Source: SPATIAL INFORMATICS GROUP LLC submitted to
COMMERCIALIZING PROBABILISTIC WILDLAND FIRE PREDICTION WITH PYRECAST
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
NEW
Funding Source
Reporting Frequency
Annual
Accession No.
1032961
Grant No.
2024-70012-43705
Project No.
CALW-2024-04687
Proposal No.
2024-04687
Multistate No.
(N/A)
Program Code
8.1
Project Start Date
Sep 1, 2024
Project End Date
Aug 31, 2026
Grant Year
2024
Project Director
Johnson, G.
Recipient Organization
SPATIAL INFORMATICS GROUP LLC
2529 YOLANDA CT
PLEASANTON,CA 945667513
Performing Department
(N/A)
Non Technical Summary
Wildfires are becoming more frequent and intense due to climate change, posing a significant threat to lives, property, and natural resources. Effective wildfire management is crucial to mitigate these risks, but traditional fire behavior models often lack the ability to accurately predict the dynamic and complex nature of wildfires. Our project aims to enhance wildfire prediction and risk management through the development of advanced fire forecasting tools, particularly the PyreCast platform. This platform uses probabilistic modeling to provide more reliable and detailed predictions of fire behavior, helping emergency managers, electric utilities, and insurance companies make better-informed decisions to protect communities and manage resources effectively.The project will integrate cutting-edge weather-fire models and advanced statistical methods to improve the accuracy and usability of wildfire forecasts. By incorporating uncertainty estimates and dynamic graphical representations, PyreCast will offer users a clearer understanding of potential fire scenarios and associated risks. The platform will be scaled to cover broader geographic areas, including Alaska and Hawaii, and will include features such as a decision support tool for electric utility companies to make informed decisions about implementing public safety public shutoffs to mitigate wildfire risks. This project will ultimately lead to improved wildfire management practices, enhanced safety for communities and firefighters, and reduced economic losses due to wildfires.
Animal Health Component
0%
Research Effort Categories
Basic
0%
Applied
50%
Developmental
50%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
12206993100100%
Goals / Objectives
The primary goal of this project is to enhance the PyreCast fire modeling software to make it marketable to the insurance, electric utility, and emergency management sectors across the United States. This enhancement aims to provide a comprehensive fire modeling capability that includes onboard forecast skill assessment, robust software support, unmatched uncertainty quantification, and public-facing products. By adhering to the principles of open science, the project seeks to foster trust, collaboration, and engagement, thus setting PyreCast apart from competition while supporting business models aimed at profit generation.ObjectivesIntegrate Phase I research into PyreCast interface: Update the software to generate uncertainty estimates and revise the user interface to incorporate uncertainty animations into the 'active fire' tool, allowing users to connect predicted burn probability with the probability of losing resources and assets.Develop and apply objective methods to evaluate and communicate the accuracy of fire behavior forecasts: This includes probabilistic forecasts, and displaying these on PyreCast. The aim is to provide quantitative forecast evaluations, including deterministic and probabilistic assessments, addressing gaps in current fire behavior modeling practices.Integrate CAWFE model into PyreCast: Automate selection, domain configuration, drawing needed inputs, execution, benchmarking, and processing of model outputs to establish feasibility for including more complex dynamic modeling.Improve PyreCast software and computer architecture: Enhance computational capacity, system speed, and reliability to ensure improved uptime, accommodating new clients and geographies for displaying weather, active fire, risk, and other features.Expand Geographic Scope of PyreCast to include Alaska and Hawaii: Develop new software code and user interface elements to support expanding the geographic extent of active fire and risk forecasts to these regions, increasing reach to additional wildfire situation awareness markets.Develop De-energization Decision Support Portal in PyreCast: Create a user-authenticated portal and interface within PyreCast to display areas of concern relative to fire weather metrics, risk forecasts, and de-energization thresholds, assisting risk managers in electric utilities.Complete Go-to-Market/Commercialization Strategic Plan: Refine and finalize the commercialization plan by expanding understanding of potential competitors, developing a marketing strategy and revenue forecasts, and strategizing staffing and support for PyreCast under different business models.These objectives are specific, focused, quantifiable, and attainable within the project's duration and available resources.
Project Methods
Conducting the ProjectGeneral Scientific Methods:Probabilistic Fire Growth Forecasts: The project employs a probabilistic framework for fire growth forecasts, which differs from traditional deterministic models. This involves conducting multiple simulations (ensembles) with varying inputs such as weather, fuel loads, and physics parameters to produce a range of possible outcomes. These methods provide a probabilistic prediction, enhancing the understanding of uncertainty in fire behavior.Coupled Weather-Fire Models: Integration of the CAWFE (Coupled Atmosphere-Wildland Fire Environment) model into PyreCast. CAWFE combines atmospheric models with fire behavior models, enabling dynamic interaction between the fire and its environment, such as fire-induced winds.Graphical Representation of Uncertainty: Dynamic graphical animations of outputs will be used to enhance the interpretability of predictions. This is a significant departure from static presentations typically utilized for modeled wildfire behavior outputs.Unique Aspects:Ensemble Simulations: The use of ensemble simulations with varied weather inputs from the National Centers for Environmental Prediction's Short Range Ensemble Forecast (SREF), probabilistic fuel information from the North American Wildland Fuels Database (NAWFD), and varied physics parameters provides a broader range of outcomes and quantifies uncertainty in fire behavior predictions.Inclusive and Accessible Design: Collaboration with graphic designers to create inclusive color palettes and presentation modes that aim to comply with Web Content Accessibility Guidelines 2.1 (AA or AAA)..Analysis, Evaluation, and Interpretation:Evaluation of Forecast Accuracy: Deterministic forecasts will be evaluated by comparing predicted area growth against observed growth, calculating statistics like the probability of detection and false alarm ratio. Object-based verification methods will also be explored.Evaluation of Probabilistic Forecasts: Spatial analysis will estimate the probability a location is burned versus the frequency it occurs in simulations. Techniques from meteorological forecast verification, such as ensemble spread and sharpness assessment, will be adapted to evaluate fire predictions.Evaluation and Quantification of ImpactTypes of Evaluation Studies:Comparative Analysis of Wildfire Models: Establishing baseline performance statistics for each model (CAWFE, GridFire, ELMFIRE) and documenting the statistical analysis in a technical report. Publishing the findings in a peer-reviewed journal.User Feedback and Adoption Metrics: Collecting feedback from users (emergency managers, utility risk managers, insurance professionals) on the usability and effectiveness of the PyreCast platform and probabilistic forecasts. Measuring the adoption rate and impact on decision-making processes.Quantitative Indicators of Success:Improvement in Forecast Accuracy: Measuring the accuracy and reliability of wildfire behavior predictions through validation against real-world fire events and statistical analysis.Key Milestones:Integration of CAWFE Model: Successful integration of the CAWFE model into PyreCast and demonstration of its improved forecasting capabilities in dynamic fire environments.Expansion of Geographic Scope: Extending the geographic coverage of PyreCast to include Alaska and Hawaii, increasing the reach and applicability of the platform.Impact Assessment:Reduction in Wildfire Impacts: Quantifying the reduction in the number of structures and acres burned, as well as the decrease in economic losses due to more accurate and timely fire behavior predictions.Enhanced Community Safety: Measuring improvements in community safety through better-informed evacuation plans and reduced fire ignitions from power lines.By employing these scientific methods, unique aspects, and evaluation strategies, the project aims to enhance wildfire management, improve decision-making, and ultimately contribute to societal benefits through advanced fire forecasting and risk mitigation tools.