Landslides and landslide dams are a major natural hazard causing high socioeconomic risk in inhabited mountainous areas. This is also true for vast parts of south-western China, which are highly prone to slope failures due to several factors, such as a humid climate with high precipitation in the summer months, geological predisposing factors with highly weathered sedimentary rocks and a high seismicity. In order to assess possible run-out distances and the potential of landslides to block rivers, it is crucial to understand which factors influence landslide propagation and how they can be quantified. Since it is often difficult or impossible to measure related geotechnical parameters in the field, their back analysis with a numerical modelling approach can be useful. In this study a numerical modelling analysis was implemented for the case of a complex landslide in south-western China, which transformed into a debris flow and blocked the river and a major road after heavy rainfall. For this purpose a quasi-3D smoothed particle hydrodynamics (SPH) model that can account for geotechnical slope parameters, run-out distance, velocities, and deposition heights was used. Based on field observations regarding initial landslide volume and final deposition volume, height, and length, the mechanical properties of the landslide were estimated in a back-analysis. Results Through stepwise parameter optimisation the best reconstruction of the observed deposition phenomena could be achieved considering an initial landslide volume of about 0.5 million cubic meters, a triggering height of 15 m, an height of the water table equal to half the soil thickness, initial pore water pressures of about 0.6 of the liquefaction value, and non-negligible bed entrainment, which resulted in a deposition with a volume of about one million cubic meters, a length of 615 m and a mean height of 11 m. Compared to models with other parameter combinations, here the total error was minimal, while the final deposition dimensions were only slightly overestimated with regard to the observations in the field. Conclusions The paper outlines the potential for quantitatively interpreting the field evidence for the propagation of a complex flow-like landslide and related cascading processes, such as landslide damming. However, the analysis should take into account multiple features of the whole processes, such as the initial conditions at the landslide source area, the propagation pattern, the total volume mobilised, and the deposition characteristics. By doing so, the estimated model parameters can be implemented in future studies for the forward modelling of events at the same site, or other sites along this slope, in order to assess the potential of future river blockings through landslide deposits. Keywords
SPH propagation back-analysis of Baishuihe landslide in south-western China
PETROSINO, STEFANO;CUOMO, SABATINO
2017-01-01
Abstract
Landslides and landslide dams are a major natural hazard causing high socioeconomic risk in inhabited mountainous areas. This is also true for vast parts of south-western China, which are highly prone to slope failures due to several factors, such as a humid climate with high precipitation in the summer months, geological predisposing factors with highly weathered sedimentary rocks and a high seismicity. In order to assess possible run-out distances and the potential of landslides to block rivers, it is crucial to understand which factors influence landslide propagation and how they can be quantified. Since it is often difficult or impossible to measure related geotechnical parameters in the field, their back analysis with a numerical modelling approach can be useful. In this study a numerical modelling analysis was implemented for the case of a complex landslide in south-western China, which transformed into a debris flow and blocked the river and a major road after heavy rainfall. For this purpose a quasi-3D smoothed particle hydrodynamics (SPH) model that can account for geotechnical slope parameters, run-out distance, velocities, and deposition heights was used. Based on field observations regarding initial landslide volume and final deposition volume, height, and length, the mechanical properties of the landslide were estimated in a back-analysis. Results Through stepwise parameter optimisation the best reconstruction of the observed deposition phenomena could be achieved considering an initial landslide volume of about 0.5 million cubic meters, a triggering height of 15 m, an height of the water table equal to half the soil thickness, initial pore water pressures of about 0.6 of the liquefaction value, and non-negligible bed entrainment, which resulted in a deposition with a volume of about one million cubic meters, a length of 615 m and a mean height of 11 m. Compared to models with other parameter combinations, here the total error was minimal, while the final deposition dimensions were only slightly overestimated with regard to the observations in the field. Conclusions The paper outlines the potential for quantitatively interpreting the field evidence for the propagation of a complex flow-like landslide and related cascading processes, such as landslide damming. However, the analysis should take into account multiple features of the whole processes, such as the initial conditions at the landslide source area, the propagation pattern, the total volume mobilised, and the deposition characteristics. By doing so, the estimated model parameters can be implemented in future studies for the forward modelling of events at the same site, or other sites along this slope, in order to assess the potential of future river blockings through landslide deposits. KeywordsI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
