Projection of non-stationary compound droughts considering internal variability of a climate model over Iran

This study investigates the combination of seasonal precipitation and temperature extremes leading to compound droughts over Iran, using data from 33 synoptic stations (1961–2021). Tail-dependent copula functions identify stations with significant likelihoods of compound droughts, with the highest occurrence in fall (45%) and lowest in summer (21%). A non-stationary assessment, accounting for climate-driven trends in temperature and tail dependence, is conducted using generalized linear models (GLMs) to project changes in occurrence probability and joint return periods from 1961 to 2100. Results indicate that ignoring dependence between temperature and precipitation substantially overestimates joint return periods and underestimates compound drought frequency. Future projections under the 100-member CESM2 ensemble (SSP3-7.0) reveal increasing return periods in some seasons due to non-stationarity, but historical dependence still leads to shorter joint return periods compared to assuming independence. These findings emphasize the importance of considering both non-stationarity and inter-variable dependence in future compound drought assessments.