Application Of BP Neural Network To Predict The Short Term Trend Of Eutrophication In West Lake

Artificial neural networks are becoming more and more common to be used in development of prediction models for complex systems as the theory behind them develops and the processing power of computers increases. A three layer Levenberg-Marquardt feedforward learning algorithm was used to model the eutrophication process in West Lake.We have established 8 sampling spots in West Lake, and collected the aquatic data (2000.1-2001.4) of West Lake by routine measurement. Selecting spot 7 which can most represent the water quality status of it as study object and filtrating the water quality parameters as the inputs for network by principal component analysis. Furthermore, the inserted method which creats the sufficient samples solved the deficiency of samples. In the paper, using Back propagation (BP) neural network, we found the most influential elements which can reflect the trend of aquatic ecology in West Lake for modeling, and established the best nework to predict the short term trend of eutrophication in West Lake. At the same time, we used the data of spot 3 to test the universality of the network, and found the outputs tallied with the measured values very well. The results showed that water temperature and chlorophyll-a affect the concentration of chlorophyll-a of next week very well. The network using them as input variables is simple and prompt, having greater advantages than other linearity numeric modelings.This indicates that artificial neural network models were able to model non-linear behavior in eutrophication process reasonably and could successfully forecast the concentration of chlorophyll-a in West Lake. It can provide the scientific basis for the control of the eutrophication of West Lake.