Application of Fuzzy Logic for Estimation of Trends of Productivity Change in Some Areas of Lake Onego

We studied the productivity level of some parts of Lake Onego. Four indicators were used: number of saprophytic bacteria, zooplankton biomass, chlorophyll- concentration, and water transparency. The analysis included samples from 14 stations located at transects from the top of Kondopozhskaya Bay to the center of Lake Onego and from the top of Petrozavodskaya Bay to the center of the lake. Two periods of 20 years (the end of the last century and the beginning of the present one) were compared. The productivity level was estimated using fuzzy logic technology according to the Mamdani algorithm. The membership functions for each indicator included two terms: “low productivity” and “higher productivity”; the ranges of values for each term were set by an expert. The output membership function included two terms of rectangular shape with the values “low productivity” and “higher productivity”. A total of 16 production rules were constructed; they, in turn, allowed to obtain 706 productivity level estimates for all stations. A bootstrapping was applied to assess the significance of differences. Comparison of data for the end of the last and beginning of the current century showed the following: There were no significant changes in productivity estimates at the transect “Petrozavodskaya Bay – the center of the lake”. For the transect “Kondopozhskaya Bay – the center of the lake”, there is an increase in productivity level, probably due to the appearance of new factors of anthropogenic impact: the trout farms. The advantage of the fuzzy logic modeling method is the clear theoretical meaning and simpler calculation technology compared to statistical classification methods.