You’ve heard this before, your whole life even: Sometime within x-amount-of years, the Bay Area will experience a major earthquake. It seems geologists have been saying this for decades, and they have. They're saying it now; the San Francisco Chronicle recently published a nice article about it. “New calculations raise the odds of a major Bay Area earthquake.” News outlets got ahold of the story and flushed it through their affiliates, at which point surely everyone’s eyes glazed over. “They’re always saying that.”
As Broke-Ass Stuart’s unofficial, self-appointed seismologist, I want to demystify that clickbaity headline. “Rising odds” feels misleading, a source of undue stress (or its antithesis, indifference). The newest headlines say “The Bay Area has a 74.1% chance of experiencing an earthquake of magnitude 6.7 or stronger over the next 30 years.” In this article, I break down the statistics so you can decide exactly how many fucks to give about earthquakes.
What that 74% chance really means
If we’re talking about rain, 74% is pretty good odds. Wear rainproof shoes that day—no Vans. But there’s also a 26% chance it won’t rain, and sometimes, the underdog wins. That’s the probability behind gambling, meteorology, and earthquakes. But the Earth holds its cards close to its chest. There is measurable data in the seas and atmosphere telling meteorologists where weather systems will form, where they’re going. The planet does not harbor such data in its interior. Even with current technology, seismologists are no closer to finding that data than they were when modern seismology began.
Either the data doesn’t exist, or it does, but takes a form we haven’t yet discovered.
So where does 74% come from? We aren’t ¾ “charged” for a quake. It’s an estimate, the result of grueling fieldwork and complex calculations that took years to complete. To start, seismologists in hardhats with headlamps descend into trenches excavated by backhoes across an active fault. They are hunting for fossils not of creatures, but major earthquakes. Digging deeper is like going back in time. Carbon deposits, misaligned fractures, and varying sediment thickness tell a story. Together, they suggest when the earthquake occurred, indicate its strength, and how much time elapsed between it and the next.
A fully excavated trench across the Neodani Fault at Madori, Japan. This fault was the site of a devastating earthquake in 1891. Creative commons.
Armed with that information, seismologists can calculate intervals between quakes going back for millennia. Repeat this process on the seven most active faults in the Bay Area and you start to understand what an undertaking this is. It’s how the USGS engineered the HayWired Scenario, which yielded a 140-year interval for the Hayward Fault. That does not mean the fault ruptures every 140 years. It means each year carries around a 1% chance of having a major earthquake. It’s merely an average. The longest estimated pause between earthquakes on the Hayward Fault is only 155 years, give or take a few decades. The same M6.7+ quake is slightly more likely to happen nowadays than it was this time twelve years ago.
Trench work across the Hayward Fault revealed eleven earthquakes in the past 2,000 years. Evidence for a twelfth large earthquake has been uncovered on the southern half. Public domain.
That said, the longer an active fault waits, the more stress it accumulates. With time, that 1% chance steadily creeps up. In 2014, the USGS assessed a 72% likelihood of a M6.7+ quake somewhere in the Bay Area within 30 years. That’s a cumulative probability, a pinch from quieter faults (Mt. Diablo, Concord–Green Valley), and more from the heavy hitters. The fault voted most likely to fail was not our big baddie, the San Andreas. We live on the fault’s northern segment, in a window between the 1906 earthquake and the next “Big One.” Seismologists agreed it was the Hayward Fault that could fail within our lifetimes, and gave that a 33.1% chance.
That number has since grown to 34.3%, still the leading fault, but the faster-moving San Andreas is catching up. It had the greatest jump, from 22% in 2014 to 25.5 percent this year. Not exactly reason to panic, but warning enough that it's wise to prepare.
The seven most active Bay Area faults. Pacific Ocean in German for scale.
So why do scientists keep repeating “thirty years?”
Faults move at different paces with messy rupture intervals that get increasingly hard to gauge as faults transfer seismic stress. Seismologists have found that the Hayward Fault fails on the order of one or two centuries, not millennia. That interval may seem set in stone, but remember, the Hayward Fault belongs to the greater San Andreas Fault system. Stresses accrue unevenly and transfer slowly or all at once, relaxing some faults and goading others. The San Jacinto Fault in Southern California, for example, is at the end of its rope.
We cannot map what is constantly shifting, but seismologists can project using parameters we know. That’s how they gave us the Uniform California Earthquake Rupture Forecast, or UCERF, a heat map of California’s most seismically prone regions. The 2026 update feeds the last twelve years to the 2014 model to get those increased odds you’re hearing about. We can know where to expect major earthquakes, how to extract and calculate intervals between them, what they’re capable of. The only thing we can’t know is when.
So why the thirty-year projection? This is the most boring part of the forecast. Does it have to do with the ever-changing face of the Earth? Or is it another mathematical head-scratcher, a misunderstanding of odds and statistics? Answer: ‘None of the above.’ Per UCERF/the USGS, “thirty years is the typical duration of a homeowner mortgage.”
That’s it. It assumes you’ve just bought a home in California and of course, paying a mortgage. This way, you can plan for your next thirty years of home ownership on land that is seismically active. Maybe that’s the inherent reason behind the miscommunication regarding the probability of earthquakes. Virtually everyone getting the message, myself included, is a renter.
