The probability density functions were used to forecast the air quality in several monitoring sites located in Worek Żytawski region. Prediction of percentile 99.8 value (the maximum allowable 30-minute pollutant concentration) was made on the basis of the frequency function, which represented the annual distribution of SO2. Analysing the historical SO2 concentration data sets identified the form of distribution. The parameters of the distribution were estimated basing upon the annual average concentration prognosis. Then, the preliminary predicted distribution of S02 was verified every month by taking into account successively upcoming concentrations values recorded during the analysed period. The results proved that the air quality estimation obtained with the presented method of verification at the end of September and the ones typically calculated at the end of the year were similarly accurate.

Meteorological parameters which are most significant for ozone forecasting were chosen in the multiple regression analysis for the daily time series. Then correlations between the variables we~e investigated, both for the daily and temporary values. There was confirmed a strong relationship between atmospheric conditions and ozone concentrations as well as autocorrelations of the temporary time series of ozone from different monitoring stations. Diversification of autocorrelation values arises probably from different receptor locations which was confirmed by the principal component analysis. There were also shown dependences between the ozone time series from different monitoring stations. Strong space-time relationships of ozone concentrations and meteorological conditions in the Black Triangle region can be used in modeling and forecasting of ozone episodes.

Time series analysis ofmonthly and daily SO2 data were considered for the detection of trends in SO2 due to possible effect of the emission abatement strategy in the Black Triangle region. Using a time series model, the main components were extracted from the original SO2 time series. Based on SO2 monitoring data from Czerniawa in Izery Mountains in Poland over the period 1993 -1998, our findings showed evidence of declining trends in SO2• A mean annual change of 14.1% was recorded in a 6-year record. It has also appeared that the exponential smoothing which considers a seasonal component and trends provided a reasonable fit to monthly mean SO2 values.