Scientists still can't predict when an earthquake will strike. But a team of mathematicians and seismologists just found a way to understand what lies beneath the surface far more quickly — and that changes everything about how we prepare.
On December 6, 2025, a magnitude 7.0 earthquake hit Alaska. The USGS records roughly 55 earthquakes daily worldwide, about 20,000 per year. Most are small. But the larger ones — and there are roughly 15 magnitude 7+ events annually — reshape landscapes and economies. In 2025 alone, an offshore magnitude 8.8 quake near Russia's Kamchatka Peninsula ranked among the ten largest ever recorded. Earthquakes now cost the United States around $14.7 billion annually, partly because more people live in seismically active regions.
We can't stop them. We can't even reliably forecast them. But we can get smarter about the ground itself.
We're a new kind of news feed.
Regular news is designed to drain you. We're a non-profit built to restore you. Every story we publish is scored for impact, progress, and hope.
Start Your News DetoxUnderstanding what's underground
Different rock types — solid stone, sand, clay — transmit seismic waves differently. When an earthquake happens, those waves behave one way through granite and another through sediment. Understanding the subsurface composition helps scientists assess where earthquakes pose the greatest risk.
The standard approach uses something called Full Waveform Inversion. Scientists run computer simulations of synthetic earthquakes, model how seismic waves would propagate through different subsurface configurations, then compare those simulations to real seismograph data from actual earthquakes. After many iterations, when the synthetic data matches the real data closely enough, they've built an accurate picture of what's underground. It works. It's also computationally expensive — the kind of calculation that demands serious processing power and time.
Kathrin Smetana, a mathematician at Stevens Institute of Technology, saw an opportunity. Working with computational seismologists from Utrecht University and the University of Twente in the Netherlands, her team developed a "reduced model" that does the same job with roughly 1000 times less computational power.
"We found a clever way to construct the reduced model while still maintaining the accuracy of the prediction," Smetana explains. The breakthrough came from an interdisciplinary approach — mathematicians and seismologists thinking through the problem together, finding where unnecessary complexity could be stripped away without losing precision.
What this unlocks
The technique can't predict earthquakes. That's still beyond reach. But faster, more practical subsurface mapping means better risk assessment. Cities in earthquake zones could generate realistic geological models with standard computing resources, not just supercomputers. Schools, hospitals, and infrastructure planners could make smarter decisions about where to build and how to reinforce.
The work also opens doors for tsunami simulation. When underwater earthquakes trigger waves, there's often an hour or more before they reach land — enough time to run simulations and issue warnings. With less computational burden, these models become feasible in real time.
"There's no way to predict earthquakes at this time," Smetana says. "But our work can help generate a realistic view of the subsurface with less computational power, which would make our models more practical and help us be more earthquake resilient." The next phase is putting this into practice — getting the technique into the hands of seismologists and urban planners who need it.
