Post-Wildfire Flood Risk Management

Abstract

Wildfires increase flow and sediment load through removal of vegetation, alteration of soils, decreasing infiltration, and production of ash commonly generating a wide variety of geophysical flows (i.e., hyperconcentrated flows, mudflows, debris flows, etc.). Numerical modelers have developed a variety of non-Newtonian algorithms to simulate each of these processes, and therefore, it can be difficult to understand the assumptions and limitations in any given model or replicate work. This diversity in the processes and approach to non-Newtonian simulations makes a modular computation library approach advantageous. A computational library consolidates the algorithms for each process and discriminates between these processes and algorithms with quantitative non-dimensional thresholds. This work presents and demonstrates a flexible numerical library framework (DebrisLib) to simulate large-scale, post-wildfire, non-Newtonian geophysical flows using both diffusive wave and shallow-water models. This research demonstrates the effectiveness of the library for predicting a post-wildfire flood risk management, utilizing the non-Newtonian library (DebrisLib), that can be used to predict downstream runout and inundation conditions for typical post-wildfire flows (e.g., debris flow, mudflows, hyperconcentrated flows). This was accomplished with U.S. Army Corp of Engineer‘s (USACE) modeling software: Hydrologic Engineering Center, Hydrologic Modeling System (HEC-HMS), the two-dimensional HEC, River Analysis System (HEC-RAS) numerical model, and the 2D Adaptive Hydraulics (AdH) numerical model. The work presented here presents a real-world demonstration of the effectiveness of the non-Newtonian library DebrisLib approaches, using the widely-used USACE hydraulic models and comparison to the Kean et al. (2019) datasets following the 09 January 2018 post-wildfire flooding in Santa Barbara, California. This precipitation event generated runoff and debris flows which displaced approximately 680,000 m³ of material and sediment with observed (e.g., Kean et al., 2019) and predicted velocities of around 4.0 m/s. The hydrology modeling approach presented is a ‘brute force‘ simplistic method that represents the current state-of-practice approach for predicting rainfall-runoff yields and routing following wildfires. Evaluation of the hydraulic numerical models versus field data collected by Kean et al. (2019) indicate that classic Newtonian physics are inadequate for predicting post-wildfire flood runout and inundation. When the HEC-RAS model was linked with DebrisLib, the model sufficiently predicted arrival times and floodplain inundation. This presentation will demonstrate the usefulness of applying conservative (fixed-bed conditions and constant sediment concentration) non-Newtonian rheology-based closures using engineering-based hydraulic models for post-wildfire flood risk management and emergency management.