Ocean mud suggests Cascadia and the San Andreas fault have broken within hours of each other

A research cruise in 1999 was meant to sample the seafloor off Oregon and Washington for traces of ancient Cascadia earthquakes. A navigational error put the ship about 90 kilometers south of Cape Mendocino instead, and the sediment core it pulled there held a pattern that Oregon State University marine geologist Chris Goldfinger and his colleagues have spent years interpreting. Their reading is that the Cascadia megathrust and the northern San Andreas fault have ruptured within minutes to hours of each other.

Read next on Morning Overview: Cadillac cut the price of its cheapest EV, which goes up to 317 miles on a charge

The finding rests on layers of mud, not on a seismograph, and the authors say the exact timing is hard to pin down.

A wrong latitude and a core off Fort Bragg

The Geological Society of America’s release says the team decided to core the site anyway, in Noyo Canyon off Fort Bragg, California. Goldfinger has said simply that the team wound up off northern California. The core held roughly 3,000 years of turbidites, the layered deposits left by underwater landslides that earthquakes often trigger. Many of them came in pairs, which the researchers called doublets and found unusual.

In some pairs the layering ran backward from the normal pattern, with coarse sand sitting above fine silt, a detail ScienceDaily’s summary flags as the clue that two separate events, not one, laid the sediment down. The summary says the cores preserve about 3,100 years of history in all.

Radiocarbon dates and three close calls

Radiocarbon dating gave the pairing its weight. According to the release, more than half of the turbidites north and south of Cape Mendocino were deposited at the same time, within dating error. Cape Mendocino is where the two fault systems meet, and the team concluded that the first unit of each doublet was triggered by a large Cascadia megathrust earthquake and the second by movement on the San Andreas.

Three cases in the past 1,500 years appear to show the two faults striking within minutes to hours of each other. The most recent is 1700, the year of the last great Cascadia earthquake, whose tsunami USGS Circular 1187 documents through Japanese records and Native American legends; Brian Atwater is the first author of the circular. The Geosphere paper, titled “Unravelling the dance of earthquakes: Evidence of partial synchronization of the northern San Andreas fault and Cascadia megathrust,” appeared in volume 21, issue 6, page 1132 of 2025 and carries the DOI 10.1130/GES02857.1. Co-authors come from Oregon State, NOAA, the University of Washington, California’s Department of Conservation, Springer Nature in Germany and the Instituto Andaluz de Ciencias de la Tierra in Spain.

The word “partial” in the paper’s title carries weight. The claim is not that every earthquake on one fault set off the other, only that more than half of the paired deposits line up closely enough in time to suggest a link, and that the three closest pairings sit within minutes to hours. For the 1700 event, the circular cites a Miyako government record from Japan saying the tsunami destroyed 13 houses, started a fire that burned 20 more, and led authorities to issue rice to 159 people.

Caveats the authors attach

One bookkeeping point helps readers weigh the coverage. The release, its ScienceDaily reprint and the later ScienceDaily summary all describe the same Geosphere paper, so they are one study told three times, not independent confirmations of the Cascadia and San Andreas link.

The release is open about the limits, and it notes that a cascade of this kind raises questions about preparedness for people and infrastructure. Timing is uncertain, and the top turbidite in a doublet could have eroded some of the sediment between the two layers, which would make the gap look shorter or longer than it was. It does not claim to prove that a Cascadia earthquake will set off the San Andreas. Goldfinger still said in the release: “There looks to me like a very high risk the San Andreas would go off next.”

His other comment, carried in a ScienceDaily reprint of the release from October 2025, was about scale: “It’s kind of hard to exaggerate what a M9 earthquake would be like in the Pacific Northwest.” That article’s wording stays tentative throughout, with “may,” “hints” and “probably” doing most of the work.

Sumatra, the only modern comparison, and the planning problem

The only documented real-world example of two big faults interacting this way is in Sumatra, where two large earthquakes struck three months apart in 2004 and 2005. That gap is far longer than the minutes to hours the Cascadia record suggests, so it works as an analogy and not as a match.

Goldfinger’s concern is emergency capacity. A single great earthquake on either fault could strain national resources, he suggested, and a paired event could put San Francisco, Portland, Seattle and Vancouver into emergencies at once. The open question the sediment cannot answer is how often the second fault follows the first. Three cases in 1,500 years is a short list, and each one is a layer of mud read by radiocarbon dates with their own margins of error.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.