Environmental federalism when numbers matter more than size.

• Author: Adelman, David E.
• Position: II. Air Pollution in the United States as a Numbers Game C. Urban Density and Air Quality through V. Conclusions, with footnotes, p. 281-328

3. Urban Density and Air Quality

   Ambient levels of criteria pollutants vary dramatically across the country, but the highest levels occur disproportionately in large urban areas. (171) Figure 6 displays the counties in nonattainment for ozone and small particulate matter ([PM.sub.2.5]), which are the criteria pollutants that affect the largest number of people and are associated with the greatest cumulative health risks. The relatively small number of areas in nonattainment obscures the large size of the affected population--123 million people, or about 40 percent of the U.S. population, for ozone and 91 million, or about 30 percent of the U.S. population, for [PM.sub.2.5]. The aggregate population in nonattainment areas for all six criteria pollutants is almost 150 million, which represented about half of the U.S. population in 2005.

   The strong correlation between high ambient levels of criteria pollutants and urbanization explains the divergence between population and geographic area--almost every nonattainment area, aside from California's Central Valley, encompasses a major metropolitan center. The cities in California, Texas, and the northeastern states, which are home to many of the largest cities in the country, stand out as areas of poor air quality for criteria pollutants. A quantitative measure of this geographic pattern is provided in Figure 7, which displays the populations of nonattainment areas by state. This grouping of the data also underscores the extent to which elevated levels of criteria pollutants are concentrated in a few states.

   Ambient levels of air toxics, and the cancer risks associated with them, likewise vary across the country. Figure 8 displays the populations, at the census-tract level, (172) exposed to different levels of cancer risks from NATA Toxics in 2005. While about 90 percent of the U.S. population was exposed to excess cancer risks of 20 to 80 per million, the distribution has a long tail that extends above 200 per million. In this tail of the distribution, 16 million people were exposed to excess cancer risks above 100 per million.

   Most of the people who are exposed to these elevated cancer risks live in urban areas. In the ten largest cities, which were home to about 27 percent of the U.S. population in 2005, the average excess cancer risk from all outdoor sources of air toxics was 68 per million (by contrast, the cancer risks from industrial sources in these cities averaged about 2 per million). More importantly, the most severe risks from air pollution are concentrated in the largest cities: 88 percent of the 16 million people subjected to cancer risks above 100 per million--a level EPA deems clearly unacceptable--live in the ten largest metropolitan areas. (173)

   At the other end of the distribution, only about 40 thousand people lived in census tracts where cumulative excess cancer risks from NATA Toxics were below 10 per million. (174) Moreover, these estimates represent a lower bound on cancer risks, as they are based on a subset of all air pollutants. This result reveals that cancer risks from air toxics exceed EPA's target risk of one per million by tenfold even in very remote areas of the country. (175) An important implication of this finding is that background risks from air toxics, whether from natural or distant anthropogenic sources, are responsible for a baseline excess cancer risk of roughly 10 per million that cannot be influenced by local regulatory efforts.

   Figure 9 displays census-tract-level data on the excess cancer risks from air toxics. (176) While elevated cancer risks are closely associated with urban areas, the impacts from industrial sources and inter-jurisdictional transport of air toxics is evident in the rural areas of the southeastern states. However, unlike the largest metropolitan areas, the cancer risks in virtually all of these areas still fall within the range of 40 to 70 per million, making them comparable to the national average of 50 per million. These results further illustrate the significance of the background sources and risks noted above.

   Similar to the observations for criteria pollutants, California and New York stand out as hotspots for toxic air pollutants. Figure 10 displays the aggregate state populations of census tracts with excess cancer risks above 100 per million for NATA Toxics. (177) While the number of people affected is collectively much smaller than the population living in nonattainment areas nationally, (178) the geographic concentration is far more skewed towards large urban areas. The results expose the extreme geographic concentration of the populations most impacted by air toxics and the strong link to intense urbanization.

   An obvious implication of this analysis is that while a great majority of the land area in the country has relatively good air quality, a large fraction of the population does not benefit from it because about 84 percent of the U.S. population lives in urban areas. This finding is also consistent with the observation that about 50 percent of Americans live in areas not meeting one or more NAAQS and that 88 percent of the population subjected to the highest excess cancer risks from air toxics live in one of the ten largest U.S. cities. (179)

4. Outliers and Hotspots of Industrial Emissions

   Any discussion of industrial sources must be conditioned on the recognition that they typically generate a small fraction of the air pollutants emitted. As noted above, industrial sources account for about 13 percent of the air toxics emitted, (180) and for facilities other than coal-fired power plants, a lower percentage of criteria pollutants. Nevertheless, in certain jurisdictions they can dominate emissions of specific pollutants--chromium emissions from steel mills, for example, are significant in areas of the Midwest and south. (181) Outside these relatively rare instances, (182) however, emissions from industrial facilities are obscured by those from other source categories and therefore must be examined separately.

   The dominance of emissions from electric utilities is the most salient observation from Figure 11, which displays the percentages of emissions of criteria pollutants by industrial source category. With the exception of VOC emissions, (183) the skewed nature of industrial emissions extends beyond electric utilities. The top five industries emitted more than 90 percent of the emissions for each of the other criteria pollutants, and electric utilities and industrial boilers together accounted for 90 percent of the S[O.sub.2], 82 percent of the N[O.sub.x], and 63 percent of the [PM.sub.2.5] emitted nationally. Industrial emissions of air toxics are also skewed towards a subset of industries, with the top five industries accounting for more than 85 percent of the NATA Toxics emitted nationally in 2010.

   The geographic distribution of industrial emissions of criteria pollutants is divided between the flatter pattern observed for VOCs versus the skewed patterns found for S[O.sub.2], [PM.sub.2.5], and N[O.sub.x] (see Figure 12). (184) Emissions of the latter three pollutants essentially tracked the locations of their most prominent sources--coal-fired power plants. The top four states, for example, each have significant numbers of coal-fired power plants and among the largest plants in the country. (185) Thus, although industrial emissions of most criteria pollutants are dominated by electric utilities, outside the Midwest the sources are geographically spread out.

   Figure 13 depicts the largest industrial sources of NATA Toxics. It highlights the degree to which NATA Toxics are concentrated in the Midwestern and Southern states, as well as the small number of large industrial sources in the western and northeastern parts of the country. This map is almost a negative image of Figures 6 and 9, which display the nonattainment areas for ozone and [PM.sub.2.5] and the overall excess cancer risks (i.e., all source categories) from NATA Toxics, respectively. Figure 13 demonstrates that industrial emissions of air toxics are weakly correlated with high levels of air toxics in large urban areas.

   The regional disparities evident in the national data for industrial sources are amplified by the aggregate state-level figures (see Figure 14). Industrial sources of air toxics are clustered in a few states and counties, (186) but Texas stands out above all of the rest with more than double the emissions of second place Louisiana. Moreover, the distribution of industrial emissions has remained consistent over roughly the past two decades--Texas has accounted for about 15 percent of the NATA Toxics emitted nationally since 1988. The Houston area (Harris County), however, is an outlier even for Texas--county-wide emissions of NATA Toxics exceeded those in all but two states and equaled those of second-place Louisiana (see Figure 14). (187) are concentrated in 15 states, and more than 90 percent of industrial emissions occur in just 300 counties (ten percent of the number nationally).

   The geographic clustering of major industrial sources extends to the local level in other areas as well. In both Texas and Louisiana, industrial emissions of air toxics in three cities collectively accounted for about 60 percent of the statewide total, (189) and similar patterns, though less extreme, are observed in other states with high emissions from industrial sources. (190) Texas and Louisiana therefore had both the highest aggregate emissions from industrial sources in 2010 and the highest geographic clustering of the facilities responsible for them.

   The data on industrial sources are remarkable for their consistency. With the exception of VOCs, industrial emissions of criteria pollutants are dominated by coal-fired power plants, which are loosely centered in several Midwestern states. Similarly, industrial sources of air toxics are concentrated in a small number of states that, with a few exceptions...