Does the Moon Trigger
Strong Earthquakes?

A 130-year-old question

"Does the Moon cause earthquakes?" sounds like a tabloid headline, but it is a serious scientific question, and people have been testing it since 1897. The idea is not crazy. The same tide that lifts the ocean by a meter also stretches and squeezes the solid rock of the Earth. That flexing puts a small, rhythmic stress on faults. The question is whether that gentle tug is ever the final nudge that tips a fault already on the edge of breaking.

The scientific literature is genuinely split. Most global studies find nothing. A few find small, specific effects, most famously a 2004 study in Science that saw a real signal in shallow faults of one particular type. What has been missing is a single, clean, global test at large scale, with the methods locked down in advance, and honest about what it can and cannot see. That is what we set out to build.

What we did

We took every earthquake of magnitude 6.0 or greater in the United States Geological Survey (USGS) global catalog from 1973 to 2026, the biggest and best-recorded events on Earth. For each one, we worked out exactly where the Moon and Sun were in the sky at that instant, from a precise astronomical almanac. That gave us four things per quake: the phase of the Moon, whether the Moon and Sun were lined up (a "syzygy," which happens at new and full moon and produces the biggest tides), where in the daily tide cycle the quake fell, and the overall strength of the tidal pull at that spot.

Then we asked a simple question: do earthquakes bunch up at any particular tide, or are they spread evenly? Crucially, we wrote down every rule of the analysis and had it reviewed before we looked at the answer. That is what stops a researcher from quietly trying twenty things and reporting the one that looked interesting.

The aftershock trap

There is one mistake that would have handed us a fake result, and avoiding it is the most important thing we did. A big earthquake is followed by hundreds of aftershocks, all clustered in the same few days, in the same place. Because they happen close together in time, they all share almost the same phase of the Moon. Leave them in, and a single large earthquake sequence can masquerade as "the Moon causing a cluster."

So we removed the aftershocks and foreshocks using a standard method, keeping only the independent mainshocks. That cut our count from 7,477 events down to 4,729 truly separate earthquakes. That smaller, honest number is what every result on this page is based on. And this step mattered: before we removed the aftershocks, there was a faint hint of a lunar pattern. After we removed them, it vanished completely. The hint was aftershocks, not the Moon.

The answer: nothing

There is no lunar signal anywhere. Not in the phase of the Moon, not at new or full moon, not in the daily tide, not in the strength of the tidal pull. Not for shallow quakes, not for deep ones, not for the very largest. The chart below says it plainly: the earthquakes are spread evenly across every phase of the Moon, sitting right on the flat line you would expect from pure chance.

What we did not rule out

Here is the honest part, and it matters. A null result is only as strong as the effect it could have caught. With 4,729 earthquakes, our test could reliably detect a lunar effect that changed earthquake rates by about 4% or more. Anything smaller than that is simply below what 53 years of strong earthquakes can reveal. So the fair conclusion is a bound, not a debunking: there is no large global lunar effect on strong earthquakes, but a small one could still be hiding.

This is why our result does not contradict that 2004 Science study. It found an effect of roughly 1%, in shallow faults of one specific type, using a more refined measure of tidal stress on the exact fault surface. That is a smaller, sharper signal than our broad global test is built to see. Our null and their detection can both be true. Settling it needs the finer test, which is exactly the follow-up we have lined up: bring in the detailed fault-orientation data and look at those shallow faults alone.

Why a null is worth publishing

It would be easy to shrug at "we looked and found nothing." But a rigorous null on a 130-year-old question is a real result. The popular version of this idea, that the Moon meaningfully governs when big earthquakes strike, does not survive contact with the full global record. Any effect large enough to matter for forecasting is not there. That is worth knowing, and worth saying carefully, with the limits drawn honestly rather than the null oversold.

Reproducibility

Every number here comes from the lunar-syzygy-seismicity workspace: the full United States Geological Survey (USGS) catalog of magnitude 6.0-and-greater earthquakes from 1973 to 2026, deduplicated and stripped of aftershocks with the standard Gardner–Knopoff method, with the Moon and Sun positions computed for each event from an astronomical almanac. All methods were frozen in writing and reviewed adversarially before the analysis was run. The tidal driver is measured reality: the recorded position of the Moon and Sun, transformed by gravity, with no model or forecast anywhere. The full results.json, scripts, and the paper are linked below.