Semiempirical modeling of the transient response of pore pressure to rainfall and snowmelt in a dormant landslide

Abstract

AbstractThis study proposes a simplified, semiempirical hydrological model of the pressure response in a landslide. Various hydrometeorological variables were measured in a dormant landslide underlain by soft sedimentary rocks in a high snowfall area in central Japan. We assessed whether the short-term response of the pore pressure to water inputs by rainfall and/or snowmelt can be predicted using a modified linear diffusion model that describes the seasonal trends of the pore pressure. We applied field observations and regression analysis to characterize the buffering function of the vadose zone. The apparent hydraulic conductivity and hydraulic diffusivity in the vadose zone were low during summer and high during winter and can be described by an empirical function of the pore pressure prior to an event. The pattern of pressure waveforms was accurately reproduced for various water inputs via empirical parameter functions. The model can also reproduce the pore-pressure peaks at different depths and locations within the same landslide with practical certainty. This modeling demonstrated a method that has the potential to predict future landslide instability as long as data from pore-pressure monitoring and water input forecasting are available.