emissions inventory, sector-specific growth assumptions, and source-specific assumptions regarding future CAA-mandated controls expected to be achieved by 2010.

In the second step, county-level baseline air-quality data for the continental United States were generated. For PM, a source-receptor matrix was first generated using the phase 2 climatological regional dispersion model (CRDM). Because the model was shown to overestimate the contribution of fugitive dust to fine PM, the source-receptor matrix was adjusted, and monitoring data were used to calibrate the matrix. Baseline annual mean PM10 and PM2.5 estimates for 2010 were then generated using the 2010 emissions data and the source-receptor matrix. PM estimates for nonmonitored counties were generated on the basis of the more complete data sets for the monitored counties. Peak-to-mean ratios were used to generate 24-hr averages. For ozone, a regional oxidant modeling (ROM) extrapolation method was used to generate county-level baseline air-quality data for ozone. Ozone air-quality monitoring data from 1990 and ROM air-quality modeling results for 2007 were used to generate ozone air-quality data for 2007. The data for 2007 were then extrapolated using 2010 emissions data and ozone modeling and monitoring data to give 2010 baseline ozone air-quality data. Data for nonmonitored counties were generated by interpolating data from surrounding monitored counties, assuming that the entire county population experienced the air pollution concentration estimated at the geographic center (or centroid) of the county.

In the third step, EPA used the PM and ozone baseline air-quality data to identify counties that would exceed the proposed or alternative standards. In the fourth step, EPA selected control strategies to implement in the nonattainment counties and then estimated the potential costs and economic impacts of the proposed and alternative standards.

In the fifth step, EPA estimated the post-control air-quality data on the basis of the control strategies selected in step four. For the partial-attainment scenario, EPA used the source-receptor matrix to estimate PM air-quality data and a quadratic rollback procedure to estimate ozone air-quality data. For the full-attainment scenario, a proportional and a quadratic rollback procedure were used to estimate PM and ozone air quality, respectively.2


Rollback procedures scale an exposure estimate by the changes modeled for the emissions estimates. Therefore, proportional rollback assumes that concentra-

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement