Landwarn: An Operative Early Warning System for Landslides Forecasting Based on Rainfall Thresholds and Soil Moisture

The Umbria Region is one of the Italian areas that are most prone to landslides and floods. For early-warning procedures aimed at the reduction of the hydrogeological risk, the rainfall thresholds represent the main component of the Italian Civil Protection System. To improve the performances of the alert system for landslide risk, the Umbria Region CFD, in cooperation with the Research Institute for Geo-Hydrological Protection (IRPI-CNR), developed and tested a continuous physically-based soil water balance model, addressed to the estimation of soil moisture conditions over the whole regional territory prior to severe landslide events. The relationship established between the maximum cumulative rainfall values and the soil moisture preceding the triggering of landslides allows to dynamically adjust the rainfall thresholds as a function of the estimated soil moisture values. The main components of the new early-warning system (named LANDWARN) for landslide risk prevention operating at the CFD are: (1) the observed and 72 h-predicted cumulative rainfall from the dense regional hydrometeorological network; (2) a local area meteorological model, known as COSMO ME; (3) the soil water balance model by Brocca et al. 2008; and (4) a web-based Information System. All these elements are arranged in a MATLAB-based flux of operations. Nowadays, the system is still under development and implemented at three different scales: (1) for a specific rockslide site, where a real-time extensometer network is available; (2) for 110 high-risk landslide sites located across the whole regional territory; and (3) for the whole regional territory over a regular grid. Further steps of the implementation are currently in progress. In particular, a GIS-based physically-based model that combines the hydrological information with a simple slope stability analysis is being validated in a test-area of the Umbria Region (Salciarini et al. Nat Hazards 9739–9742, 2011), with the aim of evaluating the possibility of developing an early-warning system capable to take into account the mechanical and physical characteristics of the slopes, besides rainfall and soil moisture.