Geotechnical and Hydrological Data Assimilation in Numerical Slope Stability Models

Numerical models of slope stability may be applied to large areas in a way that is useful for early warning applications provided that some basic requirements are met. Among them, very important is the correct assimilation of soil mechanical and hydrological parameters, such as saturated hydraulic conductivity, cohesion and friction angle. Almost every single numerical parameter used in the numerical models shows a strong variability in space and cannot be directly measured at the same resolution of the finite element model. Therefore, an averaging and interpolation scheme must be devised which relies on the field measurements and on some statistical analysis to determine suitable probability distributions to be used in Monte Carlo analysis. In this presentation we report on the application of a complete methodology for field and laboratory testing of geo-hydrological parameters, their statistical analysis in terms of probability distributions and data assimilation into the High Resolution Slope Stability Simulator (Rossi et al., 2013) with the aim of implementing a near-real-time application of an early warning system for shallow landslides induced by rainfall in the region of Tuscany, Italy. Statistics on the main model parameters are presented and discussed as related to the main soil types found in the region, including grain size, mineralogical composition, shear strength, matric suction and hydraulic conductivity. Some preliminary data on vegetation effects are also presented, with reference to root effects and apparent root cohesion implementation in the numerical model. Some simulation examples are also presented.