Comparison of univariate and bivariate regional frequency analysis of drought (case study: Part of semi-Arid climate of Fars Province)

Drought is one of the most complex and destructive climatic phenomena, which can be perceived as the least understood natural disaster (Kao and Govindaraju, 2010). The most important and challenging characteristics of drought are frequency and return period (Bazrafshan et al., 2014; and Zhang et al. 2015). This complexity is derived from the interdependence of drought characteristics that make the univariate frequency analysis inefficient (Mirakbari et al., 2012). Therefore, instead of using traditional univariate analysis, a better approach is to derive the joint distribution of drought variables (Mishra and Singh 2010). Furthermore, insufficient data at the stations and the existence of ungagged areas necessitate regional analysis. Regional frequency analysis, on the one hand, provides the possibility of analysis for ungagged regions, and, on the other hand, provides better and more comprehensive information for meteorological stations using a combination of points and regional data. The main objective of this research is regional bivariate drought analysis in the semi-arid climate of Fars Province, Iran. In this regard, the index variable based on linear moments is one of the most advanced methods ( Núñez ez et al., 2011).