Monitoring and Prediction of Drought in Western Urmia Lake Basin Rain Gage Stations by ANFIS Model

Drought as a climatic phenomenon affected many different environmental issues and generally is associated with the decreasing in average precipitation. Evaluation and monitoring of the drought is a fundamental step in proper programming of water resources management. Regarding the recent conditions water scarcity in the Urmia Lake basin, assessment of the drought index in this region is inevitable. This study was conducted to evaluation of the SPI index in western parts of the Urmia Lake basin. Rainfall data were collected from 11 rain gage stations in Tabriz and Maragheh plains. The standardized precipitation index in time scales of 9, 12, 24 months used for studying the drought features and adaptive neuro-fuzzy inference system model was applied for prediction of drought. The results indicated that the most serious (-4.07) drought was occurred in Bonab station at the 9 month time scale in October 1984. The longest drought periods were occurred in Heravi, Saiedabad, and Maragheh stations and the shortest drought periods were occurred in Zinjab, Tabriz, and Lighvan stations. The results of SPI prediction by ANFIS model indicated that the model desirably could predict the drought conditions in the study area. The highest value of the coefficient of the determination of ANFIS model was 0.829 for Maragheh station at 24 months time scale and the lowest values of r2 was 0.480 for Saiedabad station at 9 months time scale. Results also indicated that the capability of the ANFIS tend to be better at long time scales.