Background With the increasing severity of global climate change, extreme rainfall events in China have also shown a significant trend of increasing. In the karst plateau of southwestern China, most of the existing studies are based on daily, monthly or annual data analysis of the region’s extreme rainfall characteristics. Research on the spatial and temporal evolution characteristics of extreme rainfall at the hourly scale are relatively rare. Conducting research on the hourly scale extreme rainfall characteristics in the karst plateau area can provide a basis for understanding and preventing soil erosion.

Methods Based on the 0.5 h satellite rainfall data of global precipitation observation, the extreme rainfall series through the threshold values of different percentile rainfall were extracted, and Mann-Kendall analysis method, climate factor tendency rate method and other analysis methods were used to calculate the relationship between various indicators, and analyze the inter-annual and intra-year spatio-temporal evolution of 1 h rainfall extreme value in the rocky desertification area in Southwest China karst plateau.

Results 1) From 2000 to 2020, the 1 h rainfall in the study area showed a small upward trend, with a tendency rate of 0.535 3 mm/a. There were 3 distinct peak periods of 1 h rainfall with the change of year, and the peak value increased continuously while the interval time decreased. 2) Mann-Kendall test showed that the 1 h rainfall extreme value in the study area showed a significant upward trend from 2019. In addition, the maximum rainfall of 1 h in the study area ranged from 65 to 223 mm, and there was obvious spatial heterogeneity. The spatial distribution characteristics are influenced by the frequency and monthly variation of rainfall intensity. While the spatial distribution of 1 h rainfall extremes with a frequency of 95% was high in the northwest and low in the southeast, the spatial distribution characteristics of 1 h rainfall extremes below this frequency were high in the southeast and low in the northwest. 3) In addition, in spring and summer, 1 h rainfall extremum presented a spatial distribution feature of high southeast and low northwest, and the spatial distribution of 1 h rainfall extremum presented an opposite feature from the end of summer and August. The results showed that the study area will face more intense extreme rainfall hazards. At the same time, when considering the formulation of corresponding prevention and control strategies, the prevention and control standards of the areas with relatively large rainfall in 1 h should be improved.

Conclusions Based on high-resolution sub-hourly satellite data, this study explores the spatio-temporal evolution of hourly rainfall extremes in the rocky desertification area of southwest China from 2000 to 2020. The results are helpful for predicting soil erosion potential caused by rainfall under future climate change conditions. It can provide guidance for improving the risk assessment of soil erosion and secondary disasters under extreme rainfall and formulating prevention and control strategies.