The Physically-Based Approach for Numerical Modelling of Rainfall-Runoff and Debris Flow Events: New Developments and Lessons Learned from Different Case Studies

In recent decades, climate change has led to an increase in extreme precipitation events, which can make runoff events and debris flows even larger and more frequent than in the past. Depending on the geologic-geomorphologic, hydrologic, and climatic conditions of a mountainous area, various phenomena, such as flash floods, hyperconcentrated flows, debris flows, and shallow landslides can be triggered during intense or extreme rainfall events. These phenomena are generally associated with rainfall events of convective origin characterized by high intensity and short duration, affecting small catchments (a few tens of square kilometers or less) and often characterized by complex morphology. The significant hydrological processes in such cases occur on the same spatial and temporal scale as intense rainfall, with a rapid increase in the water or water-sediment mixture flow level, and high velocities of propagation, resulting in a serious geo-hydrological hazard. In this context, the EU Floods Directive recommends that to reduce the adverse consequences associated with such events, the use of appropriate practices and best available technologies that do not entail excessive costs should be carried out. In line with the above-mentioned directive, an integrated multidisciplinary approach was implemented in this Thesis for the numerical modelling of rainfall-runoff and debris flow events by means of physically-based models. Compared to other techniques, computer simulation is undoubtedly the most advanced technology for the study of rainfall-runoff and debris flows events and their back-analysis. However, the integration of spatially distributed information and specific engineering-geological field investigation data into the modelling is a key step in achieving accurate and reliable numerical modelling of rainfall runoff and debris flow events. This integrated approach was applied to three case studies located in two study areas of Italy, one in northern Tuscany (Apuan Alps) and the other one in southern Tuscany (Mt. Amiata). These two study areas present significant differences from a climatic and geological-geomorphological point of view. Thereby, the erosion phenomena that can be generated by intense or extreme rainfall events are also different. For these reasons, an in-depth study of rainfall-runoff and debris flows events that occurred in such areas, characterised by some peculiarities, offered the opportunity of applying specific methodologies and suitable (and innovative) modelling tools, with the aim of better understanding some of the key processes that characterized such events. The Apuan Alps are a mountainous chain where the elevation ranges between sea level and 2,000 m a.s.l. in a few kilometers, characterized by a rugged morphology with steep slopes and a mean annual rainfall between 2,000-3,000 mm. The analysis and numerical modelling focused on the Vezza catchment, which in turn includes the Cardoso sub-catchment, representing two of the three case studies considered. Due to the morphoclimatic conditions of this study area, heavy rains are frequent, having caused floods, landslides, and debris flows over the centuries, as the catastrophic debris flow that occurred on 19 June 1996 in the Cardoso sub-catchment. Several authors have studied the geological, geomorphological, hydraulic, and geotechnical factors controlling the slope stability of the catchment. However, specific rainfall-runoff simulations, as well as modelling of erosion, transport, and deposition rates by debris flows are lacking so far. In this respect, the objective was to model the debris flow processes of the 1996 event through two different numerical models and approaches, both based on physical conservation principles: the FLO-2D hydrological-hydraulic model and a novel experimental model based on the TRENT2D model. In the latter, the dynamics of debris flows are fully coupled with the rainfall-runoff response of a basin. For the 1996 event, rain gauges were not available within the Cardoso sub-catchment, nor were direct runoff measurements available. Hence, an in-depth hydrological analysis was performed in the Vezza catchment by modelling some well-documented recent rainfall events. For this task, the FLO-2D model was used by applying the Green-Ampt infiltration method to a well-constrained engineering-geological model of the study area. The satisfactory matching between simulated discharges of each event and the discharges recorded by the hydrometric station at the outlet of the Vezza, allowed us to check the robustness of the engineering-geological model. Both hydrological and debris flow modelling of the 1996 event was then carried out, which allowed us to highlight key processes which affected hillslopes and streams, such as widespread erosional processes or temporary deposition and subsequent erosion causing impulsive waves typical of debris flows. Results show how the used approach allowed us to gain some insight into the hydrological behaviour and debris flow formation in the Cardoso and Vezza catchments, caused by intense or extreme rainfall events. In particular, the experimental rainfall-runoff mobile bed model allowed us to simulate the variation of the hydrological response of the catchment caused by morphological changes, the widespread erosional processes, and the temporary deposition and subsequent erosion of solid material causing the development of impulsive waves typical of debris flows. Concerning the area of Mt. Amiata, located in southern Tuscany, at least historically, it was not characterised by particularly significant debris flow events or intense solid transport processes. However, probably in the context of climate change consequences, on 27-28 July 2019, extreme rainfall caused a severe channelled erosion process within the Risola catchment, upstream of Abbadia San Salvatore, a village located on the eastern slope of Mt. Amiata. The mixture of water and sediment that was generated obstructed a culvert at the entrance of the urban area and its flooding. Engineering-geological field investigations were carried out to assess the availability of debris material and its hydrological behaviour. Then, a cascade of numerical models was employed to reconstruct the debris flow. The FLO-2D model was used to simulate the rainfall-runoff process within the catchment to evaluate the liquid discharge along the hydrographic network. Subsequently, the two-phase and mobile-bed TRENT2D model was used to quantify both the erosion and deposition processes that occurred during the event. A specific modelling procedure for the numerical simulation of the culvert clogging was developed and applied. Despite the challenging framework, the results, in terms of debris volume, erosion rates, deposition area, and timing of the culvert obstruction, agree with the observed data. The back-analysis was implemented by using mostly the setting of a priori parameters, thereby proving that the proposed approach is robust and effective with good predictive capability. Therefore, it may represent a convenient tool to support further predictive hazard mapping analysis. Summarizing, the numerical modelling of the above-mentioned events allowed the reliable representation of some key aspects of rainfall-runoff and debris flow events, such as water infiltration and runoff, erosion, transport, and depositional processes. On the other hand, it also highlighted that at the current state of the art, it remains challenging to comprehensively simulate widespread erosion processes, coupling rainfall-runoff, hillslopes instability, and debris flows, as well as the interaction between anthropogenic defence works and debris flows, as occurred in the Cardoso and Abbadia San Salvatore events, respectively. The results of this Thesis, therefore, show the need for a scientific effort to improve the aforementioned modelling aspects. In any case, it is also highlighted that through a multidisciplinary approach, some of the modelling gaps can be filled by in-depth geologic-geomorphologic knowledge of the catchments, which allows reasonable choices and assumptions to be made. This approach would presumably allow for a more complete representation of extreme rainfall-runoff and debris flows phenomena in view of better territorial planning, in the context of climate change that exposes more and more citizens to geo-hydrological risk in mountain environments.