TORRENT: Timely fOrecasting and Risk management of Rapid dEbris flows in mouNTain catchments
ProgettoDebris flows are rapid movements of highly concentrated mixtures of water, sediment, and coarse debris that propagate along steep mountain channels andslopes, causing severe damage to infrastructure and buildings, as well as loss of lives. Triggered mainly by intense rainfall, they arise from runoff andinfiltration on hillslopes, which may induce slope failures. Both sediment-laden runoff and failed soil masses accelerate downslope, forming surges that traveldownstream. During routing, these surges entrain additional debris and water, increasing their volume up to an order of magnitude.
Their rapid onset, large transported sediment volumes (up to 10^5 m3), and high velocities make debris flows extremely hazardous. Their frequency is alsoincreasing significantly due to climate change, especially because intense rainstorms are becoming more common and because rockfalls and landslidessupply larger amounts of loose debris. Reliable risk-management strategies are therefore urgently needed.
This project aims to develop and validate a physics-based modelling chain for predicting debris flows, estimating the initial solid–liquid volume, andsimulating downstream routing within an uncertainty framework. The chain is integrated with an interdisciplinary component addressing the humandimensions of hydrogeomorphic risk and can support both early warning and risk assessment.
The modelling framework includes hydrological, triggering, and routing models, selected according to the type of debris flow. For channelized debris flows, ahydrological rainfall–runoff model is coupled with a mass-balance triggering model; the resulting solid–liquid hydrograph is routed downstream through ahydraulic model solving flow equations with suitable rheology. For unchannelized debris flows, a hydrological infiltration model is coupled with a slope-stability model to estimate the failed mass, which is then routed downstream using hydraulic models comparable to those used for channelized flows.Multiple hydrological and hydraulic models and parameterizations will be applied and compared.
The modelling chains will be validated through case studies where monitoring stations are operational. Rainfall inputs will come from stochastic and AI-based nowcasting models fed by radar estimates or rain-gauge observations. For hazard and risk assessment, including hazard propagation, exposureanalysis, and asset-damage estimation, uncertainties will be addressed through a multi-model probabilistic approach. This framework will also integratesocial and behavioural dimensions through quantitative vulnerability modifiers and scenario-based adjustments, allowing comparison between conventionalengineering-based damage estimates and socially informed risk scenarios.
Stakeholders and citizens will be engaged through participatory methods to demonstrate the performance of the complete risk-management chain.
Their rapid onset, large transported sediment volumes (up to 10^5 m3), and high velocities make debris flows extremely hazardous. Their frequency is alsoincreasing significantly due to climate change, especially because intense rainstorms are becoming more common and because rockfalls and landslidessupply larger amounts of loose debris. Reliable risk-management strategies are therefore urgently needed.
This project aims to develop and validate a physics-based modelling chain for predicting debris flows, estimating the initial solid–liquid volume, andsimulating downstream routing within an uncertainty framework. The chain is integrated with an interdisciplinary component addressing the humandimensions of hydrogeomorphic risk and can support both early warning and risk assessment.
The modelling framework includes hydrological, triggering, and routing models, selected according to the type of debris flow. For channelized debris flows, ahydrological rainfall–runoff model is coupled with a mass-balance triggering model; the resulting solid–liquid hydrograph is routed downstream through ahydraulic model solving flow equations with suitable rheology. For unchannelized debris flows, a hydrological infiltration model is coupled with a slope-stability model to estimate the failed mass, which is then routed downstream using hydraulic models comparable to those used for channelized flows.Multiple hydrological and hydraulic models and parameterizations will be applied and compared.
The modelling chains will be validated through case studies where monitoring stations are operational. Rainfall inputs will come from stochastic and AI-based nowcasting models fed by radar estimates or rain-gauge observations. For hazard and risk assessment, including hazard propagation, exposureanalysis, and asset-damage estimation, uncertainties will be addressed through a multi-model probabilistic approach. This framework will also integratesocial and behavioural dimensions through quantitative vulnerability modifiers and scenario-based adjustments, allowing comparison between conventionalengineering-based damage estimates and socially informed risk scenarios.
Stakeholders and citizens will be engaged through participatory methods to demonstrate the performance of the complete risk-management chain.