Undersea Internet Cables Can Detect Earthquakes—and May Soon Warn of Tsunamis

“It has always been the case that cables get laid first and then people begin trying to think of new ways to use them,” the sci-fi novelist Neal Stephenson wrote in Wired in 1996. “Once a cable is in place, it tends to be treated not as a technological artifact but almost as if it were some naturally occurring mineral formation that might be exploited in any number of different ways.”

Each cable is roughly the thickness of a garden hose, but it’s mostly a protective sheath around a dozen thin strands of glass, which are so pure that a kilometre-thick block would appear as clear as a freshly washed windshield. Today, about three hundred cables carry ninety-nine per cent of transoceanic data traffic.

Bruce Howe, an oceanographer at the University of Hawaii, has been adding scientific instruments to seafloor cables since the early nineteen-nineties. Telecom companies lay new cables roughly once every quarter century to preëmpt disruptions and incorporate more advanced materials. “Whenever a company decides to turn their cable system off, instead of abandoning it in place, as was done in those days, we thought science could use it,” he told me.

In the late aughts, Howe led the years-long installation of part of the ALOHA Cabled Observatory, which built on an old AT&T cable situated a hundred kilometres north of Oahu. He and colleagues later wrote that the team struggled to link their instruments to the cable, and the facility struggled to reach its full potential, owing in part to “still-all-too-common cable and connector problems.”

Similar attempts to co-opt mothballed cables also stumbled. In 1998, scientists added a seismometer, a hydrophone, two pressure gauges, and other instruments to an obsolete cable that linked Hawaii and California, but the system failed after just five years. One system near Hawaii developed a short circuit six months after deployment, and another was damaged by fishing activity off the coast of Japan. Commercial hand-me-downs weren’t the way forward.

Howe started to wonder whether it was possible to incorporate scientific equipment into operational telecom cables, which are meticulously maintained by the companies that profit from them. He and his colleagues designed temperature, pressure, and seismology probes that would fit snugly into cable repeaters. “The telecom people were adamant that they wanted nothing to do with us,” Howe told me. As he tells the story, they replied, “No way, because it would affect the reliability of the telecom.” This response disappointed the scientists, who would later estimate that piggybacking on cable infrastructure would give researchers data at a tenth of the cost of building their own system from scratch.

Installing a transatlantic cable takes two to three years and about two hundred million dollars, according to Nigel Bayliff, the C.E.O. of the cable operations firm Aqua Comms. A single repair can cost two million dollars. Any change to a functioning system—even a modest science package added at no cost to the cable company—could become a liability. “It’s a bit like asking for a different toilet on the space station,” Bayliff told me. “It’s, like, ‘Really, guys? Do you really want to risk the whole space station to change the toilet?’ ”

“The only business reason for these cables to exist, as far as we are concerned, is for data connectivity,” Bikash Koley, the vice-president of global networking at Google, which has laid long stretches of cable in partnership with telecom carriers, told me. The company has no intention of adding instruments to its cables, he said.

There are legal obstacles, too. Because seafloor telecom cables are treated as an essential public service, they receive certain freedoms under the United Nations Convention on the Law of the Sea, but the nebulous category of “marine scientific research” does not necessarily receive the same privileges. Bayliff worries about what could happen to telecom projects if they contribute to science.

“Is ninety-per-cent telecom, ten-per-cent science now a science cable?” Bayliff asked. We might not know until a first mover tests the legal waters. But he added that governments might be able to solve this problem by mandating collaboration between companies and researchers. “Once this becomes the norm, then it will just happen all the time and no one will worry, because the risks will all be the same for everybody,” he said.

Howe and his team ultimately collaborated with the government of Portugal, which was planning to replace its aging cable system—and which knows something about offshore earthquakes. In 1755, a massive quake southwest of Lisbon caused a tsunami and devastated the capital. Tens of thousands died.

“They’re motivated,” Howe told me. “They see this in terms not just of telecom operational costs but in human costs, and it may take governments to really balance these kinds of considerations. Companies aren’t going to do that.” The Portuguese government has approved the project, and Howe expects the appropriation of at least a hundred and twenty million euros to happen sometime this year. The cable will connect Lisbon, the Azores, and the island of Madeira; once it’s operational, in 2025, motion, pressure, and temperature sensors in the cable’s repeaters will serve as a seafloor science platform and a tsunami-warning system.

In order for scientists to break the stalemate with the cable industry, they needed ways to use data that already exists, without modifying undersea cables or repeaters. Marra’s serendipitous discovery proved that this was possible.

Then, in 2020, Google agreed to share measurements of light polarization from its fibre-optic network with a scientific team that included Zhan and other researchers from Caltech and the University of L’Aquila, in Italy. Koley told me that Google scientists were happy to collaborate—as long as they didn’t need to add instruments to their cables. “This was a set of data that you would actually throw away otherwise,” Koley said. “It has no other use to us.”

The researchers identified shifts in the polarization that occur when cables bend, twist, and stretch, and cross-referenced the changes with dozens of earthquakes that seismometers detected over a nine-month period. This approach isn’t as sensitive as Marra’s method or D.A.S., but it doesn’t require sophisticated technology in the form of an advanced laser. “Because the method is so easy to implement, we actually now have six or seven cables on board, providing data,” Zhan said.

Last year, Google gave Marra and his team access to a cable-landing station in Southport, England, where the company used a cable that extends to Dublin, and then on to Halifax, Canada. The company was willing to give the researchers temporary access to certain channels when it wasn’t using them. The researchers drove five hours from their laboratory in Teddington and installed customized lasers and detectors, as well as computers that they could access remotely. They now had the power to detect phase shifts beneath the Irish Sea and the Atlantic Ocean.

But they still needed a way to determine where the phase shifts were happening in order to figure out the exact location of seafloor movements. To solve this problem, the researchers took advantage of tiny mirrors that are built into fibre-optic repeaters, which normally help technicians diagnose problems along specific stretches of cable. The hundred and twenty-eight mirrors between Southport and Halifax allowed them to identify the specific portion of cable where a phase shift first occurred. Their approach had the potential to turn the cable into a hundred and twenty-nine localized earthquake detectors.