In order to probe Earth’s vast depths, scientists employed a machine learning algorithm that was originally devised to “analyze distant galaxies”.
Having studied seismograms from hundreds of earthquakes that took place from 1990 till 2018, a team of scientists led by Doyeon Kim, a seismologist and postdoctoral fellow at the University of Maryland, stumbled upon the presence of some sort of massive structure comprised of dense material, located between our planet’s core and the lower mantle, the Motherboard reports.
According to the media outlet, during the course of their study, the team employed a machine learning algorithm called Sequencer that was “originally developed to analyze distant galaxies”.
“This study is very special because, for the first time, we get to systematically look at such a large dataset that actually covers more or less the entire Pacific basin,” Kim remarked.
The researchers’ study of seismograms of the so called shear waves and the “echo-like” signatures they produce when hitting the structures in question has apparently indicated the presence of anomalies known as called ultra low velocity zones (ULVZs).
As the media outlet points out, at this time, no one knows how these ULVZs are formed or what exactly they’re comprised of – what is known is that they are dense and have diameters of about a hundred kilometers, and that there’s evidence of the existence of two “mega-ULVZs” which “stretch for about 1,000 kilometers or more”.