An AI map helped Antarctic researchers find a 7.6-kilogram meteorite

A map of promising Antarctic ice led researchers into the field during the BELARE 2022–2023 expedition. They returned with five meteorites, including a specimen weighing 7.6 kilograms. Their report explicitly links the choice of search areas to Veronica Tollenaar and colleagues’ machine-learning predictions.[1]

The team worked around Belgium’s Princess Elisabeth Antarctica station, searching ice fields and accumulations of rocky debris. Nils Larsenfjellet emerged as a potential new collection area. The achievement was a successful reconnaissance: researchers had tested selected places and brought back physical samples.[1]

Why the ice collects space rocks

The Antarctic landscape can concentrate material that originally fell across a much larger area. NASA planetary scientist Barbara Cohen explains that meteorites become buried in glaciers and travel with the flowing ice. When an obstacle forces ice upwards, some of those rocks return towards the surface. Dry winds remove ice and progressively expose the embedded material.[3]

A promising place therefore depends on movement and surface conditions. A rock’s discovery location may be far from where it originally landed. This is why a satellite map of the surrounding landscape can help a search team, even without showing an individual meteorite.[3][2]

Learning where to send the search team

Tollenaar’s 2022 study trained its classifier on places where meteorites had been found and a much larger set of unlabelled locations. An unsearched area could contain meteorites, so the model did not automatically treat the absence of a recorded find as evidence that none were present.[2]

The study outlined roughly 600 candidate accumulation zones across Antarctica. Those are predicted search areas, rather than 600 recovered objects. A favourable score still has to be tested by a visit.[2]

The ground can complicate that visit. Cohen’s account of a separate US expedition describes ordinary Antarctic rocks mixed with meteorites and the importance of an experienced eye. Finding unusual material also requires documenting it carefully: her team photographed candidates, recorded GPS positions and used clean bags to protect them.[3]

A sample can outlast the expedition

The BELARE team transported its recovered meteorites frozen to the Royal Belgian Institute of Natural Sciences. Controlled thawing and mineralogical classification were the next steps described in its report. The map had selected terrain; laboratory investigation would establish the specimens’ properties.[1]

The scientific value extends well beyond a rock’s size. The Field Museum’s meteorite researchers study mineral structures, chemistry and isotopes to investigate early Solar System processes. Different components preserve different evidence: some minerals formed while the young planetary system was cooling, while rare surviving grains in suitable meteorites predate the Solar System itself.[4]

Microscopic material from the region is also receiving attention. A 2025 conference abstract by Clara Zelinsky and colleagues describes research on micrometeorites from Nils Larsen and Roysane. Their programme compares mineral textures, elements and oxygen isotopes to investigate weathering, heating during atmospheric entry and possible source bodies. It is a description of developing research, rather than a completed analysis of the large BELARE specimen.[5]

That range of questions explains the appeal of improving collection. The Field Museum also studies how extraterrestrial material reaching Earth has changed through time, comparing samples preserved in geological deposits. Each well-characterised collection adds material that can be re-examined with different instruments and questions. Helping scientists choose where to search can therefore support discoveries long after the tents have been packed away.[4]

Report an error or suggest a correction