This paper summarized results of estimation of concentrations of pollutants (chemical and bacteriological) in a barrier lake on the river Łęg (Wilcza Wola near Raniżów, subkarpatian province). The second goal of the research was aimed at a trial of forecasting the pollution of apparently stagnant water (as the test bed the barrier lake mentioned was used) applying pattern recognition method in a variant called ,,look-ahead-and-back-KNN" [8, 9]. It was found, that concentrations of some contaminating species can be predicted with satisfactory accuracy. The approach tested allows to develop a dense grid, partly consisting of nodes with real (experimentally estimated) results, and partly consisting of nodes with these values predicted. Additionally, our research has been recently augmented with results achieved using data mining methodology based on virtual visualization [12].

Prognozuje się, że zasoby wody słodkiej dostępne dla ludzkości zmniejszą się o 30 proc. w ciągu najbliższych 20 lat. Jak odwrócić degradację zasobów wody i zrównoważyć zapotrzebowanie ludzi na wodę z potencjałem hydrosfery?