In partnership with the US EPA, NOAA has recently implemented ozone and smoke predictions over the lower 48 states as components of a national air quality forecast capability (NAQFC). Predictions are generated with linked weather and air quality models run operationally at NOAA's National Centers for Environmental Prediction (NCEP): the NOAA-EPA Community Multiscale Air Quality (CMAQ) model and NOAA's HYSPLIT dispersion model are driven by NOAA's operational North American mesoscale weather prediction model (NAM; currently 12km WRF-NMM) in NCEP operations. Weather observations are assimilated in real-time; emissions information is provided from the EPA's monitoring and national emissions inventory databases, and estimates of fire emissions are provided with the USFS's BlueSky model. The NAQFC is an end-to-end system providing predictions through the next day, at 12km resolution of hour-by-hour ground-level ozone concentrations (updated twice daily) and smoke at the surface and in the atmospheric column (updated once each day). Products are disseminated on NWS and EPA servers. Near-real-time verification is conducted with EPA's AIRNow monitoring observations and with NESDIS satellite observations. In general, the predictions are more accurate during episodes of higher ozone concentrations; for the urban areas in northeastern US than for those in southeastern US; and for eastern cities than for western cities

During summer 2008, two upgrades are being tested in real-time, that could improve predictions for the western US: 1) CMAQ upgrade for CONUS ozone predictions, and 2) expanded smoke predictions over Alaska. With successful demonstration of operational deployment readiness criteria for prediction accuracy, reliability and subjective feedback, these improvements will be added to the operational product suite.1) A recent CMAQ upgrade with the new CB05 mechanism is being tested in experimental ozone predictions over the CONUS. CB05 includes additional organic reaction pathways and produces systematically greater ozone concentrations than the CB-IV mechanism used in the current operational system. Preliminary experimental testing from early summer shows mixed results for large urban areas, with improvements relative to the operational predictions in California, but more over-prediction in Eastern US. 2) Expanded smoke prediction capabilities over Alaska: Although wildfire smoke is the largest contribution to poor air quality in Alaska each summer, the very short fire season poses challenges with limited real-time testing opportunities for smoke predictions. Preliminary analysis of the predictions, facilitated with smoke-column verification from specially adapted satellite observation products, shows promise.

Also in progress are efforts to develop a quantitative particulate matter forecast capability. Currently the NOAA-EPA team is developmentally testing ground-level aerosol predictions, using a modified CMAQ capability, with the new CB05 mechanism, and pollutant data from emissions inventories only. These tests do not include emissions from US wildfire smoke, or other contributions from long-range transport including smoke and dust. Test results for the CB05 mechanism, which includes additional secondary organic compounds, will be summarized. Research is ongoing to develop capabilities for real-time ingest of additional emissions sources important for quantitative predictions of particulate matter. Further developments include expanding ozone predictions over all 50 states; targeted operational deployment of nationwide ozone forecasts is within 2 years. Nationwide deployments of ozone and smoke prediction capabilities will be followed by the addition of quantitative particulate matter forecasts, and extended forecast periods - out to day 2 and beyond.