Modeling long-runout submarine landslides

Submarine landslides are known to travel down very low slopes and for very long distances relative to their elevation drop, indicating a very low frictional resistance between the slide material and sediment bed over which it travels. A number of mechanisms have been suggested for reducing this resistance. We propose that dynamic fluidization caused by rapid compaction of the sediment bed is a plausible cause of this friction drop. We use a grain-scale numerical model that incorporates the interaction between grains and pore fluid to model a small part of a landslide. The model (based on Goren et al. JGR, 2010) uses the discrete element approach to model a 2-D set of interacting grains, coupled with a finite difference model for fluid pressure in a permeable material (with pressures approximated in grid cells 2 to 3 grains across). We create a small set of idealized grains that are acted upon by gravity, grain contact forces and local gradients in fluid pressure; grain rearrangements that cause local porosity changes generate changes in fluid pressure. In a set of simple numerical experiments, we approximate the slide itself as a rapidly translating cohesive block that is just coming into contact with an undercompacted bed of cohesionless grains. As the weight of the block compacts the bed, fluid pressures rise. If the fluid pressure approaches the weight of the overlying slide, the effective resistance to motion drops quickly. Initial results show that total distance the block moves is strongly dependent on the level of pressure increase in the fluid in the top 15 grains of the bed. When pressure increase is large enough the layer becomes transiently fluidized, with frictional resistance dropping to zero. Scaling indicates that this pressure should increase with compaction rate (slide velocity) and the length scale of the slide, and decreases with the mean permeability of the sediments (grain size).