Subway systems are marvels of engineering and design, even more so for the ones (like London’s or New York’s) that were built in the early 20th century. But still, any rider of a mass transit system has no doubt cursed its design at some point. Why doesn’t the train go here instead of there? Why should an express line reach this neighborhood and not that one?
Who’s to blame (or praise) for mass transit system design? According to a French physicist named Marc Barthelemy, nobody is. Barthelemy and his fellow researchers published a paper depicting the design of subway systems as an “emergent phenomenon” of large cities. “Each network is the product of hundreds of rational but uncoordinated decisions that take place over many years,” paraphrases Scientific American. Which means, according to Barthelemy et. al., that “you can forget about details” and watch how the systems evolve toward similar design patterns regardless of where, when, and how they were built.
The three patterns that any major metropolitan subway system (with more than 100 stations) should have in common are:
- A core and branches, with core stations arranged in a ring shape above the city center
- A number of branches that tends toward the square root of the total number of stations
- About 20% of core stations contain transfers to two or more other lines.
At first this seems obvious to the point of banality (except for the square root thing). But if Barthelemy is right, why does the New York City subway not have a “core,” and why does the Moscow Metro have abnormally long branches relative to its core size, as Scientific American points out? Barthelemy insists that outliers aside, his three emergent properties still apply more often than not.
Which poses an odd question about urban planning. If the design of these subway networks is “emergent” and converges on these patterns as it grows no matter what, what does that mean for the civil engineers and designers behind them? I asked David Levinson, a transportation engineer quoted in Scientific American‘s article, how designers could use this information (provided it’s valid). “While some macroscopic properties may be similar, the various systems are far from identical,” he says. “Knowing how the system is likely to develop–its natural tendency–is something that can be supported or fought against, with different results.”
In other words, just because subway networks behave according to rough rules, like rivers or deltas, it doesn’t mean that human designers can’t influence them. “If you want development to occur at X, building a segment to X will make a difference; X will be more likely to develop, some other place Y will be less likely to develop,” Levinson explains. The “emergent” properties of a subway system and the details of its design are interdependent. So while Barthelemy’s research may have illuminated some intriguing patterns in how mass transit systems tend to evolve, our subways–love them or hate them–don’t build themselves.