Researchers awarded NSF grant for Lahaina wildfire simulation
Interdisciplinary team to combine AI, high-performance computing and modeling to better understand Lahaina fire; reduce likelihood of future disasters
Release Date: October 10, 2023
BUFFALO, N.Y. – A wildfire swept through the west Maui community of Lahaina on Aug. 8 killing at least 97 people and destroying more than 2,000 buildings.
Shortly thereafter, an interdisciplinary research team from the University at Buffalo (UB), the National Center for Atmospheric Research (NCAR), and the University of Nevada, Reno (UNR) took what information was available on the fire and began modeling the event.
Researchers knew that strong winds and other conditions helped the blaze spread rapidly, but they wanted to better understand how non-natural factors – including buildings, utility lines and other infrastructure – helped fuel the fire.
The UB researchers – Negar Elhami-Khorasani and PhD student Fernando Szasdi Bardales – used a tool they previously created called SWUIFT (short for Streamlined Wildland-Urban Interface Fire Tracing) to simulate the Lahaina fire.
“Our preliminary simulation of the overall timeline of the fire spread in Lahaina aligns with witness observations and video recordings. We are using the simulation results to understand the drivers of fire spread during the event,” says Elhami-Khorasani, PhD, associate professor in the Department of Civil, Structural and Environmental Engineering at the UB School of Engineering and Applied Sciences.
NCAR researchers generated wind input for the model.
“Given the complex topography of Maui, we used a state-of-the-art weather model to simulate the evolution of the windstorm and the resulting highly variable wind flow,” says Timothy Juliano, PhD, project scientist at NCAR. “Conducting our simulations at fine spatial scales was necessary to understand the local wind patterns that affected the island and fire spread during the event.”
The UNR team then analyzed the wind flow, which showed the tell-tale signs of a down-slope windstorm, a culprit in high-impact wildfires such as the Camp Fire in California and the Marshall Fire in Colorado.
“What is unique in this case is that after the initial westward fire run into town, the winds reverse direction and become extremely variable, which drove the fire north, south and east making escape from the fire extremely challenging,” says Neil Lareau, PhD, assistant professor in the Department of Physics at the University of Nevada, Reno.
Hamed Ebrahimian, PhD, assistant professor in the Civil & Environmental Engineering Department at UNR, says the increasing pace of wildfires has outgrown the technology that’s available to prepare for and respond to wildfires.
“There is currently a gap between available research capabilities and operational digital technologies that can be used for wildfire response simulation, prediction and situational awareness,” Ebrahimian says. “In our simulations, we observed the unfolding mechanism of the Lahaina fire disaster. Timely dissemination of the information could have been vital to enable informed decision-making to enhance emergency response management. This underscores the need to join forces and develop prime-time technologies that can move the needle.”
The image above shows a simulation of the Lahaina wildfire at 7:30 p.m. on Aug. 8. The image is taken from the University at Buffalo-developed SWUIFT (Streamlined Wildland-Urban Interface Fire Tracing) program. Credit: University at Buffalo.
NSF awards $200,000 grant to further study Lahaina fire
To improve wildfire preparedness and response, the National Science Foundation awarded a one-year, $200,000 grant to researchers from the University of Hawai’i at Mānoa (UH Mānoa), UB and UNR to further study the Lahaina fire. The grant, from the agency’s Rapid Response Research program (also known as RAPID) began Oct. 1.
Because there are limited conventional wind and environmental observations in the Lahaina wildfire area, the research team will tap into other data sources to present the fullest picture possible of the fire’s spread. This includes capturing social media posts, time-stamped photos and information collected from outreach requests by the research teams.
David Eder, PhD, senior computational scientist and graduate faculty in the Department of Physics and Astronomy at UH Mānoa, and a Maui resident, is the grant’s principal investigator.
He will lead a team that analyzes the data, using artificial intelligence, high-performance computing and other techniques. The goal is to refine and further validate the simulation of wind and fire spread. The project funds UH Maui College students to help in outreach efforts, and in the collection and analysis of the critically needed data.
Grant co-principal investigators include Elhami-Khorasani, Ebrahimian, Lareau and Sean Cleveland, PhD, associate director of cyberinfrastructure and affiliate researcher at the Hawai’i Data Science Institute.
The work will ultimately advance wildfire and atmospheric simulation technology while improving the public’s understanding of how wildfires spread from the wild to rural, suburban and urban communities.
It also will help inform policies that enable government officials and emergency responders to make decisions to reduce the likelihood of major loss of life and property damage in future wildfires and other disasters in Maui and beyond.
Media Contact Information
Cory Nealon
Director of News Content
Engineering, Computer Science
Tel: 716-645-4614
cmnealon@buffalo.edu