Estimates of PM2.5 levels in the southeastern United States for the year 2010: What else can be done?

We developed a comprehensive air quality modeling system for use in integrated assessment studies. We evaluated the ability of the modeling system to reproduce observed PM2.5 levels in Class I areas of the Southern Appalachians Mountains using IMPROVE data during 9 week-long episodes between 1991 and 1995. The mean normalized error was less than 40% for the sulfate, ammonium, elemental carbon and organic components. The error was larger for nitrates and soils but these components constitute a small fraction of PM2.5 in the Southeast. Using meteorological inputs for the same episodes and emission projections, we estimated the PM2.5 levels in the year 2010. In addition, using the direct sensitivity analysis method, we estimated the sensitivity of PM2.5 levels to SO2, NOx and NH3 emissions from various sub-regions. These sensitivities give a first-order estimate of the responses to emissions controls beyond those already factored into the 2010 projections.