An analysis of droughts in Iran during 1988–2017

Abstract

Understanding how drought changes have occurred is imperative to enhance the ability of mitigating drought impacts. This consideration encourages carrying out a comprehensive analysis of drought phenomenon in Iran using the standardized precipitation evapotranspiration index (SPEI) on various time scales based on data recorded at 100 weather stations over the period 1988–2017. We employed a nonparametric distribution i.e. the kernel density estimator instead of the commonly-used three-parameter log-logistic distribution for estimating the cumulative distribution functions of the monthly aggregated water deficits/surpluses, and found that the kernel density estimates frequencies of drought classes more accurately than the three-parameter log-logistic. Thereafter, an analysis was conducted by applying a principal component analysis and a Mann–Kendall test to spatio-temporal variables of the nonparametric SPEI at different time scales. Moreover, a bivariate risk assessment was done by calculating the joint exceedance probability using copula functions which depicts joint behavior of the drought characteristics. The spatial variability of drought in Iran revealed the existence of five coherent regions i.e. the northwestern, the southwestern, the southeastern, the northern and the central/eastern over the country. Temporal analysis indicates significant negative trends of surface water balance in all regions by producing a dry condition throughout the country, especially in two last decades (i.e. 1998–2017). However, the northwestern region is dominated by the highest risk of droughts and experienced droughts with longer durations and higher severities. This study was a forward step in considering an aggregative perspective of meteorological, agricultural and hydrological water deficit.