Modeling Levee Formation in the D-Claw Debris Flow Model

The increase in wildfire frequency coupled with the increased likelihood of debris flows on burned slopes highlights the importance of understanding the hazards that debris flows pose, especially to potential impacted downstream communities. Existing empirical models can provide a reliable prediction of the post-fire potential for initiation and magnitude of debris flows. However, debris flow runout and inundation, in both burned and unburned environments, are challenging to predict, partially due to the complex interactions between the varying composition of the debris mass and fluid mixture and the topography over which the runout occurs. Here, we simulate a field-scale debris flow at the USGS debris flow flume located near Blue River, Oregon (a 2 m wide, 95 m long channel discharging onto a flat runout pad) using the open-source D-Claw shallow flow model. D-Claw, built upon the Clawpack framework (clawpack.org), solves the two-dimensional depth-averaged nonlinear shallow flow equations for a two-phase (solid-fluid) mixture, including the coupled evolution of pore pressure based on granular dilatancy. Debris flow levee formation, that is, static material forming a channel through which the fluidized debris flow travels, is not currently included in the model. Debris flow levee formation can increase runout distance, as well as flow depth and velocity, by creating a confined flow path across otherwise unconfined topography. We evaluated the impact of the inclusion of different representations for friction and the effects of granular segregation on the presence of modeled levee formation over a flat runout topography. Further, we compared simulation results with measured inundation results from experimental flume flows. This work demonstrated that levee formation can occur with each physics representation. However, for a wide range of material properties, modeling the effects of segregation was most likely to result in modeled levee formation. Understanding which physics representations need to be simulated to allow for recreation of debris flow characteristics, such as levee formation, is an important step towards accurate debris flow runout modeling and hazard prediction.