Well it turns out that the supporting research data was slightly misinterpreted. According to the Per Square Mile blog, the UC Berkeley study associated with the interactive map reveals a more nuanced relationship between population density and carbon emissions. It turns out that people who live in the middle of nowhere (rural residents) actually have fairly low carbon footprints. Even though they’re reliant on cars, they tend to drive and consume relatively little.
And so initially, as population densities increase, so do carbon footprints. That is until it reaches about 3,000 people per square mile. At that point, carbon emissions start to drop off dramatically—roughly 35% on average from suburb to city.
The typical way to measure carbon emissions is to think about it in terms of geography. You pick a particular place, such as a country or a city. You add up all the emissions that are taking place within its boundaries. And you're then left with a territorial carbon footprint. If you've done any research on carbon emissions or climate change, you've likely encountered this method of accounting for carbon.
But there's a flaw with this logic.
The problem with this method is that it considers each geography to being more or less independent. For example, let's say you live in Philadelphia and you happen to be the owner of something called a computer. With territorial accounting, the carbon emissions associated with you powering your computer would get attributed to Philadelphia and the emissions associated with the actual production of the computer would get attributed to wherever it was made. Let's say it was China.
One of the problems with this approach is that it penalizes the places that make a lot of stuff and it privileges the places that don't make as much stuff, even if they may actually be the consumers of far more stuff. This might make you feel better about your life decisions if you happen to live in a dense urban knowledge economy that doesn't really make anything physical -- but is it entirely accurate?
An alternative measurement approach is consumption-based carbon accounting. The goal here is to capture all lifecycle emissions associated with a particular good or service, and then attribute it back to the consumer that arguably triggered the emissions. In the case of our Philadelphia computer example, the emissions associated with the production, transportation, and consumption of the computer would also get attributed locally to Philadelphia, instead of to China.
This more complex method of carbon accounting -- which is something that the University of Pennsylvania has been working on over here (hence the Philadelphia computer example) -- can be instructive for a whole host of reasons. It also has some relevance to city building.
It is widely understood that building up is more sustainable than building out. Because when you build out, you end up doing things like forcing people into cars. But the other side of this equation is that cities tend to also house a lot of rich people, and household wealth is a massive driver of carbon emissions when you account for them based on consumption. Some would argue it is more important than urban density.
In my opinion, none of this is to suggest that dense urban environments are bad. The point here is that territorial carbon emissions don't fully capture the emissions caused by high consumers who might happen to live in an otherwise efficient urban environment. You can live in a compact apartment and walk to work, but what else are you consuming? And how might these consumption patterns change based on built form?
For more on this topic, check out this report by Daniel Cohen and Kevin Ummel (of the University of Pennsylvania) called, "The case for neighborhood-level carbon footprints."
So here’s the big takeaway. If you’re looking to optimize around your carbon footprint, you need to pick a side: Either be urban or be rural. But don’t be somewhere in the middle. Don’t be suburban.
