Modele stochastyczne zanieczyszczeń powietrza w aglomeracjach przemysłowych

The Upper Silesian Industrial Region (GOP) is one of the most polluted regions in Poland. Because of the location of several important heavy industrial plants it is necessary to permanently monitor the various sort of dust and gas pollutantsconcentrations in this area. The paper presents the possibilities of stochastic air pollution modeling on the basis of data collected by monitoring stations. Several types of models were shown, including models applied in regions of big cities, like Stockholm, Vienna and Madrid, with special impact to so-called adaptive models. It was statistically proved that the formulae of the SO2 propagation model for the GOP S(t)=a+bS(t-1)+c(T-T0)2+d(v-v0)2+eQ1=eQ2. This equation was applied practically on the basis of the empirical data collected by selected monitoring stations.For the chosen monitoring station the directions of pollution flows and winds wereshown graphically (fig. 1). Nest step was derivation of the SO2 propagation model bytraditional regressive techniques (models from equations 6, 7 and 8), taking into considerationdirections of air flows, and adaptive models (fig. 3) basing on the previous model formulae. The obtained models were statistically evaluated. It occurred that the models considering air flows directions show changes of pollution propagation characteristics The advantage of adaptive models, which take into consideration data from previous periods of time, was proved, as they forecast concentration of pollution far better than the traditional regressive models.