Chasing the Dream of Earthquake Prediction

Few corners of the internet are more frustrating to Earth Scientists than those folks who think they can "predict" earthquakes. You'll see YouTube videos, TikToks, Facebook posts and more where self-proclaimed experts warn people that a large earthquake will happen somewhere on the planet during a very specific window of time. It is all garbage, every last moment of it. Earthquake prediction is not possible, especially with the accuracy that these snake oil sellers claim. Yet, a hint of a scientifically-based future for prediction might actually be out there.

If you have watched or read about these predictions, you'll see some common themes. Many of them have one foot in astrology, claiming that earthquakes can be predicted based on planetary and lunar alignments. Others like to claim that things like "earthquake weather" exist. There are even some who point at harmless experiences like HAARP as a nefarious manmade trigger for temblors.

Few offer any attempt at statistical analyses of their predictions or peer review of their methods. If anyone tries to discredit them, they circle the wagons, claim that people just want to steal their methods or that experts are merely jealous. All of this supports the notion that they are merely sources of misinformation.

Earthquake scientists are in agreement that prediction is not possible. There are just too many unknowns about fault motion processes and earthquake triggers for us to make any useful statement about the exact when and where of these events. Instead, potential hazard in the form of probability of an earthquake in a decadal or more timescale is the coin of the realm. People living in San Francisco, Tokyo or Rome just need to know that earthquakes can and will happen.

That all being said, new technology and better data might be opening the opportunity for prediction or early warning of an impending large earthquake. Global Positioning Systems (GPS) can offer us fine-scale information about fault motion in near real-time. This means we can look for any signal that could be a useful and predictable precursor to big earthquakes.

A study in Science by Quentin Bletery and Jean-Matthieu Nocquet may have started to open the prediction door a crack. They looked at GPS data collected in the two days before major earthquakes (M7 and higher) with such data to test whether these quakes are preceded by "quiet slip" -- or, motion on a fault that doesn't produce an earthquake.

What they found was a surprise to many Earth scientists: there may actually be a signal of quiet slip near the location of the earthquake in the two hours preceding the event. Not all the earthquakes they examined clearly showed this increase in slip, but a majority of the events did clearly exhibit slip. The Tōhoku M9 earthquake in 2011 had a very clear increase in quiet (aseismic) slip in the hours before the quake (see below).

On top of this, they found another signal in the slip record, one where slip rises and falls on a ~12 hour cycle. That should have perked up your ears because that is also the cycle of Earth's tides. Bletery and Mocquet suggest that the up-and-down signal is directly tied to tides -- and maybe, tides can sometimes be a trigger for earthquakes.

Now, before people get too excited, there is no evidence that tides will always trigger earthquakes or that the relative size of tides (hello, Supermoon) plays any role. Just like every match doesn't start a fire, all tides don't cause earthquakes. It is an interesting suggestion of a causation at this point.

The use in high precision and real-time, rapid GPS data is a promising idea, but we need much more infrastructure of GPS data to actually do this on major, hazardous faults. The researchers point out that we don't have the instrumentation to do this for individual earthquakes. However, finding this signal for precursory activity might attainable ... and that, just maybe, could be how we can give people warning of impending major earthquakes.