The Marshall Fire, which destroyed parts of Louisville and Superior, Colorado, in 2021, defied much of what we know about fire behavior.
It happened at the end of December, far outside the traditional fire season, and occurred not in the mountains but on the plains. And the fire was fast. Driven by hurricane-force winds, this small grass fire took the lives of two people, destroyed more than 1,000 structures, and became Colorado’s most destructive blaze in terms of structural loss—all in less than a day.
“Since 2000, fires are much more frequent, and the area burned in any given year is four times as large. And fires are not only increasing in places like forests but also grasslands,” says Dr. Virginia Iglesias, a Researcher and Interim Director at Earth Lab, based at CU Boulder. “Fires are also traveling faster. In our recent research, we found that the faster a fire moves, the more destructive it is to structures.”
There is a recognized need for high-resolution and actionable information demystifying the vulnerability of the built environment to the growing threats of fire, particularly as homes are increasingly built in fire-prone areas.
Residential Housing Development in the Marshall Fire Area - Courtesy of Dr. Maxwell Cook, Earth Lab, CU Boulder
“We have a pretty good understanding of how fires move in wildlands,” says Iglesias. “But when we think about neighborhoods, our understanding is more limited.”
Most research on wildfires has focused on wildland areas and has subsequently led to a good understanding of vegetation as fuel sources—but has left a gap in the understanding of how the composition of a built structure burns.
“Something we learned from the Marshall Fire is that fences played a huge role... houses had been destroyed because fire had traveled following fences,” Iglesias notes. This is where the importance of understanding the built environment comes into play. The built environment refers to more than just buildings and infrastructure—it also includes building materials and factors such as if a house has a porch or a nearby fence. A significant indicator of vulnerability is actually the spatial arrangement of homes and other buildings, particularly their proximity to each other.
Iglesias and her team at Earth Lab are aiming to fill this gap in knowledge through the project Mapping Vulnerability: Assessing the Built Environment's Susceptibility to Wildfires Through AI and Big Data. Partially funded by a grant from the NSF ASCEND Engine in Colorado and Wyoming, this project will:
Identify the materials, structural features, and spatial arrangements that contribute to wildfire susceptibility; and
Develop high-resolution “heat maps” of the built environment’s vulnerability in the Colorado–Wyoming region.
The Team
The team contains some of the most accomplished researchers in the region, including Dr. Jennifer Balch, the Director of the NSF-funded Environmental Science Innovation and Inclusion Lab (ESIIL). Dr. Iglesias and the rest of her team are part of Earth Lab at the University of Colorado Boulder and includes Research Scientist Dr. Lise St. Denis and Research Scientist Dr. Maxwell Cook.
Cook focuses on developing models using data from past fires to provide insight into how future wildland-urban interface (WUI) fires may behave and spread (e.g., identifying how much combustible mass is in a specific area). Cook walks me through visualizations of the Marshall Fire. The fire doesn’t move how I thought it would—with smaller conflagrations appearing seemingly out of nowhere. Severely burned areas sit right next to patches untouched by flames.
“We are trying to provide data that can augment existing models to better understand not just how fire moves down a canyon into a neighborhood, but what happens when it hits the neighborhoods,” Cook explains. “For the Engine project, what we really want to understand is: for an individual home, what is the risk based on the combustible mass of the structure and its relationship and proximity to other structures? It's about bringing these data sets together and building a more holistic understanding of how risk is being transferred from one home to another.”
Marshall Fire Damage Assessment Courtesy of Dr Maxwell Cook, Earth Lab, CU Boulder
Desired Outcomes
The primary goal of the project is to raise awareness of the vulnerability of certain structures and locations to wildfire. “From this research, we hope to reduce fire risk, and the number of homes lost to fires every year,” Iglesias says.
This research, once complete, will be open-source and available to homeowners, researchers, and policymakers. Armed with such information, policymakers can proactively change building and zoning codes to reduce fire risk, insurance companies can more accurately predict risk, and fire districts can begin to understand how to better manage their resources.
By examining and understanding how fence lines, building materials, and neighborhood layouts influence fire behavior, the Earth Lab team is filling critical research gaps that equip homeowners, policymakers, insurers, and fire districts with the actionable insights they need to safeguard communities. The resultant knowledge will serve as a catalyst for smarter building codes, more accurate risk assessments, and more strategic resource allocation—shifting wildfire response from reactive to proactive and making the growing communities of Colorado and Wyoming safer.
