Abstract: Objectives: Rainfall-induced geological disasters often result in significant human and property losses, and rainfall thresholds can effectively reduce the hazards of geological disasters. However, limited rainfall data in mountainous areas lead to low accuracy of disaster rainfall thresholds. Establishing rainfall thresholds in mountainous areas with scarce data is an unresolved problem. Methods: In this study, by downscaling Global Precipitation Measurement (GPM) data to a resolution of 1 km and verifying its validity with ground station rainfall. By combining disaster data, we have extracted rainfall events and obtain critical thresholds for different types of disasters and different precisions in Mao County based on the frequency method. Furthermore, this study also analyzed the evolution of the threshold curves over the years following the earthquake. Results: Rainfall-type landslides in Mao County are mainly controlled by long-duration weak rainfall, while debris flow disasters are primarily controlled by short-duration intense rainfall, and the critical threshold equations can be respectively: E＝4.17D^0.18 (8<D<868) and E＝3.93D^0.24 (10<D<441). The rainfall threshold after downscaling is lower than the unscaled threshold, and the critical threshold equation can be expressed as: E＝4.09D^0.19 (8<D<868) and E＝3.96D^0.21 (4<D<736). The rainfall events that triggered geological disasters in Mao County after the earthquake have shift from long-duration weak rainfall to short-duration intense rainfall control, and rainfall thresholds shows an increasing trend over time. Conclusions: Downscaling can effectively enhance the spatial resolution of rainfall products and improve their ability to capture rainfall events. The critical thresholds established using downscaled data can serve as the minimum indicators for disaster monitoring and early warning in Mao County.