Short-range forecast of atmospheric pollutants using non-linear prediction method

In this paper chaotic behavior in the nitrogen dioxide and sulphurous anhydride concentration time series at two sites in Gdansk region is investigated. To reconstruct an attractor, the time delay and embedding dimension are needed. The former is determined by the methods of autocorrelation function and average mutual information, and the latter is calculated by means of correlation dimension method and algorithm of false nearest neighbors. It was shown that the low-dimensional chaos existed in the time series under investigation. The spectrum of Lyapunov exponents was reconstructed as well as both Kaplan–Yorke dimension and Kolmogorov entropy that inversely proportional to the predictability limit are calculated. Non-linear prediction method is used for the time series. It is shown that even though the simple procedure is used to construct the non-linear model, the results are quite satisfactory.