On Why Bike Lanes Might Appear Underutilized

EUMobilityWeek1We hear complaints that bike lanes are underutilized.

This might be true, or it might appear to be true and not be true.

Let’s think about the traffic physics of this.

Elements of Access: Transport Planning for Engineers, Transport Engineering for Planners. By David M. Levinson, Wes Marshall, Kay Axhausen.
Elements of Access: Transport Planning for Engineers, Transport Engineering for Planners. By David M. Levinson, Wes Marshall, Kay Axhausen.

Imagine there is a lane of traffic full of cars, say a flow (Q) of 500 vehicles per hour, going very slowly, an average speed (V) say 5 km/h because of congestion, traffic signals, unloading trucks, and the like. The lane will appear full, because the density is high. Density (K) is vehicles per km, and the relationship between flow, density, and speed is given by Q=KV or K=Q/V. In this example, the density is 100 vehicles per km, or about  10 vehicles every 100 m, or 10 m between vehicles, which is a pretty high density. Not quite jam density (minimum vehicle spacing, maximum vehicles per km, on the order of 150), but close.

Imagine there is a parallel lane for bicycles. They are traveling at 20 km/h. The spacing is one bicycle every 40 m, or a density of 25 bicycles per km. Yet the flow is an identical 500 vehicles per hour (Q=KV). The lane looks one fourth as full (even less, because bicycles take up much less space than cars), but serves just as many vehicles as the crowded lane.

Now of course the bike lane is narrower than the car lane, so if we were to look at bicycles per square meter, accounting for a car lane of 3 m (typically 3.65m, but narrower in cities) and a bike lane of 1.5 m, we only need a density of one bicycle every 80 m to get the same flux (flux is flow accounting for the width of the lanes and vehicles). One bicycle every 80 m is about 1 bicyclist per block at a given time. In contrast that congested  lane of cars has at least 8 vehicles in it for the same length block.

(I realize I should evaluate person throughput rather than vehicle throughput. For illustration, I am assuming 1 person per motor vehicle, which is a bit pessimistic, in practice it is closer to 1.1 for work trips and 1.5 all day).

Now, I am not saying the bike lane has 500 bicyclists per bike lane per hour (or the road has 500 vehicles per lane per hour). Most have fewer. Your kilometerage may vary.  It doesn’t have to. The alternative use of the lane may have been storing cars. They have a speed of zero and a flow of 0, and a pretty high density (roughly 150 vehicles/km) for being unproductive.

Furthermore performance in terms of flow (or flux) isn’t the only question at hand. Safety is important too, and bike lanes are typically safer for bicyclists than riding in traffic, and sure feel safer.

Autonomous vehicles and the shortest path

We have given a number of reasons that autonomous vehicles will reduce congestion.

“In particular, vehicle automation, once it gets critical mass, should greatly increase road capacity, both because of shorter following distances and because of narrower lanes. New, narrow vehicle forms designed for a single passenger (which is how most cars are used) will become more widespread with automation, as safety fears diminish, and will also increase person throughput.”

Example of Route Detecting and Comparison of Alternative Paths

I have thought of one more reason autonomous vehicles will reduce traffic congestion. Currently most drivers do not take the shortest path (though we don’t really know why that is so, we have speculations). If we take humans out of the navigation decision, cars will be more likely to find the shortest path. This may not be system optimal, but will be a significant improvement over current routing decisions.

Not only should we not let people drive, we probably shouldn’t let them route themselves if we want an efficient system.