Mateo Moreno , Luigi Lombardo , Stefan Steger, Lotte de Vugt, Thomas Zieher, Alice Crespi, Francesco Marra, Cees J. van Westen , Thomas Opitz

Shallow landslides are geomorphic hazards in mountainous terrains across the globe. Their occurrence can be attributed to the interplay of static and dynamic landslide controls. In previous studies, data-driven approaches have been employed to model shallow landslides on a regional scale, focusing on analyzing the spatial aspects and time-varying conditions separately. Still, the joint assessment of shallow landslides in space and time using data-driven methods remains challenging. This study aims to predict the occurrence of precipitation-induced shallow landslides in space and time within the Italian province of South Tyrol (7,400 km²). In this context, we investigate the benefits of considering precipitation leading to landslide events as a functional predictor, in contrast to conventional approaches that treat precipitation as a scalar predictor. We built upon hourly precipitation analysis data and past landslide occurrences from 2012 to 2021. We implemented a novel functional generalized additive model to establish statistical relationships between the spatiotemporal occurrence of shallow landslides, various static factors included as scalar predictors, and the hourly precipitation pattern preceding a potential landslide used as a functional predictor. We evaluated the resulting predictions through several cross-validation routines, achieving high model performance scores. To showcase the model capabilities, we performed a hindcast for the storm event in the Passeier Valley on August 4th and 5th, 2016. This novel approach enables the prediction of landslides in space and time for large areas by accounting for static and dynamic functional landslide controls, seasonal effects, statistical uncertainty, and underlying data limitations.