When Lanes Reduce: Which Maneuver is Safer?

[nwi_navigator_1181]

Good day.

This was just a random thought that crossed my mind today; when lanes end on an interstate, which maneuver is safer: having the left lane end (making drivers merge right), or having the right lane end (with the higher likelihood of turning that lane into "EXIT ONLY" when near an interchange - regardless, drivers merge left)?

I was just curious because before the I-80/94 expansion project ended in Lake Station, IN, a "Lane Ends" sign on the right indicated the plan was to have drivers merge left. However, when striping was completed, the LEFT lane ended, so the signage changed accordingly.

As always, I thank you for your responses.

[hobsini2]

I have always believed that a lane ending from the right is the best situation because usually the left lane on the interstate is more likely to have quicker traffic. The downside to this though is having to merge in front of semis because most states have limitations as to which lane the trucks can be in.

[Truvelo]

Over here it's the opposite. The fast lane always ends with the slow lane continuing. This also applies to two lane roads where there's a passing lane up a hill. Personally I think our way is the best as it's up to the driver who's passing to complete the maneuver before the lane ends. The American way forces slower traffic into the left lane, often well before the right lane ends if there's a right lane ends 1/2 mile sign.

[1995hoo]

Then you have the setup on the southbound New Jersey Turnpike at the southern end of the quad-carriageway setup. After the carriageways join, the two center lanes merge into a single lane.

A similar setup exists on the outer loop of the Belt Parkway in Brooklyn (heading towards Coney Island) where the ramp from the Verrazano Bridge enters from the right–the left lane of the two-lane ramp and the right lane of the two-lane Belt Parkway become a single lane and the road then continues on with three lanes.

I agree with Truvelo's point about how people get over insanely early, but I suppose the mechanics of merging are a whole separate discussion (and one that often provokes heated arguments).


New Jersey Turnpike picture from AARoads.com. The pickup and the Toyota will wind up in the same lane as each other when they pass under the next gantry unless one of them crosses a solid line to change lanes.

[kphoger]

I believe Sweden used to have the fast lane merge into the slow lane, but has since switched to the other way around.  J N Winkler, can you confirm?

The old way had problems when there was a central barrier; if traffic didn't merge in time, they got pinched into the barrier.  It was determined to be better to be pinched onto the outer shoulder than the central barrier.  Even where there's no barrier, there still might be opposing traffic.

[Alps]

What I was taught in traffic engineering is that it's always safest to have the left lane end if you have sufficient distance. Next is to have the right lane end. You try to avoid EXIT ONLY situations because people will be trying to cross the line in both directions to either take or avoid the exit. If you need a lane drop, do it just after the exit, or use lane balance by having a second lane take the exit as well. (In other words, if you have a 3-lane section, split it 2/2 rather than 2/1.) The center lane merge used by the Turnpike would seem to be the least desirable option, but merging 6 lanes into 3 is a unique case to begin with.

[Brian556]

In Dallas, SB I-35E at SH 183, the right lane of I-35E and the left lane of SH 183 merge together. Very unsafe.
http://maps.google.com/maps?hl=en&ll=32.814856,-96.86716&spn=0.000009,0.006266&t=m&z=18&layer=c&cbll=32.814856,-96.86716&panoid=HAXfOSd5pK2T6GIMfvUFoQ&cbp=12,159.6,,0,0

On I-35E in Corinth, when they added a third lane, originally the left lane ended. They quickly cjanged it so that the right lane ended. I figured that engineers determined that it would be safer that way.

I've noticed that in several states when they need to close the left lane, they close the left lane, then shift traffic over. This would supporthe same theory as above. Intesesting that Steve was taught the opposite in school.

[mcdonaat]

I think that it is safer to end the right lane, and have traffic merge left. If possible, create an exit only lane, since at least the lane ending could be useful.

[Alps]

I was taught actually in the industry. School only taught people how to become professors, which didn't interest me. Anyway, the idea is that traffic is moving faster AND more uniformly in the left lane, which makes it a safer maneuver to interlock them. The right lane has a lot of varying speeds going on. Also it had something to do with being able to see more over your right shoulder than your left, I believe, or something with blind spots, that made it an inherently safer maneuver to merge right.

[1995hoo]

What I was taught in traffic engineering is that it's always safest to have the left lane end if you have sufficient distance. Next is to have the right lane end. You try to avoid EXIT ONLY situations because people will be trying to cross the line in both directions to either take or avoid the exit. If you need a lane drop, do it just after the exit, or use lane balance by having a second lane take the exit as well. (In other words, if you have a 3-lane section, split it 2/2 rather than 2/1.) The center lane merge used by the Turnpike would seem to be the least desirable option, but merging 6 lanes into 3 is a unique case to begin with.

I read somewhere, may even have been this forum, that the Turnpike Authority prefers the center lane merge because they don't want to set it up such that all, or substantially all, of the lanes on one carriageway end while the other's lanes continue, for fear of increasing the congestion on the "ending" carriageway. Of course, I suppose one mitigating factor with the current design is that it seems like every time I go through there other than very late at night the area is already so congested that the center lane merge isn't a "hazard" so much as a "nuisance" because everybody's going so slowly that the chances for a serious crash are very low!

[PurdueBill]

In Dallas, SB I-35E at SH 183, the right lane of I-35E and the left lane of SH 183 merge together. Very unsafe.
http://maps.google.com/maps?hl=en&ll=32.814856,-96.86716&spn=0.000009,0.006266&t=m&z=18&layer=c&cbll=32.814856,-96.86716&panoid=HAXfOSd5pK2T6GIMfvUFoQ&cbp=12,159.6,,0,0

On I-35E in Corinth, when they added a third lane, originally the left lane ended. They quickly cjanged it so that the right lane ended. I figured that engineers determined that it would be safer that way.

I've noticed that in several states when they need to close the left lane, they close the left lane, then shift traffic over. This would supporthe same theory as above. Intesesting that Steve was taught the opposite in school.

The 35E-183 thing is an interchange thing as opposed to a single road losing a lane--the inner lanes merging isn't all that uncommon and happens in places (e.g., both I-271/480 merges of 2+2=3, I-76/OH 59 with 3+2=4) but there are counterexamples like I-670/I-70 where they end the right lane of 670 before it would merge into 70's left lane so there is no merge.

Most of the situations in Ohio where a 3-lane section becomes a 2-lane section of Interstate involve the left lane ending, but that may also be a factor of convenience in layout since the left lane is usually the one that was added in the widening so the carriageway tapers that way.  On I-77 NB south of Akron, the road was restriped from a left lane ending to a right lane ending--this probably makes more sense overall since the right lane is added from a ramp south of there.

[vdeane]

Then you have the setup on the southbound New Jersey Turnpike at the southern end of the quad-carriageway setup. After the carriageways join, the two center lanes merge into a single lane.

The same thing happens when I-88 west merges to I-81 south.  The I-81 right lane and I-88 left lane merge into a single lane.  The I-88 right lane ultimately becomes and exit only lane for NY 17.

[TheStranger]

Then you have the setup on the southbound New Jersey Turnpike at the southern end of the quad-carriageway setup. After the carriageways join, the two center lanes merge into a single lane.

The same thing happens when I-88 west merges to I-81 south.  The I-81 right lane and I-88 left lane merge into a single lane.  The I-88 right lane ultimately becomes and exit only lane for NY 17.

Similar phenomenon in SF on the short merge between I-80 east and the ramp coming off of southbound US 101 (eastbound Central Freeway) -

https://maps.google.com/maps?q=Division+St,+San+Francisco,+CA&hl=en&ll=37.770886,-122.405596&spn=0.000946,0.000848&sll=38.549791,-121.393726&sspn=0.010589,0.013561&t=k&hnear=Division+St,+San+Francisco,+California&z=20

Prior to 1968, the left two lanes were part of I-80 east as well (when I-80 continued to Fell Street along the Central Freeway and was slated to be extended westward to Golden Gate Park), so the right lane of eastbound I-80 then immediately merged with the left lane of what was then the NB 101 to EB 80 connector (but which is now essentially the current start of 80's main lanes).

[deathtopumpkins]

The center lane merge happens with I-95 northbound where the ramps from US 1 in Peabody merge on. The right lane of 95 and the left lane of the US 1 onramp merge together, and the right lane of the US 1 ramp becomes the right lane of 95 north. This happens over a much longer distance than on the NJTP though, and involves no solid lines. Basically the lanes are next to each other, the line between them ends, and you have a super-wide lane for easily a quarter to a half of a mile that gradually narrows. Works wonderfully.

[jeffandnicole]

In keeping with the "Keep Right Except to Pass", the left lane should be the one that ends, as traffic really shouldn't be in that lane in the first place unless they're passing.  Since the lane is coming to an end, if a driver doesn't have time to pass, then they should just remain in the thru lane.

On the NJ Turnpike, the other theory is that traffic doesn't need to merge more than once at the end of the dual/dual roadway.  On a 3-3 lane set up, the inner left/center lanes merge, the inner right/outer left lanes merge, and the outer center/right lanes merge.

In this document: http://www.njturnpikewidening.com/documents/01_Chapter1.pdf , page 48 reflects the merge pattern (not to scale though).  Page 49 reflects the diverge pattern.

[Mr_Northside]

Also it had something to do with being able to see more over your right shoulder than your left, I believe, or something with blind spots, that made it an inherently safer maneuver to merge right.

Hmmm... I was actually thinking along those lines, just with the opposite lane.  Maybe it's just the majority of merging I do (and, I'd imagine, a lot of people) is ramps and such where you're used to checking for traffic on your left.  So it's a more "routine" procedure.(?)

Though, I'd have to say, as long as you're paying attention, everyone is taking their turn (if traffic is that thick), and your mirrors aren't adjusted like shit, the difference between which side merges should be very, very negligible in terms of safety.

[Duke87]

Another consideration is where the lane began. The PTC seems to have a habit of adding an auxiliary lane to the right, and then subtracting it from the left. This to me seems both undesirable and annoying since it means that a car which starts in the right lane and then stays in it when it becomes the center lane (perfectly logical when you create a spare right lane for, say, truck climbing) will after the end of the section end up in the left lane, and then have to merge right in order to go back to being in the right lane. It also makes speedy traffic cruising in the left lane needlessly have to merge right. Seems like you're creating a sort of lesser TOTSO situation, where everyone has to change lanes in order to stay in their lane.

[Alps]

I think center lane merges are something different than center lane reductions. The Turnpike is a unique case because both roads merge together - it's not like you can pick one and call it the mainline. (Well, you can, but it's a 50/50 proposition.) In most cases, that's really not a lane reduction, if anything it's a lane addition coming off the ramp.

[J N Winkler]

I believe Sweden used to have the fast lane merge into the slow lane, but has since switched to the other way around.  J N Winkler, can you confirm?

The old way had problems when there was a central barrier; if traffic didn't merge in time, they got pinched into the barrier.  It was determined to be better to be pinched onto the outer shoulder than the central barrier.  Even where there's no barrier, there still might be opposing traffic.

I cannot confirm.  I checked the drawings for Trafikverket object number 87833258 (mentioned in a Central States thread dealing with US 54 in Missouri) and it shows the fast lane merging into the slow lane.  This is a recent project (advertised around May 2012), so if design guidance has indeed changed as described, it would not seem to be applied in an uniform fashion.  (Since this project is a S2+1 upgrade on the existing alignment, I looked at it in StreetView to see if there were any special conditions that might warrant an unusual merging treatment.  There did not seem to be any; in fact the existing road is a rural two-lane with full hard shoulders for much of its length, and is reminiscent of US 54 in western Kansas.)

In regard to the question of fast into slow versus slow into fast, I am not impressed with the arguments offered on either side, since they rely on theories of driver behavior that are difficult to put to a rigorous empirical test.  The hard evidence (comparative analysis of crash experience and the like) on this point seems to be a lot thinner than for other design criteria, such as the 12' unit lane width for high-capacity rural roads.  For either type of treatment I suspect the crash experience is highly sensitive to small aspects of detail, such as how the taper is signed and striped.

[Crazy Volvo Guy]

Here in The Land of No Lane Discipline™, it should always be the right lane that ends on a freeway, never the left.

[deathtopumpkins]

Here in The Land of No Lane Discipline™, it should always be the right lane that ends on a freeway, never the left.

Yet you fail to give any reason why. Or any credentials that lend any authority to your statement that 100% of the time the right lane should end.

Fixed format error

[roadman65]

Here in The Land of No Lane Discipline™, it should always be the right lane that ends on a freeway, never the left.

It used to be that way on the Pennsylvania Turnpike climbing Sideling Hill.  At the end of the truck climb lane going both directions, the extra lane on the right did end.  Then later they made it so that the left lane ended at the top of the hills.

I am guessing this was done because, prior to the change there were signs that told merging truckers to watch out for overtaking cars and for cars to watch out for slow merging trucks.   To have both signs contridict each other  suggests that there were problems with the traffic flow between cars and trucks where the lanes narrowed down.  It must of been where the PTC felt that this was a rare exception to allow the left lane to end as this definitely does end the conflict between the slow moving trucks and the fast moving cars.  The problem with left lane traffic moving over to the right is not that much or not at all it is worth it.

[kphoger]

Here in The Land of No Lane Discipline™, it should always be the right lane that ends on a freeway, never the left.

It used to be that way on the Pennsylvania Turnpike climbing Sideling Hill.  At the end of the truck climb lane going both directions, the extra lane on the right did end.  Then later they made it so that the left lane ended at the top of the hills.

I am guessing this was done because, prior to the change there were signs that told merging truckers to watch out for overtaking cars and for cars to watch out for slow merging trucks.   To have both signs contridict each other  suggests that there were problems with the traffic flow between cars and trucks where the lanes narrowed down.  It must of been where the PTC felt that this was a rare exception to allow the left lane to end as this definitely does end the conflict between the slow moving trucks and the fast moving cars.  The problem with left lane traffic moving over to the right is not that much or not at all it is worth it.

This makes sense on a climbing lane, and I hadn't thought of that.  It seems to me that the speed differential would be greater between rightmost and center lane than between leftmost and center lane; therefore, merging should be more easily done on the left than on the right.  Hmmm....

[Crazy Volvo Guy]

Yet you fail to give any reason why. Or any credentials that lend any authority to your statement that 100% of the time the right lane should end.

Should be fairly obvious by the way I worded my post, specifically words 3-8.

The above mentioned situation about a climbing lane is about the only exception I can think of, as jamming trucks moving 25-40mph into 65+mph traffic is downright deadly.

In Europe, where people actually understand the concept of the left/inside lane, sure, the left/inside lane ending is more logical; but not here.

[deathtopumpkins]

Land of No Lane Discipline™, it

Words 3-8. My point still stands. Where is this Land of No Lane Discipline? Why is it trademarked? What makes it the indicator of ideal setups on highways nationwide? So many questions!