THE BOTTOM OF THE ATLANTIC OCEAN — Forty miles off the coast of North Carolina, the 274-foot research vessel Atlantis paced a dark, empty swath of ocean in evenly spaced lines as the crew pinged sound waves into the deep. A quarter-mile below, plumes of methane, a potent greenhouse gas, rose from the seafloor.

The underwater site, named Pea Island after an area of the Outer Banks, is one of the hundreds of active methane seeps discovered off the Atlantic coast since 2012. No human had ever explored this particular underwater world. Samantha Joye, an oceanographer and microbiologist, was about to change that.

She strolled into the ship’s computer lab at 6 a.m., a thermos of tea in hand. She looked anxious as she checked in on what the sonar had turned up.

Jason Chaytor, a marine geologist with the U.S. Geological Survey, had spent half the night mapping the ocean floor. He pointed to the columns of bubbles visible in the rainbow-colored images. The largest of the plumes extended some 250 meters from the bottom, about halfway to the surface.

“You’re going to visit this first,” Chaytor told her.

Joye leaned over his shoulder and squinted through purple-framed glasses. A mad scientist grin washed over her face.

The site is what’s known as a cold seep, an area where methane and other hydrocarbons naturally eject from the seafloor. Cold seeps are home to diverse communities of organisms, including Joye’s favorite: beggiatoa, a large, thread-like bacteria.

Along with their ability to capture energy from poisonous hydrogen sulfide gas, beggiatoa form colonies, or “mats,” that are hot spots for hitchhiking microorganisms that feast on methane. Working together, these communities of microbes act as biological filters, blanketing active seeps and limiting the amount of gas that enters the water column and, more importantly, the atmosphere.

The seeps along the Atlantic’s continental margin are not new, but recent advances in sonar imaging technology have given scientists the tools to spot them. (Hundreds more have been found in recent years off the coast of the Pacific Northwest.) The novelty of the technology means scientists lack the baseline data that would allow them to compare the amount of methane leaking today to, say, the amount leaking 20 or 200 years ago.

Methane is among the most potent greenhouse gases. And while the numerous sources of methane are well understood, what’s driving the recent surge in global emission levels remains a matter of scientific debate.

The two-week Atlantis expedition was part of project Deep Search, a five-year government-funded study to explore cold seep, canyon and coral ecosystems in this largely uncharted swath of the deep Atlantic. The team of more than 20 scientists set out from Woods Hole, Massachusetts, in mid-August 2018 with plans for a dozen manned submersible dives off the southeast Atlantic coast ― most of them to uncharted sites.

Joye, a professor at the University of Georgia, hoped the mission would further scientists’ understanding of methane seeps, their potential for contributing to global warming and the complex microbial communities that inhabit these systems.

In 2006, while exploring in the Gulf of Mexico, Joye was part of a team that discovered a massive mound of methane hydrate, a solid, ice-like form of the gas that is widespread in deep marine sediments. The feature resembled a dragon’s head and was named “sleeping dragon.” For Joye, it’s an apt metaphor for the apocalyptic situation that would unfold if a giant burst of methane into Earth’s atmosphere ever occurred.

“We are waking up the methane dragon,” Joye said. “And that’s a dragon that we really want to keep in the box.”

Methane, or CH4, is part of Earth’s natural carbon cycle, emitted from wetlands, soil, volcanoes, wildfires, rice paddies and even by termites. In the ocean, methane is produced when microorganisms or geologic processes deep in the earth’s crust break down organic matter that settles to the seafloor, including dead fish, krill and bacteria.

It is also a powerful super-pollutant, roughly 30 times more effective at trapping heat than carbon dioxide over a century in the atmosphere. Although far less prevalent in Earth’s protective shield than carbon dioxide, methane accounts for about one-fifth of human-caused planetary warming. Since 1750, methane concentrations have risen more than 150% ― a spike driven by fossil fuel production, agriculture and deforestation.

In both the deep sea and Arctic permafrost, a massive amount of methane is trapped in hydrate. This otherworldly substance, also called “methane ice” or “fire ice,” forms when methane combines with water at low temperatures and high pressure. It represents one of the largest carbon reservoirs on Earth, sequestering an estimated 16 to 20% of all carbon.

“Think about that,” Erik Cordes, the expedition’s chief scientist and a deep-sea ecologist at Temple University, said as Atlantis headed out to sea from Woods Hole. “All the forests on the planet, all the living organisms on the planet together, have less carbon in them than there is in methane hydrate.”

Hydrate remains stable under conditions like those found in the frigid deep sea. But if exposed to warmer temperatures or a drop in pressure, it can turn to gas, expanding by approximately 180 times its volume. The concern for some scientists is that as global climate change thaws Arctic permafrost and heats up the oceans, these hydrates will break down, setting off a potentially calamitous feedback loop.

Enough methane in the ocean could deplete the water of oxygen and wreak havoc on marine life, while a sharp rise in atmospheric methane would trigger rapid and cataclysmic warming.

That scenario keeps Joye up at night. She’s been studying methane seeps and hydrothermal vents for two decades and says she’s seen enough to know that these systems are poised to respond to the rapid changes now unfolding in the oceans. Climate-driven hydrate collapse, she stressed, is not some hypothetical.

“I hate to say it’s a ticking time bomb because I don’t want to scare the shit out of people,” she said. “But it scares the shit out of me.”

In 2016, a decade after first documenting the “sleeping dragon,” Joye returned to the site with a film crew from the BBC to find that the hydrate mound ― one of the largest ever documented ― had completely vanished. Similar deposits at other nearby locations were also gone, replaced by craters, or “pockmarks,” where the once-frozen methane exploded from the seabed, Joye said. The water temperature near the seafloor was several degrees above normal.

“We know it’s hydrate destabilization,” she said, adding that she and others had tried unsuccessfully to secure funding for long-term study. “We need to make people understand that we really need to be monitoring these things.”

Surges in atmospheric methane have been blamed for past planetary warming events. The most severe, the “The Great Dying,” occurred 250 million years ago and wiped out approximately 90% of all species. Among the controversial scientific theories about what may have caused it is hydrate degradation. Another is a massive bloom of methane-producing microbes, as a team of researchers at MIT detailed in a 2014 paper.

Scientists have also found signs of a large, sudden burp of methane gas from the Arctic seafloor during a period of extreme warming more than 100 million years ago, thought to be caused by hydrate destabilization. And hydrates have been implicated in a period of extreme warming 55 million years ago, called the Paleocene-Eocene Thermal Maximum, when global temperatures increased as much as 14.4 degrees Fahrenheit.

Unlike in previous episodes of climate upheaval, the activities of a single species are what’s driving the current crisis, which has the potential to affect every corner of the planet. While carbon dioxide from burning fossil fuels is the most immediate threat, the reality is that humans have little understanding of the many complex systems that could be disrupted in the process. The microbial communities found at methane seeps are just one of them.

At a depth of 500 meters, Pea Island sits at the upper limit of hydrate stability, what scientists call the “feather edge,” making it extremely susceptible to rising ocean temperatures. There are untold numbers of similar seeps around the globe.

“Pea Island is sort of the poster child of change in the oceans with respect to methane,” Joye said.

Other scientists and methane experts are less concerned about a runaway CH4 scenario from hydrate ― at least anytime soon. Carolyn Ruppel, who leads the U.S. Geological Survey’s Gas Hydrates Project, is among those who have pushed back against fears of a looming “methane time bomb.” Her research shows that the vast majority of known methane hydrate ― more than 95% ― exists in the deep ocean, below 1,000 meters, and that a large-scale release would require hundreds or even thousands of years of warming.

Additionally, ocean physics greatly limits the amount of gas that can reach the atmosphere, Ruppel explained in a phone interview. The gas dissolves into seawater on its way up through the water column, where microbes convert it into CO2. A bubble released from a depth where hydrate can exist has very little chance of retaining methane all the way to the surface, she said.

In a monumental 2016 paper, Ruppel and John Kessler, an oceanographer at the University of Rochester in New York, wrote that “there is no conclusive proof that hydrate‐derived methane is reaching the atmosphere now.” Yet they acknowledged there are many locations where methane ice is vulnerable to warming, specifically in the Arctic and on upper continental slopes, which “could be a major source of atmospheric CH4 under certain catastrophic, but unlikely, circumstances.”

It’s obvious why methane hydrate has alarmed the public. But Ruppel says it is the shallow water seeps, those on upper continental shelves and not associated with hydrate, that have a greater capacity to inject methane into Earth’s atmosphere.

“What I tell young people now is, if you really want to make a career for yourself, don’t worry so much about the deeper water seeps,” she said. “Worry about what methane is coming out of the shelves.”

Further complicating this nascent field of science is the fact that many countries, including the United States, are eyeing hydrates as a potential future energy source ― production that risks releasing even more of the gas.

Joye is not one to sugarcoat what she sees or what she makes of it. Her no-nonsense approach has earned her both praise and rebuke and has pitted her against powerful players.

When BP’s Deepwater Horizon oil rig exploded in the Gulf of Mexico in April 2010 and unleashed more than 200 million gallons of crude, Joye had already been studying natural seeps and microbial life in the region for 15 years. Within weeks of the deadly catastrophe, she organized a research team to collect samples aboard the Pelican, the first scientific vessel sent to the blowout site.

It was during that first mission that the team discovered large plumes of oil and methane forming deep in the Gulf, a sign that the spill was far worse than BP had indicated. BP insisted the plumes didn’t exist: “The oil is on the surface,” Tony Hayward, the company’s chief executive, said at the time. Other researchers later validated Joye’s finding.

Joye also sparred with the Obama administration. In August 2010, the White House released a government report that estimated 76% of the oil had dissolved or been cleaned up. Less than two weeks later, Joye co-authored a report that found nearly 80% of the oil was still in the water and a threat to the Gulf ecosystem. Government scientists maintained that their numbers were accurate.

Joye emerged from the disaster ― the largest marine oil spill in history ― as something of a scientific superhero, the brainy introvert willing to share data that many felt that the government and BP were keeping under wraps.

And she’s kept at it. Five years later, Joye co-published a study that concluded the 1.8 million gallons of chemical dispersants cleanup crews dumped into the Gulf likely made the situation worse. Rather than breaking the oil into smaller droplets that oil-eating bacteria could more easily consume, the chemicals slowed the microbes’ ability to degrade oil, she found.

“The dispersants did a great job in that they got the oil off the surface,” she told The Associated Press at the time.

Fellow scientists say she’s “a force of nature,” a researcher who has made “heroic efforts to communicate science to the general public.”

Cordes and Joye have been collaborating ever since their first cruise together in 2001. He said there are few people who can keep up with Joye’s level of energy.

“She’s one of the most creative scientists that I’ve known,” he said. “While she’s gathering data, she’s also interpreting it and throws ideas out there. And she has an amazing ability to be right more often than not.”

A little before 8 a.m., Joye and USGS microbiologist Chris Kellogg climbed to the top of a narrow staircase on Atlantis’ stern and kicked off their sneakers. Kellogg waved to fellow researchers watching from the deck. Joye flashed a modest smile. The two scooted down a small ladder into Alvin, a three-person deep-sea submersible most famous for exploring the wreckage of the Titanic in 1986. On its front end are two robotic arms, numerous cameras and a basket for stashing the samples it collects.

After the hatch on Alvin was sealed, a giant hydraulic crane plopped the sub into the sea, and the sub’s crew descended into the dark.

Atlantis, built in the mid-1990s and owned by the U.S. Navy, accommodates more than 50 people, has six onboard labs and was designed specifically to support Alvin. The ship stays in constant contact with the sub using an acoustic telephone. If you’re below deck during a dive, you can hear the radio chatter from the sub buzzing through the ship’s steel hull.

Eight hours later, Alvin dropped a large set of weights and slowly rose back to the surface, the first samples of the cruise in tow. Once the sub and samples were onboard, an excited team of scientists scrambled to unload quill worms, a pair of starfish, carbonate rocks and samples of muddy sediment and beggiatoa. Dead squid dangled from Alvin’s exterior, victims of their own curiosity.

The dive went well; good visibility, a manageable current and lots of specimens. But Joye was frustrated. Not only was the team unable to glimpse the methane plume it had seen on the radar, but several mud samples, called cores, degassed on their way to the surface, making it harder to profile the sediment inside. Joye suspected they contained chunks of hydrate, judging from how violently they had erupted.

Fortunately, the samples weren’t ruined. After a day in the ship’s cold room, a walk-in refrigerator that simulates the frigid deep ocean, the beggiatoa had wriggled their way to the top of the mud in search of oxygen, forming beautiful white geometric structures. Still sporting an astronaut-like jumpsuit for working in the cold room, Joye placed a dish underneath a microscope and instructed me to have a peek. Up close, the beggiatoa looked like hollow strings of spaghetti. Inside a few, yellow molecules of sulfur popped against a backdrop of dark mud. A translucent worm burrowed under the bacteria, making it roll and turn.

“It’s like gold from the bottom of the ocean,” Joye said. “White gold.”

The methane seeps that these bacteria inhabit are unforgiving environments. They are also diverse and vital ecosystems, part of the foundation of the ocean food web that hundreds of millions of people rely on for food and income.

Amanda Demopoulos, a deep-sea benthic ecologist at USGS, hopes to drive that human connection home.

On most evenings during the cruise, she could be found processing seafloor sediment samples in the ship’s wet lab. It’s tedious work that involves slicing cores of mud into precisely measured sections, then carefully funneling the sediment into small bottles for future analysis. A single core can contain hundreds of microorganisms, which the team identifies and documents. These tiny critters are important indicators of ecosystem health and break down organic matter that cycles to the seafloor. If they disappear, that material can create environments where nothing can live, Demopoulos said.

“We want healthy earthworms in our gardens,” she said. “We need healthy animals in the ocean, too.”

On Day 10 of the cruise, I got to join Joye on a “dragon hunt” to a gas seep more than 130 miles off North Carolina’s coast and nearly a mile and a half below the surface.

A submersible dive is like a slow-motion fall through a distant galaxy. Outside our 3-inch titanium shell of safety, a frenzy of glowing bioluminescent critters — shrimp, jellyfish and chains of egg-like animals called salps — flickered, scurried and burst as they collided with Alvin’s robotic arms. Joye described it as nature’s ultimate fireworks display.

“The Fourth of July can’t hold a candle to a submarine dive,” she said as we made our way to the bottom.

The descent into total darkness lasted 70 minutes. Your mind can’t help but run wild thinking about what creatures might be lurking just out of sight. On a murky canyon dive a few days earlier, a 20-foot sixgill shark had bumped into the front of the sub.

Pilot Jefferson Grau flipped on the sub’s exterior lights, giving us our first glimpse of the ocean floor. It was alien and beautiful, startling and mesmerizing ― made even more so by the spacey music playlist Grau had cued up. At this depth, Alvin actually shrinks slightly as the pressure outside reaches 3,200 pounds per square inch. The amount of force it would take to open the hatch is roughly equivalent to lifting a fully loaded 747 jet.

This site we were visiting, Blake Ridge, is relatively well-studied and rich in methane hydrate, with numerous active seeps. Joye directed Grau, who not only drove the sub but steered a small arm that controlled the robotic arms out front. As the two worked, I did my best to film what was happening outside using a small joystick that controlled exterior cameras, jotting down times and depths.

Extensive beds of mussels, some big enough to hold a newborn child, and piles of ghost-white clamshells littered the seafloor. A bright red Spanish dancer, a type of sea slug, fluttered by. A brittle star tossed up sand as it scampered away. Our presence seemed to confuse purple octopuses, crabs and rattail fish — a bizarre creature with bulging eyes, a long tail and a sharklike fin on its back. It was almost certainly the first time any of these creatures had seen light.

Communities of bacteria that use chemosynthesis, a process similar to photosynthesis, to convert inorganic chemical compounds like methane and hydrogen sulfide into energy fuel life in the deep sea. The mussels and clams have a symbiotic relationship with these bacteria, providing them a safe place to live in return for food.

Joye and I kept our eyes peeled for bubbles, or if we were lucky, a chunk of hydrate ― the dragon itself ― which often forms under rock overhangs.

She let out a shriek when she spotted a large bed of mussels, often a telltale sign of the presence of hydrate. “I think we’re about to hit our nirvana!” Joye said as Grau brought the sub in for a closer look.

The mussels turned out to be dead, possibly because the supply of methane in that particular spot had shut off. Cold seeps are variable systems. As Ruppel explained, they have plumbing systems that resemble tree branches below the seafloor, and the amount of gas flowing through any one pathway can fluctuate.

Joye jotted “deathbed” in her dive log, noting our depth of 2,169 meters.

We never caught a glimpse of methane bubbles or hydrate that day, but returned to Atlantis with one of the largest hauls of mud samples from the two-week cruise ― enough to keep Joye’s and Demopoulos’s labs busy for months.

While working late one evening in one of Atlantis’ cluttered labs, Joye made a startling discovery in seawater samples from Pea Island. Her students had been busy extracting gas from the water to study the content, which she then put through a gas chromatograph, a sophisticated device that separates a mixture of gases into individual components.

Her jaw dropped as the machine, which she’d nicknamed “Bucky,” kicked back the data.

One sample after another showed that the methane at and above Pea Island was off the charts. Joye wondered if her students were playing tricks on her. Or, even worse, that trusty old Bucky was broken.

She ran the samples again, but the numbers were solid. Methane concentrations at the Pea Island seep were among the highest she’d ever seen.

When she had more time back in her lab at the University of Georgia to crunch the numbers, what she found was even more alarming. Even though the microbes at Pea Island are gobbling up methane about 10 times faster than rates documented at natural seeps in the Gulf, the concentrations are so high that it would take them about 618 days to consume it all ― and that’s if the supply of methane suddenly stopped.

In other words, the microbes at Pea Island aren’t nearly keeping up. And the levels of methane were high in samples throughout the water column, all the way to the surface.

“That suggests that some of it’s going to get into the atmosphere,” Joye said. “That’s very scary.”

Over the last year, Joye has been trying to figure out what those findings mean. We know how methane in the deep ocean behaves under normal conditions, but current climate conditions are anything but normal. The global average temperature is already 1.1 degrees Celsius above preindustrial levels. The world’s oceans have absorbed an estimated 93% of the excess heat. Meanwhile, Atlantic currents have slowed by an estimated 15% since the mid-20th century.

Joye’s preliminary conclusion is that an increase in hydrate breakdown from ocean warming and a weakening of ocean circulation could lead to a marked increase in methane emissions off the Atlantic coast. She plans to make that argument in an upcoming scientific paper.

Since the cruise, Joye and her students have been exposing the methane-eating microbes to different conditions, trying to better understand what makes them tick and how they might respond under future climate scenarios. This research could prove important for possible human intervention, just as scientists are now exploring cloud seeding and geoengineering to save coral reefs from deadly bleaching events. Maybe scientists can tweak bacteria to make them more efficient at consuming methane, or maybe they can introduce an existing nutrient to stimulate their growth and activity.

Joye’s hunch is that there is a naturally occurring organism out there that, with a little help from humans, could be a fighting force against a potential future methane surge. Her research in the Gulf of Mexico is one reason she thinks that’s possible, as methane-eating bacteria flourished in the initial wake of the oil spill, consuming gas at the highest rate ever recorded in the open ocean.

“Finding that magic organism that’s able to do all the wild metabolisms that we are dreaming of having the capability to do out there ― that’s motivation for doing this work,” she said.