Atmospheric pressure compared to rainfall as landslide triggering factors along a hillslope

Landslides are one of the sources of natural hazards that cause damages and losses but also shapes the landscape. A better understanding the factors triggering or pre-conditioning landslide occurrence is therefore critical for risk assessment, with implications for hillslope erosion and landscape dynamics Triggering of catastrophic landslides is generally associated with events such as earthquakes or intense rainfalls. In Taiwan, a minimum of 22,705 landslides were reported during the typhoon Morakot in 2009 (Lin et al., 2011). Landslides triggered during storms are generally associated to the intensity and cumulated amount of rainfall, as water infiltration destabilize slopes (Iverson, 2000). However, a correlation has also been reported between slope stability and the change in atmospheric pressure (Schulz, 2009). Indeed, a change in air-pressure can lead in a readjustment in pore pressure, and cause fluid movements normal to the surface. The aim of this study is to characterize the effect of atmospheric pressure changes and define its specific contribution on slope stability when combined with rainfall

A 2-dimensional analytical model has been developed based on diffusion equations to describe the destabilization induced by water infiltration and atmospheric pressure changes induced by typhoons. As both mechanisms are function of pore pressure, and especially groundwater pore pressure, the water table within a finite-length hillslope is modelled using Townley’s (1995) analytical expression of water flow in a unconfined aquifer. The hillslope itself is a simple tilted half-space with a water divide at the top and a river at the toe forcing the water table to the surface. Slope stability is inferred through a safety factor computed using the coulomb criterion. Both rainfall infiltration and air pressure modify pore pressure through a diffusion process. While rainfall increases water table height and induce large increases in pore pressure within days or hours, , we show that atmospheric-induced pore pressure change is instantaneous and can occur even if the hillslope is fully saturated.

The model allows to separate the hillslope response into two regimes, upslope or downslope, where the destabilization is mainly linked to rainfall or to atmospheric pressure change, respectively.  Our results suggest that landslide occurring during storms in the downstream part of the hillslope are likely candidate for being triggered by atmospheric pressure change, in particular if the storm occurs with a humid initial condition. We show that the effect of atmospheric pressure changes is not negligible. On contrary, it is crucial to define the amplitude, timing and geometry of the hillslope instability, especially when combined to rainfall.