Extreme precipitation is the most destructive meteorological disaster in the Great Mekong subregion (GMS). This study employs five top-performing CMIP6 models to project end-of-21st-century changes in extreme precipitation indices under SSP245 and SSP585 scenarios, while elucidating underlying physical mechanisms through diagnostic analysis of the atmospheric water vapor balance equation. Projections reveal substantial increases in RX1day, RX5day, R95P, and R99P, which averaged over the GMS are 20 (10) mm, 40 (20) mm, 320 (160) mm, and 270 (110) mm under the SSP585 (SSP245) scenario. Spatially consistent enhancement patterns emerge across all indices, exhibiting a stronger magnitude in western GMS compared to eastern regions. The analysis identifies distinct scenario-dependent drivers for RX1day and RX5day: SSP585-induced intensification primarily stems from dynamic circulation enhancement, whereas SSP245-driven increases predominantly arise from thermodynamic moistening processes. Frequency increasing of extreme precipitations emerges as the principal contributor to R95P and R99P amplification under both scenarios, explaining around 50% of total increases. This frequency enhancement correlates with strengthened atmospheric variability across synoptic (3-8 days) and intraseasonal (10-90 days) timescales. Thermodynamic processes contribute the second most to R95P and R99P intensification, while the dynamic processes are negligible. Furthermore, topographic forcing exhibits amplifying effects in mountainous western GMS on all changes in extreme precipitation indices under both emission scenarios.
 

future change,extreme precipitation,physical processes