Year of Award

2025

Document Type

Dissertation

Degree Type

Doctor of Philosophy (PhD)

Degree Name

Computer Science

Department or School/College

Department of Computer Science

Committee Chair

Jesse Johnson

Commitee Members

Douglas Brinkerhoff, Lucy Owens, Carl Seielstad, John Hogland

Keywords

cfd, fire, fuel, lidar, wildfire

Abstract

Computational models of wildfires are an important tool for fire managers and scientists. However, representing fuel inputs with sufficient detail for process-based fire models to accurately reflect observed fire phenomena while remaining computationally tractable is a major challenge. This dissertation advances wildland fire science by applying and parametrizing such models to investigate complex fire-fuel interactions across multiple scales.

A primary contribution addresses fine-scale fuel characterization through a novel voxel-based technique converting Terrestrial Laser Scan (TLS) point clouds into three-dimensional (3D) fuel representations. This method was evaluated by comparing simulated fire behavior in the Fire Dynamics Simulator (FDS), optimized using the DAKOTA toolkit, with observed mass loss data from laboratory experiments conducted at the Missoula Fire Sciences Laboratory. Key findings reveal that while point cloud-derived fuels can accurately describe observed fire behavior, exceptionally high-resolution 3D fuel models do not necessarily improve predictive parity, suggesting an optimal balance between fidelity and computational demand. This research offers guidelines for translating LiDAR data into 3D fire models and determining appropriate fuel cell resolutions for capturing accurate fire behavior.

To overcome data limitations at broader scales, this work introduces FastFuels, a novel system generating landscape-scale 3D fuel data suitable for next-generation fire models. FastFuels integrates disparate data sources, including forest inventory data, plot imputation maps, and incorporates flexible LiDAR data assimilation to produce detailed and mutable fuel landscapes. Its utility is demonstrated through applications in evaluating fuel treatment effectiveness with FDS and simulating prescribed fire operations with QUIC-Fire, thereby enhancing decision support capabilities.

Furthermore, this dissertation contributes to a mechanistic understanding of dynamic fire behaviors by investigating convective mechanisms and wind thresholds governing junction fire formation in heterogeneous fuel treatments. Utilizing FDS simulations informed by field observations, this research reveals how fuel structure and alterations in convective heat transfer, modulated by critical wind speeds, initiate transitions between fireline slowing and acceleration regimes within fuel treatments.

Collectively, these studies enhance the fundamental understanding of fire-fuel interactions, introduce innovative methods for fuel parameterization, and deliver advanced modeling tools. The findings and developed systems contribute significantly to improving the practical utility, accessibility, and predictive power of process-based models, thereby advancing wildland fire science and supporting more effective, ecologically informed fire management strategies.

Download

Share

COinS
 

© Copyright 2025 Anthony Albert Marcozzi