On a limited access highway without obstacles, the maximum traffic capacity (vehicles per hour) of two lanes (3600 VPH) is indeed twice that of one lane (1800 VPH). However, in high density traffic, the maximum capacity of two lanes merged into one lane (about 1200 VPH) is significantly less than that of one lane (without a merge), because late merging invariably forces stop and go driving. (You cannot legislate human behavior.)

At saturation, the maximum VPH of the entire stretch of highway between any two exits is determined entirely by the most restrictive bottleneck. “Using up” empty lanes does nothing to relieve the bottleneck itself.

When truckers manage to block enough cheaters that actual merging is eliminated at the point of lane closure, the line that has already merged can and does proceed much faster (back as high as 1800 VPH), but of course only until the blocking truck (or car) arrives at the point of lane closure. Of course, some cheats are so self centered that they will even drive on the grassy median to get around a blocker. I have pictures!

When the rate of vehicles entering a segment of highway is greater than the capacity of remaining lanes adjacent to a closed lane up ahead, the most efficient thing to do would be to move the point of lane closure back as far as necessary, to a point where traffic is not (yet) at saturation and can merge without forcing stop and go driving in the process.

But wouldn’t that just force the solution some people feel is “right”?

Suppose instead we just turn on the green for the continuing lane and leave the merging lane red. On the condition that everyone obeys the lights — that is, the merge lane would sit there and patiently for the light that would never come — wouldn’t the traffic flow equal the maximum possible rate determined by up-stream issues?

Only works if you have 2 lanes to 1. 3 lanes to 2 is hard because there is no easy place to put the light for the middle lane that isn’t a hazard when things are not stop and go, unless you can string them overhead.

Power is of course an issue.

For speed, it would be best to do a 2 cars per green metering light, though people seem to have trouble understanding those, even today when they have been around for a while at highway interchanges.

It helps that no line has priority. And it all happens at pedestrian speeds. This is different from the typical “lane ends, merge left” road design.

A related issue to the vitriol toward late mergers is the vitriol toward motorcyclists who “lane-split” or “filter” through traffic. As you probably know, it’s legal in California (and nearly everywhere else outside the U.S.).

I live (and ride) on the East Coast, so I don’t lane split but I’ve heard many motorcyclists report hostility from drivers and traffic cops for doing it, even where it’s legal. Apparently a lot of them feel like it’s some sort of unethical form of “cheating traffic.”

Also, have you seen any studies that consider whether it’s safer for motorbikes to lane-split vs. sit in the lane? This is a constant source of debate, and many riders feel more vulnerable to being rear-ended while sitting in a lane than filtering through traffic like they do in Europe.

And really…you should do more blog posts on motorcycles/scooters. We’re a segment of the roadgoing public that is VERY attuned to traffic issues, and especially the hassles of dealing with “cagers” who seem to resent us (even though we take up such a small footprint on the road).

Cheers!
Joel

And if I was to generalize I would say most traffic inefficiencies are a result of inadequate road sign information. And with GPS systems I feel this is becoming even worse. So when the GPS system says they will bear left in 500 yards drivers will cram all the way over to the left and not realize the left 2 or 3 lanes will suffice. Better road signs could educate drivers of optimal lane choices and better GPS systems could also make a difference.

Late merge provides the best flow (in a highway with too much traffic to avoid collapse to stop and go) when you don’t have a minority or majority, when they are split roughly evenly.

I think this is a very important thing to point out when preaching late merge. People can see why late merge is fair if it’s 50-50, but they will obviously see it as unfair if it gets uneven.

Perhaps the right approach (nudge) is to give it a better name which implies more fairness. “Late merge” implies, to most people, that somebody is merging later than somebody else, which is unfair, even though the real meaning is that the merge occurs late in the merge zone rather than early.

So another name — “joinpoint merge,” “confluence merge”, “alternating merge”, “chokepoint merge” or something like it, might save the day.