Non-Stationary Estimation of the Return Period of Shallow Landslide Triggering

Landslides are typical non stationary processes, since they are a mass wasting process bringing toward equilibrium the earth surface. While at the hillslope scale sources of non-stationarity are inherent to the nature of the process, at the regional scale one of the potentially causes of non-stationarity of landslide frequency is climate change. In this paper, we propose a Monte Carlo based approach, to compute landslide triggering probability, and its potential variation due to climate change. The approach combines stochastic rainfall generation and hydrological and slope stability models. The stochastic model is calibrated on the basis current observations; then, factor of changes in rainfall characteristics are computed based on regional climate model (RCM) projections, and the stochastic rainfall generator is adjusted accordingly. The modified rainfall input is finally used to estimate the potential modification of landslide-triggering probabilities related to changes in precipitation patterns. Application examples conducted to a real case-study area in Italy, demonstrate the usefulness of the proposed approach, which enables to make preliminary projections of the effect of rainfall changes on the frequency of landslide triggering.