This article presents a study on the hydrometeorological thresholds of reservoir-induced landslides, based on subsurface strain observations. The research focuses on the Three Gorges Reservoir (TGR) region in China, where a giant reservoir landslide occurred. The study uses high-resolution fiber optic sensing to measure subsurface strain over a hydrological year, from February 2021. The spatiotemporal strain profile helps identify slip zones and potential drivers, indicating that high-intensity short-duration rainstorms control landslide kinematics. Considering the time lag effect, the study reexamines and quantifies potential controls of accelerated movements using a data-driven approach, revealing an immediate response of landslide deformation to extreme rainfall with a zero-day shift. A landslide prediction model is constructed using the boosting decision tree (BDT) algorithm, incorporating daily rainfall, rainfall intensity, reservoir water level, water level fluctuations, and slip zone strain time series. The results indicate that landslide acceleration is most likely to occur under mid-low water levels and large-amount, high-intensity rainfalls. The study provides a practical and reliable pathway for georisk early warning based on subsurface observations, particularly in the context of enhanced extreme weather events. The study highlights the importance of considering the synergetic effect of rainfall and reservoir water levels in determining hydrometeorological thresholds for landslide early warning. The findings suggest that subsurface strain observations can provide more accurate and reliable threshold criteria for landslide early warning compared to surface displacement measurements. The study also emphasizes the need for further research to improve the accuracy and reliability of landslide prediction models.This article presents a study on the hydrometeorological thresholds of reservoir-induced landslides, based on subsurface strain observations. The research focuses on the Three Gorges Reservoir (TGR) region in China, where a giant reservoir landslide occurred. The study uses high-resolution fiber optic sensing to measure subsurface strain over a hydrological year, from February 2021. The spatiotemporal strain profile helps identify slip zones and potential drivers, indicating that high-intensity short-duration rainstorms control landslide kinematics. Considering the time lag effect, the study reexamines and quantifies potential controls of accelerated movements using a data-driven approach, revealing an immediate response of landslide deformation to extreme rainfall with a zero-day shift. A landslide prediction model is constructed using the boosting decision tree (BDT) algorithm, incorporating daily rainfall, rainfall intensity, reservoir water level, water level fluctuations, and slip zone strain time series. The results indicate that landslide acceleration is most likely to occur under mid-low water levels and large-amount, high-intensity rainfalls. The study provides a practical and reliable pathway for georisk early warning based on subsurface observations, particularly in the context of enhanced extreme weather events. The study highlights the importance of considering the synergetic effect of rainfall and reservoir water levels in determining hydrometeorological thresholds for landslide early warning. The findings suggest that subsurface strain observations can provide more accurate and reliable threshold criteria for landslide early warning compared to surface displacement measurements. The study also emphasizes the need for further research to improve the accuracy and reliability of landslide prediction models.