In several previous posts, I summarized the results of 12 GHG inventories that have been conducted in Missouri. In this post I will discuss what I think they mean.
The first graph shows per capita GHG emissions for each of the 12 jurisdictions. For most jurisdictions, GHG emissions range between 20 and 30 MTCO2e per capita. However, there are some exceptions: Creve Coeur has significantly higher per capita emissions, while Maplewood and Wildwood have significantly lower per capita emissions. These exceptions show that population is not the only driver of GHG emissions.
The profile varies between Missouri communities. The commercial sector is the most important sector driving this variation. Clayton and Creve Coeur have particularly large per capita commercial emissions – more than twice as large as Richmond Heights, the third largest. In fact, their commercial emissions are larger than total emissions in Wildwood, a community with more than twice their population. Clayton and Creve Coeur are the two communities that have the largest increase in daytime population. In fact, their percentage increases in daytime population were the second and third largest in the state in 2005, trailing only Fenton. (Fenton was the home of the Chrysler Assembly Plant, which has since closed.) Wildwood and Lee’s Summit on the other hand, lose a significant fraction of their population during the day.
Creve Coeur and Richmond Heights have the largest per capita transportation emissions. A couple of factors are at work here. Both communities are crossed by significant highways and roads, and this increases emissions. But in addition, both are destination communities with significant business and retail districts. Many people who live in other communities do lots of driving in these two communities, especially compared to the size of their residential population.
Creve Coeur, Columbia and St. Louis County have the largest per capita residential emissions, and the pattern between jurisdictions is somewhat unexpected. Per capita residential emissions are probably a function of the number of people living in each home, building energy efficiency, building size, and the amount of electronic equipment used therein. It would seem that Creve Coeur might be the location of some large residences with lots of electronic equipment, but so might Clayton and Wildwood. Alternatively, one might expect that Creve Coeur, with an aging population, might have relatively few people per residential building. But so might Clayton, and one might expect Columbia to be the opposite, with its housing for Mizzou students. One might expect that Creve Coeur, with its ranch style houses constructed in the 1960s, has an energy inefficient housing stock. But the housing stock in Clayton is probably as big or bigger and significantly older. This is an interesting finding that calls out for further understanding, however it seems to be a less important driver of per capita emissions than are commercial and transportation emissions.
The second graph shows per capita emissions graphed against daytime change in population. In this graph, the per capita computation is made using residential population, and the daytime population is shown as the percentage change in population during the day. The data points do not form a perfect line, but they show a definite trend from lower left to upper right. This suggests a positive relationship between the two variables.
Population is the primary driver of GHG emissions. If you compute a regression of total emissions on population for the 11 cities and counties (a statistical technique that yields an overall estimate of how much emissions depend on population), you find that population almost entirely predicts emissions (R-squared = .997). But that analysis is distorted by the large-scale differences between communities: Kansas City and St. Louis have 35 times the population of Maplewood and Richmond Heights. You’d be surprised if there wasn’t a large difference in total community emissions. That’s why the per capita analysis is so important.
Does population predict per capita emissions? If it did, then we would be expecting that as communities grew larger, it would reliably affect the amount of GHG each resident emitted. Some might argue, for instance, that large communities are inherently more (or less) energy efficient. For these 11 communities, the answer is “no.” Population does a very poor job of predicting per capita emissions (R-squared = .003). What, then, does predict per capita emissions?
Daytime population change is a large part of the answer. If one conducts a regression, one finds that daytime population change is a strong predictor of per capita emissions (R-squared = .664). To host a large daytime population increase, a community has to have lots of commercial buildings to house workers and shoppers. These buildings all cause GHG emissions. In addition, people drive to and from these buildings, emitting GHGs inside the community boundaries as they do so.
An R-squared of .664 means that about 2/3 of the variation in per capita emissions can be explained by the daytime population change. The remaining 1/3 is left unexplained. We don’t really know what accounts for that remaining 1/3, although we might suspect it could involve factors such as the energy requirements of the businesses in the community, and the energy efficiency of the building stock. Other factors might include the energy practices of the residents (like driving less or using high mpg vehicles). We have no evidence, however, that large differences in energy practices exist in the communities studied, and until such evidence emerges, I would be skeptical.
It appears that people emit GHG emissions. Where there are lots of people, there are lots of emissions. If they leave their community to work or shop in another community, then they take some of their GHG emissions with them to that other community. It is unlikely that differences in GHG emissions between communities should be attributed to community energy practices. At least, not yet.
Well, I’ve just offered an opinion, not something I do often in this blog. Do you agree? Also, I’ve reported on all the GHG inventories I know about in Missouri. If you know of one that I’ve missed, let me know.
For links to the GHG inventories from the 12 jurisdictions, see my previous posts on each one. Also, see the 4 previous posts in this summary series. Here’s a link to a list of all previous posts.
Missouri population is from Part 1. Population of the United States and Each State: 1790-1990, http://www.census.gov/population/www/censusdata/Population_Part1.xls.
County populations are from Table 1. Annual Estimates of the Resident Population for Counties of Missouri: April 1, 2000 to July 1, 2009, http://www.census.gov/popest/data/counties/totals/2009/tables/CO-EST2009-01-01.xls.
Municipal populations are from Table 4. Annual Estimates of the Resident Population for Incorporated Places in Missouri: April 1, 2000 to July 1, 2009, https://www.census.gov/popest/data/cities/totals/2009/SUB-EST2009-4.html.
Daytime population changes are from Daytime Population Changes in Missouri Counties and Selected Cities, Missouri Economic Research & Information Center, December, 2005, http://www.missourieconomy.org/pdfs/daytime_population.pdf.
Geographic areas are from the Wikipedia article for each location.
Regressions were computed using StatPlus Mac.