When we think of an earthquake, we imagine the shaking of the ground and the destruction it can cause on the surface.
But shaking is just one small part of the story. A new study by MIT geologists shows that most of an earthquake’s energy doesn’t actually go into the tremors we feel—it vanishes deep underground as heat.
Earthquakes happen when tectonic plates—the giant slabs of Earth’s crust—grind against each other.
Over millions of years, rocks under pressure build up stress. Eventually, the stress becomes too much, and the rocks slip suddenly along a fault line.
This slipping sends seismic waves rippling through the ground, which we experience as shaking.
Seismometers can measure this motion, but they can’t easily reveal what happens to all the energy that is released underground.
Until now, scientists didn’t know exactly how much earthquake energy is spent on shaking, how much on breaking rocks, and how much simply becomes heat.
To answer this puzzle, MIT researchers created miniature earthquakes in the laboratory—what they call “lab quakes.”
These small-scale experiments allowed them to track the energy released during a quake in detail.
Their results, published in AGU Advances, reveal that earthquakes are surprisingly inefficient at producing the shaking we feel. In fact, only about 10 percent of a quake’s energy goes into ground motion.
Less than one percent breaks rocks apart. The vast majority—around 80 percent—turns into heat.
That heat isn’t gentle warmth, either. In some experiments, the researchers saw temperatures spike from room temperature to about 1,200 degrees Celsius in just microseconds—hot enough to briefly melt rock.
Once the slip stopped, the heat dissipated just as quickly. This shows that earthquakes are mostly violent heat engines, burning through energy deep in the crust.
To perform these experiments, the team used samples of granite, a rock type commonly found in the Earth’s crust where earthquakes start. They ground the granite into powder and added tiny magnetic particles that could act as thermometers. These particles change their magnetic field when heated, giving the researchers a way to measure temperature changes after a quake.
The powdered rock was placed between pistons, wrapped in a gold jacket, and squeezed under pressures similar to those found 10 to 20 kilometers beneath the Earth’s surface.
As the stress built up, the samples suddenly slipped—just like faults in the Earth. During these “microquakes,” sensors recorded how much the samples shook, while the magnetic particles revealed how hot the rocks became.
Afterward, the scientists examined the samples under microscopes to see how many grains had fractured. With all this information, they could reconstruct the full energy budget of each quake.
The findings also showed that a region’s “deformation history” matters. Rocks that have been shifted, stressed, or broken in the past behave differently in new quakes.
Their memory of past stress influences how energy is divided among heat, shaking, and fracturing. That means two earthquakes of the same size can behave very differently depending on the history of the rocks involved.
This research may eventually help seismologists better understand and predict earthquake hazards. For instance, if scientists know how much shaking a past quake produced, they might estimate how much heat was also released underground, and how that heat may have altered the rocks. That knowledge could reveal how vulnerable a fault is to future quakes.
Of course, these lab quakes are tiny compared to the real thing. But by isolating the basic physics in controlled conditions, scientists hope to bridge the gap between small experiments and Earth’s massive seismic events.
“We can’t reproduce the complexity of the Earth in the lab,” says MIT geophysicist Matěj Peč, “but we can strip things down to their fundamentals and try to scale them up.”
The new results also highlight just how sudden and extreme earthquakes are. A slip that lasts only microseconds can unleash enormous heat and energy. Understanding that balance may lead to better earthquake models and, ultimately, strategies to reduce the damage when the ground shakes.
As lead researcher Daniel Ortega-Arroyo puts it: “The deformation history—what the rock remembers—really influences how destructive an earthquake could be.”
In other words, Earth’s crust has a memory, and learning how it stores and releases energy may be the key to unlocking safer futures in quake-prone regions.
