Studying drought and evaluating its prediction possibility in Ardabil province by using SPI index and ANFIS model

Drought affects on all aspects of human actirities seriously as a climatic phenomenon. However, the studies related to this phenomenon based on suitable method are very few. Studying the drought features and its prediction can be effective on decreasing of resulted losses or damages. Therefore, drought and evaluation of its prediction possibility are studied for some stations of Ardabil province. The data used in this research is the rainfall value as montly during 23 years statistical duration. The standardized precipitation index in time index of 1, 3, 6, 12 monthes used for studying the drought features and adaptive neuro-fuzzy inference system model for prediction of drought. Based on the research results, the most serious drought has occurred in Ardabil province in Ardabil station during six monthes time index, in December 2010 by 2.47 index. Generally, the highest percentage of drought occurance is seen in Pars Abad station and the lowest rate is seen in Khalkhal station. Except pars abad station, the linar trend of SPI index is decreasing trend in 3 and 6 monthes. In other words, the drought increases. Based on clustering analysis, Khalkhal and Ardabil stations are located in one group in each of four time index. Therefore the drought feature is similar in them and pars abad station is located in other group. The results of prediction of index by ANFIS model indicated that the prediction errors are considerable in most of eases and the model hasnot acceptable efficiency in this prediction.