Live scientific debate announcement

Register now for a debate about earthquake precursors on February 18

In 2023, Quentin Bletery and Jean-Mathieu Nocquet claimed to have observed accelerating aseismic slip over the hours leading up to large earthquakes around the world. This result was published in the journal Science. Four days later, we published a re-analysis of their study, in which we claimed that the putative signals do not actually exist — the results are due to common GPS noise that was not recognized or treated in their study. The discussion has since gone back and forth, with a new paper arising in Seismica last month.

Next week, the four of us will participate in a live debate hosted by Temblor, and moderated by Temblor CEO Ross Stein, with a question-and-answer session at the end.

Spots are limited, so register soon — and show up early! If you can’t make it, the recorded debate will also appear online.

Tuesday, February 18, 11 AM Eastern Standard Time
https://us02web.zoom.us/meeting/register/55KMGj8NTducRQxodC-XUQ#/

Here on the blog, most of our posts are in response to recent events — analyzing earthquakes, and giving context about past and future hazard.

Some of our posts are different: careful (we think), often critical (we are told), analyses of research results published in high-profile journals. We consider these posts as a form of extended peer review, meant to contextualize new or commonly cited papers, and written for an interested but non-expert public audience. These posts are forever free to read and are open to comment: no subscription necessary.

Most of these critical posts focus on one general topic: earthquake prediction. Specifically, we make an effort to review papers that claim to be able to predict large earthquakes, or which identify signals that show that large earthquakes are impending. (In fact, these studies are all entirely retrospective — they only claim to have seen something happen in the past, and may or may not claim the power to predict future events.)

We have received many messages from the scientific community about our analyses of these studies. The majority are appreciative, from scientists who felt uncomfortable with the published claims, but did not have the time to dig into them themselves. Some are critical — not of our science (although we welcome any corrections or objections), but of the public nature of our criticism, especially when the authors are young scientists. We recognize the unusual nature of our approach. However, we are firm in our belief that public discussion of these issues is appropriate.

In this, we call your attention to the words of Clarence R. Allen, who delivered the presidential address to the Seismological Society of America in 1976. At the time, it appeared that earthquake science was on the verge of practical earthquake prediction, and Allen was providing guidance on how the scientific community should handle public discussion. Below, we have excerpted some of his comments.

There is no question, moreover, but that the scientist working in the field of earthquake prediction does have an overriding obligation to the public interest; lives, property, and the public well-being are directly at stake, whether we like it or not.

The total effect of this state of affairs, in my opinion, is that a scientist who is not willing to see his or her work subject to prompt and intense public criticism and evaluation — in ways quite different from those we are accustomed to — should not be in the field of earthquake prediction. There is no question but that this exposure in the public arena will have the effect of leading some scientists — unfortunately perhaps the young ones in particular — somewhat embittered and bloodied after their work has been subject to open criticism. […] But if one is unhappy with this environment and prefers the sequestered criticism of colleagues in smoke-filled rooms of the traditional scientific meetings, he or she should choose a different research area.

[…]

Only by offering hypotheses and data for testing, the very basis of the scientific method, will we ever achieve the ultimate goal of routine and consistent earthquake prediction.

Over the decades following this address, it became clear that many observations that appeared to relate to upcoming large earthquakes were in fact nonexistent or unrelated. The optimism of practicing scientists that earthquake prediction was on the horizon has since faded. Today, it remains unclear whether the Earth provides any signals that can forewarn of upcoming large events. None has yet been detected.

And yet, periodically, scientists point to new observations that were made (or rather, could have been made) before large earthquakes in the past — the implication being that if the same type of data were collected today, it might warn of a dangerous upcoming event. The vast majority of these papers appear in fringe journals and are largely ignored. However, occasionally they leap to prominence in a major journal.

One such paper was published in the journal Science in 2023. Titled "The Precursory Phase of Large Earthquakes,” the paper was authored by Quentin Bletery and Jean-Mathieu Nocquet. Bletery and Nocquet claimed to find a signal in high-rate GPS data over the two hours prior to ninety M7+ earthquakes worldwide: the motions of the crust revealed a precursory phase of slow slip leading up to large earthquakes.

When the paper was published, the first author announced it on Twitter:

It is hard to overstate the importance of an observational discovery that many, most, or even all large earthquakes begin quietly, accelerating toward rupture for about two hours in advance. First, it would answer fundamental questions about the physics of faulting in Earth. Furthermore, the claim that this behavior has already been detected by existing GPS networks (even if in retrospect) suggests that real-time monitoring could eventually give reliable early warning of Earth’s most dangerous earthquakes.

An improvement of earthquake early warning times from the minute or less of current systems to tens of minutes or more, would allow effective evacuation well in advance of any shaking. This would prevent most loss of life due to great fault ruptures, and allow for effective interventions that would also limit infrastructural damage. This is not just an intellectual exercise; the stakes are pretty high.

And so, following the recommendation of Clarence Allen, we subjected the study to “prompt and intense public criticism and evaluation.”

Four days after the Science paper came out, we published our response on this Substack. The two-hour signal, we found, was not related to any upcoming earthquakes: it was noise. The networks that Bletery and Nocquet were using exhibited a large-scale jitter, with the whole network moving back and forth simultaneously in irregular ways. While the source of this GPS noise is not understood, it is broadly recognized in GPS data and is commonly corrected by subtracting out the average displacement of the whole network at each time step — a process known as removal of a common mode.

When we applied common mode correction to the data, the precursory slip signal disappeared. In fact, all signals of putative tectonic origin presented in the original paper simply vanished.

Since our initial post, Bletery and Nocquet have:

• acknowledged that there is common mode noise in the data, which they were not aware of when they first published.

• agreed that subtracting the common mode removes the proposed accelerating precursory slip signal.

• further agreed that an additional signal that they initially identified — a sinusoidal back-and-forth movement interpreted as a type of tectonic tremor — is not real.

But they have not withdrawn their original claim that there is a likely precursory slip signal in the data. In June 2024 they published a preprint arguing that common mode noise correction might remove a real tectonic signal, and concluded:

[T] interpretation of the signal as indicative of precursory slip acceleration remains entirely plausible.

We responded to that preprint here in July 2024, arguing that their presented calculations were misleading, and in fact the vast majority of a precursory slip signal would be preserved through common mode noise correction. Their preprint was eventually published in Seismica in January 2025.

Because we have clearly not achieved consensus on this critical topic, we have all been invited by Temblor to present our cases in a live debate. Each side will have only 15 minutes to explain their position, followed by 5 minute responses, and then a 15 minute audience question-and-answer session. Questions will be submitted by private chat with the debate organizer, and will be moderated for content and relevance.

We encourage our interested readers to read the papers, posts, and arguments, which are all clearly laid out in print, and then attend the online debate if possible. We understand that there is a limit of 300 attendees, so if you want to watch live and send in your questions, please register and arrive early!

References:

Allen, C.R., 1976. Responsibilities in earthquake prediction: To the seismological society of America, delivered in Edmonton, Alberta, May 12, 1976. Bulletin of the Seismological Society of America, 66(6), pp.2069-2074. https://doi.org/10.1785/BSSA0660062069

Bletery, Q. and Nocquet, J.M., 2024. Do large earthquakes start with a precursory phase of slow slip? Preprint at EarthArXiv, https://doi.org/10.31223/X5RT3N

Bletery, Q. and Nocquet, J.M., 2025. Do large earthquakes start with a precursory phase of slow slip? Seismica, http://doi.org/10.26443/seismica.v3i2.1383

Bletery, Q. and Nocquet, J.M., 2023. The precursory phase of large earthquakes. Science, 381(6655), pp.297-301. https://doi.org/10.1126/science.adg2565

Bradley, K., Hubbard, J., 2023a. Earthquake precursors? Not so fast. Earthquake Insights, https://doi.org/10.62481/310cc439

Bradley, K., Hubbard, J., 2023b. Update on apparent GPS detection of earthquake precursors. Earthquake Insights, https://doi.org/10.62481/479c2ea4

Bradley, K., Hubbard, J., 2024. Precursory slip before large earthquakes - signal or noise? Earthquake Insights, https://doi.org/10.62481/0ff960fa