Air pollution due to high levels of O3 and PM2.5 has been a research topic for many regions worldwide. While we have gained relatively good understanding on their spatial and temporal distributions and formation mechanisms, the relative importance of atmospheric processes such as emissions, photochemistry, microphysics, and transport in their formation and fate has not yet been quantified due to the lack of an adequate tool for such a process analysis. Process analysis (PA) imbedded in the U.S. EPA Models-3/Community Multiscale Air Quality (CMAQ) modeling system provides a valuable tool for an in-depth understanding of atmospheric processes and chemical reactions as well as their relative importance to the formation and fate of air pollutants. In this work, a full-year simulation with CMAQ has been conducted for 2001 at a 36-km grid resolution over the contiguous U.S., southern Canada, and northern Mexico. Model predictions of chemical concentrations and radiative properties such as aerosol optical depth (AOD) are evaluated against surface and satellite measurements. Controlling processes are identified and quantified for major air pollutants. The net export from the planetary boundary layer (defined as 0-2.9 km in this study) to the free troposphere is estimated based on PA outputs.

Our results have shown that CMAQ model performance in terms of surface species and column CO and NO2 predictions is reasonably good; larger biases exist in its column predictions for tropospheric O3 residuals and AOD. Vertical transport, chemistry, aqueous processes, and dry deposition are the major processes contributing to the formation and fate of total odd oxygen (Ox), and those for PM2.5 include emissions, vertical transport, PM processes, and aqueous processes. Export of Ox from the continental U.S. occurs only in summer and at a rate of 0.66 Gmol day-1. In contrast, export of PM2.5 occurs for all seasons and at rates of 29.32, 34.18, 26.56, and 25.68 Gg day-1 for winter, spring, summer, and fall, respectively. These results indicate a need to use a different control strategy for O3 and PM2.5, namely, monitor and control PM2.5 throughout the year, rather than for high O3 seasons (May-September). Simulations have also been conducted to examine the sensitivity of CMAQ predictions to model treatments (e.g., cloud chemistry) and inputs (e.g., volatile organic compound emissions). Under future emission scenarios, surface O3 mixing ratios will increase in summer, surface PM2.5 concentrations may increase or decrease, depending on the emission scenarios and the geographical regions. Under future climate conditions with 0.7 ˚C increase in surface temperatures, surface O3 mixing ratios will increase in summer and surface PM2.5 concentrations will decrease in the eastern U.S. but increase in the western U.S.