Crash prone 'modern roundabouts'

[ET21]

 

[jakeroot]

Engineers typically build things with multiple redundancies, to avoid catastrophic failure (look up all the redundancies for the Tower of Terror...really amazing stuff). Traffic lights technically have a fail-safe operation: if someone approaches an intersection without noticing the signal, they can, hypothetically, pass through it unharmed because there are no obstacles directly in front of them other than other cars (which there would be at roundabouts too, so it's still a fair comparison).

Sorry, I have a hard time agreeing with the notion that traffic lights have fail-safe operation. Russian roulette-style fail-safe, maybe. Miss the signal and you are either really lucky, or fail in a really ugly fashion. At least with a roundabout, you have the chance the splitter island curb will slow-down or redirect a errant vehicle (granted, those chances decrease on multilane approaches).

But traffic lights aren't red all the time. The comparison is not red light vs roundabout. The comparison is traffic signal vs roundabout. At a roundabout, if you approach one without paying attention, you will crash, in some way. It is impossible not to. But at a traffic light, assuming you're approaching it without paying attention, there's a decent chance that either A) the signal will be green when you arrive, allowing you to pass through unharmed (assuming you can at least stay in your lane), or B) the light will be red but there won't be any cars in the intersection. On the flip side, C) the light could be red and you'll whack into the backside of waiting traffic, or D) the light is red and you T-bone another car.

Of course, scenarios C and D can be absolutely horrific crashes, usually the kind of crash when people end up either seriously injured or killed. And both are likely the reason that roundabouts became popular state-side (to curb those serious T-bone collisions). What I don't think roundabout engineers have fully considered, is how many people are distracted by something they're doing in the car (literally anything from adjusting the radio, to putting on makeup, or using their phone). More often than not, these people are not going to kill themselves approaching a roundabout, but they could by A) crashing into a concrete central island, B) T-boning a car going around the central island (assuming they don't make it to the central island), or C) taking last-second evasive action and rolling their car (also possible at a signal, to be fair, although more likely at roundabouts because one cannot simply drive through a central island without crashing).

Despite everything I've said, I'm actually still a proponent of roundabouts. But I feel there is good roundabout design, and bad roundabout design. I'm not trying to be Dieter Rams, and I'm certainly not an engineer with loads of evidence on my side, but hear me out: good roundabout design would be,

* large enough central island that could accommodate a crashing car
* approaches that do not feature chicanes (promotes lane drifting, side-swiping)
* right-angle approaches (approaches tangential to the circulating roadway promote failure to yield and high speeds, but still require driver action to avoid crashing...may as well require hard-right turns)

This roundabout in Federal Way, WA (opened 2004) fits my criteria pretty well; the historic trees in the middle are surrounded by bushes that might help slow approaching traffic, should someone approach at high speed; I would personally remove the trees in favor of dense, forgiving foliage, but it's otherwise quite good:

[tradephoric]

Sorry, I have a hard time agreeing with the notion that traffic lights have fail-safe operation. Russian roulette-style fail-safe, maybe. Miss the signal and you are either really lucky, or fail in a really ugly fashion. At least with a roundabout, you have the chance the splitter island curb will slow-down or redirect a errant vehicle (granted, those chances decrease on multilane approaches).

There are lots of inaccuracies in this paragraph.  DaBigE compares 'missing the signal' (ie. blowing through a red light) to a game of Russian roulette.  In Russian roulette there is a 83% chance that you survive each pull of the trigger.  Using DaBigE's analogy, if a driver has an 83% chance of making it through a red light without getting hit, it would actually be 'really unlucky' if they got hit (which is opposite of what was being argued).  And who says that when a crash occurs that it will be a "fail in a really ugly fashion"?  I've witnessed a t-bone crash at a traffic signal that resulted in nothing more than a love tap.  I'll admit that horrendous accidents can occur at traffic signals, but it's foolish to suggest they always do.  Lastly, the curbing at roundabouts redirecting an errant vehicle isn't necessarily a safety feature... many motorcyclists have died after striking the curbing inside a roundabout. 

[DaBigE]

Engineers typically build things with multiple redundancies, to avoid catastrophic failure (look up all the redundancies for the Tower of Terror...really amazing stuff). Traffic lights technically have a fail-safe operation: if someone approaches an intersection without noticing the signal, they can, hypothetically, pass through it unharmed because there are no obstacles directly in front of them other than other cars (which there would be at roundabouts too, so it's still a fair comparison).

Sorry, I have a hard time agreeing with the notion that traffic lights have fail-safe operation. Russian roulette-style fail-safe, maybe. Miss the signal and you are either really lucky, or fail in a really ugly fashion. At least with a roundabout, you have the chance the splitter island curb will slow-down or redirect a errant vehicle (granted, those chances decrease on multilane approaches).

But traffic lights aren't red all the time. The comparison is not red light vs roundabout. The comparison is traffic signal vs roundabout. At a roundabout, if you approach one without paying attention, you will crash, in some way. It is impossible not to. But at a traffic light, assuming you're approaching it without paying attention, there's a decent chance that either A) the signal will be green when you arrive, allowing you to pass through unharmed (assuming you can at least stay in your lane), or B) the light will be red but there won't be any cars in the intersection. On the flip side, C) the light could be red and you'll whack into the backside of waiting traffic, or D) the light is red and you T-bone another car.

True, a traffic light isn't red all the time; however, that still does not eliminate the possibility of a fatal collision at a traffic light. You could have the green light and be the innocent victim of someone running the red light for the other roadway. Neither control is perfect, but I'd rather take my chances at a roundabout, where statistically, I am more likely to be able to walk away from a crash.

Despite everything I've said, I'm actually still a proponent of roundabouts. But I feel there is good roundabout design, and bad roundabout design. I'm not trying to be Dieter Rams, and I'm certainly not an engineer with loads of evidence on my side, but hear me out: good roundabout design would be,

* large enough central island that could accommodate a crashing car
* approaches that do not feature chicanes (promotes lane drifting, side-swiping)
* right-angle approaches (approaches tangential to the circulating roadway promote failure to yield and high speeds, but still require driver action to avoid crashing...may as well require hard-right turns)

Unfortunately, there is a finite limit to how large a central island can be and still have effective speed control. The larger the central island, the more distance circulating traffic has to speed up, the more severe a circulating crash can be. Likewise, while most good roundabout designers try to avoid chicanes, they may be a necessary element of approach speed control, due to site constraints. Hard right turns on the entry have the worst sight conditions for the yielding driver. Too flat of an entry promotes vehicles speeding up upon entering the circle (what some designers refer to as a 'hockey stick' design).

This story is floating around numerous engineering forums I follow as well. There is some discussion that the final design for the CA roundabout includes sloped soil/landscaping on the outside of the concrete ring, ultimately limiting the exposure of the concrete face. If that is true, and there was no medical event prior to the crash, we may be looking at a poor construction staging. There's the chance that the final design might have resulted in the vehicle only getting stuck in the central island; depending on the speed, it could result in someone vaulting. Until we know the results of the crash report, everything is pure speculation.

If this intersection was up to me (and I was told it had to be a roundabout), the use of concrete in the central island would limited to curb and gutter and the truck apron only.

[DaBigE]

^^^^^^^^^^^
Unfortunately, Caltrans planning seems to be going through a prolonged phase of concept over reality; putting a roundabout on CA 12 anywhere between I-5 and I-80 -- a "shortcut" from the North Bay/"wine country" to Valley N-S arterials that avoids much of Bay traffic -- was and is a recipe for disaster, as evidenced by the fatality.  It's as if the prevailing "wisdom" (right!) is to attempt to micromanage rather than expedite open-road travel -- in this case, forcing high-speed through traffic into sporadic slowdowns to achieve -- well, what? -- a drop in the aggregate speed; a punitive measure for actually being out on the road to begin with!? -- the rationale, or lack thereof, behind putting obstacles in the way of rural through traffic is astonishing!  These are connecting corridors, folks, not parts of a miniature golf course!  I've been through 12/113 (pre-roundabout) numerous times; a channelized and signalized intersection -- with plenty of advance notice, particularly on 12 in both directions, would have been appropriate, not a center-of road structure.  I understand the desire of some planners to attempt to get folks out of their cars as much as possible -- but not through their windshields!

Based on your description, a signal is no better of a solution. Having people come to a complete stop in the middle of a rural 55-mph highway is no better than a roundabout. By your description, a grade separated interchange is the best solution, as any other form of intersection control has no choice but to negatively impact throughput on the mainline at some point.

I don't recall seeing one person defend the vertical concrete surface in the middle of the roundabout; can we finally move on?

When it comes to a signalized intersection at the 12/113 junction, I clearly mentioned quite a bit of advance notice; previous Caltrans installations of this type have placed warnings of a red signal phase in the form of flashing notification to that effect at least 3/4 mile back from the interchange itself.  That has proven to be quite effective at getting traffic to slow down if and when a red signal is imminent; ample time & space to drop from 65-70 to zero.  I doubt whether advance notification of this sort is placed for a roundabout; the ones I've encountered simply have stepped speed drops.  I'm just questioning the appropriateness of this format in high-speed non-urban situations; it seems like the "garbage can" method -- a solution searching for a problem!

With any rural roundabout projects that I've been involved with,* there is plenty of advanced warning. At the very least, there are two sets of warning signs, but depending on the site, flashing beacon lights and rumble strips are also used. That is in addition to the roadway geometry queues. While I cannot speak for Caltrans decision methods, I can say that the states I've worked in have formal decision reports (aka Intersection Control Evaluations: "ICE Reports"). The decision to install a roundabout is not as thoughtless as some believe it to be.

*none have been on a facility with speeds higher than 55 mph

[tradephoric]

While I cannot speak for Caltrans decision methods, I can say that the states I've worked in have formal decision reports (aka Intersection Control Evaluations: "ICE Reports"). The decision to install a roundabout is not as thoughtless as some believe it to be.

Regardless of the decision making process to install a roundabout, once installed agencies don't want to admit that it may have been the wrong decision.  I recently heard a presentation from Paul C. Ajegba, the director of MDOT, discussing the 14 Mile & Orchard Lake roundabout.  He actually admitted he was initially skeptical of the roundabout design but since it opened he has been impressed with how it has operated.  This is the same roundabout that saw a huge spike in total crashes with no significant drop in injury crashes when converted from a traffic signal to a roundabout in 2015.  In 2018 alone the roundabout saw 144 crashes including 20 injury crashes.  Yet you have the director of MDOT praising how well it's operating?  It's like he's willfully ignorant to the crash problem at the roundabout.  No mention of crashes at all.

[kphoger]

A couple of years ago, I ran a red light and crashed into a car going the other way.  We were both going less than 30 mph, I wasn't fiddling with the radio or using my cell phone or on medication or anything like that.  I was simply concerned that I had missed my turn and was looking the other way to confirm or deny my suspicion, rather than paying attention to the color of the stoplight.  Even at such a low speed, my car pushed his up onto the sidewalk, the whole side of his car was caved in, his airbag deployed, and he was having trouble breathing due to an apparent asthma attack.  The front end of my car's frame was bent and eventually ended up being totaled out by insurance.  Ultimately, both of us were OK, and I didn't even suffer any injury at all.  On the one hand, I'm glad I hit a car in motion rather than a concrete wall.  But on the other hand, I'm also glad we weren't going 45 mph instead of less than 30.

And who says that when a crash occurs that it will be a "fail in a really ugly fashion"?  I've witnessed a t-bone crash at a traffic signal that resulted in nothing more than a love tap.

I've never seen a T-bone crash result "in nothing more than a love tap", and I doubt many others have either.  But I've seen the aftermath of plenty of much more severe T-bone accidents, have witnessed a couple of them happen firsthand, and, as already stated, have been involved in one myself.

Nothing about a traditional stoplight does anything to slow a driver down.  There's a stoplight intersection between Andover and Augusta (KS), for example, on a four-lane US Route with a 65mph speed limit.  Drivers encounter two yellow warning signs in advance (one with a flashing yellow beacon that stays on all the time) but, other than that, there's nothing to slow a driver down.  Other than signs and lights, nothing about it encourages a driver to slow down at all.

In contrast, roundabouts on high-speed roads are commonly constructed to include (as DaBigE said, "two sets of warning signs ... flashing beacon lights ... rumble strips ... roadway geometry".  And that's not to mention the fact that there's something visible directly in front of you as you approach one.  Everything about a roundabout screams "SLOW DOWN".

the curbing at roundabouts redirecting an errant vehicle isn't necessarily a safety feature... many motorcyclists have died after striking the curbing inside a roundabout. 

You were replying to a post related to the curbing of a splitter island and then countered with a statement about the curbing inside a roundabout.  Which one are you talking about?  Also, just because some people wreck or die while striking a safety feature, that doesn't mean it isn't a safety feature.  If I run my car into a guardrail, roll my vehicle, break my neck, and die–that doesn't mean guardrails aren't "necessarily a safety feature".  It just means safety features don't save everyone, and every so often have the opposite effect.

Regardless of the decision making process to install a roundabout, once installed agencies don't want to admit that it may have been the wrong decision.

As opposed to how willing they are to admit they made the wrong decision with other types of installations?

[MNHighwayMan]

If drivers are going to ignore or miss two sets of warning signs, other traffic control features, and the appearance of the roundabout itself, then fuck 'em. No DOT should be beholden to try to accommodate that level of obliviousness. I get that people want roads to "fail safe," but there's a certain level at which you just have to throw your hands up and say "I give up."

[sparker]

^^^^^^^^^^^
One of the basic differences between a signalized intersection and one that's been retrofitted with a roundabout is the fact that with the former -- depending, of course, on the traffic flow of each of the intersecting roads -- a signal can be set to normally display a green/through aspect for the roadway with the greatest traffic flow, with traffic either from the crossroad or a left-turning motion from that main road (which would generally entail only a traffic interruption for the opposite direction) triggering a red/"stop" signal aspect.  From my own experience with the CA 12/113 intersection, that would mean that about 75-80% of the time the traffic flow on CA 12 would not be impaired.  Contrast that to the roundabout, where 100% of the time through traffic on CA 12 is forced to a crawl in order to negotiate the configuration.  Caltrans has plenty of experience with installations with advance warnings of a stop aspect; installing one on CA 12 wouldn't be rocket science, just another application of a method they've used in the past (pioneered, in fact, on CA 99 in Merced and Stanislaus Counties prior to full freeway completion).  IMO, the concept of placing such installations on higher-speed rural routes is either (a) a "sop" to agency personnel (or political oversight) who have an inherent bias against automotive travel in general, or (b) an application of a temporal "one size fits all" approach to such issues -- with the current "fan favorite" being the roundabout.  It's one thing to attempt to make auto travel in the city a less favorable experience vis-à-vis non-automotive alternatives -- but to try to extend that rationale to the heretofore open road seems gratuitous at best and downright obnoxious at worst.  I say the following a bit jokingly -- but given the recent spate of mechanical failures of the lift mechanism of the CA 12 Rio Vista/Sacramento River bridge (which has caused a series of backups on that highway), perhaps the roundabout was considered a way to create a "holding pattern" in the EB direction to minimize the potential for rear-enders if and when a stationary line of cars in the travel lane is encountered.  If so, that seems to be similar to trying to install a picture-hanging hook on a wall with a sledgehammer!  CA 12 is the main artery from southward I-5 and CA 99 to the Napa Valley and other North Bay regional destinations for folks looking to bypass Bay Area congestion; dropping obstacles in that traffic path will just result in a number of folks looking to make up time by doing 75-80+ on that mostly two-lane road; turning it into a social experiment may not be the best use of scarce construction funds!

[DaBigE]

^^^^^^^^^^^
One of the basic differences between a signalized intersection and one that's been retrofitted with a roundabout is the fact that with the former -- depending, of course, on the traffic flow of each of the intersecting roads -- a signal can be set to normally display a green/through aspect for the roadway with the greatest traffic flow, with traffic either from the crossroad or a left-turning motion from that main road (which would generally entail only a traffic interruption for the opposite direction) triggering a red/"stop" signal aspect.  From my own experience with the CA 12/113 intersection, that would mean that about 75-80% of the time the traffic flow on CA 12 would not be impaired.  Contrast that to the roundabout, where 100% of the time through traffic on CA 12 is forced to a crawl in order to negotiate the configuration.  Caltrans has plenty of experience with installations with advance warnings of a stop aspect; installing one on CA 12 wouldn't be rocket science, just another application of a method they've used in the past (pioneered, in fact, on CA 99 in Merced and Stanislaus Counties prior to full freeway completion).  IMO, the concept of placing such installations on higher-speed rural routes is either (a) a "sop" to agency personnel (or political oversight) who have an inherent bias against automotive travel in general, or (b) an application of a temporal "one size fits all" approach to such issues -- with the current "fan favorite" being the roundabout.  It's one thing to attempt to make auto travel in the city a less favorable experience vis-à-vis non-automotive alternatives -- but to try to extend that rationale to the heretofore open road seems gratuitous at best and downright obnoxious at worst.  I say the following a bit jokingly -- but given the recent spate of mechanical failures of the lift mechanism of the CA 12 Rio Vista/Sacramento River bridge (which has caused a series of backups on that highway), perhaps the roundabout was considered a way to create a "holding pattern" in the EB direction to minimize the potential for rear-enders if and when a stationary line of cars in the travel lane is encountered.  If so, that seems to be similar to trying to install a picture-hanging hook on a wall with a sledgehammer!  CA 12 is the main artery from southward I-5 and CA 99 to the Napa Valley and other North Bay regional destinations for folks looking to bypass Bay Area congestion; dropping obstacles in that traffic path will just result in a number of folks looking to make up time by doing 75-80+ on that mostly two-lane road; turning it into a social experiment may not be the best use of scarce construction funds!

How can you make operational claims before the project is even completed? Have you seen any of the operational analyses? Do you know what the projected delays are for the roundabout? Does anyone have access to the traffic volumes for this intersection? If the approaches are fairly balanced, the roundabout generally will have better operations (to about 40K entering VPD); however, if one road is significantly heavier, then a signal typically will be the better operational choice.

Can you define what you mean by "impaired"? Dropping from 55 mph to 20 mph over the size of the roundabout adds a minuscule amount of travel time at free flow conditions (off-peak), while also significantly reducing (eliminating?) unnecessary side street control delay that would occur if the same intersection was signalized.

Being a rural intersection, trying to inhibit travel by auto doesn't seem to apply here. I don't think I've ever come across such perceived "benefit" of a rural intersection project. The fact that they're making any improvements contradicts the anti-car conspiracy theory. If they truly wanted to discourage travel by car, zero lane improvements would be made, except for a bus stop or bike lanes.

[tradephoric]

If drivers are going to ignore or miss two sets of warning signs, other traffic control features, and the appearance of the roundabout itself, then fuck 'em. No DOT should be beholden to try to accommodate that level of obliviousness. I get that people want roads to "fail safe," but there's a certain level at which you just have to throw your hands up and say "I give up."

I understand the "I give up"  mindset when something bad happens at a roundabout.  We have been told that roundabouts reduce fatal crashes by 90% so when something bad happens many try to excuse it away and assume it was an outlier.   Overhead advanced warning flashers and warning signs existed at the flasher that preceded the roundabout yet the same "I give up"  mindset was absent then.
   

In contrast, roundabouts on high-speed roads are commonly constructed to include (as DaBigE said, "two sets of warning signs ... flashing beacon lights ... rumble strips ... roadway geometry".  And that's not to mention the fact that there's something visible directly in front of you as you approach one.  Everything about a roundabout screams "SLOW DOWN".

Having visible objects in the central island can be beneficial to reduce the likelihood of a vehicle plowing through the middle of a roundabout, but do those objects need to be rigid?  The 96th Street and Westfield Blvd roundabout in Carmel has been the site of multiple fatal crashes and in both cases the driver blew through the middle of the roundabout and crashed into a concrete retaining wall.  Instead of a concrete wall, why not just have some visible brush/plantings like this to alert the driver they got to slow down?  At least if they still don't slow down they only blow through some flower gardens...



EDIT:  The picture above was taken at the 96th Street and Westfield Blvd roundabout in Carmel.

[jakeroot]

^^
Glad you brought up that 96th/Westfield roundabout. In this BBC article from 2011, where they went to Carmel and interviewed the mayor and made comparisons to Milton Keynes, they drive through this roundabout (in the attached video, skip to 0:56):

"This used to be stoplight. This is actually one of the conversions. Someday we'll [put a] sculpture in the middle -- that's a budget item".

I've noticed that the roundabout does now indeed have  a sculpture in the middle. Because the concrete wall wasn't enough of an obstacle.

[kphoger]

Does anyone have access to the traffic volumes for this intersection? If the approaches are fairly balanced, the roundabout generally will have better operations (to about 40K entering VPD); however, if one road is significantly heavier, then a signal typically will be the better operational choice.

2017 numbers

[jakeroot]

I was hoping to find some GIS data from Caltrans, which might help illustrate TOD operations, but their website is currently offline while they upgrade their systems.

[tradephoric]

^^
Glad you brought up that 96th/Westfield roundabout. In this BBC article from 2011, where they went to Carmel and interviewed the mayor and made comparisons to Milton Keynes, they drive through this roundabout (in the attached video, skip to 0:56):

"This used to be stoplight. This is actually one of the conversions. Someday we'll [put a] sculpture in the middle -- that's a budget item".

I've noticed that the roundabout does now indeed have  a sculpture in the middle. Because the concrete wall wasn't enough of an obstacle.

Thank you for this additional info Jake.  Designing the roundabout with a concrete retaining wall makes a little more sense assuming the city had always planned a sculpture there.  That being said, it took over a decade before the sculpture was added and both fatal crashes occurred before the sculpture was installed.  In the video you linked from that 2011 BBC article, the retaining wall is largely obscured by the foliage surrounding it (just like it is in the streetview image i linked).  The foliage hides the very real dangers of that fixed object retaining well.  At least with the large sculpture now installed it is much more prevalent that a fixed object exists in the central island.  I wish no fixed object existed in the central island but at least now it stands out.

BTW, i knew the city installed that sculpture in 2017 (it was discussed previously in this thread) i just didn't realize the mayor of Carmel was talking about it way back in 2011. 

[sparker]

Does anyone have access to the traffic volumes for this intersection? If the approaches are fairly balanced, the roundabout generally will have better operations (to about 40K entering VPD); however, if one road is significantly heavier, then a signal typically will be the better operational choice.

2017 numbers

I'd say that a 275% difference in traffic flow (derived from the smaller of the CA 12 numbers) qualifies as significant.   I'll simply reiterate that IMO the present Caltrans penchant for utilizing roundabouts as their default favored configuration is at best misguided and at worst gratuitous and potentially dangerous.   Let's just hope that common sense (and San Benito County residents) can convince D5 that installing one of these at the CA 25/156 intersection is not only wrongheaded but will likely produce a plethora of unintended consequences (and a spike in collisions in the process).   

[DaBigE]

Does anyone have access to the traffic volumes for this intersection? If the approaches are fairly balanced, the roundabout generally will have better operations (to about 40K entering VPD); however, if one road is significantly heavier, then a signal typically will be the better operational choice.

2017 numbers

Thanks for posting that. Any chance they have the turn movement counts? On the surface, it looks unbalanced, but since CA 12 intersects another state highway, the peak hour turning patterns may tell a slightly different story. If there are a significant number of left turns from CA 12 (EB?) to CA 113, that tends to create more gaps for traffic entering from CA 113, and offset some of the bias of CA 12 through traffic.

[sparker]

Does anyone have access to the traffic volumes for this intersection? If the approaches are fairly balanced, the roundabout generally will have better operations (to about 40K entering VPD); however, if one road is significantly heavier, then a signal typically will be the better operational choice.

2017 numbers

Thanks for posting that. Any chance they have the turn movement counts? On the surface, it looks unbalanced, but since CA 12 intersects another state highway, the peak hour turning patterns may tell a slightly different story. If there are a significant number of left turns from CA 12 (EB?) to CA 113, that tends to create more gaps for traffic entering from CA 113, and offset some of the bias of CA 12 through traffic.

CA 12, while mostly 2 lanes (with periodic passing lanes in both directions) is, except for one narrow bridge over the old Sacramento Northern railroad tracks (which can't be replaced without a far-flung bypass since it's on the grounds of the Bay Area Railroad Museum), a facility with wide banked curves designed for travel at the state 65mph surface-road speed limit.  OTOH, CA 113 is a narrow 2-lane rural facility with a couple of sharp right-angle curves that tend to inhibit travel by larger commercial vehicles.   Not much in the way of overall consistent volume south of the town of Dixon.  That being said, the primary use seems to be an ersatz "bypass" of Sacramento (which experiences "normal" commute congestion), particularly used by UC Davis students from points south (the sporadic mechanical failures of the CA 12 Rio Vista bridge notwithstanding!) -- but that traffic still doesn't come close to crowding the section of CA 113 from CA 12 to I-80.  Traffic using it as a bypass is relatively minimal (contained easily within that 5100 AADT); most of the consistent traffic is related to the agricultural area (CA's prime tomato-producing region) through which it passes.  Trust me on this one, you don't want to be behind an open-bed truck hauling tomatoes when it hits a bump or pothole; you'll get squashed tomatoes on your windshield and in your grill!  There is some tomato traffic heading south on 113 to east 12 (a few processing plants in the Lodi/Stockton area), but most transport the fruit from somewhere near 113 northward to I-80 to head for the major processors in SE Sacramento ("Sacramento" brand tomato juice, ubiqutous on the East Coast, hails from there, along with much of commercial/OEM juice production).   Maybe during next years' harvest season I'll get my ass out to the 12/113 roundabout to see if trucks negotiating that facility has resulted in squashed tomatoes either within the circle center or around the periphery!

[jakeroot]

For fun, I prepared a redesign of the intersection.

* traffic approaching from the south must turn right and then U-turn to continue onto northbound CA-113 or westbound CA-12;
* traffic approaching from the west must go straight and then U-turn to access northbound CA-113;
* traffic approaching from the north may go in any direction (left lane would be left turn only);
* traffic approaching from the east may go in any direction (changing one lane to the right to continue onto northbound CA-113 unimpeded).

Operational notes:

* traffic from the south would have a flashing right red arrow (except for the slip lane) during the oncoming green arrow phase, assuming this is allowed;
* the left turn from westbound CA-12 could operate with a flashing yellow arrow, if Caltrans were to be so bold (the visibility would be perfectly acceptable);
* the U-turn would operate as yield-only (it appears to be a quiet movement);
* the slip lanes in the SW and SE corners operate with yield signs;
* the slip lanes in the NW and NE corners would be free-flow maneuvers;
* I did not draw any pedestrian facilities, but could easily modify the design to incorporate these if so desired.

[sparker]

For fun, I prepared a redesign of the intersection.

* traffic approaching from the south must turn right and then U-turn to continue onto northbound CA-113 or westbound CA-12;
* traffic approaching from the west must go straight and then U-turn to access northbound CA-113;
* traffic approaching from the north may go in any direction (left lane would be left turn only);
* traffic approaching from the east may go in any direction (changing one lane to the right to continue onto northbound CA-113 unimpeded).

Operational notes:

* traffic from the south would have a flashing right red arrow (except for the slip lane) during the oncoming green arrow phase, assuming this is allowed;
* the left turn from westbound CA-12 could operate with a flashing yellow arrow, if Caltrans were to be so bold (the visibility would be perfectly acceptable);
* the U-turn would operate as yield-only (it appears to be a quiet movement);
* the slip lanes in the SW and SE corners operate with yield signs;
* the slip lanes in the NW and NE corners would be free-flow maneuvers;
* I did not draw any pedestrian facilities, but could easily modify the design to incorporate these if so desired.

That is one nice design!  It keeps the most essential turning movements:  S>E and W>N intact as simple directional motions while addressing the less-utilized movements -- all while maintaining the primacy of through traffic on CA 12.  Much better than (a) the original or (b) the roundabout replacement.  Because of the slightly hilly terrain, which extends east along CA 12 for a mile or two, an advance safety indicator (signage + flashing yellow?) on WB CA 12 prior to the U-turn may be appropriate; otherwise, the cited measures should be sufficient 

[tradephoric]

Good job Jake, that design of CA 12/CA 113 looks great.  I may not have extended the left turn lane through the intersection for traffic traveling EB CA 12 as that left-turning movement is likely pretty low, but at the same time I'd rather have too much queue space than not enough!  The safety benefits of innovative intersections that eliminate direct left turns often seems overlooked.  Past research has found 20-50% reductions in total crashes at Median U-Turn intersections compared to conventional intersections.  That's why i take issue with the IIHS blanket statement that "roundabouts are a safer alternative to traffic signals and stop signs".  That's the first thing you read when you go to the IIHS roundabout website.  Roundabouts may be a safer alternative to conventional traffic signals but not necessarily innovative intersections like what Jake drew up.

[kphoger]

Does anyone have access to the traffic volumes for this intersection? If the approaches are fairly balanced, the roundabout generally will have better operations (to about 40K entering VPD); however, if one road is significantly heavier, then a signal typically will be the better operational choice.

2017 numbers

Thanks for posting that. Any chance they have the turn movement counts? On the surface, it looks unbalanced, but since CA 12 intersects another state highway, the peak hour turning patterns may tell a slightly different story. If there are a significant number of left turns from CA 12 (EB?) to CA 113, that tends to create more gaps for traffic entering from CA 113, and offset some of the bias of CA 12 through traffic.

I didn't see any turning movement counts, but I did see peak counts.

[jakeroot]

That is one nice design!  It keeps the most essential turning movements:  S>E and W>N intact as simple directional motions while addressing the less-utilized movements -- all while maintaining the primacy of through traffic on CA 12.  Much better than (a) the original or (b) the roundabout replacement.  Because of the slightly hilly terrain, which extends east along CA 12 for a mile or two, an advance safety indicator (signage + flashing yellow?) on WB CA 12 prior to the U-turn may be appropriate; otherwise, the cited measures should be sufficient

Good job Jake, that design of CA 12/CA 113 looks great.  I may not have extended the left turn lane through the intersection for traffic traveling EB CA 12 as that left-turning movement is likely pretty low, but at the same time I'd rather have too much queue space than not enough!  The safety benefits of innovative intersections that eliminate direct left turns often seems overlooked.  Past research has found 20-50% reductions in total crashes at Median U-Turn intersections compared to conventional intersections.  That's why i take issue with the IIHS blanket statement that "roundabouts are a safer alternative to traffic signals and stop signs".  That's the first thing you read when you go to the IIHS roundabout website.  Roundabouts may be a safer alternative to conventional traffic signals but not necessarily innovative intersections like what Jake drew up.

Thanks guys.

If I had to do the redesign again, I would have also opted for a U-turn lane that started after the intersection, rather than before, but I wasn't sure what the turning movements were, so I opted for a more liberal design. I think a warning sign for westbound traffic would certainly be warranted, as they might not be aware of that U-turning traffic. The U-turn could also be relocated if visibility concerns persist.

I do feel like more innovative intersections are what we need in this country. We seem to be applying roundabouts and DDI's far more than I think may be truly necessary. They both have their places, but I'm not 100% sure its where many of them are built, or are being built. Intersections that maybe combine a bit of everything really need more consideration, at least in areas where there is enough ROW.

For this intersection, I knew that the southbound to eastbound maneuver was really heavy, so I opted to design it as regular left turn. But due to the offset with the private road, this required some alternative considerations for the left turn from that road, as well as the left turns off CA-12. Combining a couple of those turns into the U-turn, and leaving two of the original left turns, got me to what you see above.

[tradephoric]

Even if the concrete wasn't there, a basic mounded central island might not have changed the ultimate outcome. Depending on speed (which we still don't know), they could have been sent flying.

It seems like race car drivers have more common sense than some of the engineers designing these roundabouts.  After flipping wildly late in the race at Talladega today, Brendan Gaughan jumped out of his car unscathed and had this to say...

https://www.youtube.com/watch?v=oxcapOXpKxs&lc=z231dxfzwyzlfdktlacdp4341katy34bhutrultcfwpw03c010c

"The easiest thing you can do in a race car is go upside down... nothing hits hard.  So that's the nicest thing about it..."

While a wild flip may look spectacular, each impact is dissipating energy and you don't get that 'dead stop'.  That's not to say flipping can't be deadly (and Brendan was making a pretty big generalization) but I'd take my chances over hitting a wall.  If Brendan was designing a rural roundabout he wouldn't have designed it with a concrete wall in the central island... just asking to kill someone when struck.  Here is a video of 100 Nascar flips, none of which were fatal.

https://www.youtube.com/watch?v=a5xc8pDMA7c

[tradephoric]

A driver was killed last Tuesday at a roundabout in Liberty Lake, WA after striking a concrete retaining wall in the central island of a roundabout.  They should change the headline to read "Alcohol, speed, and a concrete retaining wall believed to be factor in man's fatal crash into Liberty Lake roundabout". 

Alcohol, speed believed to be factor in man's fatal crash into Liberty Lake roundabout
https://www.khq.com/news/police-alcohol-speed-believed-to-be-factor-in-man-s/article_a40e0da6-e9c4-11e9-af1e-d7c910840c47.html