Stochastic regional air pollution modeling

Existing air pollution simulation models have not been validated with actual ambient air pollution data. The two major reasons for the lack of correlation between simulated and actual data are: 1) the physical system is far more complex than the model, and 2) available sensors are unreliable and inaccurate. A stochastic model is proposed in lieu of a deterministic model. The stochastic model is simple in structure and adaptive by nature, i.e., the parameters in the model are "tracked" by the data processing algorithm. A method to determine a priori parameter estimates is presented along with a method to "track" parameter changes and detect pollution violations statistically. Finally a method to optimally control the air quality in the regional area is discussed. The methodology proposed for use by APCDs in the 1970s has been extensively used by the defense and aerospace industry throughout the 1960s. There is every reason to believe that the methodology outlined in this paper for control of air quality in APCDs will become a "standard" for the 1970s.