Forecasting Inundation from Debris Flows That Grow By Entraining Sediment

Abstract:

Destructive debris flows often grow, and extend their runouts, by entraining sediment as they travel. However, incorporating varied entrainment processes into physics-based flow routing models is challenging. As an alternative, we developed a relatively simple, automated method for forecasting the inundation hazards posed by debris flows that entrain sediment and coalesce from multiple flows. Within a drainage network, we amalgamate the effects of many possible debris flows with each flow volume proportional to an entrainment rate scaled by the upslope contributing area, and then use these volumes in the USGS GIS-based inundation model LAHARZ. Our approach only requires estimates of two parameters: spatial entrainment rate & maximum entrainment area or maximum volume. Our procedure readily integrates various sediment sources and it can portray different inundation hazard levels on a GIS-based map by varying our two parameters.

We applied this approach to part of the Coast Range, southern Oregon, USA. Using aerial photography, we mapped debris flows triggered by a large 1996 rain event on a LiDAR-derived topographic base, and identified initiation locations, travel paths, and areas of channel erosion and deposition. Many catchments experienced multiple debris flows that coalesced downstream and about 95% of the debris flows entrained sediment as they traveled. Flows typically stopped entraining sediment before the upslope contributing area reached ~500,000 m². We used pre- and post-debris-flow stereo photos to estimate spatial entrainment rates in four clear-cut catchments having both channel erosion and coalescence of flows; these rates varied from 0.12 to 0.2 m³/m². GIS-based inundation maps, using our automated methods, are quite similar to the mapped flow paths and deposits. Given appropriate parameters, our approach could be applied to a variety of steep, channelized environments where entrainment is important, such as alpine and post-wildfire slopes.