Project

Discipline

precursor events; catchment hydrology; self-organized criticality; landslides

Lead
Lay summary
Steep catchments are very heterogeneous structures consisting of different soil types, structures, bedrock formations, and surface topographies that are at least partially unknown. These uncertainties hamper predictions of abrupt shallow landslide that may be triggered by progression of local failures culminating to mass release. In this project we apply a new model that describes landslides as chain reaction of failing elements at different scales, to test systematically the effect of heterogeneities on loading and failure patterns. With this model we can capture abruptness of landslide release and the wide size distribution reported for landslide inventories that are not captured by traditional models of landslide triggering. As final objective we intend to provide ‘a map’ for different combinations of slope heterogeneities and rainfall patterns with high landslide susceptibility.

Direct link to Lay Summary

Last update: 21.02.2013

Active participation

Self-organised

Number	Title	Start	Funding scheme

Rapid shallow landslides triggered by heavy rainfall are ubiquitous hazardous phenomena in steep catchments, causing thread to live and infrastructure on an annual basis. Complex architecture and failure processes at various scales aggravate predictive modeling of time, location, and size of landslide. Due to the abruptness of landslide without clear indication of imminent mass release, it is not possible so far to anticipate time of failure from measured precursor events. In this proposal we intend to apply a new triggering model that can reproduce abruptness of mass release by simulating landslides as progressive failures. For that purpose we implemented concepts of Self-Organized Criticality (SOC) into a hydro-mechanical modeling framework, taking into account organization and interaction of many hillslope elements (soil columns). As a first benchmark of the new model concept we will reproduce frequency/magnitude distributions for landslide inventories collected in the Swiss Pre-Alps. We will relate simulated power-law statistics to catchment geometry and soil properties and provide rules to assess risks of large hazardous mass release applicable for any catchment. Another level of predictability could be obtained by defining thresholds separating between ‘stable’ and ‘unstable’ conditions in a catchment. We will define a phase space spanned by various geomorphological and hydrological properties and define the critical subspace with high landslide risk in a series of numerical experiments. Finally, to simulate precursor events at appropriate scale we will implement fiber bundle models to represent local failures at base and between soil columns. The model could be applied to reproduce initial loading patterns and predict development to mass release. With the three model applications planned for the final year of the thesis (landslide frequency/magnitude, critical parameter space, modeling precursor statistics) we are confident to provide a modeling framework required for better understanding and prediction of landslide triggering.