Antarctica is now understood to be more seismically active than previously thought, with deep intraplate earthquakes at mid-crustal depths prompting new questions about how stresses operate within continental interiors.

This hidden seismic activity could reflect that deep continental earthquakes are more common than assumed, potentially reshaping theories about tectonic activity in historically quiet regions.

The findings also show the power of machine-learning tools to extract signals from older seismic records, revealing activity that traditional methods missed.

To reach these conclusions, researchers reanalyzed two seismic datasets from the early 2000s and the 2010s using deep learning to identify earthquakes that had been overlooked.

The earthquakes cluster where cold, rigid crust meets warmer, softer rock at depth, creating a sharp contrast in tectonic strength that enables deep seismic activity within the continental interior.

Lead authors explain that the boundary between very different rock environments underlies the deeper quakes, tied to the juxtaposition of East and West Antarctica’s lithospheric properties.

The David Glacier region, a vast 1100-kilometer ice stream, may influence stress conditions that contribute to deep seismicity nearby.

The detected events are intraplate intermediate-depth earthquakes, with magnitudes between 1.6 and 3.5, occurring at depths beyond 43 miles beneath the crust.

These earthquakes originate at depths around 100 to 145 kilometers under the David Glacier, inside a plate rather than at plate boundaries, challenging conventional plate tectonics expectations.

Though typically linked to subduction zones, intermediate-depth quakes inside a continental plate are unusual and hard to fit into standard models.

Locations 60 to 90 miles beneath David Glacier lie where East and West Antarctica meet contrasting rock environments, concentrating tectonic activity at a boundary.

Although these deep quakes are too weak to threaten ice sheets or ecosystems, understanding them helps improve models of Antarctic dynamics and sea-level rise predictions, and demonstrates AI’s value in parsing old seismic data.