Advanced modeling and forecasting of pollutant concentrations temporal dynamics in the atmosphere of an industrial city (gdansk region)

Abstract

In the paper we present the results of an advanced investigation of dynamics of variations of the atmospheric pollutants (sulphur dioxide) concentrations in the air basins of Polish industrial cities (Gdansk region) by using the improved non-linear prediction and chaos theory methods. Chaotic behavior of the sulphurous anhydride concentration time series at two sites in the city of Gdansk has been computed. As usually, to reconstruct the corresponding chaotic attractor, it is necessary to determine time delay and embedding dimension. The former is determined by the methods of autocorrelation function and average mutual information, and the latter is calculated by means of the correlation dimension method and algorithm of false nearest neighbours. Further, the Lyapunov exponents’ spectrum, the Kaplan-Yorke dimension and the Kolmogorov entropy and other invariants are calculated. An existence of a low-D chaos in the cited system is confirmed and using polynominal algorithm with neural networks block allows making an improved short-term forecast of the atmospheric pollutant fluctuations dynamics.