This next story, honestly, it is the kind of thing that could have been done by anyone with an amazing grasp of machine learning. Research in the journal Science Advances and led by Veronica Tollenaar reports that a new machine learning algorithm has identified more than 600 locations where ice movements in Antarctica may have piled up meteorites — meteorites scientists want to get their hands on so they can study the chemistry and geology of other worlds. 

These space rocks are essentially free samples or at least free compared to the costs of launching a spacecraft to collect samples directly from asteroids.

While meteors streak through all the skies of our world, finding them in most places is exceedingly difficult because they are hard to distinguish from local rocks, and they can get lost in the foliage, buried over time, or they may just land in water.

Antarctica, with its amazing glaciers, collects meteorites on ice flows, and over time, can pile up rocks collected over the breadth of the continent into locations where the glaciers uplift materials or come to a melty end. The trick comes in finding these places, and while satellite data can be used, looking through all the images can be time-consuming.

So they trained a computer to do it by feeding the algorithm information on ice’s velocity, thickness, surface temperatures, and the shape of the bedrock below, and by showing it where known meteorite stranding zones are located. According to the release related to this work: To date, about 45,000 meteorites have been plucked from the ice. But that’s a fraction of the 300,000 bits of space rock estimated to lie somewhere on the continent’s surface. 

Hopefully, with a new list of more than 600 places to search, these free space rock samples will be found a whole lot easier in the future.

More Information

Machine learning points to prime places in Antarctica to find meteorites (Science News)

“Unexplored Antarctic meteorite collection sites revealed through machine learning,” Veronica Tollenaar et al., 2022 January 26, Science Advances