As reported by national and global natural disaster databases, numerous damaging events linked to floods and landslides occur in Italy every year, confirming that the country is among the most exposed in Europe to hydro-geomorphological hazards. Since rainfall is the main triggering factor of these phenomena and its regime is changing due to climate change, the study of these processes is becoming increasingly relevant, particularly in Italy. Cities located downstream of mountain catchments are especially vulnerable to fluvial floods, where sediment yield significantly amplifies the hazard in terms of both economic losses and casualties. Mountain catchments produce large amounts of sediments following slope erosion, and the accumulation of such material in critical river sections can alter hydraulic behaviour and increase flood risk. Therefore, the quantification of erosion rates, sediment yield and water discharge is crucial for urban planning, and for the development of mitigation strategies and emergency plans. Over the years, different erosion models have been developed to estimate sediment processes. Some of them are empirical, such as the USLE and RUSLE models, and are based on topographic variables, land use and vegetation cover. In recent times, many numerical and physically-based models have also been conceived, including OpenLISEM and SWAT, which simulate both sediment erosion and transport. However, these models present limitations, as they require several input parameters that are often spatially and temporally variable, and not always easy to measure or estimate. Sometimes the hydraulic processes are even decoupled from the sediment processes, with sediment yield provided only as a boundary condition. Moreover, most existing physically-based models were designed for rangeland or agricultural areas rather than mountain catchments, reducing their accuracy in predicting runoff and sediment fluxes. Another limitation is that many models are not spatially distributed across the entire catchment but instead focus on channel scale or outlet results, frequently requiring a prior delineation of drainage networks and water bodies. To overcome these issue and provide a complete tool, the Politecnico di Milano has developed an innovative physically-based and spatially distributed model named SMART-SED (Sustainable MAnagement of sediment transpoRT in responSE to climate change conditions). SMART-SED simulates hydrological processes such as water discharge, erosion, sediment transport, infiltration, evapotranspiration and snowmelt at catchment scale, adopting a coherent and efficient semi-implicit numerical approach for fluid dynamics. Conceptually, the model improves on traditional methods by automatically identifying drainage zones instead of relying on a static a priori division between slope and channel cells. It requires only a limited set of simple input data and can statistically downscale soil maps to the desired resolution. In this thesis, SMART-SED was further developed to enhance its applicability to a range of geological and climatic contexts and to extend its use beyond sediment dynamics in mountain catchments to fields such as dam sedimentation management and climate-related hazard assessment. The model was first calibrated and validated in a well-monitored case study in Northern Italy, and subsequently applied to diverse catchments in the Alps, Apennines and Pyrenees. Results demonstrated its robustness and adaptability, successfully reproducing sediment transport regimes dominated by bedload in steep carbonate and crystalline terrains, as well as suspended load in fine-grained basins. Moreover, the SMART-SED model showed strong potential in reservoir management applications, by predicting sediment deposition patterns and estimating reservoir filling rates, an increasingly relevant task where storage capacity, hydropower efficiency or infrastructure safety are threatened by sedimentation. To support the modelling outcomes with field evidence, a geomorphological analysis of the case studies was carried out coupled with comparisons with data collected by both traditional and innovative sediment monitoring techniques. Among these, the analysis of a pebble tracing dataset offered valuable insights into particle motion and formed the basis for the development of a river-specific bedload transport model. A preliminary landslide module was also developed and integrated into SMART-SED, enabling the estimation of the factor of safety at catchment scale. This represents a first step toward simulating landslide triggering and propagation mechanisms. In its current state, the model already provides a promising basis for hazard mitigation planning. Future developments will further enhance its capacity to assess landslide susceptibility, ultimately allowing SMART-SED to offer a comprehensive and integrated approach to hydro-geomorphological hazard modelling at catchment scale. Together, these efforts contribute to advancing process understanding, model performance and practical applications in sediment management and hazard mitigation.
Come riportato da banche dati nazionali e globali sui disastri naturali, ogni anno in Italia si registrano numerosi eventi dannosi legati a inondazioni e frane, confermando che il Paese è uno dei più esposti in Europa al dissesto idrogeologico. Considerando che il principale fattore scatenante di questi fenomeni è la pioggia, e che il regime delle precipitazioni sta cambiando a causa dei cambiamenti climatici, lo studio di questi processi naturali sta diventando sempre più rilevante, soprattutto in Italia. In particolare, le città situate a valle di bacini montani sono specialmente vulnerabili alle inondazioni fluviali, poiché la portata di sedimenti aumenta significativamente il rischio sia in termini di perdite economiche, sia di vittime. I bacini montani infatti producono grandi quantità di sedimenti a seguito dell’erosione dei pendii, e l’accumulo di tale materiale in tratti critici dei fiumi può alterarne il comportamento idraulico e aumentare il rischio di alluvioni. Pertanto, la quantificazione dei tassi di erosione, della produzione di sedimenti e della portata idrica è fondamentale per la pianificazione urbana, lo sviluppo di strategie di mitigazione e i piani di emergenza. Negli ultimi anni sono stati sviluppati diversi modelli di erosione e trasporto solido per valutare l'entità di questi processi a diverse scale spaziali e temporali. Alcuni di questi sono empirici, come i modelli USLE e RUSLE, e si basano su variabili topografiche, uso del suolo e copertura vegetale. Recentemente sono stati sviluppati anche modelli numerici e fisicamente basati, tra cui OpenLISEM e SWAT, che simulano sia l’erosione dei sedimenti sia il loro trasporto. Tuttavia, questi modelli presentano dei limiti, in quanto richiedono numerosi parametri di input spesso variabili nello spazio e nel tempo e non sempre facilmente misurabili o stimabili. I processi idraulici sono frequentemente separati dai processi sedimentari, e la produzione di sedimenti, in alcuni casi, viene fornita solo come condizione al contorno. Inoltre, la maggior parte dei modelli fisico-basati esistenti è stata ideata per pascoli o aree agricole piuttosto che per bacini montani, riducendo la loro accuratezza nella previsione della portata e del flusso di sedimenti. Un ulteriore limite è che molti modelli non sono distribuiti spazialmente sull’intero bacino, ma si concentrano invece sulla scala del canale o sui risultati nella sezione di chiusura del bacino, richiedendo spesso in entrata la rete di drenaggio e i corpi idrici. Per risolvere queste criticità, il Politecnico di Milano ha recentemente sviluppato un modello innovativo di erosione e trasporto solido, fisicamente basato e distribuito spazialmente, chiamato SMART-SED (Sustainable MAnagement of sediment transpoRT in responSE to climate change conditions). SMART-SED simula processi idrologici quali portata, erosione, trasporto dei sedimenti, infiltrazione, evapotraspirazione e scioglimento della neve a scala di bacino, adottando un approccio numerico semi-implicito coerente ed efficiente per la dinamica dei fluidi. Concettualmente, il modello migliora i metodi tradizionali identificando automaticamente le zone di drenaggio, invece di fare affidamento su una divisione statica a priori tra celle di pendio e celle di canale. Richiede solo un set limitato di dati di input semplici e può aumentare la risoluzione delle mappe granulometriche con metodi statistici. In questa tesi, SMART-SED è stato ulteriormente sviluppato per aumentarne l’applicabilità a diversi contesti geologici e climatici, e per estenderne l’uso, includendo ambiti come la gestione della sedimentazione nei bacini artificiali e la valutazione dei rischi legati al cambiamento climatico. Il modello è stato inizialmente calibrato e validato in un caso ben monitorato nel Nord Italia, e successivamente applicato a bacini diversi nelle Alpi, negli Appennini e nei Pirenei. I risultati hanno dimostrato la sua robustezza e adattabilità, riproducendo con successo regimi dominati dal trasporto dei sedimenti al fondo in substrati caratterizzati da rocce carbonatiche e cristalline e in contesti caratterizzati da alte pendenze, così come in bacini dominati dal trasporto in sospensione di sedimenti fini. Inoltre, il modello SMART-SED ha mostrato un forte potenziale nelle applicazioni sugli invasi artificiali, stimando il loro tasso di riempimento, un compito sempre più rilevante dove la capacità di stoccaggio dell'acqua, l’efficienza idroelettrica e la sicurezza delle infrastrutture sono minacciate dalla sedimentazione. Per supportare i risultati della modellazione con evidenze sul campo, è stata condotta un’analisi geomorfologica dei casi studio, accoppiata a confronti con dati di monitoraggio raccolti tramite tecniche sia tradizionali sia innovative. Tra queste, l’analisi di un dataset di tracciamento di ciottoli ha fornito informazioni preziose sul movimento dei sedimenti e ha costituito la base per lo sviluppo di un modello specifico per il trasporto di materiale al fondo in uno dei torrenti oggetto di studio. È stato infine sviluppato e integrato in SMART-SED un modulo frane preliminare, consentendo la stima del fattore di sicurezza a scala di bacino. Questo rappresenta un primo passo verso la simulazione dei meccanismi di innesco e propagazione delle frane. Nella sua configurazione attuale, il modello fornisce quindi già una base promettente per la pianificazione della mitigazione del rischio. Sviluppi futuri ne miglioreranno ulteriormente la capacità di valutare la suscettibilità alle frane, consentendo infine a SMART-SED di offrire un approccio completo e integrato alla modellazione dei rischi idro-geomorfologici a scala di bacino. Nel complesso, questi sforzi contribuiscono ad avanzare la comprensione dei processi, le potenzialità della modellazione e le applicazioni pratiche per la gestione dei sedimenti e la mitigazione del rischio.
Innovative approaches to soil erosion and sediment transport modelling
Corti, Monica
2024/2025
Abstract
As reported by national and global natural disaster databases, numerous damaging events linked to floods and landslides occur in Italy every year, confirming that the country is among the most exposed in Europe to hydro-geomorphological hazards. Since rainfall is the main triggering factor of these phenomena and its regime is changing due to climate change, the study of these processes is becoming increasingly relevant, particularly in Italy. Cities located downstream of mountain catchments are especially vulnerable to fluvial floods, where sediment yield significantly amplifies the hazard in terms of both economic losses and casualties. Mountain catchments produce large amounts of sediments following slope erosion, and the accumulation of such material in critical river sections can alter hydraulic behaviour and increase flood risk. Therefore, the quantification of erosion rates, sediment yield and water discharge is crucial for urban planning, and for the development of mitigation strategies and emergency plans. Over the years, different erosion models have been developed to estimate sediment processes. Some of them are empirical, such as the USLE and RUSLE models, and are based on topographic variables, land use and vegetation cover. In recent times, many numerical and physically-based models have also been conceived, including OpenLISEM and SWAT, which simulate both sediment erosion and transport. However, these models present limitations, as they require several input parameters that are often spatially and temporally variable, and not always easy to measure or estimate. Sometimes the hydraulic processes are even decoupled from the sediment processes, with sediment yield provided only as a boundary condition. Moreover, most existing physically-based models were designed for rangeland or agricultural areas rather than mountain catchments, reducing their accuracy in predicting runoff and sediment fluxes. Another limitation is that many models are not spatially distributed across the entire catchment but instead focus on channel scale or outlet results, frequently requiring a prior delineation of drainage networks and water bodies. To overcome these issue and provide a complete tool, the Politecnico di Milano has developed an innovative physically-based and spatially distributed model named SMART-SED (Sustainable MAnagement of sediment transpoRT in responSE to climate change conditions). SMART-SED simulates hydrological processes such as water discharge, erosion, sediment transport, infiltration, evapotranspiration and snowmelt at catchment scale, adopting a coherent and efficient semi-implicit numerical approach for fluid dynamics. Conceptually, the model improves on traditional methods by automatically identifying drainage zones instead of relying on a static a priori division between slope and channel cells. It requires only a limited set of simple input data and can statistically downscale soil maps to the desired resolution. In this thesis, SMART-SED was further developed to enhance its applicability to a range of geological and climatic contexts and to extend its use beyond sediment dynamics in mountain catchments to fields such as dam sedimentation management and climate-related hazard assessment. The model was first calibrated and validated in a well-monitored case study in Northern Italy, and subsequently applied to diverse catchments in the Alps, Apennines and Pyrenees. Results demonstrated its robustness and adaptability, successfully reproducing sediment transport regimes dominated by bedload in steep carbonate and crystalline terrains, as well as suspended load in fine-grained basins. Moreover, the SMART-SED model showed strong potential in reservoir management applications, by predicting sediment deposition patterns and estimating reservoir filling rates, an increasingly relevant task where storage capacity, hydropower efficiency or infrastructure safety are threatened by sedimentation. To support the modelling outcomes with field evidence, a geomorphological analysis of the case studies was carried out coupled with comparisons with data collected by both traditional and innovative sediment monitoring techniques. Among these, the analysis of a pebble tracing dataset offered valuable insights into particle motion and formed the basis for the development of a river-specific bedload transport model. A preliminary landslide module was also developed and integrated into SMART-SED, enabling the estimation of the factor of safety at catchment scale. This represents a first step toward simulating landslide triggering and propagation mechanisms. In its current state, the model already provides a promising basis for hazard mitigation planning. Future developments will further enhance its capacity to assess landslide susceptibility, ultimately allowing SMART-SED to offer a comprehensive and integrated approach to hydro-geomorphological hazard modelling at catchment scale. Together, these efforts contribute to advancing process understanding, model performance and practical applications in sediment management and hazard mitigation.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi_Monica_Corti.pdf
accessibile in internet per tutti a partire dal 24/09/2028
Descrizione: Tesi
Dimensione
380.33 MB
Formato
Adobe PDF
|
380.33 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/244091