Beneath the surface of every earthquake lies a record — and now, for the first time, a machine has read two million of them at once. Researchers trained an artificial intelligence in the physics of seismic waves and unleashed it on decades of planetary data, where it found six structural zones deep in Earth's mantle that no human team had ever formally identified. The discovery is less a single answer than a new set of questions: about how the planet's interior has evolved over billions of years, and about what it means when a machine becomes the instrument through which nature finally reveals
AI Discovers Six Hidden Zones in Earth's Mantle Using 2 Million Earthquake Records
Six hidden zones in Earth's mantle, revealed by two million earthquake records
Mark
So the AI didn't discover these zones by inventing something new—it just found patterns in data that already existed?
Mimi
Exactly. Two million earthquakes have been recorded over decades. The seismic waves from all of them contain information about Earth's interior. The AI was trained in physics and asked to find zones with consistent seismic signatures. It's like having a massive library of fingerprints and finally having a system fast enough to compare them all.
Luke
But I want to know: how confident are we that these six zones are real and not artifacts of the algorithm? What's the validation process?
Mimi
That's the right question. The zones were identified based on seismic wave patterns, but they'll need independent confirmation—probably through other geophysical methods and detailed modeling.
Mark
And what do we actually know about what these zones are made of?
Mimi
That's still open. The seismic signatures tell us they're structurally distinct, but composition and temperature require more investigation.
Luke
So we have six new regions identified, but we don't yet know their material properties or their function in Earth's dynamics?
Mimi
Correct. The discovery is the identification itself. Understanding what they mean comes next.
Mark
How does this change what we thought we knew about the mantle?
Mimi
It suggests the mantle is more structurally complex than previous models indicated. These zones may help explain heat flow and convection patterns we've observed but couldn't fully account for.
Luke
And the AI—was it trained on earthquake data from all over the world, or specific regions?
Mimi
The source indicates it analyzed two million earthquake records, but I'd want to know the geographic distribution and time period covered.
Mark
What's the practical importance? Does this affect how we predict earthquakes or volcanic eruptions?
Mimi
Not immediately. But understanding mantle structure better could eventually improve our models of Earth's internal dynamics, which does influence seismic and volcanic activity.
El Pulso
- Two million earthquakes worth of seismic data sat largely unread at the scale needed to reveal the mantle's hidden architecture — a volume of information that simply exceeded human processing capacity.
- An AI trained in physics, not just pattern-matching, tore through tens of millions of wave recordings and surfaced six distinct structural zones near Earth's core that had never been formally mapped.
- Each zone carries its own seismic signature, suggesting the mantle is not a uniform mass but a layered history — evidence of billions of years of cooling, movement, and internal reorganization frozen in rock.
- The scientific community now faces the harder work: determining what these zones are made of, how they formed, and whether they hold answers about volcanism, plate tectonics, and Earth's magnetic field.
- The breakthrough reframes AI's role in science — not as a tool that replaces theory, but as one that applies known physics at a scale that makes entirely new patterns visible for the first time.
Beneath the surface of every earthquake lies a record — and now, for the first time, a machine has read two million of them at once. Researchers trained an artificial intelligence in the physics of seismic waves and unleashed it on decades of planetary data, where it found six structural zones deep in Earth's mantle that no human team had ever formally identified. The discovery is less a single answer than a new set of questions: about how the planet's interior has evolved over billions of years, and about what it means when a machine becomes the instrument through which nature finally reveals itself.
Somewhere beneath your feet, two million earthquakes have left their mark in seismic waves — ripples of energy carrying information about the planet's hidden architecture. A research team trained an artificial intelligence in the principles of physics and set it loose on this vast archive, looking for patterns human geologists might have missed. What emerged was unexpected: six distinct structural zones in Earth's mantle, near the planet's core, never formally identified before.
The mantle — the thick layer of hot rock between Earth's crust and core — has long resisted direct study. Seismic waves pass through it constantly, and the way they bend, slow, or scatter reveals something about the material they move through. But parsing two million earthquake records by hand, searching for subtle variations across the entire mantle, exceeds human capacity. The AI could do what no individual researcher could: process the full dataset at once and identify zones with consistent seismic characteristics.
Each of the six zones appears to transmit and alter earthquake waves in its own particular way. Their existence tells a story about how Earth's interior has evolved — the mantle is not static, and these zones are evidence of billions of years of movement, cooling, and reorganization, preserved in the rock itself.
The barrier has always been volume. A single earthquake generates recordings across hundreds of stations worldwide. Two million earthquakes means tens of millions of individual data points. Finding meaningful signal in that noise required either years of coordinated human effort, or a machine trained to apply known physics systematically at scale — the difference between knowing a forest exists and actually mapping every tree.
The discovery now opens into harder questions. Researchers will need to determine what material fills each zone, how they formed, and what role they play in Earth's internal dynamics — potentially illuminating how heat moves through the mantle, how volcanoes and tectonic plates behave at the surface, and how the planet's magnetic field is shaped. The AI has opened a door. The work of walking through it has only just begun.
Somewhere beneath your feet, two million earthquakes have left their mark in the form of seismic waves—ripples of energy that travel through rock and stone, carrying information about the planet's hidden architecture. A team of researchers trained an artificial intelligence system in the principles of physics and set it loose on this vast archive of seismic data, looking for patterns that human geologists might have missed. What emerged was unexpected: six distinct structural zones in Earth's mantle, regions near the planet's core that had never been formally identified before.
The mantle is the thick layer of hot rock that sits between Earth's crust and its core, and it has long been difficult to study directly. Seismic waves from earthquakes pass through it constantly, and the way those waves behave—how fast they travel, how they bend and scatter—reveals something about the material they move through. Denser regions slow waves down. Cooler zones affect their path. But parsing two million earthquake records by hand, looking for subtle variations in wave behavior across the entire mantle, is work that exceeds human capacity. The AI, trained to recognize the physical principles governing how seismic energy moves through rock, could do what no individual researcher could: process the entire dataset at once and identify zones that shared consistent characteristics.
The six zones the system identified represent distinct structural regions in Earth's deep interior. Each one appears to have its own seismic signature—a particular way of transmitting and altering earthquake waves that sets it apart from the surrounding mantle. The discovery is significant not just because these zones exist, but because their presence tells a story about how Earth's interior has evolved. The mantle is not static. Over billions of years, it has moved, cooled, heated, and reorganized itself. The zones that the AI found are evidence of that history, frozen in the rock itself.
This kind of discovery has been possible in theory for decades. Seismologists have long known that the mantle contains variations in temperature and composition. But the sheer volume of data required to map those variations with confidence has been a barrier. A single earthquake generates seismic waves that are recorded by hundreds of stations around the world. Two million earthquakes means tens of millions of individual wave recordings. Finding meaningful patterns in that noise requires either an enormous team of researchers working for years, or a machine that can process the information faster and more systematically than any human could.
The breakthrough points to a broader shift in how science works. Artificial intelligence, when trained properly in the underlying physics of a problem, can become a tool for discovery rather than just analysis. The system didn't invent new physics or contradict existing theory. Instead, it applied known principles to data at a scale that made new patterns visible. It's the difference between knowing that a forest exists and actually being able to map every tree in it.
What happens next is equally important. The six zones now need to be studied in detail. Researchers will want to understand what material makes up each zone, how they formed, and what role they play in Earth's internal dynamics. The zones may help explain how heat moves through the mantle, which in turn affects volcanic activity and plate tectonics at the surface. They may reveal something about the planet's magnetic field, which is generated in the outer core and influenced by mantle convection. The AI has opened a door; now comes the work of walking through it.
Quieres la nota completa?
Lee el original en Google News ↗
Technology