A simplified scheme for predicting peak ozone concentrations in the central United States

With the impending requirements for an 8-hour ozone standard (since remanded by the courts), many municipalities found themselves on the brink of violating a new standard which in many cases was more stringent than the existing one-hour NAAQS. In order to continue in attainment, or to return to that status, some communities elected to develop citizen action plans of voluntary reductions of precursor emissions. These actions would be most beneficial on days with a high potential for violating the ozone standard. Thus arose a requirement for forecasting such days with sufficient lead time to alert the public and other participants in the emission reduction effort. In St. Louis and Kansas City, MO, forecast of the upcoming day's ozone potential was prepared by private forecasters for dissemination through the media. A forecasting system was devised which used as input openly those meteorological variables that an operational forecasting could routinely be expected to forecast (cloud cover, temperature, dewpoint, wind speed and direction, rainfall, air mass type). Five years of ozone and meteorological data (1994-1998) were used to develop a simple statistical forecasting scheme for both cities. The technique demonstrated considerable promise in identifying those days on which there existed a high potential for violating the ozone NAAQS. The technique recognizes that regional transport tends to dominate locale production in this area, necessitating the recognition of such transport regimes were. In addition to a review of the ozone climatology of the region, the technique will be described, hindcasting skill scores presented, and initial results from the 1999 season will be reviewed.