO, the fact that the Lincoln roundabout was dramatically downsized (from a 3x2 to a 2x1 roundabout) was the reason there was such a dramatic reduction in crashes and the fencing had little to no effect.  The M-5 and Pontiac Trail roundabout is a 3x2 roundabout in Commerce Michigan and it too had fencing installed along the splitter islands in an attempt to reduce the high number of crashes occurring, but unlike the Lincoln roundabout the configuration of the roundabout wasn't touched (remained a 3x2).  When comparing the crashes from the 6 months before the fencing to the 6 months after, the crash rate was identical.

Whenever i've seen a dramatic drop in crashes at these problem roundabouts, it seems to always include a downsizing of the roundabout.  However, when the roundabout isn't downsized and other "safety features" are tried, it's often ineffective at reducing the crash rate.

[kphoger]

That still doesn't address the actual question part of my question:  "does the firm even know that's the case?"  Is anyone outside this forum aware of the statistical correlations you've been drawing?

[MNHighwayMan]

That still doesn't address the actual question part of my question:  "does the firm even know that's the case?"  Is anyone outside this forum aware of the statistical correlations you've been drawing?

I would love to see a peer review of his data done by actual traffic engineers.

[tradephoric]

I would love to see a peer review of his data done by actual traffic engineers.

Derek Leuer and John Hourdos out of Minnesota have done a lot of research specific to the safety of multi-lane roundabouts.  Based on their research there has been a 6% increase in total injury crashes at the full multi-lane roundabouts analyzed in Minnesota.  Here is a link to their study.  This study is unique in that it analyzes each type of roundabout separately and splits it up by single-lane, unbalanced multi-lane, and full multi-lane. 

A Study of the Traffic Safety at Roundabouts in Minnesota
http://www.dot.state.mn.us/trafficeng/safety/docs/roundaboutstudy.pdf

The good news is that Type-A crashes dropped from 3 to 0, a 100% reduction in serious injuries.  But if you plug in ALL the numbers into FHWA's comprehensive crash costs by injury severity level it's clear that the dual-lane roundabouts had a higher social cost than the intersections they replaced.


http://www.dot.state.mn.us/trafficeng/safety/docs/roundaboutstudy.pdf

BEFORE ROUNDABOUT CRASH COSTS = $4,882,600
Fatality (K)           0 X $4,008,900 = $0
Disabling Injury (A)   3 X  $216,000 = $648,000
Evident Injury (B)   15 X $79,000 = $1,185,000
Possible Injury (C)   46 X $44,900 = $2,065,400
PDO (O)          133 X $7,400 =  $984,200

AFTER ROUNDABOUT CRASH COSTS = $6,546,900
Fatality (K)           0 X $4,008,900 = $0
Disabling Injury (A)   0 X  $216,000 = $0
Evident Injury (B)   15 X $79,000 = $1,185,000
Possible Injury (C)   53 X $44,900 = $2,379,700
PDO (O)          403 X $7,400 =  $2,982,200

[MNHighwayMan]

So a reduction in actual disabling injuries, and an increase in what might be injuries. That's somehow worse than before?

[TheHighwayMan3561]

So a reduction in actual disabling injuries, and an increase in what might be injuries. That's somehow worse than before?

It's like we're trying to find the magic number of how much money a life is worth.

[kalvado]

So a reduction in actual disabling injuries, and an increase in what might be injuries. That's somehow worse than before?

It's like we're trying to find the magic number of how much money a life is worth.

Low values are $2M, high are $10M. Nothing magic, pretty well-known stuff.

[tradephoric]

So a reduction in actual disabling injuries, and an increase in what might be injuries. That's somehow worse than before?

That's correct according to the FHWA.  If you really want to dive deep into how they come up with the numbers you should read the FHWA publication "Crash Costs for Highway Safety Analysis". 

https://safety.fhwa.dot.gov/hsip/docs/fhwasa17071.pdf

A few posts ago you are questioning the data i'm citing throughout this thread and wanted "real traffic engineers" to do a peer review of it.  Now i cite you a study looking at the safety of dual-lane roundabouts (which is a running theme throughout this thread) and now you seem to be questioning the FHWA's data too.  At this point i don't know what to tell you.

EDIT:  When i say "that is correct according to the FHWA" i'm speaking to this specific example of calculating the social crash costs of dual-lane roundabouts in Minnesota.  If you do the same analysis of single-lane roundabouts in Kansas, the roundabouts would probably fair quite well. 

[6a]

I have absolutely no citations to back this up, only word of mouth, but it was widely believed that when these roundabouts were designed, school bus traffic wasn't taken into consideration.

Are there a lot of school buses going through that area? Even more than usual should be outweighed by the sheer number of cars. Unless the school buses are stopping immediately after exiting the roundabout. In that case, roundabouts are pretty annoying, as stopped cars tend to lock up the roundabout (as drivers seem to be less attentive to box-blocking at roundabouts compared to signals).

It's just the sheer number of schools in that immediate area. Each with its corresponding bus traffic. I've circled schools in blue, the roundabouts in question in red.

[MNHighwayMan]

So a reduction in actual disabling injuries, and an increase in what might be injuries. That's somehow worse than before?

That's correct according to the FHWA.  If you really want to dive deep into how they come up with the numbers you should read the FHWA publication "Crash Costs for Highway Safety Analysis". 

https://safety.fhwa.dot.gov/hsip/docs/fhwasa17071.pdf

A few posts ago you are questioning the data i'm citing throughout this thread and wanted "real traffic engineers" to do a peer review of it.  Now i cite you a study looking at the safety of dual-lane roundabouts (which is a running theme throughout this thread) and now you seem to be questioning the FHWA's data too.  At this point i don't know what to tell you.

EDIT:  When i say "that is correct according to the FHWA" i'm speaking to this specific example of calculating the social crash costs of dual-lane roundabouts in Minnesota.  If you do the same analysis of single-lane roundabouts in Kansas, the roundabouts would probably fair quite well.

Well, I guess I'm just confused about how it is they're valuing these different injuries. I'll take a look at that report, thanks.

[tradephoric]

Here is a list of the 10 most dangerous roundabouts in Michigan for 2018.  Total crashes ranged from 49 crashes (Lee Road at Whitmore Lake Road roundabout) to 144 crashes (Orchard Lake Road at W 14 Mile Road roundabout). 

These are the top 10 most dangerous roundabouts in Michigan
https://www.wxyz.com/news/these-are-the-top-10-most-dangerous-roundabouts-in-michigan

[kalvado]

Here is a list of the 10 most dangerous roundabouts in Michigan for 2018.  Total crashes ranged from 49 crashes (Lee Road at Whitmore Lake Road roundabout) to 144 crashes (Orchard Lake Road at W 14 Mile Road roundabout). 

These are the top 10 most dangerous roundabouts in Michigan
https://www.wxyz.com/news/these-are-the-top-10-most-dangerous-roundabouts-in-michigan

Proper way to present these news would be comparing with 10 most dangerous conventional intersections, or trying to normalize that per traffic counts and acceptable numbers (5 per million, if I remember correctly?) - or anything else to put numbers in context...

[tradephoric]

The IIHS published a report looking at the long-term crash trends at single and double-lane roundabouts in Washington State.  They concluded that annual crash counts declined "significantly" by 8.8% at double-lane roundabouts and increased nonsignificantly over time at single-lane roundabouts.  Here is a link to the study:

Long-term crash trends at single-lane and double-lane roundabouts in Washington State
https://www.iihs.org/api/datastoredocument/bibliography/2180

[jakeroot]

The IIHS published a report looking at the long-term crash trends at single and double-lane roundabouts in Washington State.  They concluded that annual crash counts declined "significantly" by 8.8% at double-lane roundabouts and increased nonsignificantly over time at single-lane roundabouts.  Here is a link to the study:

Better link: https://www.iihs.org/news/detail/safety-at-two-lane-roundabouts-improves-over-time-new-study-shows

Are you going to request a copy? Would be nice to know if the crash rates, after so many years, are more or less than the intersection these roundabouts replaced.

[tradephoric]

Are you going to request a copy? Would be nice to know if the crash rates, after so many years, are more or less than the intersection these roundabouts replaced.

That is the real question.  If you look back at the State & Ellsworth roundabout, the very first roundabout discussed on this thread, it has seen similar crash reductions to that IIHS report.  But the roundabout still had 123 crashes last year which is 6x higher than the pre-roundabout condition.  Also, after the initial drop in crashes between year 1 and year 2, the total number of crashes has appeared to have plateaued.  To me the IIHS study is just trying to put lipstick on a pig by saying "hey look, double-lane roundabouts are seeing 9% annual crash reductions!".   

[kalvado]

The IIHS published a report looking at the long-term crash trends at single and double-lane roundabouts in Washington State.  They concluded that annual crash counts declined "significantly" by 8.8% at double-lane roundabouts and increased nonsignificantly over time at single-lane roundabouts.  Here is a link to the study:

Better link: https://www.iihs.org/news/detail/safety-at-two-lane-roundabouts-improves-over-time-new-study-shows

Are you going to request a copy? Would be nice to know if the crash rates, after so many years, are more or less than the intersection these roundabouts replaced.

Paper is not published yet; but looks like it will be an open access one once it goes through.
I do remember a paper from the same authors (on Boston lowering speed limit 30 -> 25 ); back then I was unimpressed with their data processing.

[kalvado]

And since my fishing expedition have started, here are some catches:
Master thesis on roundabouts in AZ: https://repository.asu.edu/attachments/170314/content/Souliman_asu_0010N_15713.pdf
Single lane ones are cool, double lanes significantly increase crash rates with some decrease in injuries.

https://www.tac-atc.ca/sites/default/files/conf_papers/weber_p_-_solutions_to_driver_errors_at_multi-lane_roundabouts.pdf - how to make roundabouts safer. I like their first suggestion:

A single-lane instead of a multi-lane roundabout

[jakeroot]

https://www.tac-atc.ca/sites/default/files/conf_papers/weber_p_-_solutions_to_driver_errors_at_multi-lane_roundabouts.pdf - how to make roundabouts safer. I like their first suggestion:

A single-lane instead of a multi-lane roundabout

Two suggestions I also like:

1) higher entry angles, so entry into roundabouts is more of a right turn. I can understand failing to yield when you can just drive straight across the roundabout, but if it becomes a sharp-right, and then a gradual left turn, there might be better safety numbers.

2) circulatory lines be removed or not installed. The paper suggests that the lines make roundabouts appear as having four (or more) individual intersections, and that removing the markings make the roundabouts appear as one large intersection, increasing driver hesitation. Makes total sense to me, even if it is counter-intuitive relative to typical North American marking practices.

[kalvado]

https://www.tac-atc.ca/sites/default/files/conf_papers/weber_p_-_solutions_to_driver_errors_at_multi-lane_roundabouts.pdf - how to make roundabouts safer. I like their first suggestion:

A single-lane instead of a multi-lane roundabout

Two suggestions I also like:

1) higher entry angles, so entry into roundabouts is more of a right turn. I can understand failing to yield when you can just drive straight across the roundabout, but if it becomes a sharp-right, and then a gradual left turn, there might be better safety numbers.

2) circulatory lines be removed or not installed. The paper suggests that the lines make roundabouts appear as having four (or more) individual intersections, and that removing the markings make the roundabouts appear as one large intersection, increasing driver hesitation. Makes total sense to me, even if it is counter-intuitive relative to typical North American marking practices.

Basically reduce throughput to improve safety. As traffic totals are not very elastic, for higher traffic cases this means deflecting traffic to other intersections (if possible); or creating backups.
So yes, looks like roundabouts are great in cases of low traffic, replacing 2-way stop. Over here in NY they are considered as high traffic solution for complex intersections. See why I don't like them? 

[Rothman]

Driver hesitation = driver confusion.  That doesn't sound like a good thing.

[kalvado]

Driver hesitation = driver confusion.  That doesn't sound like a good thing.

welcome to The Brave Roundabout World.
Drivers must hesitate before entering the circulation.
Drivers should not see what is on the other side.
Ignorance Is Strength.
War is Peace.

[jakeroot]

Driver hesitation = driver confusion.  That doesn't sound like a good thing.

welcome to The Brave Roundabout World.
Drivers must hesitate before entering the circulation.
Drivers should not see what is on the other side.
Ignorance Is Strength.
War is Peace.

Right. Some of the tactics seem (as I said) counter-intuitive, but a confused/hesitant driver is usually one that is slower and more observant. Last I checked, confused drivers weren't usually the ones flying up the right lane going 20 over the limit! Of course, slow driving is not always safe, but everyone (even confident drivers) have to slow down at a roundabout (assuming they see it coming), so that differential in speed is less of a concern.

This confused/hesitant tactic has been in place across Western Washington for decades. Many of our urban two-lane intersections are unsigned/unmarked. Drivers basically approach each one with hesitation; you don't stop unless necessary, but you slow down and quickly glance before moving on.

[kalvado]

Driver hesitation = driver confusion.  That doesn't sound like a good thing.

welcome to The Brave Roundabout World.
Drivers must hesitate before entering the circulation.
Drivers should not see what is on the other side.
Ignorance Is Strength.
War is Peace.

Right. Some of the tactics seem (as I said) counter-intuitive, but a confused/hesitant driver is usually one that is slower and more observant. Last I checked, confused drivers weren't usually the ones flying up the right lane going 20 over the limit! Of course, slow driving is not always safe, but everyone (even confident drivers) have to slow down at a roundabout (assuming they see it coming), so that differential in speed is less of a concern.

This confused/hesitant tactic has been in place across Western Washington for decades. Many of our urban two-lane intersections are unsigned/unmarked. Drivers basically approach each one with hesitation; you don't stop unless necessary, but you slow down and quickly glance before moving on.

At which point its a good idea to stop and think: what is the purpose? Why are we doing whatever we're doing?
The ultimate goal of road design is, from my perspective, is to move traffic safely, efficiently, without undue stress. There are other factors, like emergency vehicle access, but lets not go into that for now.
So, roundabouts bring:
-mixed results in terms of safety
-throughput? My impression is that it is reduced compared to alternatives
-now we need to cut throughput down even further to improve safety
-we need to increase the level of stress for safety at a roundabout (and whatever ripple effects result from that; the stressed driver is the bad driver)
-we lose intersection scalability; sort of OK in NY as the state seems to be planning for decline

And.. did we gain anything, other than a happy contractor with a great construction contract?