Climatological changes in rainfall distributions at different rain-rates under Qinghai-Tibet Plateau warming during 1981-2060

Climatological changes in rainfall affect the Tibetan Plateau's eco-hydrological processes and water resource regimes. To better understand these changes, we examined the historical and future changes in rain amount distribution (RAD) and rain frequency distribution (RFD) under Tibetan Plateau warming. Daily rain accumulation from two observed datasets (Climate Prediction Center and Climate Forecast System Reanalysis) and CMIP6 models under SSP126, SSP245, and SSP585 scenarios are used to estimate RAD and RFD at each grid point. To capture the full range of rainfall from light to heavy, we used logarithmically distributed rain-rate bins, whose coordinates enable mathematical operations. This means that each rain-rate bin is a percentage wider in log coordinates than the previous one, with the same bin center. Results showed that RAD magnitudes varied from 0.74 to 0.93 mm/day at moderate rain-rates, with the broadest and narrowest curves occurring in 2011-2020 and 1991-2000, respectively. These distributions changed strongly from moderate to heavier rain-rates at 2.8 mm/day/K. Furthermore, RFD magnitudes ranged from 13.03% (1991-2000) to 13.51% (2011-2020), with a negative response at light rain-rates (- 10.33%/K) and a positive response at moderate rain-rates (10%/K). Under future warming, RAD magnitudes were around 0.42 mm/day for the SSPs at moderate rain-rates, which shifted to heavier rain-rates, implying that the RAD curve will become larger with more rain at heavier rain-rates. Besides RAD whose response was around 2.85%/K under SSP126 at light rain-rate, most SSP126 responses were negative at light rain-rate but shifted positively and increased with SSP245 and SSP585 at heavier rain-rates. Overall, the rain distributions are moderately skewed and platykurtic, producing fewer and less extreme outliers than the normal distribution. Notably, there was no significant mean difference between the distributions; however, the change responses showed statistical differences (p < 0.05). This study will raise awareness of unexpected future events by showing how rain distribution changes with global warming.