Assessment of future changes in drought characteristics through stochastic downscaling and CMIP6 over South Korea

Assessments of future droughts are essential tools due to the potential for serious damage to the environment, economy, and society, particularly under climate change. This study proposes a framework for assessing drought characteristics at different scales, periods, and emission scenarios modeled by phase 6 of the Coupled Model Intercomparison Project. The four drought characteristics were determined by applying the Run theory to a standardized precipitation index time series, and the severe drought areas were detected by the Jenks Natural Breaks and Kriging methods. The study produced four main findings. (1) A stochastic weather generator, AWE-GEN, captures the variability of precipitation with inter- and intra-annual stochastic properties, and presents naturally occurring variability as an ensemble. (2) According to the ensemble average of drought characteristics, future droughts are projected to become less frequent with similar durations and intensity due to future rise in precipitation. However, the ensemble (stochastic or natural) and spatial variabilities are expected to increase, making drought management difficult (e.g., future decrease of 18% in \({DE}_{max}\) ${\mathit{DE}}_{\mathit{max}}$ for END585). (3) Different temporal scales can affect the detection and characterization of drought events. Smaller temporal scales identify mild drought events of short duration and low severity, while larger scales merge and extend drought events, resulting in more prolonged and severe droughts. (4) Severe drought areas can expand compared with a control period for drought duration and severity, but may decrease for drought interval and frequency especially for the END period (e.g., 24% and 17% increase for \({DD}_{max}\) ${\mathit{DD}}_{\mathit{max}}$ and \({\left|DS\right|}_{max}\) ${\left|D,S\right|}_{\mathit{max}}$ , and 85% and 78% decrease for \({DI}_{mean}\) ${\mathit{DI}}_{\mathit{mean}}$ and \({DE}_{max}\) ${\mathit{DE}}_{\mathit{max}}$ for SPI3 and END585). The framework proposed in this study is expected to provide important information for the building of strategies required to adapt to and mitigate the potential impacts of drought in the future.