Long-term modeling of regional ozone concentrations and control strategies

The capability to accurately model the long-term frequency of occurrence of high and low ozone concentrations in the urban and regional atmosphere is developed. A full year of hourly ozone concentration predictions is generated throughout Southern California using a photochemical airshed model driven by automated assimilation of routine meteorological observations. Statistical measures of airshed model performance for the full year are comparable to results achieved by episodic models for 2-3 days of simulation that are driven by data from expensive special field measurement programs. The effect of alternative emission control strategies on the long-term frequency distribution of daily peak ozone concentrations in Southern California is determined. Ozone isopleth diagrams are constructed in terms of the number of days per year that the current U.S. Federal 8 h average and former 1 h average ozone standards will be exceeded at all feasible combinations of air basin-wide reactive organic gases and oxides of nitrogen emissions. It is found that the frequency of violation of the former 1 h average ozone standard set at 0.12 ppm ozone could be reduced to approximately 25 days per year through stringent emissions controls under 1987 meteorological conditions. No practical way is evident to even come close to meeting the new 8 h average standard set at 0.08 ppm ozone. It is shown that the days with the highest historically observed ozone concentrations are not necessarily the hardest days to bring below the air quality standard. Use of synthetic meteorological data is explored as an efficient means to meet the input data requirements of airshed models that must operate over long periods of tine. A semi-Markov process is used to generate a time series of synoptic weather conditions that statistically resembles the occurrence and persistence of historically observed weather. Local weather variables then are drawn for each day that are representative of the meteorological potential for ozone formation. These values are used as the initial conditions for a prognostic mesoscale meteorological model that generates the meteorological fields needed by a photochemical airshed model. The procedure produces daily peak 8 h average ozone concentration predictions that compare well with historical observations.