LAUSR

Future changes in drought events are critical for risk assessment and associated policymaking. In this study, the future changes in meteorological droughts in Henan Province, China are explored. Random forests downscaling model is first constructed based on ERA5 reanalysis data and meteorological observations. The model is validated using evaluation indices such as R2 and RMSE, and is shown to be able to capture the relationship between large-scale predictors and monthly precipitation. The validated random forests downscaling model is driven by multiple global climate models (GCMs) from the Phase 6 of the Coupled Model Intercomparison Project (CMIP6) under three emission scenarios for projecting three future drought characteristics (duration, frequency, and intensity). Results show that drought frequency decreases in most areas of Henan while drought duration and intensity increase in various degrees. Some differences are also observed among different emission scenarios, especially under SSP2-4.5, where the magnitudes of changes in drought duration and intensity are lower relative to other scenarios. The decrease in drought frequency in most areas is found to be caused by increases in monthly mean precipitation in this study. Changes in drought duration and intensity are related to a combination of increases in precipitation mean and variability.