GSA Connects 2021 in Portland, Oregon

Paper No. 148-1
Presentation Time: 8:15 AM

MULTI-MODEL COMPARISON OF COMPUTED DEBRIS-FLOW RUNOUT AND INUNDATION FOR THE 9 JANUARY 2018 MONTECITO, CALIFORNIA POSTFIRE EVENT (Invited Presentation)


BARNHART, Katherine1, JONES, Ryan P.1, GEORGE, David L.2, MCARDELL, Brian W.3, RENGERS, Francis1, STALEY, Dennis1 and KEAN, Jason W.1, (1)Geologic Hazards Science Center, U.S. Geological Survey, Box 25046, MS 966, Denver Federal Center, Denver, CO 80225, (2)U.S. Geological Survey, Cascade Volcano Observatory, 1300 SE Cardinal Ct, Vancouver, WA 98683, (3)Swiss Federal Institute for Forest, Snow and Landscape Research, Birmensdorf, 8903, Switzerland

Reducing public exposure to geologic and hydrologic hazards requires an assessment of (a) the degree of susceptibility of an area to the hazard and (b) the potential impact of that hazard on human lives and infrastructure. Hazard assessments for postfire debris flows in the steep, recently burned terrain of the western United States have focused on the susceptibility of upstream basins to generate debris flows. Several runout models have been developed, but their use for postfire runout and inundation hazard assessment has not been extensively tested. To this aim, we compared the ability of three models (RAMMS, FLO2D, and D-Claw) to simulate runout and inundation for the most destructive flow paths from the 9 January 2018 Montecito, California event. We broke the inundated alluvial fan area into three domains (Montecito Creek, San Ysidro and Oak Creeks, Romero and Buena Vista Creeks), and for each model-domain combination we performed a numerical sampling study to identify the flow volume and material property values needed to best match the observed inundation extent and flow depth. We assessed model performance using inundated area and maximum flow depth. Comparing the best simulations from each model, we tested the hypothesis that one model's governing equations may better capture debris-flow run-out and inundation than another. We found that all models can accurately simulate the event with comparable results. Simulation performance was most sensitive to flow volume and least sensitive to material properties. Our results emphasize the importance of flow volume for estimating postfire debris-flow runout and inundation; however, further analysis of model accuracy in other locations is warranted given our small sample size.