Prediction of long-term monthly precipitation using several soft computing methods without climatic data

摘要: Accurate estimation of precipitation is an important issue in water resources engineering, management and planning. The accuracy four different soft computing methods, adaptive neuro-fuzzy inference system (ANFIS) with grid partition (GP), ANFIS subtractive clustering (SC), artificial neural networks (ANN) support vector regression (SVR), investigated predicting long-term monthly without climatic data. periodicity component, longitude, latitude altitude data from 50 stations Iran are used as inputs to the applied models. ANFIS-GP model found perform generally better than other models precipitation. SVR provides worst estimates. maximum correlations be 0.935 0.944 for ANFIS-SC Fasa station, respectively. highest ANN 0.964 0.977 Bam Tabas (Zabol) stations. minimum 0.683 0.661 Urmia station while provide 0.696 0.785 Sari Bandar Lengeh stations, comparison results show that precipitations any site can successfully predicted by weather annual also mapped evaluated using optimal study. maps revealed amounts occur north, southwestern west regions, lowest values seen east southeastern parts Iran.