Doctor of Philosophy (PhD)
Wildfire effected regions of the western U.S. frequently produce non-Newtonian floods (or floods that have a non-linear relationship between stress and deformation) in response to moderate to severe precipitation events. This research presents the development, evaluation, and demonstration of a post-wildfire hydrodynamic one-dimensional and two-dimensional diffusive wave and shallow-water numerical modeling approach that can be used to predict post-wildfire downstream runout of debris flows and floodplain inundation conditions. While researchers have developed a variety of Non-Newtonian approaches to simulate debris flows and mudflows, there has been very limited application to post-wildfire flooding. This can make it difficult to understand the assumptions and limitations in any given model or replicate work, making the modular non-Newtonian computation library approach presented in this dissertation advantageous. This research was conducted with two widely used U.S. Army Corp of Engineer’s (USACE) hydraulic modeling software: the two-dimensional Hydraulic Engineering Center (HEC) River Analysis System (HEC-RAS); and the two-dimensional Adaptive Hydraulics (AdH). First, this research demonstrates the effectiveness of the non-Newtonian library, DebrisLib, to predict a variety of non-Newtonian flow conditions using flume experiments from Hungr (1995), Haldenwang (2003), and Iverson et al. (2010). The modeling approach was then validated using Kean et al. (2019) datasets collected following the 2017 Thomas Fire and 09 January 2019 post-wildfire debris flow events near Santa Barbara, California. The evaluation of the numerical modeling approaches versus observed laboratory and field data indicates that both HEC-RAS and AdH, when linked with the non-Newtonian library DebrisLib, can adequately predict a range of non-Newtonian flow conditions. The shallow-water hydraulic models with non-Newtonian library DebrisLib were able to replicate the downstream runout and floodplain inundation demonstrating the effectiveness of the numerical modeling framework. The results from both flume and the 09 January 2019 flood events demonstrates that Newtonian physics-based models are not capable of predicting these types of events. The post-wildfire non-Newtonian approaches presented here provide an improvement to the existing state-of-practice for predicting post-wildfire flood risk, specifically, the ability to predict downstream runout and floodplain inundation.
Floyd, Ian Eli, "Non-Newtonian Model Development for Post-Wildfire Flood Risk Management" (2021). LSU Doctoral Dissertations. 5513.