A model building strategy is tested to assess the susceptibility for extremeclimatic events driven shallow landslides. In fact, extreme climatic inputs such as stormstypically are very local phenomena in the Mediterranean areas, so that with the exceptionof recently stricken areas, the landslide inventories which are required to train any stochasticmodel are actually unavailable. A solution is here proposed, consisting in training asusceptibility model in a source catchment, which was implemented by applying the binarylogistic regression technique, and exporting its predicting function (selected predictorsregressed coefficients) in a target catchment to predict its landslide distribution. To test themethod, we exploit the disaster that occurred in the Messina area (southern Italy) on 1October 2009 where, following a 250-mm/8-h storm, approximately two thousand debrisflow/debris avalanches landslides in an area of 21 km2 triggered, killing 37 people andinjuring more than 100, and causing 0.5 M € worth of structural damage. The debris flowsand debris avalanches phenomena involved the thin weathered mantle of the Varisican lowto high-grade metamorphic rocks that outcrop in the eastern slopes of the PeloritaniMounts. Two 10-km2-wide stream catchments, which are located inside the storm corearea, were exploited: susceptibility models trained in the Briga catchment were testedwhen exported to predict the landslides distribution in the Giampilieri catchment. Theprediction performance (based on goodness of fit, prediction skill, accuracy and precisionassessment) of the exported model was then compared with that of a model prepared in theGiampilieri catchment exploiting its landslide inventory. The results demonstrate that thelandslide scenario observed in the Giampilieri catchment can be predicted with the samehigh performance without knowing its landslide distribution: we obtained, in fact, a very poor decrease in predictive performance when comparing the exported model to the nativerandom partition-based model.
|Numero di pagine||39|
|Stato di pubblicazione||Published - 2014|
All Science Journal Classification (ASJC) codes