Abstract
Chronic exposure to fine particulate matter (PM2.5) has been linked to increased mortality rates. Chemical transport models (CTMs) are often used to evaluate PM2.5 regulations in the United States, but are computationally expensive and require significant technical expertise. Reduced-complexity models (RCMs) for PM2.5 are computationally-efficient air quality models designed to estimate ambient PM2.5 concentrations attributable to emissions of PM2.5 and PM2.5-precursors. We evaluate the ability of three widely-used RCMs (AP3, EASIUR, and InMAP) to reproduce changes in PM2.5 concentrations during 1990-2010. Compared to changes in observed, speciated PM2.5 concentrations, RCMs demonstrated strong correlation (R=0.61–0.66) and low mean error (0.84–1.03 μg m-3), performing comparably to a CTM (R=0.72, error=0.74 μg m-3). Most RCM applications assume that concentrations respond linearly to emissions; over the 20-year period investigated, during which most PM2.5 precursor emissions declined ~37–65%, we find only modest evidence of nonlinearities. These findings support the use of RCMs for many policy analysis and other applications for which the slight loss of accuracy compared to CTMs is tolerable, especially given RCM benefits. Common benefits of RCMs include increased computational efficiency and accessibility of air pollution modeling.

This publication has 40 references indexed in Scilit: