Thanks to AI in a new study from the Lowell Observatory in Flagstaff, Arizona, hidden insights into meteor falls that have long gone unnoticed in observational data are finally revealed.
This is detailed in a new paper published in . IcarusThe researchers used data from the Lowell Observatory Camera for All-Sky Meteor Monitoring (LO-CAMS) network. This network is part of the Global Meteor Network, which features more than 28,000 meteor events.
Their work contributes significantly to expanding the parameters used to classify meteors and provides a deeper understanding of why these space rocks are so unique.
Meteor explanation
A meteor is an object that flares up in the sky in a bright streak across the sky, and is commonly referred to as a shooting star or shooting star. A meteorite is a piece of space rock that somehow joins together and lands on the Earth’s surface.
The terminology used to describe these objects can be confusing, as the distinction between meteoroids, shooting stars, and meteorites is not always clear-cut. While still in space, these rocks are called meteoroids, but the difference becomes noticeable as soon as they enter Earth’s atmosphere.
Meteors are a common sight in the night sky and have fascinated humans for thousands of years, but there’s still much to learn about them, according to researchers at Lowell Observatory.
“Meteors have been observed for centuries, but only recently have datasets been large and detailed enough to apply modern machine learning techniques,” said lead author Sam Hemmelgarn. “This allows us to extract physical information that was previously hidden in the data.”
Evolving meteor observation
Traditionally, only a few parameters have been used to characterize meteors. The new study expands this to 13, including speed, brightness, duration, height, and atmospheric density.
“Our goal was to go beyond traditional classification systems,” said co-author Nick Moskowitz. “Modern meteor networks provide a wealth of observational information, so we needed a framework that could make the most of it.”
The team combined multiple machine learning algorithms to identify natural groupings in the data that reflect existing physical meteor models. This analysis reveals three important factors that determine the behavior of meteors upon entry into the atmosphere. These were its size and shape, ease of heating, also known as “activation,” and its path of travel.
“One of the most interesting results was how clearly the asteroid material was separated from the cometary material by the ‘activation’ behavior,” Hemmelgarn explained. “This shows that we’re capturing something fundamentally physical, not just a statistical pattern.”
New classification scheme
As a result of the study, the researchers developed a new classification system, Hclass, to identify the hardness of meteors. The hardest end of the new scale contains dense material with high iron content, commonly associated with asteroids, while the far end contains brittle, porous material that likely came from comet debris.
The scheme of this new classification system is multi-layered, allowing for more general or more detailed classifications depending on the needs of the researcher. Moreover, it handles a wide range of datasets, ranging from single orders of magnitude to millions of observations.
“Hclass provides a more nuanced view of the composition of meteors,” Hemmelgarn says. “This bridges the gap between classical meteor theory and the reality of modern large-scale observations.”
The research team tested the new scale by fitting data from known meteor showers to it to see how those materials behave in real-world observations. The validation was successful and the meteor behaved as expected based on the classification.
“This study shows that machine learning is not just about processing big data,” Moskowitz said. “We use these data to work toward a physical understanding of where this material comes from and how our solar system functions.”
The paper titled “Machine learning approach to meteor classification” is Icarus April 27, 2026.
Ryan Whalen covers science and technology for The Debrie. He holds a master’s degree in history and a master’s degree in library and information science with a certificate in data science. Contact him at ryan@thedebrief.org and follow @mdntwvlf on Twitter.
