A new prediction method for the occurrence of landslides based on the time history of tilting of the slope surface

In recent decades, early warning systems using tilt sensors to predict the occurrence of landslides have been developed and employed in slope monitoring due to the simple installation and low cost of these systems. However, few studies were carried out to investigate the tilting behaviors of landslides, and the prediction methods for the occurrence of slope failure based on tilting measurements also demand detailed investigations. In this paper, pre-failure tilting behaviors of slopes were investigated by performing a series of model tests as well as a field test. The test results reveal a linear relationship between the reciprocal tilting rate and time during the acceleration stage of tilting before slope failure. Furthermore, an equation for this linear relation was also proposed. By approximating the reciprocal tilting rate to be 0/(o), the slope failure time can be forecasted using the proposed equation, and the predicted failure time is consistent with the actual slope failure time recorded in this study. Additionally, the correlation between the tilting rate and remaining time before slope failure in logarithmic space was also studied, and most of the data is situated in a region with boundaries. Based on this region, it is possible to anticipate the remaining time before slope failure at an arbitrary tilting rate in the acceleration stage. Conclusively, this paper provides comprehensive investigations on the correlation between the pre-failure tilting behaviors and duration time before landslides, and also introduces a method to potentially predict the occurrence of slope failure based on the slope tilting measurement.