The physicist Geoffrey West and his collaborators famously found that cities obey some universal scaling laws in the same way that much of biology does. Take the population of a city anywhere in the world, and these laws can predict with startling accuracy the size of the city’s road network or its output in patents or its quantity of crime. This is because so much about urban life scales to population. The larger a city gets, the less infrastructure it needs per capita. At the same time, as populations increase, problems like crime and benefits like intellectual property proliferate at an even greater rate than one-to-one (when a city doubles in size, for instance, it produces more than twice the patents).
In these reliable, mathematical ways, cities mimic nature. And yet, ironically, there is one element of cities that doesn’t seem to follow any scaling law: nature itself.
Nature – and it’s an admittedly slippery term that you can define multiple ways – is spread within and across cities with wild variability. Some small cities have lots of public parks, or lots of greenery as seen from satellites, while some larger cities have notably less of it. There’s no neat relationship between city population and parks per capita, or tree cover or even space that’s simply unpaved.
"Whenever you see that great variation in data, it begs for analysis and possible explanation," says Jon Christensen, an adjunct assistant professor in the history department at UCLA.
He and collaborators at Stanford began studying the question using a large set of updated data from the Trust for Public Land’s ParkScore project, which looked at park metrics in America’s 40 largest cities. Christensen and Stanford digital humanities research developer Karl Grossner are presenting some of the findings this week at the annual meeting of the American Association of Geographers. More relevant for the rest of us, they’ve also created an intricate, interactive public exploration of the question online.
The project, City Nature, offers comparisons of urban nature across cities, as well as an analysis of 2,661 "naturehoods," or neighborhoods within 34 cities where you can contrast the presence of greenery against home values, population density, and racial diversity. The project also contains a textual analysis of planning documents for mention of nature and greenery.
Why, though, does the question of scale matter? In his book The Triumph of Cities, Edward Glaeser argued that the conservation and preservation of open space in cities gets in the way of their economic growth (among other things, driving up the cost of scarce housing in places like San Francisco). Some of Christensen’s earlier research in Silicon Valley suggested this wasn’t true. “I thought,” Christensen says, “there’s something not quite right about this theory of the city, of nature being an impediment to the efficient economic development of cities for people.”
But if there’s no clear relationship between the presence of nature in cities and their population, it’s hard to make the argument that nature impedes development. To illustrate this, using City Nature’s tools, this is a chart of park area in 38 large American cities compared to population:
That big teal circle at far right is New York City (the dots are colored by geographic region and sized by population). The little lime-green outlier up top is Jacksonville. Chicago, a city with a seemingly deep history of park planning dating to Daniel Burnham, is that large blue circle with more than 2 million people but less than 20,000 acres of park area.
Here is a comparison of ParkScores from the Trust for Public Land with the land area (in square miles) of cities:
Christensen and Grossner have tried to come at the question from multiple angles, assessing numerous variables (weather, geographic region, even the historic influence of the Olmsted Brothers landscape design firm) and multiple measures of nature ("greenness" and "unpavedness" in addition to the presence of parks).
"There doesn’t seem to be any one factor that can explain that variation," Christensen says. Park area and total land area correlate somewhat, although it isn't surprising that as cities would grow in physical size, so, too would the physical mass of parkland.
Patterns similarly disappear in comparing the presence of nature with home values or home-ownership down to the neighborhood scale. "There’s weak correlation at best between any of these social variables," Grossner says, "and the degree of greenness and/or ‘pavedness’ and/or park need, and/or percent park."
All of this essentially means that while no hard rules or correlations govern the presence of nature in cities, historians have to look instead for the very specific ways that individual communities prioritized and planned for its inclusion throughout history. And this is a much bigger task than the 40-city statistical analysis.
"That’s the way in which this broadly comparative project also becomes a historical project to try to understand what it is that explains this variation," Christensen says. "What are the choices that have been made in different cities, and how has that shaped cities?"