Using D-Claw to model landslides, debris flows, water bodies, and their interactions.

Abstract:

D-Claw is an open-source software package that solves a two-phase, depth-averaged model for motion of granular-fluid mixtures such as those in landslides and debris flows. D-Claw and its underlying governing equations were developed with the aim of seamlessly modeling landslide initiation, subsequent downslope flow, and deposition. It can simulate both highly mobile debris flows and slow landslides that stabilize after traveling a short distance. The model, its software, and initial model tests are described in detail elsewhere (Proc. Roy. Soc. Lond. A, 2014, doi: 10.1098/rspa.2013.0819 and doi:10.1098/rspa.2013.0820).

Recent extensions of the D-Claw model increase its applicability to a wider class of surface flow problems by incorporating additional physics. A core model property is evolution of the solid volume fraction (m(x,y,t)), which, in the vanishing limit (m→0) reduces the model equations to the shallow water equations for clear-water flow over topography. Additionally, incorporating entrainment of erodible bed material into debris or water flows greatly extends D-Claw’s capabilities. Together, these newer features enable the application of D-Claw to hybrid problems that may involve multiple evolving and interacting flows whose behavior depends upon varying sediment loads, including landslide generated tsunami propagation, dam breaching and erosion, and overland floods that entrain copious sediment and thereby form debris flows. Traditionally, modeling of these complex phenomena has entailed arbitrary coupling of disparate models of individual processes, but seamless simulations with D-Claw eliminate the need for such coupling and thereby ensure that mass and momentum are rigorously conserved as multiple processes interact.

Our recent applications of D-Claw address real-world problems involving hybrid and cascading hazards at volcanoes in the Northwest US. For these problems, we derive landslide source models and compute their initiation, downslope flow, and subsequent interactions with water bodies in the flow-path. A current effort of ours is the extension of D-Claw to debris flows that arise entirely from runoff and erosional processes, such as post-wildfire flows. This involves the incorporation of new entrainment models and initiation mechanisms.