In a retired shore station for transpacific communications cables on the western coast of Vancouver Island sits a military computer in a padlocked cage.
It's the sort of cage you might otherwise use to lock up automatic rifles or expensive electronics at a big box store, but this cage is protecting data—classified signals intelligence gathered from underwater microphones called hydrophones that sit on the ocean floor. These hydrophones are part of an undersea Internet-connected scientific research network of sensors and video cameras called NEPTUNE, operated by the nonprofit group Ocean Networks Canada. Much to the delight of researchers world-round, the hydrophones record the distinct sounds of whale songs, earthquakes, and volcanic activity. But to the chagrin of the United States and Canadian militaries, they detect the passing movements of military submarines through the Juan de Fuca Strait, too.
And so, on occasion, someone in a nearby Canadian military base, sometimes by U.S. request, will push what I like to imagine is a big red button, and the hydrophones deep off the coast of Vancouver Island effectively go dark—hydrophone data is re-routed from NEPTUNE's scientists and researchers to the military through the computer in the locked cage.
It is during these diversion periods, some of which can last from hours to days a few times each month, that the military essentially listens for the engine sounds and propeller whirs unique to navy ships. That’s the stuff they keep. Sometimes they push the button—there’s never advance warning—just because, so that you can never tell based on the pattern of diversions whether the military is or isn't performing exercises nearby. It was on a recent trip to Victoria, British Columbia that Ocean Networks Canada marine microbiologist Kim Juniper showed me a picture of the computer in the cage. It was hard not to laugh at the absurdity of it all.
"It's going to come to a point in the future where this is no longer going to be feasible for the navies to put resources into sorting all this data," he later says. The hydrophones alone generate 200 gigabytes of raw data each day, and there are other, similar networks of Internet-connected sensors that already exist, or are soon to come online. (Indeed, the navy is having similar discussions with the National Science Foundation regarding the Regional Scale Nodes.) But it's not NEPTUNE's scientists the military is necessarily worried about. Because the hydrophones are connected to the internet, and the raw data they collect is archived and made available for anyone to access in near-real time. It's how non-scientists might use the data that's of primary concern.
The navies are, essentially, worried about being surveilled by us—and, presumably, by the militaries of foreign states.
* * *
Though the art and science of submarine hunting is decades old, broadcasting this information on the Internet is new. It's Cold War concern with a cyber-era twist. Because, in the age of big data, little stays hidden. Not even the things beneath the surface of the sea.
In the later years of World War I, submarine surveillance was one of the main reasons hydrophones were used. Notes on Submarine Hunting Using Hydrophones, a 1918 book for British Naval officers, notes how hydrophones could be used to determine the speed of an engine and, in turn, an enemy submarine's position and distance. A book from 1920, Hush, or The Hydrophone Service, describes rather poetically how a submarine's electric motors might sound like, "the rotary motion of a spoon in a tea-cup," or "a rhythmical sawing of wood."
But in reality, any sound was of value or concern during the war, depending on which side of the hydrophone you were on. That's how sensitive hydrophones are. "Do not throw things about the deck or against the hull of the ship," the 1920 book warns submarine operators in hiding. "Do not break up coal or use a hammer: any noise is certain to give you away."
The U.S. Navy, too, began to install its own system of cabled hydrophones off the northwest pacific coast in 1951. The SOund SUrveillance System (SOSUS) was used, according to the Woods Hole Oceanographic Institution "for classified antisubmarine warfare and surveillance during the Cold War." Following its declassification after the Cold War's end, dual-use access was granted to the National Oceanic and Atmospheric Administration (NOAA) where it was repurposed to monitor undersea oceanic phenomena for scientific study, much as NEPTUNE is used today. But unlike NEPTUNE, “Access to SOSUS is restricted,” reads NOAA’s website, “both in the sense that the data are classified and can only be used in a secure facility.”
"There's a long tradition of the ocean being the exclusive domain of the militaries and the fishing community, and we're more or less interlopers in this world,” says Juniper, the microbiologist who showed me the photo of the computer in the cage. “The world is changing."
While the Canadian military has yet to return a request for comment, the U.S. Navy reminds me that naval ship movements are classified information, and the fact that those movements might potentially be broadcast on the internet is obviously of concern. “The value of having a cabled system is that it releases data live to the internet," says U.S. Navy oceanographer Wayne Estabrooks. “But there are some times where we want to protect information, so we have to do diversions.”
Basically, any instrument that can digitally eavesdrop on the military's stuff is of concern, including seismometers, which measure vibrations so low most wouldn’t typically consider them sound. Both Canadian and U.S. militaries want updates on each of NEPTUNE's five active instrument sites—they want to know what devices are in the water, what needs to be removed for maintenance and repair, and what's yet to be deployed—and Ocean Networks Canada meets with military officials to discuss this twice a year. Though officials have never vetoed an instrument from being deployed, Juniper says the “mutually agreeable relationship” between the two groups is that the military can monitor data generated by these instruments of concern.
"They were also concerned about physical things," said Benoit Pirenne, Ocean Networks Canada’s associate director of digital infrastructure. "I would call it the Tom Clancy scenario. They were afraid that some foreign power would be able to come and plug in their own instruments." Due to the way NEPTUNE was designed, that wouldn't be possible, Pirenne said. Besides, NEPTUNE's deepest instrument site sits a mile and a half beneath the waves.
But it’s clear that the navies—or, the U.S. Navy anyhow—also recognize the importance and even public-safety implications of the research aspect. Not all data produced by the hydrophones is diverted, for example, just data within the same frequency range as ships and submarines. The range is purposefully narrow so that earthquakes and tremors can still be detected—and in the rare event that scientists need urgent access to data during a diversion-in-progress, there’s now a hotline to the military too. Of the data that’s diverted, only a few percent is actually kept by the military (the U.S. says less than 5 percent), and what remains is returned to scientists within the following days, with the intention of eventually declassifying all of it. This all has to be done manually, because, according to Estabrooks, "it's very difficult to come up with some sort of computer algorithm that can scrub the data." In the future, scientists may even choose to install more advanced seismometers that can perform computations on the devices themselves, allowing those at shore to collect specific types of readings without requiring access to the raw data of military concern.
"Just like with fisherman, I think it's important to realize that we are in the ocean, we work in the ocean, and we share the ocean just like anybody else—including the military, whether its the Canadian military or U.S. Navy,” says Pirenne. “So we have to sort of live with the expectations of the other groups, and it goes both ways.”
Still, I can't stop thinking about that computer in the cage—that it even exists in the first place, and how an Internet-connected hydrophone is at the same time a tool for science, but also a potential national security threat.
While it sometimes seems that government surveillance knows no bounds, it's some small comfort to think that, in this case, there is fear in us surveilling them. Even in the age of open data, of Internet-connected underwater microphones, you never know who might be listening.
We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.