During 2003, NOAA's National Weather Service (NWS) and Office of Atmospheric Research (OAR) developed a prototype ozone prediction capability for the Eastern U.S. The NWS/ National Center for Environmental Prediction (NCEP) Eta model at 12 km was used to drive the EPA Community Multi-scale Air Quality (CMAQ) model to produce up to 48 hour ozone predictions. The CMAQ system simulates various chemical and physical processes that are important for modeling atmospheric trace gas transformations and distributions. CMAQ includes chemical mechanisms to simulate various air quality constituents including tropospheric ozone, fine particles, toxic, acidic deposition and visibility degradation. For this study, CMAQ predictions were evaluated against the EPA's surface ozone monitoring network to provide temporal and spatial guidance on model performance. This presentation will review the forecast model system and present preliminary quantitative results from the surface ozone evaluation system. The verifications of the surface ozone forecasts include both mean absolute and root mean square errors of the hourly and 8-h average ozone values, as well as categorical evaluations of certain ozone threshold events.

Comparisons will be made on model performance in urban versus rural areas and model evaluation during several episodes when high ozone was observed. The Eta model will also be evaluated with special emphasis on key meteorological (for example, vertical boundary layer mixing, clouds and radiation and land surface) predictions that influence air quality. Boundary layer profilers and surface radiation budget stations deployed for the New England High Resolution Temperature Program (NEHRTP) will be used to further diagnose errors in the Eta-CMAQ prediction system The NEHRTP was a targetted program to investigate methods to improve meteorological forecasts for the energy sector. New observational platforms were deployed to diagnose meteorological model error. These platforms will be used to help diagnose Eta model errors that may have contributed to ozone forecast errors.