Drought Estimation of Şanlıurfa Station with Artificial Neural Network

Abstract

Early estimation of the drought may help reduce the potential adverse effects of drought. Indices developed for this purpose provide the information for the past drought events. Estimating the drought using indices and precipitation data from past periods should allow early warning systems to be established. In this study, the Standardized Precipitation Index (SPI) values calculated with the rainfall data of Şanlıurfa station between 1938-2014 years are estimated by using past rainfall and drought index values. Droughts are predicted using Feed Forward Back Propagation Neural Network (FFBPNN) method. The drought indices and precipitation values between 1937-1990 are used for training, and the values between 1991-2014 were used for test data. To estimation of 1, 3, 6 and 12 months drought index values, 16 models are used for each time scale. It has been determined that artificial neural network method is applicable for estimating 6 and 12 month SPI indexes for Şanlıurfa station.

Keywords

• Drought Analysis
• Drought estimation
• Artificial neural network
• Standardized precipitation index