A Seasonal Modeling Study of Air Quality in Central California

Tropospheric ozone has been a continuing research and regulatory concern due to its adverse health effects and its importance in atmospheric chemistry. Photochemical air quality models integrate scientific understanding of how pollutants evolve in the atmosphere at regional or larger scales and have played an important role in developing air quality management plans. Current practice to develop control strategies for ozone precursors is based on simulating short ozone episodes with the 'worst case' weather conditions. There are concerns about the representativeness of such episodes, and about models being tuned to perform well by adjusting input data and model parameters. In our research, we seek a more comprehensive evaluation of air quality model performance, and new insights into questions such as appropriate air pollution control strategies and inter-basin transport of ozone and its precursors, through application of an air quality model to Central California for an entire summer season. Ozone air pollution problems in Central California are severe and not improving. Here we report results from application of the Community Multi-Scale Air Quality model (CMAQ) to the Central California Ozone Study period in 2000 (June to Oct.). Gridded meteorological and emission inputs are developed to reflect variability occurring on diurnal, weekly, and seasonal time scales. Driven by these inputs we assess model skills at predicting 1-h and 8-h average ozone concentrations, as well as ozone precursors, across a range of days and locations, with wide air quality variations seen in both space and time over the entire summer season. We compare modeled and observed ozone responses to changes in meteorological fields (temperature and flow patterns) and emissions. The model shows consistent performance in the San Joaquin Valley in terms of characterizing observed ozone variabilities during different pollution episodes and across the summer season. Ozone sensitivity regimes are reasonably reproduced in the valley. Model performance in the coastal regions is more sensitive to uncertainties in the wind fields under different synoptic conditions.