An M7.3 off Chiapas Splits the Seismologists: Megathrust, or a Break Inside the Bending Cocos Slab?
A little before nine in the morning, the seafloor off southern Mexico lurched. Within minutes the U.S. Geological Survey had a magnitude, a location, and an automated damage forecast on the wire; a tsunami alert went out across the Pacific and was called back once the water rose barely a third of a metre. By the standards of a great subduction zone, it was almost an anticlimax. And then the interesting problem started — because the people who study these earthquakes for a living began to disagree about what they were even looking at.
The dispute is not about how big the quake was. It is about which fault moved, and where. That question sounds academic until you notice that the two leading answers imply two different hazards for the same stretch of coast.
Two ways to break a subduction zone
Off Chiapas, the Cocos plate slides beneath North America. Most people picture the danger there as the megathrust: the shallow contact surface between the two plates, the fault that stores centuries of strain and releases it in the largest earthquakes on Earth. When a megathrust slips, the seafloor above it heaves upward, and that vertical shove is what builds an ocean-crossing tsunami.
But the subducting slab does not simply slide. As it plunges, it bends, and the bending cracks it from the inside. Earthquakes that rupture within the descending plate, rather than along its top surface, are called intraslab events. They tend to sit deeper, and because they can involve the slab pulling apart under its own weight, they often show normal-faulting motion rather than the thrust of a megathrust.
The distinction is not a footnote in Mexican seismology. The 2017 Mw8.2 earthquake that devastated parts of Oaxaca and Chiapas (the largest to strike the country in a century) was an intraslab rupture, a tear inside the bending Cocos plate roughly 100 kilometres west of this month's event. It still produced a damaging tsunami. So "intraslab" does not mean "harmless," and "not the megathrust" does not mean "no wave." That 2017 quake is the cautionary precedent hanging over any quick read of this one.
What the mechanism says, and where it goes quiet
The USGS solution for the July 17 event describes reverse, or thrust, faulting on or near the plate interface, with the Cocos plate driving under Mexico. On paper that is the signature of a megathrust: a well-constrained moment tensor, a nodal plane striking through the coast and dipping toward the interior, motion consistent with the two plates grinding past each other. Read alone, it points to the shallow contact fault.
The independent rapid analysis by seismologists Judith Hubbard and Kyle Bradley, published on their Earthquake Insights newsletter, agrees that the mechanism is close enough in location and orientation to be a plausible megathrust rupture. But they stop short of calling it one. The reason is a smear on the cross-section: the zone of seismicity around the event is, in their words, quite wide. That width can mean two things, and neither is comfortable. Either the location itself is uncertain by enough to blur which fault is in play, or there are several fault sources tangled together near the interface. When the earthquakes near a plate boundary do not fall cleanly onto one plane, you cannot cleanly say which plane broke.
This is why a headline that reads "megathrust earthquake" would be running ahead of the evidence. The megathrust interpretation is genuinely plausible. It is not, as of now, settled. It is one reading among a couple that the same data will support, and the aftershock sequence over the coming days is what will start to separate them.
The 60-kilometre problem
Sitting underneath the mechanism argument is a more basic disagreement: the two agencies do not place the earthquake in the same spot. The USGS puts the epicentre about 58 kilometres west-southwest of Puerto Madero, at a depth near 18 kilometres. Mexico's national seismological service, the Servicio Sismológico Nacional, records its epicentre roughly 60 kilometres to the southwest of that, a substantial offset for the same event.
Sixty kilometres is not a rounding error; it is the whole question in disguise. Move the rupture farther offshore and you move it to a different position on the subducting slab, which nudges the interpretation of which fault is more likely. And you change the distance from the rupture to the towns along the coast, which is exactly what governs how violently the ground shakes at a given place.
That second point has a visible consequence. The USGS automated impact system, PAGER, flagged the event at its Yellow level and estimated that a large population felt strong shaking. Yet the on-the-ground reports describe far less damage than that estimate implied. The rapid analysts offer a tidy explanation: if the Mexican location is the better one (and a denser regional network usually is closer to the truth near its own coast), then the earthquake happened farther from people than the USGS coordinates suggested, and the shaking model over-counted. The location gap and the damage-forecast miss may be the same discrepancy seen from two angles.
Why the uncertainty is the story
None of this is a failure of seismology. It is what the first day of a large earthquake actually looks like from the inside. Magnitudes get refined, epicentres migrate as more stations report, focal mechanisms firm up or shift, and an automated damage forecast built for speed sometimes overshoots when the input location is off. Every number attached to this event (the magnitude, the depth, the epicentre, the PAGER level) is a preliminary estimate, and any of them may be revised as the analysis matures.
What makes the Chiapas M7.3 worth a scientist's attention is precisely that it refuses, for now, to resolve into a single clean answer. Megathrust or intraslab; USGS location or SSN location; over-forecast shaking or genuinely light damage — these are not competing errors so much as competing readings of a still-forming picture. The honest position, days after the shaking stopped, is that the coast off Chiapas produced a large earthquake whose fault has not yet confessed which one it was.