Modelling and Forecasting of SO₂ Concentration in Atmospheric Air – A Case Study of the City of Krakow

The paper presents a statistical analysis of SO₂ air concentration in Krakow for the period from 2007 to 2016. The source data were obtained from two measuring stations of the Provincial Environmental Protection Inspectorate in Krakow. SO₂ concentration data had daily average values. The average monthly values and quarterly averages were used for the analysis. The analysis started with the determination of the trend function and the time series of the moving average. Then indices of monthly and quarterly changes were determined. Considering the whole period of ten years, a slight downward trend was noted, even though a somewhat rising trend was observed in the first five years. In Krakow, the level of sulphur dioxide pollution is approximately two to 2.5 times higher than in the least polluted area of the region the capital of which is Krakow. To create a mathematical model, the following two series were identified: cyclic variability and random variation. A multiplicative model composed of four components was obtained. This model interpolates the source data. The methodology of time series modelling was also applied to monthly and quarterly data. The wind directions prevailing in the city centre were also defined for the last four years. Statistical analysis and multiplicative models of SO₂ variability for these directions were made. For each of the seven selected wind directions, the basic concentration statistics were determined and all components of the multiplicative models were identified. No significant effect of the wind direction on SO₂ concentration in Krakow was observed. This allows formulating a hypothesis of the dominating impact of the relatively uniformly distributed local sources of surface emission on the magnitude of air pollution by sulphur dioxide in a large urban agglomeration, which is confirmed by some selected scientific sources. The models presented were used to obtain quarterly and monthly forecasts for SO₂ concentrations in Krakow. The discussed methodology is universal and can be used in the study of variability and forecasting of concentrations of other air pollutants in other cities. (paper)