Process based models of landslides' triggering. My talk at Cosenza on November the 7th

Last November the 7th, I was invited to give a talk about process-based models of landslides triggering.  My talk (you can find it here) was mainly about the key elements for describing hillslope hydrology and stability, and inherits a lot on my previous talk at the Praia a Mare Summer School.
However there is some new material, and especially I mentioned the work made for embedding GEOtop into OMS, and the recent work we made for the IWL Round Robin.

At the seminar there were other three interesting topics by Giovanna Capparelli, Fausto Guzzetti (see also the website of his group) and Roberto Greco. The latter, one of the author of the IWL organisers and designer of the IWL3  "round robin". At the round robin (we did not arrive in time for presenting our work) some Spanish guys won (with the best predictions of the suctions of the experiment) by running a model that solves the coupled Richards and Stress equations.  As it will result for giving a look to the slides, our trial was not successful: but because we tried to reproduce contemporarily both suction and water content. We fail. This can only means that the soil water retention curves used were not appropriate. We do not know however if the case is that the vanGenuchten parameterisation is failing (and/or Mualem that gives hydraulic conductivity) or if the parameterisation is actually time-dependent, as the strategy used by the Spanish guys seems to suggest. 

In my opinion, it is reasonable that soil water retention curves vary after soil starts to move, but I  am not sure that a dynamical parameterisation would be required before the failure time.