Implementation and evaluation of a continental US air quality forecast system

Baron Advanced Meteorological Systems, LLC, executes a real-time experimental air quality forecast system. The system generates continental US (CONUS) 5-day air quality forecasts at 45-km resolution using both the Community Multiscale Air Quality (CMAQ) and the Multiscale Air Quality Simulation Platform (MAQSIP-RT) models. The air quality forecasts are based on the PSU/NCAR mesoscale model (MM5) meteorological forecasts and Sparse-Matrix Operator Kernel Emissions (SMOKE) emissions estimates. This paper examines operational issues and performance evaluation of both models for portions of 2005. Ozone performance will be evaluated for both models, and fine particulate matter (PM 2.5) will be evaluated for CMAQ.

Preliminary results indicate that CMAQ generally underestimates daily peak 1-hr and 8-hr average ozone and 24-hr average PM 2.5. MAQSIP-RT, on the other hand, generally overestimates daily peak 1-hr and 8-hr average ozone.