Author(s): Stefania Sansone; Giorgio Rosatti; Daniel Zugliani; Lorenzo Giovannini; Thomas Amplatz; Giulia Dallapozza

Linked Author(s): Stefania Sansone, Giorgio Rosatti, Daniel Zugliani

Keywords: TRENT2D; Snow avalanche; Sensitivity analysis; Back-analysis

Abstract: In the night between January 31 and February 1, 1986, during a snowfall, a snow mass of an estimated volume of 31000 m³ was released from the Foscac mountainside (Trentino-Alto Adige, Italy), thus involving a ski slope and causing damage to several lodges and the Ciampié refuge. The purpose of this study consists in reconstructing this snow avalanche using the numerical model for dense snow avalanches TRENT2D❄ (Zugliani and Rosatti, 2021) in order to verify that the model can reproduce real cases. The model is based on the mass and momentum balance principles of a single-phase fluid, where the flow resistance is modelled using the classical Voellmy-friction law composed of a velocity-independent Coulomb term and a velocity squared dependent term. As a preliminary step for the back-analysis, historical testimonies, such as pictures, measurements near buildings, and a map that delimits the area involved by the flow, were collected to find possible information useful for the study. Moreover, due to the lack of information concerning the release, it was necessary to estimate both the release snow depth and area. While the last was identified thanks to a cartographic analysis, the release snow depth was estimated by reconstructing the meteorological event that causes the avalanche. This reconstruction was performed using the Weather Research and Forecasting model. The back-analysis of the 1986 snow avalanche was carried out in two stages. The first step consisted in performing a sensitivity analysis for the friction-law parameters, the release volume and the position of the release area. In this way, it was possible to identify how all these variables affect the flow dynamics of the simulated avalanche. As regards the second step, the flow parameters were calibrated carefully by comparing the numerical results with the historical testimonies and by minimizing the related differences expressed in terms of snow depth, involved area and runout distance. In conclusion, the results of the back-analysis of the 1986 Foscac snow avalanche show that the TRENT2D❄ model can reproduce snow avalanches accurately.

DOI: https://doi.org/10.3850/IAHR-39WC2521711920221005

Year: 2022