GSA Annual Meeting in Phoenix, Arizona, USA - 2019

Paper No. 2-1
Presentation Time: 8:15 AM

RAINFALL MICROSTRUCTURE AND DEBRIS-FLOW GENERATION IN A RECENTLY BURNED WATERSHED (Invited Presentation)


STALEY, Dennis M., SMITH, Joel B., KEAN, Jason W. and RENGERS, Francis K., U.S. Geological Survey, Box 25046, MS 966, Denver Federal Center, Denver, CO 80225

Post-fire debris flows are initiated by runoff, erosion, and sediment transport processes occurring during intense rainfall. The close temporal correlation between rainfall intensity and debris-flow occurrence has resulted in the operational use of rainfall intensity-duration (ID) thresholds for early warning. Although a useful forecasting tool, ID thresholds suffer from two limitations. First, ID thresholds are an empirical construct, lacking any physical basis for the explanatory power of the threshold model. Second, there is a range of rainfall intensities that produces both non-hazardous runoff and destructive debris flows, even within the same watershed. Here, we analyze rainfall microstructure (drop size, velocity, and kinetic energy) to address limitations in the threshold model. We hypothesize that differences in the kinetic energy of raindrop impact may explain varying watershed response at similar rainfall intensities. Acceptance of the research hypothesis would support the importance of raindrop-impact-induced erosion processes in post-fire debris-flow generation for the analyzed events, thereby providing a physical basis for the predictive accuracy of ID thresholds.

Rainfall microstructure and watershed response were recorded in a study watershed during the winter following the 2016 Fish fire, California, USA. Three of the nine debris-flow events were initiated at rainfall intensities less than a pre-defined triggering threshold. Two non-hazardous runoff events involved peak rainfall intensities below threshold, but higher than those that triggered the sub-threshold debris flows. The analyzed debris-flow and non-hazardous runoff events were generated by rainfall events of similar drop size, velocity, and kinetic energy. Our results suggest that erosion from rainsplash detachment and overland flow may not be a defining feature of post-fire debris-flow initiation for the limited number of analyzed events. Instead, runoff-related erosion and transport processes in rills, gullies, and stream channels are the primary mechanism by which sediment laden runoff transition to debris flow. Physics-based and numerical modeling of post-fire debris-flow initiation mechanics should focus on areas dominated by fluvial erosion and transport, rather than hillslope erosion processes.