Mario L. V. Martina

Implementation of a Probabilistic Model of Landslide Occurrence on a Civil Protection Alert System at Regional Scale

In this work we describe the deployment of a probabilistic model for landslide occurrence to support the decisional process within the Civil Protection Agency of the Emilia-Romagna Region (Italy). The model is based on the Bayes’ theorem and computes the conditional probability to have a landslide in response to a rainfall event of given duration and intensity. Unlike most the methods used to date to define rainfall thresholds, our probabilistic approach takes into account of the uncertainty in the data and provides a quantitative assessment of the threshold reliability. The model was recently applied to evaluate the landslide probability in the Emilia-Romagna region, taking advantage of the landslide historical archive that includes about 3,000 landslides whose date of occurrence is known with daily accuracy. For the operational use, we fitted the isolines of landslide probability on the rainfall duration-intensity plane and created two easy-to-read probability charts useful at two different scales: the rain gauge reference area (about 65 km2) and the regional alert zone (about 2,600 km2). The user combines the rainfall data recorded by the automated rain gauge network and the rainfall forecasts to characterize the rainfall event, and plot the rainfall path on the probability charts to evaluate the probability of failure. A sample application of the model is described.