Lane Merges: the Good, the Bad, the Ugly, and the “WTF Is That?!”

[johndoe]

Before 2009, the right-most lane ended on the mainline. Now it's this right lane ending at this ramp shown. There is no weaving in any situation.

Weird...could you share the google link?

[andrepoiy]

Alright.

Ramp Streetview:

2009: 3 lanes ramp

https://goo.gl/maps/PqoodSeLMyuXjk7ZA

today: 2 lanes ramp, with third lane ending before the bridge

https://goo.gl/maps/JLQ3XYkarMJyd47b6

Mainline Streetview:

The red indicates the ramp lanes that are joining the mainline, and the blue indicates the mainline lanes.

2009: right-most lane of mainline ends, with 3 lanes coming from the ramp

https://goo.gl/maps/jxJvHrZcbWSm8d4r6

today: 2 lanes coming from ramp, mainline maintains 3 lanes

https://goo.gl/maps/TGHy31ax2evtdezj6

[johndoe]

Thanks for the links.  See how WB has 5 lanes over Dixon Road, and lane 6 comes on from Martin Grove?  Since that lane goes all the way to the 427 diverge, that's a weaving section.  So maybe they thought if they eliminated the option lane it'd be tougher for Martin Grove vehicles to get on and find gaps. (401 to 427 will push further right)

I wonder why that 3rd lane didn't continue til about Eglinton- looks there is a pretty decent straight (tangent) section there.

[andrepoiy]

Thanks for the links.  See how WB has 5 lanes over Dixon Road, and lane 6 comes on from Martin Grove?  Since that lane goes all the way to the 427 diverge, that's a weaving section.  So maybe they thought if they eliminated the option lane it'd be tougher for Martin Grove vehicles to get on and find gaps. (401 to 427 will push further right)

I wonder why that 3rd lane didn't continue til about Eglinton- looks there is a pretty decent straight (tangent) section there.

Ohhhhhhhh I see what you mean now. Yeah you're probably right. As for the 3rd lane continuing, maybe because they wanted to avoid drivers trying to merge on a curve? Since if you drive this section regularly, you'd probably try to switch out of the lane early

[bcroadguy]

Interior lane merges are terrible.  By design, they leave little room for drivers to be able to get into proper position.  I understand when 2 2-lane roads merge into a 3-lane road that this type of merge is done so as not to give priority to either route - but it is still a problem.  Give MassDOT some kudos for fixing the merge from hell on MA 24, as mentioned above.

Perhaps the only version of interior lane merge that was done safely was the merge on the NJTP just south of exit 8A, before the road was widened some years ago.  The right lane of the left roadway merged in with the left lane of the right roadway to form the new middle lane of the 3-lane NJTP that is south of the neckdown.  Traffic on the other lanes was restricted from interfering with these two lanes (no changing lanes) and they provided a good amount of room to make the merger.  Safe merge when traffic was light, horrible merge that cause miles-long backups during busy times.  Thank goodness they widened this a few year ago to keep the dual roadway setup for several more miles south.

Here's a GSV from 2012:

https://www.google.com/maps/@40.3377398,-74.4793688,3a,75y,212.07h,88.02t/data=!3m7!1e1!3m5!1sP1CUNHc-LygPoCiHgVOcIw!2e0!5s20120601T000000!7i13312!8i6656

Here's an example of an interior lane merge that I hate, largely because it is so unnecessary.  In West Sacramento, the 2-lane CA-275 (Tower Bridge Gateway) merges in with the 3-lane US 50 to create a 4-lane US 50.  Unlike when two freeways come together, there isn't that much traffic on CA 275.  CA 275 should form one lane prior to merging with US 50 and then just merge into US 50's fourth lane.  I lived in the area, and when I was driving on CA 275, I'd always drive only in the right lane to avoid this point of merger.


https://www.google.com/maps/@38.576595,-121.5293029,3a,75y,274.33h,90.87t/data=!3m6!1e1!3m4!1stsDtfhT9YFHc8tAynmjR8w!2e0!7i16384!8i8192!5m1!1e3

I was going to post this interior lane merge, because somehow it is one of only two I've ever encountered and I hate it, but that Sacramento example is sooo much worse.

[jakeroot]

I'm still very much on the fence about interior merging. I can see why they are hated, but I really think they're not as bad as we're thinking.

The main advantage (I think I mentioned this before) is that an interior merge allows for two continuous lanes, whereas ending a lane does not. If a two lane on-ramp shrinks to one lane prior to the merge, the effective capacity of that merge is limited by the capacity of that single-lane segment (something else ahead or prior notwithstanding). With an interior merge, both lanes can continue onto the freeway without "losing" a lane: the left lane merges into through traffic, and the right lane is simply added on. There is also the unintentional advantage of traffic "staying right" to avoid using the interior merge. Mainline capacity is slightly hindered, but this could be advantageous depending on the situation*.

On the flip side: there is no "escape" for the interior merge. I can understand this argument, but would rebut (A) traffic still has two options for merging (left or right, although I recognize the Vancouver example above does not permit this...sigh) and (B) if they are long enough, there should be more than enough room to complete a merge anyway.

In short: properly-designed interior merges may actually be my favorite (I-69 @ Indy's ring road near Fishers is the example I know best), but they need to be designed well. If they're not, lane dropping may be superior. Excess mainline capacity could be used by an interior merge as well.

*Dropping a mainline lane is another option, but I think the flexibility of the interior merge is still better since it allows traffic to respond to current conditions. As opposed to a permanent lane drop, which could be undesirable when the merge is quiet but the mainlines are busy. Exceptionally busy on-ramps may also benefit from a three or four lane meter.

[Dirt Roads]

I'm still very much on the fence about interior merging. I can see why they are hated, but I really think they're not as bad as we're thinking.

The main advantage (I think I mentioned this before) is that an interior merge allows for two continuous lanes, whereas ending a lane does not. If a two lane on-ramp shrinks to one lane prior to the merge, the effective capacity of that merge is limited by the capacity of that single-lane segment (something else ahead or prior notwithstanding). With an interior merge, both lanes can continue onto the freeway without "losing" a lane: the left lane merges into through traffic, and the right lane is simply added on. There is also the unintentional advantage of traffic "staying right" to avoid using the interior merge. Mainline capacity is slightly hindered, but this could be advantageous depending on the situation*.

On the flip side: there is no "escape" for the interior merge. I can understand this argument, but would rebut (A) traffic still has two options for merging (left or right, although I recognize the Vancouver example above does not permit this...sigh) and (B) if they are long enough, there should be more than enough room to complete a merge anyway.

In short: properly-designed interior merges may actually be my favorite (I-69 @ Indy's ring road near Fishers is the example I know best), but they need to be designed well. If they're not, lane dropping may be superior. Excess mainline capacity could be used by an interior merge as well.

*Dropping a mainline lane is another option, but I think the flexibility of the interior merge is still better since it allows traffic to respond to current conditions. As opposed to a permanent lane drop, which could be undesirable when the merge is quiet but the mainlines are busy. Exceptionally busy on-ramps may also benefit from a three or four lane meter.

There used to be an interior merge in downtown Charleston, West Virginia where I-64 westbound splits from I-77 northbound.  Originally, the [straightline] had two lanes coming from I-77/I-79 and the [onramp] was the two lanes of I-64.  Since most of the traffic was on the I-64 side and they were slowed by the Y-curve beneath the [straightline], the interior merge seemed to work quite well.  I'm guessing that they had some problems, because its now switched over so the [straightline] drops a lane to allow unrestricted flow from I-64.  I'm sure this is problematic at times, as the slow lane from I-77/I-79 is forced all the way over to the far left lane, and any traffic headed for the Fort Hill exit (Exit 59A for US-119, now called the Southridge exit by locals) has got a heck of chore to make it all the way over.  Since all of this is on a viaduct system (including the I-64 onramp), it's a wild ride here during icy conditions.

[sprjus4]

My issue with interior merges is they often occur unsigned, and for an out-of-state motorist unfamiliar with the setup, could create a false sense of security there's two lanes as they are driving alongside another motorist, little knowing that lane dumps into the other, which could result in a sideswipe collision.

[roadfro]

My issue with interior merges is they often occur unsigned, and for an out-of-state motorist unfamiliar with the setup, could create a false sense of security there's two lanes as they are driving alongside another motorist, little knowing that lane dumps into the other, which could result in a sideswipe collision.

This. The couple that I've encountered were in California (I don't recall locations) someplace I was unfamiliar with, and the signage either didn't suggest an interior merge or was non-existent. In one case, it was stop-and-go traffic so I was able to figure it out without incident, but I wasn't prepared for it.

Interior merges are somewhat against typical driver expectation, so I'd rather they not be used. But in certain circumstances, they could be fine if designed well–clear signage and markings are definitely needed.

[mrsman]

My issue with interior merges is they often occur unsigned, and for an out-of-state motorist unfamiliar with the setup, could create a false sense of security there's two lanes as they are driving alongside another motorist, little knowing that lane dumps into the other, which could result in a sideswipe collision.

This. The couple that I've encountered were in California (I don't recall locations) someplace I was unfamiliar with, and the signage either didn't suggest an interior merge or was non-existent. In one case, it was stop-and-go traffic so I was able to figure it out without incident, but I wasn't prepared for it.

Interior merges are somewhat against typical driver expectation, so I'd rather they not be used. But in certain circumstances, they could be fine if designed well–clear signage and markings are definitely needed.

https://www.google.com/maps/@34.0645381,-118.2505999,3a,75y,330.14h,72.59t/data=!3m6!1e1!3m4!1swl_Vo5tkgsVl9g6Cj6XEDA!2e0!7i13312!8i6656

On US 101 just north of the 110 freeway, just oustdie of Downtown LA, there is an inside lane merge as well.  The leftmost lane coming from the 110 combined ramp merges in with the rightmost lane of US 101.  From the vantage point of the GSV posted above, it appears that each movement will come into its own lane.  But as you advance, you see that there is an interior merge.  But as this area is so frequently stop and go, it is generally not a safety problem, but is probably part of the reason why this is so constantly congested.  If they did not do an interior lane merge, it might lead to even more congestion, particularly if you limit traffic from 110 to only two lanes instead of three.

The reason why the Sacramento one [that I posted earlier] is so insidious in my opinion, is because it is simply unnecessary.  The on-ramp is coming from the defunct CA-275, which was once a mini-freeway from Downtown Sac to this merge, but it has since been downgraged into a boulevard.  The traffic level is so much lower now, so there is no reason why the two lanes of the boulevard couldn't merge into one lane and then later merge as one lane onto US 50.  The interior lane merge only provides danger here - there is no congestion benefit to speak of.

[fwydriver405]

On US 101 just north of the 110 freeway, just outside of Downtown LA, there is an inside lane merge as well.  The leftmost lane coming from the 110 combined ramp merges in with the rightmost lane of US 101.  From the vantage point of the GSV posted above, it appears that each movement will come into its own lane.  But as you advance, you see that there is an interior merge.  But as this area is so frequently stop and go, it is generally not a safety problem, but is probably part of the reason why this is so constantly congested.  If they did not do an interior lane merge, it might lead to even more congestion, particularly if you limit traffic from 110 to only two lanes instead of three.

The reason why the Sacramento one [that I posted earlier] is so insidious in my opinion, is because it is simply unnecessary.  The on-ramp is coming from the defunct CA-275, which was once a mini-freeway from Downtown Sac to this merge, but it has since been downgraded into a boulevard.  The traffic level is so much lower now, so there is no reason why the two lanes of the boulevard couldn't merge into one lane and then later merge as one lane onto US 50.  The interior lane merge only provides danger here - there is no congestion benefit to speak of.

For the same reasons bolded above... I wonder why the I-80 / US 101 interchange in downtown San Francisco is always congested or if even the interior lane merge is even needed for all movements. Maybe the interior lane merge for this interchange was chosen to prevent lane diving from these movements for drivers wanting to exit within a short distance?

- Traffic coming from I-80 W wanting to follow US 101 Northbound via a TOTSO movement, or to Mission St / Dubois St, ~500 m
- Traffic coming from US 101 SB / Central Fwy to I-80 EB wanting to exit to Exit 1 (Seventh St), ~300 m

[roadfro]

^ Ah, thank you. Pretty sure that I-80 WB to US 101 NB interior merge was one of the ones I've encountered. And it's an incredibly short merge with no signs whatsoever. Sure, there's a yellow beacon, but that does a crapload of good if there's not a sign explaining the beacon...

[Tom958]

On the Schuylkill Expressway, the two two-lane onramps for Penrose Avenue and Passyunk Avenue-Oregon Avenue combine with the two-lane mainline for I-76 (which is on the right!), each compressing from two lanes to one and forming a single three-lane roadway... in less than 600 feet. And the lane that both lanes of I-76 cram into is dropped at the next exit, another 2/3 mile downstream. WTF, indeed. Well, this is the Schuylkill Expressway.



Here's the view in the opposite direction. It has the same configuration but is easier to comprehend.

I don't understand why they don't at least stripe the two onramps down to one lane apiece before the convergence.

[mrsman]

^ Ah, thank you. Pretty sure that I-80 WB to US 101 NB interior merge was one of the ones I've encountered. And it's an incredibly short merge with no signs whatsoever. Sure, there's a yellow beacon, but that does a crapload of good if there's not a sign explaining the beacon...

It seems to me that the design of the I-80/US 101 interchange was one that involved symmetry.  There was an expectation with a full Central Freeway (and a Panhandle Freeway) that each of the three legs of that interchange would have more or less equivalent traffic and that one way of equalizing the effects was to incorporate the interior merge.  If two 2-lane ramps merge together into a 3-lane road, how can one favor one route over the other, so both routes will merge together into the middle lane.  And this seems to be designed with interior merges at all three merging in points.

But given the current realities that the north leg of US 101 is not going all the way to the GG Bridge as a freeway (and in fact was even shortened considerably from what was constructed all the way to Turk/Gough has been truncated by more than a mile to Market/Octavia), that it makes sense to rethink the layout of the interchange.  The 101 north leg is now quite short and is basically a glorified onramp/offramp.  There is not as much need for this leg to be served by this branch to have two lanes on each segment as it comes into the merge.  So if the interchange  lane lines were repainted so that the 101-80 approach were considered a thru freeway from Oakland to San Jose and the 101 north approache were considered an exit and entrance to Market/Octavia, then all of the interior merges here can be eliminated.  Any two lane ramp coming from Market/Octavia should merge together into one lane prior to joining the 101-80 thru traffic, on the merges in both directions.  Likewise, the 80 WB two lane ramp to US 101 north should merge into one lane prior to joining the two lanes coming from San Jose.

[Finrod]

Where the southbound Edens Expressway in Chicago (I-94 East) joins the Kennedy Expressway (I-90 East) headed towards downtown Chicago, the leftmost lane of 94 and the rightmost lane of 90 have to merge immediately:

https://www.google.com/maps/@41.9610668,-87.7440178,3a,75y,131.58h,97.19t/data=!3m6!1e1!3m4!1sTvaWhdG6blkBt4MhlUHQzA!2e0!7i16384!8i8192

On the north side of Atlanta, the ramp from I-285 East to Georgia 400 North is a left merge with zero taper; it's a leftover from when 400 didn't go south of 285.  At least this one is being fixed; here's a shot of it from 2016 before the construction began.

https://www.google.com/maps/@33.9107101,-84.3567697,3a,75y,90t/data=!3m7!1e1!3m5!1sI-ouxTMURPuoWA9XyxJyNQ!2e0!5s20161001T000000!7i13312!8i6656

[andrepoiy]

On the topic of center-lane merges, I encountered my first one somewhere in Montreal when I was there like a day ago. I think it was on QC138, but I don't remember exactly. I had no idea and no warning either, however it was at night so I didn't actually deal with any traffic.

[Bruce]

This one on 44th Avenue in Mountlake Terrace, WA always gets me:

[mrsman]

^^^^^

It seems like the sign conflicts with the road markings.  Does the right lane yield to the left on merging, or is it a zipper merge where both lanes are supposed to take turns?

[bcroadguy]

^^^^^

It seems like the sign conflicts with the road markings.  Does the right lane yield to the left on merging, or is it a zipper merge where both lanes are supposed to take turns?

Yeah that sign is very strange. Based on the pavement markings it looks like a typical right lane ending situation to me. The only somewhat weird (but not really) thing about it, other than the sign, is that the through lane shifts a bit to make room for a left turn lane and bike lane.

[jeffandnicole]

^^^^^

It seems like the sign conflicts with the road markings.  Does the right lane yield to the left on merging, or is it a zipper merge where both lanes are supposed to take turns?

Both lanes are always supposed to take turns. If the road is jammed, it's not expected traffic in the right lane will wait until the left lane is completely cleared.

[Revive 755]

Where the southbound Edens Expressway in Chicago (I-94 East) joins the Kennedy Expressway (I-90 East) headed towards downtown Chicago, the leftmost lane of 94 and the rightmost lane of 90 have to merge immediately:

https://www.google.com/maps/@41.9610668,-87.7440178,3a,75y,131.58h,97.19t/data=!3m6!1e1!3m4!1sTvaWhdG6blkBt4MhlUHQzA!2e0!7i16384!8i8192

Yet the entrance ramp from Montrose gets an auxiliary lane to the Kedzie Avenue/Iriving Park Exit.  Seems backwards to give that ramp a better merge than mainline freeway traffic gets, especially  when there is another entrance only about a third of a mile to the north.

[roadman65]

Interesting border marking on the lower right.


Interstate Bridge on I-5 North at OR/WA Line.

[Scott5114]

On northbound I-35 in OKC, the right lane ends just after the gore at both the ramp to I-240 and the ramp to S.E. 66th Street. This isn't signed at all, by the way, so it catches drivers not familiar with the situation unaware. Why not just do a typical Exit Only situation?

https://www.google.com/maps/@35.3864017,-97.4950737,143m/data=!3m1!1e3

[Ned Weasel]

On the Schuylkill Expressway, the two two-lane onramps for Penrose Avenue and Passyunk Avenue-Oregon Avenue combine with the two-lane mainline for I-76 (which is on the right!), each compressing from two lanes to one and forming a single three-lane roadway... in less than 600 feet. And the lane that both lanes of I-76 cram into is dropped at the next exit, another 2/3 mile downstream. WTF, indeed. Well, this is the Schuylkill Expressway.

Here's the view in the opposite direction. It has the same configuration but is easier to comprehend.

I don't understand why they don't at least stripe the two onramps down to one lane apiece before the convergence.

I remember noticing that a long time ago when I took an airport shuttle in Philly, and I definitely had an internal "WTF" take when I noticed two lanes suddenly merging in the middle of the other lanes.

[fwydriver405]

Reviving a dormant thread, but noticed this the last time I drove into Boston. I normally take Exit 63 instead of 66 coming from the north via I-95 to get to US Route 1, but took Exit 66 the last time around and noticed this:

Before the change, traffic coming from I-95 S Exit 66 had two thru lanes and traffic coming from US Route 1 SB merged with the exiting I-95 traffic. An example of the former configuration is here. Sometimes US 1 traffic would drive on the shoulder if traffic was heavy (and the fact this is a short merge), but that's a completly different problem.

Sometime in 2022, the configuration was changed so that the right lane coming from I-95 SB Exit 66 ends as US Route 1 traffic is merging (Overview, Streetview). This new configutation gives US Route 1 traffic an added lane instead of having to merge with exiting I-95 traffic. What caught me offguard was, although there were signs saying the right lane ended, I did not expect the right lane to end right at the gore point of merging US 1 traffic. Curious to know why the lane ends here instead of further back up the ramp before US Route 1 SB re-adds a lane onto the mainline.