The study objective was to derive a susceptibility model for shallow landslides that could include process-related non-stationary variables, to be adaptable to climate changes. We selected the territory of the Mont-Emilius and Mont-Cervin Mountain Communities (northern Italy) as the study area. To define summary variables related to landslide predisposing and triggering processes, we investigated the relationships between landslide occurrences and intense rainfall and snowmelt events (period 1991-2020). For landslide susceptibility mapping, we set up a Generalized Additive Model. We defined a reference model through variable penalization (relief, NDVI, land cover and geology predictors). Similarly, we optimized a model including the climate variables, checking their smooth functions to ensure physical plausibility. Finally, we validated the optimized model through a k-fold cross-validation and performed an evaluation based on contingency tables, area under the receiver operating characteristic curve (AUROC) and variable importance (decrease in explained variance). The climate variables that resulted as being statistically and physically significant are the effective annual number of rainfall events with intensity-duration characteristics above a defined threshold (EATean) and the average number of melting events occurring in a hydrological year (MEn). In the optimized model, EATean and MEn accounted for 5% of the explained deviance. Compared to the reference model, their introduction led to an increase in true positive rate and AUROC of 2.4% and 0.8%, respectively. Also, their inclusion caused a transition of the vulnerability class in 11.0% of the study area. The k-fold validation confirmed the statistical significance and physical plausibility of the meteorological variables in 74% (EATean) and 93% (MEn) of the fitted models. Our results demonstrate the validity of the proposed approach to introduce process-related, non-stationary, physically-plausible climate variables within a shallow landslide susceptibility analysis. Not only do the variables improve the model performance, but they make it adaptable to map the future evolution of landslide susceptibility including climate changes.
Keywords: Aosta Valley; Climate variables; Flowslides; Generalized Additive Models; Slides in soil; Snow Water Equivalent.
Copyright © 2021 Elsevier B.V. All rights reserved.