RESUMO
The Er-longshan reservoir, located in Harbin city, Heilongjiang province, plays a significant role in development of economy and society. It's one of the most important fresh water sources for drinking as well as the safeguard of flood and soil erosion. Thus, as the monitored place in this study, its water quality was measured and predicted through a fuzzy model. The objective of this paper is to illustrate how to set up an appropriate subjection function model to solve the fuzzy problem. Most environmental monitoring data that can not be compared may be mapped into subjection degrees, and analyzed for their weight coefficients, yielding the best situation in multi-objective comprehensive exponential decision-making matrix.An appropriate subjection function model was set up to solve the fuzzy problem. Five kinds of pollution sources were investigated: the point source, the plane source, the entering river (Feiketu river), precipitation and falling dust and touring pollution around the reservoir area respectively. The distribution of floating algae in Er-longshan reservoir was also examined. Farmland plane source pollution was found to be the major controllable pollution source by monitoring TN and TP pollution loads, which occupy 84.8% and 84.0% of the controllable pollution source respectively. When we evaluate data of water quality, the concentration of part pollutant factors increases while others may decreases, then whether the whole water quality is to increase or decrease, from the monitor data is not easy to judge. In this study we used the fuzzy theory to analyze the trends of water quality fluctuations in the Er-longshan reservoir. COD(Mn), BOD, TN and TP were selected as main contamination factors. The results showed that the primary pollutants were nitrogen and phosphorus by calculating the weight coefficient e(i) of contamination factors from 1996 to 2005. According to fuzzy comprehensive exponent z(j) evaluating water quality of reservoir from 1996 to 2005, the ten-year water quality dynamic trend was studied.
