Space-time landslide predictive modelling
Citation
Lombardo, L., Opitz, T., Ardizzone, F., Guzzetti, F., Huser, R. (2020). Space-time landslide predictive modelling. Earth-Science Reviews, 209: 103318. Link to paper
Abstract
This paper proposes a Bayesian modelling framework for the prediction of the spatio-temporal occurrence of landslides caused by weather triggers, building on a Log-Gaussian Cox Process (LGCP) by assuming individual landslides are the result of a point process described by an unknown intensity function. The framework has two stochastic components: a Poisson component modeling observed landslide count for given landslide intensity, and a Gaussian component accounting for spatial distribution of environmental conditions influencing landslide occurrence. The framework was tested in the Collazzone area, Umbria, Central Italy, with a multi-temporal landslide inventory spanning 1941-2014 along with lithological and bedding data. The model successfully captures both spatial patterns driven by terrain characteristics and temporal clustering of landslide events associated with rainfall episodes. This approach represents a significant advancement in space-time landslide forecasting, providing a probabilistic framework that can inform risk management and early warning systems by explicitly modeling both where and when landslides are likely to occur.