BYU grad develops groundbreaking algorithm for wildfire prediction

PROVO, Utah — A recent Brigham Young University math graduate is breaking barriers in wildfire prediction. Her research helped develop a computer algorithm that better predicts where the wind could push fires. Jane Housley developed the algorithm using artificial intelligence, which makes predicting wind patterns of wildfires seven times faster while maintaining the same level of accuracy that leading researchers are used to seeing in high-end wildfire prediction models. When fires are burning, that time is valuable. On June 25, Housley showed KSL a brief demonstration of her algorithm. For the last year and a half, she'd been working with WindNinja, the leading wind prediction software used across the country. The program had two wind models—one was fast and one that was more accurate. Housley said she taught AI how to combine the two. In her eyes this opens the possibilities to learn even more. "So, the next steps are kind of like where else can we use machine learning in this big field of wildfire modeling to improve wildfire modeling as a whole? I've done this one part of wind, but there's so many other parts we can explore. We've kind of shown this does work and works really well," Housley said.

What's next in Housley's wildfire prediction research?

Housley said she plans to hand off her findings to another BYU student. They will complete the research and share it with WindNinja. Then, the program's developers can incorporate it in their software. Right now, there isn't a specific timeline, but it's happening quickly. Housley expects her research to be published in a new paper by the end of July. The hope is to get the program into the hands of crews who need it as soon as possible. It's wildfire season in Utah. Read our latest reporting: A tree in Pine Valley was struck by lightning. Here’s how the Forsyth Fire took over 10 days later

Related topics