Improving the performance of the SPEI using four-parameter distribution function

Abstract

One of the main challenges in the present era is competition for access to water resources. Iran is also on attention due to its geopolitical and strategic location. Water scarcity is a problem which will bring the country into the next dimensions of the challenges. Reducing water resources in this country is affected by global climate change and droughts. Meteorological drought is studied by researchers using multiple indices. The Standardized Precipitation Evapotranspiration Index (SPEI) is also one of themostwidely used indices in this field. The aimof this study was to investigate the meteorological drought and identify dry and wet months in the eastern stations of Iran using the SPEI. In this regard, it has been tried to select a function proportional to precipitation minus potential evapotranspiration by examining continuous and discrete statistical distribution functions. Among the 65 distribution functions examined, the results of goodness of fit tests of Anderson-Darling, Kolmogorov-Smirnov, and chisquare tests, introduced the four-parameter Burr distribution function (BDF) as the best distribution function. The results showed that the four-parameter BDF has higher accuracy than the conventional log-logistic function. The results of the extraction of SPEI showed that drought intensity in the eastern regions of Iran during the statistical period of 1973–2011 has increased and almost 26% of the months examined at all stations have faced drought. Finally, according to the results of this study, it is suggested to examine various distribution functions or use the proposed distribution function for the extraction of SPEI values. Also, as well as the existing climate change, the results of the MSPEI index appear to be better than the SPEI index.