Submarine sleuthing exposes cause of 1918 Puerto Rico tsunami

Researchers found that an ancient, mineral-encrusted landslide was too old to have caused the 1918 tsunami. A nearby fault was to blame instead.

By Lauren Milideo, Science Writer (@lwritesscience)

Citation: Milideo, L., 2023, Submarine sleuthing exposes cause of 1918 Puerto Rico tsunami, Temblor,

On Oct. 11, 1918, an earthquake of estimated 7.2 magnitude occurred off Puerto Rico’s northwest coast. A tsunami reaching up to 20 feet (6 meters) high then hit multiple locations along the island, reportedly breaking undersea cables and causing $4 million in damage (in 1918 dollars). The earthquake caused 116 deaths, with 40 of those blamed on the subsequent tsunami. A 1919 report on the events left room for doubt about whether the quake caused the tsunami. Then, in 2008, a study revealed, through boat-mounted multibeam bathymetry, the presence of an underwater landslide deposit at the same site where breakage of submarine cables had been reported during the 1918 disaster. Scientists then theorized that the submarine landslide caused the tsunami.

Damaged building in Mayaguez, Puerto Rico. Credit: Jose Luis Rodriguez, Rodriguez Archives LLC, Connecticut, via NOAA/NCEI
Damaged building in Mayaguez, Puerto Rico. Credit: Jose Luis Rodriguez, Rodriguez Archives LLC, Connecticut, via NOAA/NCEI.


Now, a team of researchers has found that the underwater landslide blamed for the 1918 tsunami could not have been the cause: It was too old. Instead, they pinpointed a nearby fault as the likely culprit, which they reported in a new study in the Bulletin of the Seismological Society of America.

Submarine views

To get a closer look at the landslide, Uri ten Brink, a geophysicist at the U.S. Geological Survey and co-author on the new study, and his colleagues used a remotely operated vehicle (ROV) loaned to them by the nonprofit Ocean Exploration Trust. The footage from the ROV let the researchers examine the landslide’s surface in 3-D, revealing far greater detail than the boat-mounted bathymetry had provided. Special software let the researchers analyze the ROV footage frame-by-frame and build a 3-D model of the landslide’s surface.

Viewing the images, ten Brink and his team realized the landslide was covered in a ferromanganese crust — a hard shell of minerals that takes hundreds, if not thousands, of years to form. The tsunami was in 1918. There was no way this landslide could have caused it, the team concluded.

Sediments filling the landslide scar, sampled using the ROV’s arm, provided additional evidence that the landslide was much older than the tsunami. Radiocarbon dating returned an age of around 440 years before present, more than 400 years before the tsunami.

The team had conclusively determined that the landslide couldn’t have been the tsunami’s trigger, they reported.

Who’s at fault?

The team turned to the other disaster from that day: the earthquake. Could the fault rupture alone have caused the tsunami?

“Puerto Rico has very, very complicated tectonics,” says Noel Bartlow, a geology professor at the University of Kansas who was not involved in the study. The region “likely can experience both tectonically-sourced tsunamis and landslide-sourced tsunamis,” she says.

The team turned to a nearby fault in the Mona Rift area off the northwest Puerto Rican coast. The fault lay in an area that had previously been studied, and its bathymetry, seismic data and a historical catalog location of the 1918 earthquake and aftershocks suggested it was a possible candidate as a source for the quake. Descriptions of the original earthquake also seemed to line up with the fault’s characteristics, including its two segments, which would correspond to the description of the original quake — one incidence of intense shaking, followed two minutes later by another, less intense occurrence.

Using bathymetry and coastal and near-shore lidar data from the National Oceanic and Atmospheric Administration, ten Brink and the team modeled what a tsunami from a quake on that fault might look like.

The model suggests a magnitude-7.2 earthquake that ruptured two segments along the fault “fit it pretty well, so we said, ‘well, that could be the source of that 1918 tsunami,’” ten Brink says. In addition, they reported seeing “fresh vertical slickenslides” — scratches on fault surfaces that indicate seismic movement — with the ROV. The evidence added up.

Scientists used an ROV to examine the seafloor off the coast of Puerto Rico. They found that a landslide scar was much older than previously thought, so it couldn't be the source of a 1918 tsunami. Credit: Uri ten Brink.
Scientists used an ROV to examine the seafloor off the coast of Puerto Rico. They found that a landslide scar was much older than previously thought, so it couldn’t be the source of a 1918 tsunami. Credit: Uri ten Brink.


The combination of a high-tech ROV, 3-D modeling, and simple geologic observations led the team to conclusions that underwater modeling alone wouldn’t have allowed.

“By going out and just doing some really basic field observations, looking at clues to the age of [the landslide] on the seafloor, I think they pretty conclusively found evidence that [the landslide was] not young enough to have been the source of that [tsunami],” says Harold Tobin, a seismologist at the University of Washington who was not involved in the study. “It’s a great example of geological sleuthing.”

“These guys are basically doing paleoseismology on the seafloor,” Tobin says. “There’s not very much of that, and I think it needs to grow.”

Changing hazards elsewhere?

The finding could make waves in the submarine hazards world. Following the 2008 study that pinned the submarine slump to the tsunami, researchers in other locations published studies tying landslides to tsunamis and calling for concern about tsunamis in locations where landslides could occur, says ten Brink.

This research suggests that “maybe not all tsunamis that are attributed to submarine landslides are what scientists thought they were,” says Tobin. “One study doesn’t prove it, but this is one piece of evidence in the big picture,” he says.

Even though the 1918 tsunami is no longer tied to this particular landslide, the relationship between tsunamis and landslides is not negated by the study, Bartlow notes. “There is still a landslide scar there,” she says. “The modeling done previously showed that it could explain a similar tsunami, so what that says is that landslide occurred, and likely did generate a tsunami when it occurred — it just didn’t occur in 1918.”

“It’s a good example of a place where as scientists, we have to look back at our evidence sometimes and in the light of new evidence reevaluate old conclusions,” Bartlow says. “I see science self-correcting here, which is always a good thing.”

“I don’t want to say [landslide-induced tsunamis] won’t happen somewhere else,” ten Brink says, but it’s good to “set the record straight” on what happened at this site.


ten Brink, U., Chaytor, J., Flores, C., Wei, Y., Detmer, S., Lucas, L., Andrews, B., and Georgiopoulou, A. (2023). Seafloor Observations Eliminate a Landslide as the Source of the 1918 Puerto Rico Tsunami, Bull. Seismol. Soc. Am. 113, 268-80, doi: 10.1785/0120220146