Researchers have discovered something interesting about the powerful magnitude 8.2 earthquake that devastated Mexico in 2017.

Not only did it occur at a location where earthquake models say it shouldn’t have, but it was so powerful that it split through the tectonic plate.

2017 Southern Mexico Earthquake

On Sept. 7, 2017, Mexico’s Chiapas state, which borders on Guatemala, was struck with a powerful 8.2 magnitude earthquake that killed nearly100 people and injured hundreds more. It was initially believed that the earthquake occurred when the Cocos ocean plate was overridden by a continental plate because such megaquakes typically occur at the subduction zone, which is close to where the tops of the plates converge.

However, over a year after the earthquake, researchers of a new study published in the journal Nature Geoscience reveal that that was not the case at all, and is actually a most curious case that defies current earthquake models. In fact, the researchers discovered that the epicenter of the earthquake was actually much deeper at 28 miles deep in the Cocos plate.

While this is something that can happen, it typically occurs in older and cooler subduction zones such as the earthquake in Sanriku, Japan in 1933 that eventually caused a 94-foot tsunami, claiming the lives of 1,522 people and the destruction of over 7,000 homes. The one that caused the 2017 Mexico earthquake, however, is much younger and warmer.

Tectonic Plate Split

Another incredible discovery by the 13-member research team is that the Cocos plate actually split as a result of the massive energy that was released in just a matter of seconds. As study lead Diego Melgar describes it, the rupture made a big, gaping crack and that evidence suggests that the crack ran through the entire width of the plate.

So why did this happen? The researchers believe it is possible that seawater got into the plate and sped up the cooling process, thereby making it more susceptible to earthquakes. This, however, is merely a hypothesis.

“Our knowledge of these places where large earthquakes happen is still imperfect,” said Melgar. “We still need to do a lot of work to be able to provide people with very accurate information about what they can expect in terms of shaking and in terms of tsunami hazard.”