Team Members: Yuying Ren, University of Colorado Boulder, PhD Student, Aleksander Berg, University of Colorado Boulder, PhD Student
Project Supervisor: Stefan Leyk, Professor of Geography, University of Colorado Boulder, Professor of Geography
Project Description: Our group has published high-quality and peer-reviewed maps of wildland urban interface (WUI) and fire exposure probability maps through time (1990 to 2020) for WUI properties in the contiguous United States. These maps present a detailed view into how U.S. WUI areas have evolved through time and what present wildland fire exposures they may face. A key gap is the dissemination of our high-fidelity information to decision makers and stakeholders. We propose a unified, decision support dashboard, “WUI-LLM,” that collates these data into a clear presentation of how built environments across the U.S. interact with flammable wildlands. WUI-LLM will map structure level WUI classifications and fire exposure probabilities and allow users to interact with these data using a large-language model enhanced query system to allow for on-the-fly spatial analysis. Our dashboard will remove technical barriers in quantifying community risk to wildland fire and how that risk has changed over time.
WUI White Paper Knowledge Gaps Addressed: Our proposed WUI-LLM decision support dashboard combines peer reviewed high-resolution WUI and wildfire probability layers tied to individual structures across the conterminous U.S. Our proposed work, directly addresses key gaps identified in the white paper, specifically the “lack of standardized methodologies for the calculation of quantitative values of parcel-level fire exposure...”, the need for “categorization of fire exposure scenarios…based on their likelihood of occurring,” and “a need for peer reviewed risk assessment tools at the WUI property level” (Section 3.4). Overarchingly, our temporal WUI (1985-2020) data fill the need for systematic historical datasets to validate wildfire risk and vulnerability (Section 4.4), while our structure-level exposure data respond to the need for accessible parcel-level risk assessments (Risk Assessment at the Parcel Level). WUI-LLM further bridges the gap between complex fire engineering tools and decision-making by enabling interactive, multi-scale analysis through an LLM-driven platform.
Team Members: Adetola Ololade Nicole Koiki, University of Maryland College Park, PhD Student, Andres Felipe Rivas Bolivar, University of Maryland College Park, PhD Student
Project Supervisor: Fernando Raffan-Montoya, Assistant Professor & Co-Director Fire Testing and Evaluation Center (FireTEC), University of Maryland College Park
Project Description: Wildfires in the wildland-urban interface (WUI) often spread through airborne embers (firebrands), which can ignite homes far ahead of the main fire front. Despite their importance, there is currently no method to measure how many firebrands are present, where they travel, or whether they remain hot enough to ignite structures. This project develops a novel sensing system that combines stereo imaging and thermal measurement techniques to track firebrands in three dimensions and estimate their surface temperature over time. By identifying which firebrands remain thermally hazardous when they land, the system provides critical insight into ignition risk. The platform is designed for both ground-based and UAV-supported deployment, enabling safe data collection near active fires. The resulting data will help improve wildfire risk assessment, validate fire spread models, and support more effective mitigation strategies in vulnerable WUI communities.
WUI White Paper Knowledge Gaps Addressed: The SFPE Foundation WUI White Paper identifies a critical gap in the lack of quantitative frameworks for firebrand exposure, specifically noting the absence of methods to measure firebrand flux, transport distance, deposition, and ignition potential (Section 3.1.3). It further highlights the need for improved fire exposure characterization and data to support hazard and vulnerability assessment at the parcel scale (Section 3.4), as well as a broader lack of systemized data collection for WUI fire exposure (Executive Summary). This paper directly addresses these gaps by developing a measurement system that captures time-resolved firebrand trajectories, velocities, spatial flux, and surface temperature. By linking firebrand motion with thermal state, the work provides a pathway toward quantifying which firebrands remain capable of ignition, enabling improved exposure characterization, model validation, and future integration into WUI risk assessment frameworks.
Team Members: Hannah Odia, University of Waterloo, Undergraduate Student, Johnathan Hernes, University of Waterloo, Undergraduate Student, Linnea Townsend, University of Waterloo, Undergraduate Student, Keon Senez, University of Waterloo, PhD Student
Project Supervisor: Elizabeth Weckman, Professor, University of Waterloo Fire Research Group
Project Description: Canada's plants have been surviving wildfires for thousands of years. This project asks: what can we learn from them? By testing several species from major vegetation regions, a Fire Performance Index will be developed to classify plants by their fire behaviour – including classifications of resiliency. Combined with Indigenous knowledge and regional fire case studies, findings will inform a Wildland-Urban Interface framework that puts the right plants in the right places, turning ecological resilience into a practical tool for community fire risk reduction.
WUI White Paper Knowledge Gaps Addressed: Wildfire risk assessment in Canada has a recognized gap: we lack effective tools for evaluating how mitigation strategies perform within a community risk assessment framework. This project directly addresses that challenge, identified in section 4.4 under community wildland fire protection, by making the case for strategic plant selection and placement as a quantifiable, evidence-based mitigation strategy.
Using cone calorimeter-derived fire performance data, including heat release rate, ignition characteristics, and fuel properties alongside regrowth data, Indigenous ecological knowledge, and regional fire case studies, this project will develop a practical mitigation framework tailored for community risk assessment. The framework will integrate vegetation type and fuel characteristics to evaluate how plant selection and placement influence fire behaviour and overall risk at the community scale. The result is a science-backed, actionable tool that could help communities understand and reduce their wildfire exposure through informed, vegetation-based mitigation planning.
