Generated by Codex with GPT 5.6 Sol High

Earthquakes do not become catastrophic simply because a fault begins to slip. The rupture must keep racing along the fault, releasing stored tectonic stress as seismic waves. Jacek Krywko’s Scientific American article examines new evidence for the underground barriers that can abruptly halt this process—and for the distinctive shock wave created when they do.

How a rupture becomes a megaquake

An earthquake begins when stress along a fault overcomes the friction holding blocks of rock in place. Once the fault slips, a rupture can spread rapidly through Earth’s crust. It may fade when it reaches a region with little stored stress, or it may collide with a physical obstacle such as a change in rock composition or a pocket of high-pressure fluid.

These obstacles do not always stop an earthquake. If a rupture still carries enough energy, it can break through one barrier and continue into the next fault segment. That cascade can turn a local event into a megaquake. If the barrier holds, however, the rupture can end almost instantly—more like a speeding train hitting a wall than a vehicle gradually braking.

The sudden halt sends a shock wave backward along the fault. Seismologists call this signal a “stopping phase.” Someone standing above the barrier would first feel the ground move in the direction of the advancing rupture and then snap sharply the other way, much as a passenger lurches when a car’s brakes engage.

The mechanism had been predicted, but clean observations were scarce. Earth scientists Jesse Kearse of Victoria University of Wellington and Yoshihiro Kaneko of Kyoto University searched seismic and geodetic records from 12 large earthquakes around the world. Five had sensors distributed densely enough along the fault for the researchers to isolate a stopping phase.

A map of underground checkpoints

Finding this signature matters because it offers a way to reconstruct barriers hidden deep underground. Researchers could search past earthquake records for stopping phases, locate the obstacles that produced them and estimate how much rupture energy each barrier absorbed. In principle, those maps could help show where an earthquake of a given strength is likely to stop—and where it might punch through and grow.

The data also showed that geology near the surface can alter the danger. Soft rock layers above a stopping point may amplify its seismic signal, producing stronger shaking at ground level. A barrier that limits the total length of a rupture can therefore still sit beneath an area that experiences severe local motion. Hazard models need to account for both the deep structure that governs rupture growth and the shallower layers that modify what people feel.

The result does not yet make megaquakes predictable. Kearse and Kaneko studied only strike-slip earthquakes, in which crustal blocks slide horizontally past one another, because those events offered the richest observations. The magnitude 7.7 earthquake off northeastern Japan that frames the article was instead a thrust event, involving vertical motion and a greater tsunami risk. The researchers expect stopping phases to occur in thrust earthquakes as well, but they could not confirm that from this study.

The practical promise is therefore real but preliminary. Stopping phases give seismologists a new observational handle on why some ruptures stay small while others cross one fault segment after another. Turning that handle into dependable forecasts will require many more well-instrumented earthquakes, broader tests across fault types and models that connect underground barriers to surface shaking. For now, the study replaces an inferred “stop sign” with a signal scientists can begin to find and map.

Source: Jacek Krywko, “Seismic Stop Sign,” Scientific American, September 2026, pp. 15–16.