No abstract

Guidance for the performance evaluation of three-dimensional air quality modeling systems for particulate matter and IMPLICATIONS The National Ambient Air Quality Standards for particulate matter (PM) and the federal regional haze regulations place some emphasis on the assessment of fine particle (PM 2.5 ) concentrations. Current air quality models need to be improved and evaluated against observations to assess the reliability of model simulations. The guidance presented here provides the necessary framework for conducting rigorous performance evaluations of PM and visibility models. The costs associated with the field programs needed to obtain the data necessary for such performance evaluations are estimated to be $15 million for data collection (1-year program with an intensive program of 15 days, over 200,000 km 2 ) and $10-$20 million for planning, emission inventories, data analysis, and modeling. visibility is presented. Four levels are considered: operational, diagnostic, mechanistic, and probabilistic evaluations. First, a comprehensive model evaluation should be conducted in at least two distinct geographical locations and for several meteorological episodes. Next, streamlined evaluations can be conducted for other similar applications if the comprehensive evaluation is deemed satisfactory. In all cases, the operational evaluation alone is insufficient, and some diagnostic evaluation must always be carried out. Recommendations are provided for designing field measurement programs that can provide the data needed for such model performance evaluations.

No abstract

No abstract

The SARMAP air quality model, enhanced with aerosol modeling capability, and its associated components were developed to understand causes of ozone (O3) and particulate matter exceedances in the San Joaquin Valley of California. In order for this modeling system to gain increasing acceptance and use in guiding air quality management, it is important to assess how transportable this modeling system is across geographic domains. We describe the first application of the modeling system outside the "home" domain for which it was developed and evaluated. We have chosen the August 27-28, 1987, intensive monitoring period of the Southern California Air Quality Study to evaluate the performance of the modeling system and to assess its sensitivity to emission control options. The predicted surface concentrations of O3 and other gas-phase species were spatially and temporally correlated with measured data. The fractional normalized absolute error was 0.32 to 0.36 for O3, and somewhat larger for other species. The fractional normalized bias for O3 on August 27 and 28, 1987, was 0.02 to 0.04. The simulated PM2.5 mass and constituent species concentrations reproduced the magnitude and variability of the observed daytime concentrations at most locations; however, nighttime PM2.5 concentrations were overpredicted by the model. The model's response to emission control options was consistent with other models of the same genre.