Home · Science · Sep 6 archive

Unusual Structures Identified at Earth's Core-Mantle Boundary with Deep Learning

Confirmed

Science Desk

In Short: Scientists have used advanced machine learning techniques to identify six mysterious structures deep within the Earth, potentially shedding light on seismic activity.

Researchers at the Chinese Academy of Sciences have identified six unusual structures at the core-mantle boundary using data from over two million earthquakes and a machine learning algorithm, according to a new study published in the Journal of Geophysical Research: Solid Earth.

These structures, identified through the analysis of 'PKP precursors'—seismic signals that arrive before the main earthquake—could be related to subduction zones, where material is dragged deep below the surface and chemically altered by extreme pressure and temperature.

The study, which analyzed data from 1990 to 2024, suggests that these ancient structures could help scientists better understand the formation of the Earth's layers and the movement of tectonic plates, potentially aiding in predicting future seismic events.

By mapping these structures, scientists hope to gain insights into the processes that occur deep within the Earth, which could have implications for understanding and mitigating the effects of seismic activity on the surface.

The findings could also contribute to the broader goal of unraveling the mysteries surrounding the Earth's interior, including the dynamics of the core-mantle boundary and the mechanisms that drive tectonic activity.

These discoveries, which have identified approximately 175,000 PKP signals, are about ten times as many as previously identified in all previous studies combined, highlighting the potential of deep learning in advancing our understanding of the Earth's interior.

What's confirmed

What's still developing

Sources