AI‑Driven Seismic Survey Reveals Six Unusual Structures at Earth’s Core‑Mantle Boundary

Scientists employed artificial‑intelligence techniques on seismic recordings and uncovered six separate, previously unnoticed formations at the interface between Earth’s liquid outer core and solid mantle, situated about 2,900 km below the surface.

The finding, described in a paper in Journal of Geophysical Research: Solid Earth, results from an innovative method that feeds deep‑learning networks thousands of earthquake records. By training the system to detect minute changes in seismic wave speed and direction, the researchers mapped heterogeneities that traditional techniques overlooked.

Since no instrument can directly access the core‑mantle boundary, investigators depend on earthquake‑generated seismic waves to deduce its characteristics. While traversing Earth’s interior, these waves are modified by variations in composition, temperature and phase. The AI‑based examination pinpointed six areas where the waveforms consistently diverged, implying unusual entities like localized compositional anomalies or thermal plumes.

The results contribute to an expanding literature that depicts the core‑mantle interface as a dynamic, heterogeneous region instead of a smooth, uniform layer. Earlier work has documented extensive low‑shear‑velocity provinces and ultra‑low velocity zones, yet the six structures reported here are finer and more isolated, suggesting a level of complexity previously unrecognized.

These insights bear on multiple geophysical enigmas, such as the forces behind mantle convection, the origin of Earth’s magnetic field, and the planet’s long‑term thermal evolution. Boundary anomalies may alter heat flow from the scorching core into the mantle above, possibly modifying the intensity of mantle plumes that emerge as volcanic hotspots.

The lead authors warn that although the deep‑learning model yields a strong statistical signature, additional confirmation using independent datasets and different analytical methods is required. Upcoming studies might incorporate data from dense seismic networks deployed across the United States and Europe to sharpen the spatial resolution of the detected structures.

This research highlights the expanding influence of machine learning in Earth sciences, serving as a potent supplement to conventional seismology. As computational techniques advance, scientists expect to reveal further concealed features deep inside the planet, enhancing our grasp of the processes that sculpt Earth’s interior.