GPS Watched Kamchatka's Fault Lock Up. Then a Magnitude 8.8 Ruptured Exactly There.

Geophysicists at the University of California, Riverside have published an algorithm that uses GPS measurements of ground motion to identify the specific stretches of a fault where the next great earthquake is most likely to begin — and they tested it against a magnitude 8.8 that had not yet happened when the analysis was set up.

The method, developed by Gareth Funning and Axel Periollat and published in Geophysical Research Letters, works from a simple physical premise. A subduction zone, where one tectonic plate slides beneath another, does not slide smoothly along its whole length. Parts of the interface are locked — welded together by friction — while the plates on either side keep converging at a few centimeters a year. That mismatch bends the crust, and the bending shows up as measurable displacement at GPS stations on the surface. The locked patches are called asperities, and they are where the accumulated strain is eventually released.

"Earthquakes capture headlines when they happen, but for years beforehand the fault is quietly accumulating strain," Funning said. "This strain can be measured."

The Kamchatka subduction zone off eastern Russia was the test case. Running the algorithm on the region's deformation data identified a strongly locked region along the interface. In 2025 a magnitude 8.8 earthquake — one of the largest instrumentally recorded anywhere — ruptured that zone, and the slip concentrated where the model said the strain had piled up. The team also found that the 2025 event produced a smaller tsunami than the magnitude 9.0 that struck the same region in 1952, which they attribute to a different distribution of slip: less movement in the shallowest part of the fault, closest to the trench, where slip does the most to displace the water column above.

The limitations are explicit and the authors state them plainly. The method says nothing about timing. It cannot forecast how much of a locked patch will fail in a given event, how far a rupture will run once it starts, or how large the resulting tsunami will be. What it produces is a map of where the energy is stored, which is a different and more tractable problem than the one seismologists have failed to solve for a century.

The group is now applying the approach to subduction zones off Japan, Mexico, New Zealand and the Pacific Northwest, plus California's Hayward Fault, which runs through the East Bay beneath roughly two million people. The main obstacle to extending it is that GPS stations sit on land while subduction zones sit offshore, so the most important part of the fault is the part the network sees worst. Improving that means seafloor geodesy — instruments on the ocean bottom — which is expensive and, in most of the world, barely deployed.