(Photo: Atlantacitizen via Wikimedia commons)
A new paper out of the University of Michigan revisits an old debate we’ve covered here: trends in vehicle travel in rural versus urban areas. The paper's abstract reports (presumably based on data from the Federal Highway Administration) that vehicle travel in urban areas increased by 33 percent between 2000 and 2016, while rural vehicle travel decreased by 12 percent.
These numbers sound compelling, but as we’ve noted before, the definitions of “rural” and “urban” shift over time as metropolitan areas grow, and those changes are applied inconsistently by the states (and FHWA) to vehicle travel data. Some of the increase in urban driving, therefore, is due to honest-to-goodness changes in the operation of our cities and of individual behavior, but some has to do with reclassifying driving that had previously been “rural” as “urban” when the boundaries of urbanized areas change.
The problem is that it is hard to know exactly how much of the change in recent years is real versus the result of boundary shifts. To ascertain that, you would need to map data on highway travel volume onto urban area boundaries of different vintages.
We attempted something close to this in research for our 2016 paper, A New Way Forward: Envisioning a Transportation System without Carbon Pollution. In that paper, we estimated transportation carbon dioxide emissions for the largest urbanized areas across the United States, based on GIS analysis of data from the Database of Road Transportation Emissions (DARTE). Out of curiosity, as a side project, we ran the numbers for 74 urbanized areas based on their boundaries from the 1990, 2000 and 2010 Censuses.
In 1990, transportation carbon dioxide emissions from those 74 urbanized areas – based on their 1990 boundaries – totaled 435 million metric tons. In that same year, emissions occurring within those areas’ expanded 2000 boundaries totaled 469 million metric tons – an addition of 8 percent. If we had used the 2010 boundaries, emissions in 1990 would have been 488 million metric tons, a further addition of 4 percent.
Even in 1990, therefore, the amount of driving taking place on “rural” roads adjacent to urbanized areas was enough to produce significant amounts of carbon dioxide. Shifting emissions from this preexisting driving from the “rural” to the “urban” category when urbanized area boundaries are redrawn is enough to move the needle on overall emissions from that urbanized area.
It is an open question whether adding preexisting rural driving and rural populations would be enough to skew per-capita VMT comparisons. That question is more difficult and time-consuming to answer.
There is little reason to doubt that urban driving has increased faster than rural driving over the last 15 to 25 years, based on patterns of population growth and the continued sprawling of many U.S. metropolitan areas. But the FHWA data alone are not enough to tell us how much urban driving is increasing relative to rural driving, and analysts should be aware of the shortcomings of these data.
Ultimately, though, the fault is not with researchers and analysts who are trying their best to understand the changes occurring in transportation behaviors given the information at hand, but rather with state Departments of Transportation and the FHWA for putting out information that makes such analysis more difficult and less conclusive than it should be. There are many important questions to be asked about transportation trends. We need better data if we want to answer them.