Abstract: To improve the forecasting accuracy of monthly precipitation and deal with the determination of the number of hidden neurons in neural networks,this paper introduced a particle swarm optimized wavelet multiple neural network model,which was applied to the prediction of monthly precipitation in Dongting Lake Basin.The standardized monthly precipitation and large-scale climate index time series were first decomposed at different temporal scales as predictors.Then the standardized monthly precipitation subseries were forecasted,respectively,under different time scales by using the cascade-forward(CF) neural networks in which the number of hidden neurons is optimized by particle swarm optimization(PSO).Finally the monthly precipitation was forecasted by combining all predicted subseries and using the inverse transform of standardized monthly precipitation.The results show that PSO-based wavelet multiple neural network model provides more accurate forecasts than wavelet single neural network model for monthly precipitation in Dongting Lake Basin and improves the prediction accuracy of extreme monthly rainfall.

Key words: Wavelet neural networks,Particle swarm optimization,Dongting Lake Basin,Monthly rainfall prediction