AI Brings Clarity to the Aurora Borealis, Enhancing Space Weather Forecasts

Harnessing the power of machine learning, scientists have transformed a vast dataset of auroral images into a treasure trove of insights about Earth's magnetosphere and solar wind interactions.

Research: Automatic Detection and Classification of Aurora in THEMIS All-Sky Images. Image Credit: Manzooranmai / Shutterstock

The aurora borealis, or northern lights, is known for a stunning spectacle of light in the night sky, but this near-Earth manifestation, which is caused by explosive activity on the sun and carried by the solar wind, can also interrupt vital communications and security infrastructure on Earth. Using artificial intelligence, researchers at the University of New Hampshire categorized and labeled the largest-ever database of aurora images that could help scientists better understand and forecast disruptive geomagnetic storms.

A Breakthrough in Analyzing Massive Data Sets

The research, recently published in the Journal of Geophysical Research, utilized a self-supervised learning method called SimCLR, enabling efficient annotation of a vast data set. The artificial intelligence and machine learning tools successfully identified and classified over 706 million images of auroral phenomena in NASA's Time History of Events and Macroscale Interactions during Substorms (THEMIS) data set. The data set, which was collected by twin spacecraft studying the space environment around Earth, provides images of the night sky every three seconds from sunset to sunrise from 23 different stations across North America.

"The massive data set is a valuable resource that can help researchers understand how the solar wind interacts with the Earth's magnetosphere, the protective bubble that shields us from charged particles streaming from the sun," said Jeremiah Johnson, associate professor of applied engineering and sciences and the study's lead author. "But until now, its huge size limited how effectively we can use that data."

A sample of images from the Oslo Aurora THEMIS data set (OATH). From left to right, the manually assigned ground–truth labels are “Arc,” “Diffuse,” “Discrete,” “Cloudy,” “Moon,” “Clear.” Note that these images are false color images; they have been converted to RGB with intensity information recorded in the green channel. 

Novel Algorithm and Enhanced Insights

The researchers created a novel algorithm to sort through the THEMIS all-sky images (ASI) from 2008 to 2022. They efficiently annotate them using six distinct categories: arc, diffuse, discrete, cloudy, moon, and clear/no aurora. By employing SimCLR, the team overcame the challenge of limited labeled data by leveraging a small manually labeled subset and generalizing to the entire data set. This algorithm makes filtering, sorting, and retrieving valuable information more manageable.

"The labeled database could reveal further insight into auroral dynamics, but at a very basic level, we aimed to organize the THEMIS all-sky image database so that the vast amount of historical data it contains can be used more effectively by researchers and provide a large enough sample for future studies," said Johnson.

The research also highlighted correlations between auroral phenomena and solar wind parameters, interplanetary magnetic field values, and geomagnetic indices. These insights enhance our understanding of the interaction between the solar wind and Earth's magnetosphere.

Collaborative Effort and Future Potential

Co-authors on the study include Amy Keesee, associate professor of physics and astronomy in UNH's Space Science Center; Doğacan Su Öztürk, Donald Hampton, and Matthew Blandin, all from the University of Alaska–Fairbanks; and Hyunju Connor from NASA Goddard Space Flight Center. The research was funded by NASA's heliophysics division and the National Science Foundation.

The University of New Hampshire inspires innovation and transforms lives in our state, nation, and world. More than 16,000 students from 50 states and 87 countries engage with an award-winning faculty in top-ranked programs in business, engineering, law, health and human services, liberal arts, and the sciences across more than 200 programs of study. A Carnegie Classification R1 institution, UNH partners with NASA, NOAA, NSF, and NIH and received over $250 million in competitive external funding in FY24 to further explore and define the frontiers of land, sea, and space.