Argonne National Laboratory

04/25/2024 | Press release | Distributed by Public on 04/25/2024 11:16

Unraveling the song of ice and fire across the American landscape with machine learning

April 25, 2024
A new study explores the intriguing relationship between wildfires in the West and hailstorms in the central U.S.
A small plant surrounded by hail after a hailstorm. Fires in the western U.S. enhance large hailstones in the central U.S. (Image by Shutterstock.)

In the rugged terrain of the western U.S., where wildfires rage unchecked, a surprising connection emerges with the tumultuous skies of the central U.S. A recent study published in Advances in Atmospheric Sciences explores the intriguing relationship between wildfires in the West and hailstorms in the central U.S. At the core of this pioneering study led by Jiwen Fan, who was at the U.S. Department of Energy's (DOE) Pacific Northwest National Laboratory and is currently at DOE's Argonne National Laboratory, lies the innovative application of machine learning (ML) techniques to illuminate the hidden link between seemingly disparate phenomena.

Machine learning algorithms, including Random Forest and Extreme Gradient Boosting, are employed to analyze vast datasets spanning two decades, from 2001 to 2020. These MLmodels are trained to predict the occurrence of large hail in central U.S. states based on a multitude of variables, including meteorological conditions in the fire region, wind patterns and characteristics of wildfires themselves.

"We are now able to paint a vivid picture of the intricate relationship between fire and hail across the American landscape. Wildfires in the western U.S., exert a far-reaching influence on atmospheric conditions, shaping the trajectory of severe weather events thousands of miles away - something that we never thought before." - Jiwen Fan, Argonne atmospheric scientist

Through meticulous analysis and data processing, the MLmodels achieve remarkable accuracy, with predictions exceeding 90% in some cases. By identifying key variables and patterns, these models unveil correlations between wildfires in the western U.S. and hailstorms in the central U.S., providing invaluable insights into the remote impacts of wildfires on severe weather events thousands of miles away.

"We are now able to paint a vivid picture of the intricate relationship between fire and hail across the American landscape. Wildfires in the western U.S., exert a far-reaching influence on atmospheric conditions, shaping the trajectory of severe weather events thousands of miles away - something that we never thought before," said Fan. ​"Meteorological variables like westerly wind, the temperature and relative humidity in the fire region and the intensity of wildfires emerge as key players in this climatic symphony."

Yet, amidst the marvel of discovery, challenges abound. Attempts to predict the daily count of large hail events reveal the complexities of nature's whims, reminding us of the unpredictable nature of weather phenomena. As researchers continue to refine their models and confront data imbalances, the quest for understanding presses onward.

The utilization of MLtechniques represents a significant advancement in atmospheric science, allowing researchers to navigate complex datasets and extract meaningful patterns that may have eluded traditional statistical methods. With MLas their guiding light, scientists embark on a journey to unravel the mysteries of Earth's interconnected systems and pave the way for more accurate predictions and proactive measures in the face of evolving climate dynamics.

Original story from the Institute of Atmospheric Physics, Chinese Academy of Sciences.

Argonne National Laboratory seeks solutions to pressing national problems in science and technology. The nation's first national laboratory, Argonne conducts leading-edge basic and applied scientific research in virtually every scientific discipline. Argonne researchers work closely with researchers from hundreds of companies, universities, and federal, state and municipal agencies to help them solve their specific problems, advance America's scientific leadership and prepare the nation for a better future. With employees from more than 60 nations, Argonne is managed by UChicago Argonne, LLCfor the U.S. Department of Energy's Office of Science.

The U.S. Department of Energy's Office of Science is the single largest supporter of basic research in the physical sciences in the United States and is working to address some of the most pressing challenges of our time. For more information, visit https://​ener​gy​.gov/​s​c​ience.

Tags:Computer Science and Engineering, Machine Learning, Environmental and Earth Science, Atmospheric Science, Climate Science, Argonne Climate News