Managing Future Risk of Increasing Simultaneous Megafires

This research effort tackles the challenges that simultaneous megafires currently pose to decision makers and stakeholders, and supports proactive planning for future scenarios to mitigate risk (NSF Growing Convergence Research #2019762). Megafires are fires that are unusually large or that require a complex and aggressive firefighting response because of dramatic threats to lives, property, and/or infrastructure. When multiple megafires occur simultaneously, firefighting resources may be strained beyond capacity with catastrophic results. To successfully advance the frontiers of fire science and management to mitigate risk at the intersection of natural and human systems, we are developing a highly convergent approach in a team comprised of researchers from University of Washington, NCAR (National Center for Atmospheric Research), and University of California, Merced. We bring expertise in decision science, climate science, statistics, and fire science to our collaboration with on the ground decision makers including fire managers, fire ecologists, and land managers for tribal and US government agencies.

We aim to strengthen risk management related to wildfire impacts with improved climate projections in support of decisions regarding land use, fuel and land management, and wildfire suppression, thereby helping to safeguard against the future loss of life, property, infrastructure, and natural resources.

Please contact Alison Cullen alison@uw.edu with questions or to inquire about partnership opportunities.

NSF Logo
NCAR UCAR logo
University of California Merced Logo

Data Visualizations

Historic Data

Median number of fires in each Geographic Area Coordination Center (GACC) in each calendar month - comparing 1984-1993 and 2010-2019. Each row of charts corresponds to a different fire size. ALF represents all large fires 1,000+ acres, VLF represents 75th percentile fires, and ELF represents 90th percentile fires. Fire sizes are binned using percentile cutoffs within each GACC over the entire period to determine fire size categories. Color indicates GACC according to the key on the right. Data on simultaneous wildfire occurrence come from the Monitoring Trends and Burn Severity (MTBS) project. This visualization was made as part of NSF Growing Convergence Project 2019762.
Median number of fires in each Geographic Area Coordination Center (GACC) in each calendar month – comparing 1984-1993 and 2010-2019. Each row of charts corresponds to a different fire size. ALF represents all large fires 1,000+ acres, VLF represents 75th percentile fires, and ELF represents 90th percentile fires. Fire sizes are binned using percentile cutoffs within each GACC over the entire period to determine fire size categories. Color indicates GACC according to the key on the right. Data on simultaneous wildfire occurrence come from the Monitoring Trends and Burn Severity (MTBS) project. This visualization was made as part of NSF Growing Convergence Project 2019762.

Data Projections

One map and seven graphs showing the change ni the number of simultaneous 1000+ acre wildfires
Change in the number of simultaneous 1000+ acre wildfires in each Geographic Area Coordination Center (GACC) in the Western US comparing 1985-2015 and 2030-2060. These results are projected by a statistical model applied to fire indexes output from regional climate models (RCMs) from the NA-CORDEX data archive. Each panel shows boxplots, for an ensemble of 13 RCM simulations, of the probability of exceeding the 1-year return level in simultaneity for that GACC on a biweekly basis. This visualization was made as part of NSF Growing Convergence Project 2019762.

Research Goals, Questions, Hypotheses

Our overall research goal is to understand current and future wildfire characteristics to support fire-related decisions throughout the 21st century in the face of resultant suppression resource scarcity and competition in the US. This project will model future patterns and uncertainty in the simultaneous co-occurrence of megafire events to inform risk management.

 Research Questions

  • How will climate change alter future patterns of wildfire, particularly co-occurring megafires?
  • What implications does this hold for risk management decisions?

We are developing statistical models to represent relationships between biogeophysical and human factors (e.g., ignitions, suppression policy, land and fuel management) and firefighting resource demand at geographical scales relevant to firefighting management decision-making. These resource demand and risk management models will be based on wildfire characteristics, climate, weather, and land history covariates. We are evaluating climate change impacts on ignition patterns and on wildfire risk with regional climate model projections from NA-CORDEX, and observations from the gridMET dataset, by looking at fire danger indices in concert with projected spatiotemporal patterns in anthropogenic activities associated with human-caused fires, as well as diagnostics for lightning activity.

Hypotheses:

  1. We hypothesize that ignition efficiency will increase further with warming, facilitating increased lightning-ignitions, and consequently increases in simultaneous wildfire events.
  2. We hypothesize that a positive feedback may occur where fire suppression resources at the national level become strained, reducing the efficacy of managing active fires and new ignitions, and further increasing resource strain and relative burned area.
  3. We hypothesize that short term fire management decisions (e.g., both fuel management and fire suppression) have significant delayed impacts, and demand innovative scientifically supported decision tools that explicitly account for climate change and the continuing interaction of natural and human systems.

Publications

Project Team

Presentations

Policy Messages

Press Coverage

Software, Data Products

All software and data products generated by this project are publicly available. The code and documentation that we have developed for calculating fire indices in support of climate projections is publicly available on GitHub. Visualizations appear in both published papers and above on this page.

burn_viz is a web application that allows users to visualize relationships between wildfire occurrence, the location of human populations, and different landcover types. Understanding these relationships is crucial for risk assessment and policymaking, but it can be challenging to navigate data from disparate sources that may be provided in different formats and spatial resolutions. Integrating these elements into a interactive visualization creates an opportunity to explore how wildfire, population, and landcover interact over space and time.