Triple "Left Turn Only" Lanes

[jeffandnicole]

There is a pair of triple left turn only lanes proposed for the Algonquin Road legs of the Randall Road/Algonquin Road intersection in McHenry County, Illinois.

See Page 3 of this exhibit

The road width suggests to me that the diagram is actually of an intersection in Florida.

Huh?  Only Florida has roads 8 lanes wide with turn lanes?  That'll be news to nearly every other state, territory, province and country in the world.

My comment was a poor attempt at humour. You probably don't know this, but I come from an area where four through lanes is absolutely nuts, let alone a three left turn lanes. The state that I often (mockingly) associate with wide roads is Florida.

Except, as Kacie Jane points out one page 1 of this thread, you do have a triple left turn lane near your area:

The ramp from I-5 southbound at SR 512 has one.  The interchange used to be a full cloverleaf, but the SB-EB movement moved to a traffic signal about a decade ago. Google Maps

I remember reading a newspaper article - probably somewhere between 2 and 5 years ago - saying that one and a then-brand new one on or near SR 410 in the Sumner/Buckley/Enumclaw area were the only ones in the state, but I don't recall exactly where that one is.

[jakeroot]

There is a pair of triple left turn only lanes proposed for the Algonquin Road legs of the Randall Road/Algonquin Road intersection in McHenry County, Illinois.

See Page 3 of this exhibit

The road width suggests to me that the diagram is actually of an intersection in Florida.

Huh?  Only Florida has roads 8 lanes wide with turn lanes?  That'll be news to nearly every other state, territory, province and country in the world.

My comment was a poor attempt at humour. You probably don't know this, but I come from an area where four through lanes is absolutely nuts, let alone a three left turn lanes. The state that I often (mockingly) associate with wide roads is Florida.

Except, as Kacie Jane points out one page 1 of this thread, you do have a triple left turn lane near your area:

The ramp from I-5 southbound at SR 512 has one.  The interchange used to be a full cloverleaf, but the SB-EB movement moved to a traffic signal about a decade ago. Google Maps

I remember reading a newspaper article - probably somewhere between 2 and 5 years ago - saying that one and a then-brand new one on or near SR 410 in the Sumner/Buckley/Enumclaw area were the only ones in the state, but I don't recall exactly where that one is.

Yes, and as I point out on the last page, there's another one in Federal Way:

Federal Way, WA. SR 18 West, left towards Kits Corner Road: http://goo.gl/SrYT6O

I'm not suggesting (at least intentionally) that there aren't any triple-lefts. I'm saying they're rare, and in comparison to Florida (which, to me, seems to have plenty), we have so few, that some people north in Snohomish or Skagit counties probably have never seen one, and to them, a triple-left is in fact unheard of mythical.

I'm now interested to see if the Vancovuer (BC) area has any.

[clong]

There are 2 I can think of in Birmingham, one is at Riverchase Parkway @ Valleydale Rd, the other is at Summitt Blvd @ US 280.

Since this post, there was a project to improve flow on 280 - http://aldotapps.dot.state.al.us/US280/doc/280%20Access%20Management.pdf.
As you can see in the link, this created additional triple lefts onto US 280 at...
Rocky Ridge Rd
Dolly Ridge Rd
Inverness Pkwy

[tradephoric]

SYNCHRO models of Bruce Browns Blvd and Florida 56 (both have the same upstream geometry and traffic volumes):

https://www.google.com/maps/@28.185609,-82.3535353,339m/data=!3m1!1e3

The triple left turn model runs a 200 second cycle length with pedestrian crossings up to 225 feet.  A long cycle is needed to satisfy the pedestrian interval requirements found in the MUTCD.  The Median U-turn model runs an 80 second cycle length with short pedestrian crossings.  It's a simple 2-phase intersection that maximizes throughput. 

Multiple left turn lanes aren't always the answer to maximize throughput.

[jeffandnicole]

If you are able to view the Mummers Parade in Philly today, you get a great look at Philly's quadruple right on Market Street at City Hall, which is where the TV broadcast takes place.

[cl94]

SYNCHRO models of Bruce Browns Blvd and Florida 56 (both have the same upstream geometry and traffic volumes):

https://www.google.com/maps/@28.185609,-82.3535353,339m/data=!3m1!1e3

The triple left turn model runs a 200 second cycle length with pedestrian crossings up to 225 feet.  A long cycle is needed to satisfy the pedestrian interval requirements found in the MUTCD.  The Median U-turn model runs an 80 second cycle length with short pedestrian crossings.  It's a simple 2-phase intersection that maximizes throughput. 

Multiple left turn lanes aren't always the answer to maximize throughput.

If it gets to the point where you need more than a double left and the road requiring such a left isn't one-way, you should be considering something other than standard turn treatment. The current setup has an 11-lane cross section on its widest approach with no refuge island. As you said, the cycle length is excessive, as clear time is needed in addition to the long pedestrian interval. Assuming a pedestrian speed of 2.5 mph (conservative estimate for walking speed if there are a significant amount of elderly users), the minimum clear pedestrian interval is 62 seconds. While the median U-turn proposal only produces a slight reduction of lanes at the intersection, a refuge island will allow the pedestrian interval to be halved.  Additionally, as there are fewer phases at the intersection, there's less lost time built into the cycle.

If you have the space for a triple left turn, you almost certainly have the space for a median U-turn, which is almost always a better option, especially if two are opposing, as in this case.

[Revive 755]

The triple left turn model runs a 200 second cycle length with pedestrian crossings up to 225 feet.  A long cycle is needed to satisfy the pedestrian interval requirements found in the MUTCD.  The Median U-turn model runs an 80 second cycle length with short pedestrian crossings.  It's a simple 2-phase intersection that maximizes throughput.

Few nitpicks:

1) Many traffic signals run much shorter phases when there are no pedestrian calls, and only run the super-long ped phase when a pedestrian pushes the ped button or the button jams.

2) I would assume that one of those roads there has the signals coordinated which would probably require some other cycle length than the one that optimizes the intersection in question?

3) It appears many of the u-turning vehicles are not proceeding to turn right when they return to the main intersection.  I hear that Synchro/Simtraffic is not as good as VISSIM in handling this, but I would be curious to see how this changes the right turn queues.

[tradephoric]

1) Many traffic signals run much shorter phases when there are no pedestrian calls, and only run the super-long ped phase when a pedestrian pushes the ped button or the button jams.

Let's assume, based on the vehicle demands, the signal could run a 120 second cycle length when there are no pedestrian calls.  This cycle length may provide perfect "green to green"  coordination with the surrounding signal (which in this case is 1500 feet to the south at Williamsburg Drive).  What happens when a pedestrian actuation occurs?  Instead of running a 120 second cycle, the signal would need to run a 200 second cycle just to fit the pedestrian intervals and the once perfect coordination turns 180 out.  It's not a great option when the coordination of a corridor is dependent on whether a pedestrian actuates a pushbutton.

2) I would assume that one of those roads there has the signals coordinated which would probably require some other cycle length than the one that optimizes the intersection in question?

The intersection being modeled, which includes triple left turn lanes at all 4-legs, would likely be the critical intersection within the zone.  There is no doubt that the optimized cycle length of the critical intersection may not match the optimal cycle length of the surrounding intersections (which could very well run shorter cycle lengths).  A determination would need to be made whether the surrounding signals would benefit from coordinating to the critical intersection or if they should run isolated.

3) It appears many of the u-turning vehicles are not proceeding to turn right when they return to the main intersection.  I hear that Synchro/Simtraffic is not as good as VISSIM in handling this, but I would be curious to see how this changes the right turn queues.

I don't put that much stock in any model to be honest.  Synchro does have a tough time accurately modeling the intersection paths and some quirky things can be seen when tracking individual vehicles.  That said, there are several real world examples of Michigan left corridors that efficiently run at low cycle lengths during the main rush hours without much right turning queue issues.

[roadfro]

3) It appears many of the u-turning vehicles are not proceeding to turn right when they return to the main intersection.  I hear that Synchro/Simtraffic is not as good as VISSIM in handling this, but I would be curious to see how this changes the right turn queues.

I don't put that much stock in any model to be honest.  Synchro does have a tough time accurately modeling the intersection paths and some quirky things can be seen when tracking individual vehicles.  That said, there are several real world examples of Michigan left corridors that efficiently run at low cycle lengths during the main rush hours without much right turning queue issues.

VISSIM is better at modeling vehicle paths, because in order to run the model you have to define all the vehicle paths in the modeled network. With Synchro you just build the network and SimTraffic interprets it, although there are ways to reduce abnormal and unlikely vehicle paths.

[cl94]

3) It appears many of the u-turning vehicles are not proceeding to turn right when they return to the main intersection.  I hear that Synchro/Simtraffic is not as good as VISSIM in handling this, but I would be curious to see how this changes the right turn queues.

I don't put that much stock in any model to be honest.  Synchro does have a tough time accurately modeling the intersection paths and some quirky things can be seen when tracking individual vehicles.  That said, there are several real world examples of Michigan left corridors that efficiently run at low cycle lengths during the main rush hours without much right turning queue issues.

VISSIM is better at modeling vehicle paths, because in order to run the model you have to define all the vehicle paths in the modeled network. With Synchro you just build the network and SimTraffic interprets it, although there are ways to reduce abnormal and unlikely vehicle paths.

Having experience with both, I agree completely. Synchro is relatively easy to set up (especially for the novice user), but there are quite a few interesting paths. For one class this past semester, I modeled a roundabout in Synchro and there were a bunch of U-turns that don't currently happen, even though I plugged in traffic counts. Is it good for simplistic cases, such as a "traditional" intersection between 2 single carriageways? Yes. But not for much more.

[roadfro]

Having experience with both, I agree completely. Synchro is relatively easy to set up (especially for the novice user), but there are quite a few interesting paths. For one class this past semester, I modeled a roundabout in Synchro and there were a bunch of U-turns that don't currently happen, even though I plugged in traffic counts. Is it good for simplistic cases, such as a "traditional" intersection between 2 single carriageways? Yes. But not for much more.

Yeah, there is a way to prevent things like this. I had to model a small grid for a traffic class a few years ago (I-80 at Exit 13/Virginia St in downtown Reno) and had similar issues--one car going forward at Virginia/Maple and making three consecutive left turns in a maneuver that ultimately had the same origin/destination result as if the car had just taken a right at Virginia/Maple. Our instructor pointed this out and showed us an adjustment that prevented (most of) these random movements from occurring--I haven't played with Synchro for quite some time, so I forget the exact setting...but probably documented that somewhere in the project write-up.

[SignGeek101]

I just remembered one close to where I live.

http://goo.gl/maps/fGLXJ

Satellite

http://goo.gl/maps/t1YW4

Not sure why its a triple turn, there's not much there except fields, my old high school, and a city park.

[jakeroot]

I just remembered one close to where I live.

http://goo.gl/maps/fGLXJ

Satellite

http://goo.gl/maps/t1YW4

Not sure why its a triple turn, there's not much there except fields, my old high school, and a city park.

That intersection has a lot of interesting features. Triple left turn, ugly BGSs, etc.

[Revive 755]

It's not a pure triple left turn only, but Northbound Schuetz Road at Page Avenue in St. Louis County, Missouri has been widened to have a dual left, a shared thru-left, and a dedicated right turn lane.  Streetview

[jakeroot]

It's not a pure triple left turn only, but Northbound Schuetz Road at Page Avenue in St. Louis County, Missouri has been widened to have a dual left, a shared thru-left, and a dedicated right turn lane.  Streetview

I was going to complain about the lack of a second through signal, but GSV shows that setup in place since before 2008, before the requirement for two through heads for all straight movements came about (2009 MUTCD).

[kj3400]

Got one at Connecticut Av (MD 185) at Georgia Av (MD 97) near Aspen Hill.

https://goo.gl/maps/TJtk7dm8Efo

[jakeroot]

Got one at Connecticut Av (MD 185) at Georgia Av (MD 97) near Aspen Hill.

https://goo.gl/maps/TJtk7dm8Efo

Wasn't sure how many there were in the Northeast. Nice find!

This intersection in NYC, where 495 ends and meets 9A (West St.), features 4 Right turn lanes and 3 Left turn lanes.  http://goo.gl/maps/0iBUK

Apologies for quoting such an old post.

That right turn is a bit odd. There are four right turn lanes, but only three through lanes. Sounds like a very New York-ish situation (where lanes are suggestions, IIRC).

Found a third one in Chicagoland:  Streeter Drive at Grand Avenue just west of Navy Pier:
https://maps.google.com/maps?hl=en&ll=41.891832,-87.611369&spn=0.001261,0.002064&t=h&z=20

This one has been removed since this posting. Not sure why.

[doorknob60]

Vista Ave and Capitol Blvd in Boise, ID: https://goo.gl/maps/reTqj2Lbxgw
There is no through movement, but it is still definitely a triple left.

This one almost counts so I'll post it, since it's the only other triple turn I can think of in the area. Idaho St and 16th St. The far right lane is a straight/left option lane. https://goo.gl/maps/PAAM6Q12MaD2

I was going to post the one on WA-18 in Federal Way since I drove through it last week, but Jakeroot already did.

[bcroadguy]

There is a pair of triple left turn only lanes proposed for the Algonquin Road legs of the Randall Road/Algonquin Road intersection in McHenry County, Illinois.

See Page 3 of this exhibit

The road width suggests to me that the diagram is actually of an intersection in Florida.

Huh?  Only Florida has roads 8 lanes wide with turn lanes?  That'll be news to nearly every other state, territory, province and country in the world.

My comment was a poor attempt at humour. You probably don't know this, but I come from an area where four through lanes is absolutely nuts, let alone a three left turn lanes. The state that I often (mockingly) associate with wide roads is Florida.

Except, as Kacie Jane points out one page 1 of this thread, you do have a triple left turn lane near your area:

The ramp from I-5 southbound at SR 512 has one.  The interchange used to be a full cloverleaf, but the SB-EB movement moved to a traffic signal about a decade ago. Google Maps

I remember reading a newspaper article - probably somewhere between 2 and 5 years ago - saying that one and a then-brand new one on or near SR 410 in the Sumner/Buckley/Enumclaw area were the only ones in the state, but I don't recall exactly where that one is.

Yes, and as I point out on the last page, there's another one in Federal Way:

Federal Way, WA. SR 18 West, left towards Kits Corner Road: http://goo.gl/SrYT6O

I'm not suggesting (at least intentionally) that there aren't any triple-lefts. I'm saying they're rare, and in comparison to Florida (which, to me, seems to have plenty), we have so few, that some people north in Snohomish or Skagit counties probably have never seen one, and to them, a triple-left is in fact unheard of mythical.

I'm now interested to see if the Vancovuer (BC) area has any.

There's two intersections I know of in the Vancouver area that sort of count (two left turn only lanes and a left / through lane):

https://www.google.com/maps/@49.2300177,-122.689431,3a,45.5y,356.47h,88.26t/data=!3m6!1e1!3m4!1sb7Kw2ZqfSEw6ck8X2scGiA!2e0!7i13312!8i6656

https://www.google.com/maps/@49.0901466,-123.0576636,3a,42.6y,90.96h,89.74t/data=!3m6!1e1!3m4!1sTRoOCTh62WlDGSJtXNcitg!2e0!7i13312!8i6656?hl=en

It looks like they never updated the signals after changing the lane configuration at both locations  .

There was a real triple left turn but it was removed a few years ago when they built an interchange:

https://www.google.ca/maps/@49.228272,-122.8193543,3a,39.4y,201.91h,89.3t/data=!3m7!1e1!3m5!1slFUYcV6650j4ANF0KReMgQ!2e0!5s20110801T000000!7i13312!8i6656 (2009 Streetview link)

[US 89]

I can think of three off the top of my head in Utah, but I know there are more:

Harrison Blvd at US 89, Ogden
Riverdale Rd at 1900 West, Roy
Bluff St at Sunset Blvd, St. George

[jakeroot]

I'm now interested to see if the Vancovuer (BC) area has any.

There's two intersections I know of in the Vancouver area that sort of count (two left turn only lanes and a left / through lane):

https://www.google.com/maps/@49.2300177,-122.689431,3a,45.5y,356.47h,88.26t/data=!3m6!1e1!3m4!1sb7Kw2ZqfSEw6ck8X2scGiA!2e0!7i13312!8i6656

https://www.google.com/maps/@49.0901466,-123.0576636,3a,42.6y,90.96h,89.74t/data=!3m6!1e1!3m4!1sTRoOCTh62WlDGSJtXNcitg!2e0!7i13312!8i6656?hl=en

It looks like they never updated the signals after changing the lane configuration at both locations  .

There was a real triple left turn but it was removed a few years ago when they built an interchange:

https://www.google.ca/maps/@49.228272,-122.8193543,3a,39.4y,201.91h,89.3t/data=!3m7!1e1!3m5!1slFUYcV6650j4ANF0KReMgQ!2e0!5s20110801T000000!7i13312!8i6656 (2009 Streetview link)

I was thinking the Mary Hill Bypass might have had one, but I never got to see it pre-interchange. That same area also had the only ramp meter in BC (I don't know of any, though I wish they were more common (unpopular opinion, I know)).

I think I may have driven through that first one, but didn't notice it. I don't mind the sparse signals, but I think more than two for a three lane split-phased signal might be wise. Perhaps an extra signal head on the right, as has become more common.

I can think of three off the top of my head in Utah, but I know there are more:
...
Bluff St at Sunset Blvd, St. George

I like that the sign actually specifies "3 LEFT TURN LANES". You'd normally see "LEFT 3 LANES" for such an approach (or no signage at all, just arrows).

[Super Mateo]

Jackson Street in Chicago has one of these at its east end at Lake Shore Drive.

[TXtoNJ]

SYNCHRO models of Bruce Browns Blvd and Florida 56 (both have the same upstream geometry and traffic volumes):

https://www.google.com/maps/@28.185609,-82.3535353,339m/data=!3m1!1e3

The triple left turn model runs a 200 second cycle length with pedestrian crossings up to 225 feet.  A long cycle is needed to satisfy the pedestrian interval requirements found in the MUTCD.  The Median U-turn model runs an 80 second cycle length with short pedestrian crossings.  It's a simple 2-phase intersection that maximizes throughput. 

Multiple left turn lanes aren't always the answer to maximize throughput.

If it gets to the point where you need more than a double left and the road requiring such a left isn't one-way, you should be considering something other than standard turn treatment. The current setup has an 11-lane cross section on its widest approach with no refuge island. As you said, the cycle length is excessive, as clear time is needed in addition to the long pedestrian interval. Assuming a pedestrian speed of 2.5 mph (conservative estimate for walking speed if there are a significant amount of elderly users), the minimum clear pedestrian interval is 62 seconds. While the median U-turn proposal only produces a slight reduction of lanes at the intersection, a refuge island will allow the pedestrian interval to be halved.  Additionally, as there are fewer phases at the intersection, there's less lost time built into the cycle.

If you have the space for a triple left turn, you almost certainly have the space for a median U-turn, which is almost always a better option, especially if two are opposing, as in this case.

Does anyone know definitively why Florida makes these design choices as consistently as they do?

[UCFKnights]

If there is major left turn movement, I can't imagine a median U turn would provide sufficient capacity. Is that capacity really comparable? Triple left also probably is still the cheapest.

[tradephoric]

Orlando has many wide boulevards throughout the city and probably has the highest concentration of long pedestrian crosswalks in America.  Orlando isn't the only city in America with wide boulevards - Detroit and New Orleans have wider boulevards on average — but Orlando intersections are designed to get pedestrians the entire way across the intersection and don't split them up into two shorter crossings (like you routinely see at wide intersections in Detroit and New Orleans).  Long continuous crosswalks naturally lead to high cycle lengths, as enough time needs to be given to allow pedestrians to cross the street.  The higher the cycle length the more left-turn queuing space is needed to prevent left-turners from backing up into the through lanes.  Long continuous crosswalks throughout Orlando is one of the main reasons why you see so many massive intersections with double and triple-left turn lanes.  Here are all the intersections in Orlando with continuous crosswalks of 140 feet or more.  I doubt you will find another city in America that comes close to this many long continuous crosswalks: