Most of the world's cities started from an important marketplace or town square. Over time, they developed multiple centers where people could work, shop and play. But why? Some economists have suggested that cities fragment because of agglomeration—businesses that spring up in clusters increase their chances of success.
Yet physicists have arrived at a slightly different explanation: traffic jams. Marc Barthelemy and Rémi Louf, both at the Institute of Theoretical Physics in France, designed a mathematical model to explain how cities and their surrounding suburbs evolve. Their research suggests that as a city grows and congested roadways make it increasingly difficult to get to the center, subcenters emerge along the outskirts. “It's an interplay between how attractive the place is and how much time it takes to go there,” Barthelemy says. Cities with accommodating transportation networks remain centralized longer, he adds.
The physicists validated their ideas using data from 9,000 U.S. cities and towns of different sizes.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
A better understanding of how metropolitan areas evolve could prove useful, considering that two thirds of the world's population is expected to live in urban areas by 2050, notes David Levinson, a transportation engineer at the University of Minnesota. “There's a lot of urbanization left to happen,” Levinson says. “If planners imagine a city to take a particular form, but that's not the way the city wants to behave, we'll be making unwise investments.”
Barthelemy believes the model could also come in handy for estimating traffic delays, gas consumption and carbon dioxide emissions. “I think that this opens up the path to some really quantitative insights about cities,” he says. “We can take simple mechanisms, simple ingredients, and in the end predict how important properties are scaling with population.”