Objective The mechanism of super large landslides in hydropower station reservoir areas is extremely complex. It is difficult to accurately predict the deformation stage of landslides using existing early warning and prediction methods.

Methods In order to conduct more accurate early warning and prediction research on it, this article adopts the borehole deep displacement monitoring method. An improved early warning indicator of kinetic energy and kinetic energy change rate for inclinometers was proposed, which based on the monitoring principle of inclinometers. On the basis of deep displacement monitoring data of the DH super large landslide, the trend and time limit values of cumulated displacement depth curve, deformation rate, deformation acceleration, kinetic energy and its variation rate were analyzed. A method was established to identify the stages of landslide deformation evolution.

Results The research results indicate that landslides have multiple sliding surfaces in different zones, and there is an overall sliding trend above the landslide sliding surface. During the process of landslide from uniform deformation stage to accelerated deformation stage, the rate of kinetic energy change significantly increases and fluctuates significantly. The identification method for landslide deformation evolution stages based on deep displacement kinetic energy and kinetic energy change rate warning indicators is established.

Conclusions This method is more accurate in dividing each stage and identifying warning time compared to the rate and acceleration curves. Not only can it better identify the deformation stage of super lager landslides in hydropower station reservoir areas, but it can also provide a reference basis for determining the deformation stage of such landslides.