Increasing multiscale variability in extreme precipitation under global warming in the Tienshan Mountains, Central Asia

Changes in precipitation variability not only have a significant impact on water cycle processes, but also pose an additional challenge to society's climate resilience. Extreme precipitation is more severe, more abrupt, and more sensitive to temperature changes than mean-state precipitation. However, our understanding of the features of extreme precipitation variability over a broad range of temporal scales, as well as the differences between wet and dry seasons, is limited. In this study, we perform filtering on a daily (2-5 days) to an interannual (2-8 years) scale, using daily precipitation data from detrended APHRODITE and bias-corrected CMIP6 based on zero-phase Butterworth filters. We then analyze the changes in climatological extreme precipitation variability at different time scales, looking at the features along geographical gradients and in wet and dry seasons. At various warming levels, extreme precipitation variability is projected. The findings reveal that the longer the time scale, the higher the overall variability of both extreme and mean-state precipitation, with extreme precipitation showing greater variability. From daily to interannual scales, the variability of R95p increases from 142.79 mm to 875.05 mm, representing a 6.13-fold increase in volatility, but the variability in mean precipitation increases just 1.43-fold, from 2.27 mm to 3.24 mm. The variability in the eastern Tienshan Mountains (80 degrees E - 95 degrees E) is greater than that in the western Tienshan Mountains (66 degrees E - 80 degrees E). Furthermore, along longitudinal and latitudinal gradients, extreme precipitation variability exhibits considerable time-scale differences, with the more extreme the event, the greater the variability. R10mm variability increases from 15.28 d to 76.74 d on a daily to interannual scale. The variability increase (61.46 d) is roughly twice that for R5mm. Seasonally, the wet season is more variable than the dry. The total variability of R10mm was 4.57 days in the wet season and 1.78 days in the dry season. Compared with the reference period (1976-2005), there is an overall increase in extreme precipitation variability at different warming levels, along with an increasing sensitivity to temperature. With the exception of consecutive dry days, the degree of response of filtered extreme precipitation variability to temperature (absolute value of response rate) shows obvious increases with time scale and accounts for a greater proportion of the total variability response rate.