Interesting but disturbing report today on pyhs.org about how a typical Iranian earthquake on October 5 of last year was almost immediately mischaracterized as a possible underground nuclear test on social media during a time of heightened geopolitical tension between Iran and Israel:
A new study debunks claims that a magnitude 4.5 earthquake in Iran was a covert nuclear weapons test, as widely alleged on social media and some mainstream news outlets in October 2024, a period of heightened geopolitical tensions in the Middle East.
Led by Johns Hopkins University scientists, the study warns about the potential consequences of mishandling and misinterpreting scientific information, particularly during periods of international conflict. The findings appear in the journal Seismica.
"There was a concerted misinformation and disinformation campaign around this event that promoted the idea this was a nuclear test, which is not something you often see happen with an earthquake," said Benjamin Fernando, a Johns Hopkins seismologist who led the study. "This shows how geophysical data played an important role in a geopolitical crisis."
...
Fernando's team analyzed seismic signals from the event, identifying natural seismic activity caused by the earthquake.
Using publicly available data from seismic monitoring stations, the researchers concluded that the earthquake originated along a gently sloping fault where Earth's crust was being deformed by the collision between Arabia and Eurasia. The process aligns with the geophysical forces that characterize the region's tectonically active interior and rules out any connection to a particularly unusual source or nuclear test, the study concludes.
As a former seismic analyst working for the predecessor of the CTBTO’s International Data Centre, now located in Vienna, Austria, and who has literally looked at millions of these seismograms to help distinguish between natural earthquakes and underground nuclear explosions, there is absolutely nothing about either of the seismic events recorded in Iran that day to suggest they might be anything other than natural earthquakes — even though like a significant fraction of Iranian quakes, these were fairly shallow (~10 km below the surface).
Aside from a reliable depth estimate (anything more than 10 km below the surface can almost automatically be ruled out as a nuke), the two main methods of discriminating between natural and artificial seismic events are:
1) fault plane solution — earthquakes involve two blocks within the crust sliding past each other, meaning there will be a mix of compressional and dilatational “first motions” recorded by different seismic stations depending on their orientation to the actual fault plane, but all explosions (and particularly underground nukes) have a compressional wave radiating outward in every direction; and
2) waveform analysis — earthquakes typically have a much lower amplitude ‘P’ (Primary) wave relative to their slightly slower ‘S’ (Secondary or Shear) wave, which also forms the basis of real-time earthquake alarms, but nukes and other explosions are almost all ‘P’ wave with a much smaller ‘S’ wave. There are also significant differences between the two in terms of surface waves, but that was never my particular specialty. Suffice it to say, the waveforms presented up top from the original CTBTO report look nothing like what we would expect to see for an actual underground nuke (and I’ve seen many hundreds of those too during my career as a seismic analyst).
What was particularly disturbing about the phys.org story was just how and why this mischaracterization of the Iranian earthquake came about:
Despite the clear scientific evidence of natural seismic activity, claims that the earthquake was a nuclear test began spreading rapidly on social media just 17 minutes after the event. Initial tweets misinterpreted seismic data, which soon gave way to misinformation and likely active disinformation, Fernando said.
The first suggestion that this was a nuclear test appeared on Twitter/X 27 minutes after the earthquake. Over the following hours, misinformation escalated, with some posts citing seismic data from an entirely different earthquake in Armenia earlier that day to substantiate the nuclear test narrative.
The misinformation gained traction as conspiracy theories linked the Iranian earthquake to a supposed seismic event in Israel the same evening. While the study noted that it is difficult to confirm deliberate disinformation, the sustained engagement and specialized use of seismology data on social media suggest potential human authors with expertise.
One of the most widely shared posts promoting the nuclear test theory came from an account tied to Russian-supported disinformation campaigns, the researchers found.
What a surprise — not!
Within hours of the event, the false narrative moved from social media to news reports worldwide. Indian English-language media were the most active in reporting the nuclear test claims, often referencing each other's stories, and citing the incorrect seismic data. The researchers also identified media reports from the United States, Israel, Pakistan, Zimbabwe, France, and the United Kingdom.
In contrast, Persian-language media generally described the event accurately as a natural earthquake. These reports drew on local expert commentary and official seismic data more often than English-language media, offering a more precise understanding of the event.
Interesting that Persian-language media was much more accurate than a lot of foreign English-language media in this case, with Indian sources once again being one of the principal amplifiers of the mis/disinformation. One reason I will never cite an Indian media source on just about anything without a huge amount of corroborating evidence from other sources — and then I’ll just cite those other sources instead.
As a personal aside, one of my proudest moments as a seismic analyst came when I was doing a quality control check on the newest automated search algorithms used by the IDC to detect seismic events around the world through their International Monitoring System that has now grown to over 300 different seismic stations (though back in my day it was only half that size).
Normally the computer is allowed to assemble the preliminary catalogue for each day’s seismic events, with a human analyst checking to see if any bogus events made it in that need to be tossed, if any real events incorporated false signals (or failed to incorporate real ones) that could help to more precisely define an event’s actual location and/or depth, and most importantly — checking to see if the automated system missed any real events it presumably should have detected but didn’t.
As it happened, I found one event in particular that the computer never should have missed — well above the detection threshold, and smack dab within our Lop Nor warning box (i.e. the area where China has conducted all of its nuclear testing). Needless to say, for all of the reasons elucidated above, there was never any suggestion it was anything but a naturally occurring small earthquake, but it was definitely a wakeup call for the IDC to better fine-tune its algorithms!