Studio dei meccanismi di deposizione dei Debris Flow: integrazioni tra esperienze di laboratorio, analisi di campo e modellazioni numeriche

This research is related to the study and modelling of the depositional processes of debris flows and an understanding of the related rheological behaviour. The main aims of this study are to analyse the depositional features of a debris flow, to verify the possible contribution of laboratory tests, carried out by means of a tilting-plane rheometer, to determine the rheological parameters and to investigate the reliability and limitations of computer models employed for debris flow simulation and rheological parameter determination. The laboratory tests were been carried out at the Institute for Hydrological and Geological Protection of the Italian National Research Council (CNR IRPI) in Padova. The physical model consists of a 2 x 1 m tilting plane with inclination from 0° to 38°, on which a steel tank with a removable gate has been installed. A fixed horizontal plane (1.5 x 1 m), with an artificial roughness to simulate natural basal friction, served as the deposition area. In total, 93 laboratory tests were carried out: 62 tests simulated the quasi-static formation of a fan (with the tank installed at the lower end of the tilting plane), the remaining 31 examined dynamic fan formation by the means of a flume. The steel tank, with a removable gate facing the deposition plane, is parallelepiped with a square base (15 x 15 cm and 33 cm high) having a maximum volume of 7 dm3. The laboratory tests were conducted using three different materials: plastic cylinders with a diameter and height of 3 mm; medium to fine gravel; and debris-flow matrix (with a diameter < 19 mm), varying amounts of water were added to the debris-flow matrix to test solid concentrations in a range between 0.45 and 0.67. Data analysis included the development of semi-empirical equations for runout distance, the maximum width of the deposit and total travel distance. An energy balance approach was tested in order to determinate the rheological parameters of the debris-flow matrix, this method is based on the comparison between the potential energy of the mass stored in the tank and the work made in the process of deposit formation. The field sites related to the debris-flow events are located upstream of the town of Cortina d'Ampezzo (Fiames locality, Belluno, Italy), where an intense rainstorm triggered six debris flows during the afternoon of 5th July 2006. Immediately following the event, field surveys were carried out in the study area. These field surveys made it possible to measure several features, including the debris-flow deposits, main channels and initiation areas. Samples taken from the debris-flow deposits have been used for laboratory tests. Total travel distances and the runout distance on fans measured in the field were compared with formulas found in the literature (empirical/statistical and physically oriented) and also compared to the results of the laboratory tests. An estimation of shear stress from the field site was calculated using Johnson's (1970) formula. The Fiames debris-flow event of 5th July 2006 and the laboratory tests (dynamic runs) were simulated using FLO-2D, while RAMMS (Rapid Mass MovementS) was used solely to simulate the Fiames event. FLO-2D (O'Brien, 2003) is a two-dimensional flood routing model with a rigid bed (debris-flow simulations) or a mobile bed (sediment transport simulations), it simulates water flows in wide rivers as well as non-Newtonian flows over alluvial fans. FLO-2D numerically routes a flood hydrograph while predicting the area of inundation, the maximum depth and the flow velocity in each cell of the square grid system derived from input topographic data. RAMMS was developed in 2005 by the Swiss Federal Institute for Forest, Snow and Landscape Research (WSL), Birmensdorf and the Swiss Federal Institute for Snow and Avalanche Research (SLF), Davos. RAMMS uses a one-phase approach based on Voellmy rheology (Voellmy, 1955; Salm et al. 1990). The input file combines the total volume of the debris flow located in a release area with a mean depth. The model predicts the area of inundation, the maximum depth and the flow velocity in each triangular-shape cell of the input DTM. The limitations of the numerical simulations relating to the laboratory tests were investigated to verify to what extent rheological parameters could be determined using this indirect method. The analysis of the Fiames event enabled understanding of the rheological behaviour related to the dolomitic debris-flows and the influence exerted by the input parameters on the final results. Finally, some improvements to the RAMMS model were proposed to obtain simulations in keeping with the events observed in the field.