Prediction of Winter Rainfall Using Adaptive Fuzzy Neural Networks: A Case Study of Khorasan Razavi Province, Iran

This study aims to predict winter rainfall of Khorasan Razavi Province using Adaptive fuzzy neural networks. To this goal, first average regional rainfall was calculated using Kriging method. In the next stage, rainfall correlation was analyzed by climatic predictors in different intervals. After identifying the effective predictors on regional rainfall, adaptive fuzzy neural network model was trained in 1970-1997 period and finally the rainfall was predicted for 1998-2007 period. The results show that adaptive fuzzy neural networks are able to predict the rainfall amount in an acceptable accuracy. Root mean square error was obtained 7.4 mm for the model.