Do earthquakes have a season? What new Caltech research means for facilities in California.

Earthquakes don’t follow a calendar. That’s been the standard answer for decades, and it’s still mostly true. You can’t predict when an earthquake will happen.

But a study published March 25 in Science Advances by researchers at Caltech found something worth understanding: in parts of Northern California, seasonal changes in groundwater levels correspond to a measurable increase in seismic activity. Up to 10%, with the peak arriving about two weeks after peak groundwater fluctuation.

It’s not a prediction tool. But it adds nuance to how we think about earthquake risk, and for facilities operators, nuance is worth paying attention to.

What the study found

The research, led by Krittanon Sirorattanakul in Jean-Philippe Avouac’s lab at Caltech, looked at how different types of stress affect earthquake rates in California. They examined three: tectonic stress (the slow, constant push of plate boundaries), hydrological stress (seasonal expansion and contraction of groundwater), and tidal stress (gravitational pull from the moon and sun).

The finding that stands out: in areas where groundwater levels swing up and down dramatically with the seasons, earthquake rates went up by as much as 10%. The effect lagged about two weeks behind the peak rate of groundwater change.

Tides, interestingly, didn’t show the same effect. Even though the stress from tides is comparable in magnitude to seasonal groundwater changes, the 12-hour tidal cycle is too fast to influence fault behavior. Groundwater shifts develop over months, which gives the stress time to build and interact with existing fault pressure.

Why this matters for facilities, not just scientists

If you manage a building, a campus, or a critical operation in California, this doesn’t change your day to day protocols. You can’t (and shouldn’t) try to time earthquake preparedness around wet and dry seasons.

But it does reinforce something that a lot of facilities operators underestimate: earthquake risk is not uniform. It varies by location, by geology, and apparently by hydrology too. The areas where groundwater fluctuates the most, like parts of the Central Valley and Northern California, may carry a slightly different risk profile than their seismic history alone would suggest.

For organizations that take earthquake readiness seriously, this kind of research matters because it refines the picture. It’s one more reason to stop treating earthquake preparedness as a checkbox activity and start thinking about it as ongoing risk management.

The human activity angle

There’s another layer to this study that’s easy to miss. If natural groundwater fluctuation can influence earthquake activity, so can human activities that affect subsurface water levels. Agricultural pumping and reservoir management are the obvious examples. The researchers noted that “removing subsurface fluids atop faults” can modulate seismic behavior.

This has implications for facilities in areas where large-scale water extraction is common. It doesn’t mean pumping groundwater causes earthquakes in any direct sense. But it means the stress environment around faults is more dynamic than a hazard map might suggest.

What to take away

None of this makes earthquakes predictable. You still need systems and protocols that assume the shaking could start at any moment.

What it does change is the conversation. If you’re evaluating earthquake risk for a new facility, or reassessing preparedness for an existing one, the science now accounts for more variables than it used to. The ground under your building is affected by more than plate tectonics.

For a primer on how early warning systems fit into facility preparedness, start here.

Sources: Caltech/Science Advances (March 25, 2026), Caltech Seismological Laboratory