Hazard and risk assessment of landslides with potentially long run-out is becoming more and more important. Numerical tools exploiting different constitutive models, initial data and numerical solution techniques are important for making the expert’s assessment more objective, even though they cannot substitute for the expert’s understanding of the site-specific conditions and the involved processes. This paper presents a depth-integrated model accounting for pore water pressure dissipation and applications both to real events and problems for which analytical solutions exist. The main ingredients are: (i) The mathematical model, which includes pore pressure dissipation as an additional equation. This makes possible to model flowslide problems with a high mobility at the beginning, the landslide mass coming to rest once pore water pressures dissipate. (ii) The rheological models describing basal friction: Bingham, frictional, Voellmy and cohesive-frictional viscous models. (iii) We have implemented simple erosion laws, providing a comparison between the approaches of Egashira, Hungr and Blanc. (iv) We propose a Lagrangian SPH model to discretize the equations, including pore water pressure information associated to the moving SPH nodes.
Snow erosion and entrainment processes in avalanches are classified according to their mechanisms, the flow regimes in which they occur, and their spatial position within the avalanche. Simple, but process-specific, models are proposed for erosion by impacts, abrasion, plowing and blasting. On the basis of order-of-magnitude estimates, the first three mechanisms are clearly expected to be important. The fourth mechanism stipulates that the compaction of the snow cover ahead of the avalanche leads to the flow of escaping air just in front of the avalanche that may disrupt the snow cover and support formation of a saltation layer. The effects of this hypothetical mechanism resemble those of the plowing mechanism. All mechanisms depend strongly on the snow properties, but, with plausible parameter values, erosion rates at or above the experimentally found rates are obtained. The entrainment rate of an avalanche is most often limited by the shear stress needed to accelerate the eroded snow to avalanche speed.
17 August 2005 in Natural Hazards and Earth System Sciences
Recent experimental and theoretical work on the dynamics of submarine debris flows is summarized. Hydroplaning was first discovered in laboratory flows and later shown to likely occur in natural debris flows as well. It is a prime mechanism for explaining the extremely long runout distances observed in some natural debris flows even of over-consolidated clay materials. Moreover, the accelerations and high velocities reached by the flow head in a short time appear to fit well with the required initial conditions of observed tsunamis as obtained from back-calculations. Investigations of high-speed video recordings of laboratory debris flows were combined with measurements of total and pore pressure. The results are pointing towards yet another important role of ambient water: Water that intrudes from the water cushion underneath the hydroplaning head and through cracks in the upper surface of the debris flow may drastically soften initially stiff clayey material in the "neck" of the flow, where significant stretching occurs due to the reduced friction at the bottom of the hydroplaning head. This self-reinforcing process may lead to the head separating from the main body and becoming an "outrunner" block as clearly observed in several natural debris flows. Comparison of laboratory flows with different material composition indicates a gradual transition from hydroplaning plug flows of stiff clay-rich material, with a very low suspension rate, to the strongly agitated flow of sandy materials that develop a pronounced turbidity current. Statistical analysis of the great number of distinguishable lobes in the Storegga slide complex reveals power-law scaling behavior of the runout distance with the release mass over many orders of magnitude. Mathematical flow models based on viscoplastic material behavior (e.g. BING) successfully reproduce the observed scaling behavior only for relatively small clay-rich debris flows while granular (frictional) models fail at all scales. For very large release masses, hydroplaning or significant softening of the shear layer due to water incorporation must be invoked to recover the observed scaling behavior; a combination of both effects likely will give the most realistic description of the phenomenon. Detailed studies of the neck behavior and the compositional dependence of the material properties are needed to arrive at a quantitative model. Other related and important open questions concern the rheological model appropriate for sandy debris flows and the suspension rate from the dense body into the associated turbidity current.
Tsunami generation from subaqueous landslides is controlled by landslide kinematics, which in turn is governed by the material properties of the slide mass. Yet the effect of the material properties on tsunami genesis is poorly understood. Geomorphological observations of landslide runout put constraints on the landslide dynamics. In addition, observations of tsunami runup heights can improve our understanding of how the landslide material transforms from initiation to final runout. The giant prehistoric Storegga Slide off the mid‐Norwegian coast caused a well‐documented ocean‐wide tsunami that offers a unique setting for coupling landslide material models to tsunami generation models. In this study we simulate the dynamics of the Storegga Slide and tsunami using the depth‐averaged landslide model BingClaw, which implements visco‐plastic rheology and remolding, and couple it to a standard tsunami propagation model. A broad sensitivity study varying the landslide material strength parameters in BingClaw shows that the initial soil yield strength and remolding rate are most important for the tsunami genesis but that the residual strength determined the final runout distance. BingClaw parameters were further optimized to obtain the observed runout distance and to minimize the relative error of the tsunami runup heights. As detailed time‐dependent three‐dimensional representations of landslide parameters cannot be determined through a field investigation of the landslide itself, these simulations of the Storegga Slide and tsunami can help in the selection of plausible parameter ranges for prognostic modeling in quantitative hazard assessments.