The drive to the little white dome on the northern slope of Mauna Loa is a bumpy one. Mauna Loa, the “Long Mountain,” is a colossal volcano that covers half of the island of Hawaii. The rocky terrain, rusty brown and deep red, crunches beneath car tires and jostles passengers. Up there, more than 8,000 feet above sea level and many miles away from the sounds of civilization, it doesn’t feel like Earth. It feels like another planet. Like Mars.

For the past five years, small groups of people have made this drive and moved into the dome, known as a habitat. Their job is to pretend that they really are on Mars, and then spend months living like it. The goal, for the researchers who send them there, is to figure out how human beings would do on a mission to the real thing.

In February of this year, the latest batch of pioneers, a crew of four, made the journey up the mountain. They settled in for an eight-month stay. Four days later, one of them was taken away on a stretcher and hospitalized.

The remaining crew members were evacuated by mission support. All four eventually returned to the habitat, not to continue their mission, but to pack up their stuff. Their simulation was over for good. The little white dome has remained empty since, and the University of Hawaii, which runs the program, and NASA, which funds it, are investigating the incident that derailed the mission.


The mission that began in February was the sixth iteration of the Hawaii Space Exploration Analog and Simulation, or HI-SEAS. The durations have varied, from four months to a full year, and participants come from all over the world and different fields.

HI-SEAS is a social experiment, and the participants are the lab rats. They wear devices to track their vitals, movements, and sleep, answer countless questionnaires about their own behavior and their interactions with others, and journal several times a week about their feelings.

Psychology researchers take all that data and use them to tease out information about what works and what doesn’t when you stick people in a tiny space they can’t escape. (Hint: They get on each other’s nerves—a lot—as documented in a recent podcast series, The Habitat. There’s also a little romance.)

Meanwhile, the crew members live as much as possible like they are on Mars. They eat freeze-dried food, use a composting toilet, take 30-second showers to conserve water, and never step outside without a space suit and helmet. They don’t communicate with anyone in real time, not even family. An email to mission support or their loved ones takes 20 minutes to get there. Receiving a response takes another 20 minutes. They’re not allowed to see anyone outside of the mission.

The habitat is a tight squeeze. The ground floor, which includes a kitchen, bathroom, a lab, and exercise spaces, measures 993 square feet. The second floor, where the bedrooms are, spans 424 square feet.

“You really do get the sense, when you’re going to sleep and you’re closing your eyes at night, that this could be a distant planet,” says Ross Lockwood, a physicist from Edmonton, Canada, and one of the members of mission two. “This could be Mars.”

But sometimes, Earth finds a way of sneaking in, of breaking the fuzzy boundary between simulation and reality.

Mission six arrived at the habitat on February 15. The crew waved goodbye to the researchers gathered outside the dome, felt the breeze on their faces for the last time for a long time, and piled in. The doors closed. Michaela Musilova, one of the crew members, described their first moments in an interview in April with The Cosmic Shed, a science podcast. (Musilova declined an interview with The Atlantic.)

“Our commander cited part of [the 2011 novel] The Martian. I think it was the very first line of The Martian, like, ‘Oh, we’re fucked now,’ or something along those lines,” recalled Musilova, an astrobiologist from Slovakia. “And so we just gave each other a big hug and like, ‘Okay, we can do this.’”

The first few days were cloudy, which can be a problem on the volcano. The habitat and its systems run on a battery bank that is charged each day through a large solar array on the grounds. On cloudy or rainy days, it can be difficult for the batteries to bounce back. When this happens, the crew is supposed to suit up, go outside, and turn on a car-size backup generator that runs on propane.

“We really make it as primitive as some farm in Vermont,” said Bill Wiecking, the HI-SEAS tech-support lead and the energy-lab director at the Hawaii Preparatory Academy. Mission support receives text-message alerts when the habitat’s life-sustaining systems reach dangerous levels, but for the most part, it’s up to the crew to manage their use.

As stubborn clouds hung over the habitat, the crew tried to minimize their energy use. They dimmed most of the lights, kept kitchen appliances unplugged, and stayed off the treadmill.

On the morning of February 19, Lisa Stojanovski, a science communicator from Australia, woke up to find that the power in the habitat had gone out. “We must have used too much power, I guess,” she told me.

Stojanovski and another crew member initiated the procedures for leaving the habitat. They shimmied into their space suits, stepped outside, and headed for the backup propane generator, located nearby on the grounds. Stojanovski and her partner would flip a switch to bring the generator to life, while the two other crew members would flip a switch on a circuit breaker inside the habitat. This maneuver would shift the power source from the dead batteries over to the generator, Stojanovski said.

When it was done, Stojanovski came back inside. “I was elated that we were on track to solve the problem, and I was pretty bouncy and excited,” she said. “It was a little jarring at first, when the two crew members who were inside didn’t quite share the excitement. That was my first gut feeling that something was not quite right.”

One of the crew members was typing furiously at a computer. The other looked stricken, pale. They said they didn’t feel well.

They said they had sustained an electric shock.

Nothing like this had ever happened before inside the habitat. Kim Binsted, the HI-SEAS principal investigator and a professor at the University of Hawaii, told me that injuries during previous missions ranged from bruises and scrapes, acquired during treks across the rocky landscape, to “household accidents.” “The kinds of things that you can hurt yourself doing at home are also the kinds of things that you can hurt yourself doing at the hab,” Binsted said.

Stojanovski said she suspects the electric shock may have occurred because the crew member’s fingers brushed against live wiring. “In a regular household circuit breaker, you have a safety panel that covers all the live wiring that’s behind the switches,” Stojanovski said. “Unfortunately, our circuit breaker didn’t have one of those.”

The injured crew member was shivering. They lay down on the floor. The others covered them in blankets.

The crew placed several calls to the mission’s on-call doctor on an emergency cellphone in the habitat, which works in real time, but there was no answer.

The person designated as the crew commander then called 911 on the emergency line. The crew wasn’t supposed to have contact with people outside of the habitat. If first responders came to the dome, the simulation would be compromised. Stojanovski said the commander told her he wasn’t calling to summon an ambulance, but only to ask for medical advice. This took her aback. Stojanovski believed they needed an ambulance, and they needed it now.

The crew commander, Sukjin Han, an assistant professor of economics at the University of Texas at Austin from South Korea, told me he signed off on most of the important decisions during the mission, but that he made sure to “hear the thoughts and opinions of all crew members beforehand and reflect them in the decisions.” In the tense moments after the accident, Han went with the majority.

“The majority of the members—including the member who experienced the incident—decided that we [wanted] to ask for medical advice from 911, before asking for an ambulance. I don’t remember if Lisa had the same opinion but I do remember that she never objected to the plan,” Han said. “I have never thought and don’t think that maintaining the simulation is more important than the safety of the crew.”

During their training, HI-SEAS crew are told often that their well-being comes first. Safety is paramount. But so is maintaining the simulation. No one involved in HI-SEAS wants to jeopardize the data by breaking the sim, as they sometimes call it. They don’t want to give up before it’s over, either. Leaving the habitat would mean throwing away hours and hours of physical, social, and emotional investment. For participants who came from outside of the United States, it even means visa troubles.

“We all left our regular lives, quit our jobs in some cases, left our loved ones to go spend eight months doing this,” said Laura Lark, a software developer in New York who participated in mission five. “So we’re all pretty committed to getting high-quality data out of it.”

The thought of abandoning the simulation becomes more painful the longer the mission goes on. The crew members of mission six were faced with this dilemma just four days in. What if it were four months?

In one of the early missions, a crew member unwittingly turned on an internet access point that interfered with the HI-SEAS network, causing a communications blackout between the mountain and mission support. Wiecking had to go up the mountain to fix it, a move that could have jeopardized the integrity of the crew’s isolation. As Wiecking fiddled quietly with hardware a few feet from the dome, he could hear the voices of the crew through the tentlike walls. “That was so close to breaking the simulation, we had to have a big review over it,” he said.

During Lockwood’s mission, the second mission of the HI-SEAS project, a crew member decided to withdraw because of a chronic medical issue. “We struggled with the idea of what we would do if we really were on Mars,” Lockwood said. They decided to pretend that the crew member died. They imagined that they would leave his corpse out in the Martian atmosphere, where it would not decompose as it would on Earth, in the hopes of bringing it back to Earth for burial.

And they actually acted this all out. Lockwood said they had the departing crew member step into the vestibule that separates the habitat from the outside, the simulated “airlock.” The person stood there for five minutes, as they all would do before doing an extravehicular activity (EVA), and waited, pretending the airlock was depressurizing, evening out the pressure inside and outside, so they could safely exit. Then the crew member opened the door and walked outside, where mission support staff picked them up and took them down the mountain.

This time, in mission six, the danger was real. As the crew tried to figure out what to do, Stojanovski started to get worried about the injured crew member. “They were going downhill,” she said. “They had a tight chest and some pain behind the shoulder blades. I’m not a doctor or anything like that, but those kinds of symptoms freaked me out a little bit. I was pretty worried that they were going to have a heart attack or something.” The crew had received first-aid training, but the situation seemed to require more than that.

She called Binsted, the HI-SEAS principal investigator, and told her what happened. No one could reach the on-call doctor. Stojanovski said Binsted told the crew to call 911 again. This time, they asked for an ambulance.

“Throughout all our training, we’d been told, ‘Don’t worry, emergency services knows where you are, they know who you are, and they know how to get to you,’” Stojanovski said. “I was like, ‘My name’s Lisa, I’m from the HI-SEAS project, we would like an ambulance please, this is where we are.’ And they were like, ‘You’re from what project? Where are you located?’”

Stojanovski’s call to 911 had been picked up by Hawaii County dispatchers, but help would arrive from elsewhere.

The Pohakuloa Training Area is a U.S. Army base of several hundred people, located less than 15 miles from the habitat. Its jurisdiction stretches from Mauna Loa to Mauna Kea—and the HI-SEAS habitat sits nearly in the middle. Like the habitat, Pohakuloa is isolated from the rest of the world. The remoteness requires the military base to operate like a city, complete with a fire department and EMTs.

“We got the call that morning that there was a potential electrocution, that the individual was awake and conscious, but [they were] breathing heavily and [they] needed to be checked out,” said Eric Moller, the fire chief of the Pohakuloa Training Area, about the call from Hawaii County. “They were afraid about hypertension, elevated blood pressure.” The Army base dispatched an ambulance carrying four responders.

On a clear day, the drive from the base to the habitat takes 35 to 45 minutes. According to a response report from Pohakuloa obtained by The Atlantic, the drive took 43 minutes on February 19. Inside the habitat, the minutes seemed to drag like hours. At one point, Stojanovski said one of the responders called the habitat to say they were lost.

Moller said Pohakuloa’s fire captain phoned the habitat because responders were concerned about the conditions of the roads, which are unpaved, but they weren’t lost. When they reached a gate along the route to the habitat, they found that the lock was jammed. This added to their response time.

“Our guys go up and down that mountain all the time,” says Gregory Fleming, Pohakuloa’s deputy garrison commander, often to rescue lost hikers in flip-flops. And military staff know not to disturb their neighbors, the fake astronauts. They have been advised that any interaction risks wrecking their delicate reality.

When the crew finally heard tires grinding over rock outside, Stojanovski turned toward the exit, ready to greet the first responders. Han stopped her, she said, warning that whatever happened next would break the simulation. “I actually lost my temper at this point,” Stojanovski said. “I don’t remember exactly what I said, but there were some curse words involved.”

Han said he remembers Stojanovski moving quickly to the door. “I correctly remember that at least two of the members, including myself, called her name, almost simultaneously,” he said. “At least for myself, it was partly in order to calm her down, because she has suddenly become very emotional at that very moment, and give [her] at least one second to think about her reaction.”

Stojanovski could have ignored the others. HI-SEAS participants receive specific roles, like commander or communications specialist or health officer, but compliance is not compulsory as it would be, for example, in a military mission. “They have to fulfill those roles, but ultimately as they come together as a team, that’s something the crew have to figure out on their own,” said Joseph Gruber, the mission-support coordinator for HI-SEAS, and one of the people who regularly communicates with crews over email. “There’s structures in place and we give them guidelines on how best to do this, but it’s up to them. They’re the ones up there.”

Stojanovski decided to heed Han’s request. She didn’t go outside.

Stojanovski opened the door and waved the first responders into habitat. They loaded the injured crew member into the ambulance and checked their vitals. The ambulance drove down the volcano as far it could go; after about 20 miles, the vehicle neared the edge of Pohakuloa’s jurisdiction, a line the first responders aren’t allowed to cross. If they travel beyond this region, the reasoning goes, they leave the Pohakuloa residents at risk.

A hospital ambulance met the Pohakuloa ambulance at this edge, grabbed the crew member, and sped toward Hilo Medical Center, about 30 miles east of the habitat.

“It was really surreal when the ambulance drove away and there was kind of just silence,” Stojanovski recalls. “Like, wow, what just happened?”


Back at their training base, a house in Kona, Stojanovski compiled a list of safety concerns about the habitat and sent it to Binsted, who confirmed she received it. Binsted wanted to continue the mission after getting approval from the university and NASA. Stojanovski said she did too, but only after mission support addressed her concerns and implemented some fixes.

Stojanovski sought some reassurance, but Binsted couldn’t make any guarantees, at least not before an investigation. “I kind of sat there and thought, You know what? I’m not okay with this,” Stojanovski said. “I’m not okay with the culture and the attitude toward safety.” Now that she was off the mountain and out of the bubble, her perception of the mission changed. She decided to withdraw from it altogether.

Binsted, the principal investigator, said she could not discuss the specifics of the incident until the institutional review boards, one at the University of Hawaii and one at NASA, concluded their investigations and issued reports and recommendations.

Musilova, Han, and the fourth crew member, Calum Hervieu, an astrophysicist and systems engineer from Scotland, declined extensive interviews but provided The Atlantic with a joint statement, saying, in part, “We prefer not to discuss this topic with the media” until the University of Hawaii and NASA complete their reviews. They point to press releases from February, which say only that a crew member was hospitalized, treated at the hospital for a few hours, and then released.

Stojanovski said mission support was understanding and professional about her decision. Her fellow crew members were shocked and tried to persuade her to stay. If Stojanovski left, they all had to. HI-SEAS protocol prohibits a crew smaller than four, which produces fewer data for the researchers. There’s also the matter of maintaining the habitat and its various systems—power, water, food, the toilet—which requires several sets of hands.

They couldn’t replace Stojanovski with a backup, either; HI-SEAS missions are set up to investigate the evolution of one particular crew over time, and besides, finding someone willing to fly out to Hawaii for an eight-month mission on such short notice would be difficult.

Every HI-SEAS mission since the first one in 2013 has had a crew of six. Mission six started out that way, too, but two people were removed from the program, one of them just days before she said she was scheduled to fly from Australia to Hawaii. Binsted said she couldn’t comment on why mission six went ahead with four.

Brian Shiro, a geophysicist at the U.S. Geological Survey’s Hawaiian Volcano Observatory, who has worked on HI-SEAS since its inception, said staff deliberated about whether to move forward with a smaller crew. “At any point along this timeline, there could have been a hard decision to delay the mission or cancel, but that’s not what they decided,” Shiro said. “I was on the side of the fence to delay. I didn’t want to start this mission because of the crew size. I said, ‘Guys, let’s find more people, let’s wait a few months at least.’ But I was overruled.”

He added: “This crew was very, very impressive, very professional, very motivated. But there were only four of them, and that left them vulnerable.”

During a real Mars mission, crew members will face a panoply of risks. People can, and probably will, get injured. They may die. Simulations like HI-SEAS attempt to forecast reactions to some of these threats, ranging from the things we cannot control, like the poisonous air outside, to those that we can only intuit, like the ideal way to organize a crew.

“We have things that we know we don’t know,” said Jenn Fogarty, the chief scientist at NASA’s Human Research Program, the office that provides financial grants to HI-SEAS. “The ‘I don’t know what I don’t know’ is the scary space.”

Long before we send the first humans to Mars and keep them happy and healthy, we’ll have to figure out how to do that here—and it starts with deciding who should be on the mountain, which isn’t easy.

“You can select a crew all you want, get the right fit and mix, but there’s too many variables when it comes to human beings,” said Raphael Rose, the associate director of the Anxiety and Depression Research Center at UCLA, who was set to study stress management and resilience on mission six. “It’s just really hard to predict how we’re going to perform in all situations.”


Mission six arrived on Mauna Loa after the customary, rigorous application process that requires written essays, reference checks, Skype interviews, and, perhaps most important, the same kind of psychological screenings that NASA gives its astronauts. With each iteration of HI-SEAS, researchers and mission personnel learn a bit more about crew composition and what types of people work well together.

Steve Kozlowski, an organizational psychologist at Michigan State University who studies team effectiveness, said HI-SEAS finalists are scored on five personality traits, known in the field as the Big Five: extroversion, agreeableness, conscientiousness, neuroticism, and openness to experience. Kozlowski said they want conscientious people, but up to a point. Conscientiousness can veer toward passivity. Some degree of extroversion is valuable, until it’s too much. Outgoing people can morph into domineering people. In other words, it’s all about balance.

“There’s no magic formula,” Kozlowski said.

Psychological screenings can only predict so much. “Sometimes people look really good on paper and they might even interview well, but if there’s a big red flag on that screening, it gives one pause,” Shiro said. “There’ve been people we’ve ruled out because of that.”

During the mission, crews make regular trips outside of the habitat to explore the volcanic terrain in their space suits. To prepare them for this EVA, Shiro leads participants on excursions across the rugged landscape soon after they arrive in Hawaii. “Three days in the field in those conditions is a good way to get to know people,” he said. “There’s people that I think, Eh, I wonder how that one’s gonna do. Usually, that gut feeling, there’s something to it.”

A real mission to Mars would likely require crews to train together for months, maybe even years—far longer than the nine days of training that mission six had, Shiro said. Crew members would be put through a multitude of stressful situations to test their reactions. “You would tease out any red flags before you even left Earth,” Shiro said.

Shiro said one of his gut feelings kicked in during the fieldwork training for mission six. “There was this one person who was not as comfortable in the field,” he said. “That’s the sort of thing you don’t know until you get out there. Still did it, did all the training—a little slower, but did it all. But when the incident happened that ultimately led to the cancelation of the mission, that’s the person who quit. And it was not a surprise to any of us because we said, ‘Yeah, you know, she was a little more timid out there.’”

Stojanovski rebuffed Shiro’s assessment of her training. “I actually enjoyed being out in the field,” she said. “In fact, I was the first to volunteer to go outside on an emergency space walk on the morning of the incident.”

The HI-SEAS staff say the habitat is a safe environment.

“We’ve learned all the ways that you can kill yourself on Mars, and we’ve learned to prevent those things,” Wiecking said. “So it’s been very, very valuable, because it’s way better to do it here, where you can drive up and go, ‘Oh gosh, a water valve opened up and now you don’t have any water.’ Instead of on Mars, where it’s like, ‘You don’t have any water, you guys are gonna die in a couple of days.’”


Round-trip communication between Earth and Mars will take about 40 minutes. Astronauts won’t have the luxury of sitting around and waiting for commands or approval from Earth. HI-SEAS has mission support, rather than mission control, for that reason. The first astronauts on Mars will choose for themselves, for the most part, how they will live and work. In case of an emergency, they will need to decide what to do. And there’s no guarantee the  astronauts won’t choose to take matters into their own hands.

“That’s the complexity of humans. They are going to do things on their own, maybe outside of the mission rules. They are going to try to make things work on their own, and they’re inventive and smart, and that’s the reason you picked this crew,” Fogarty said. “So thinking you can keep them in this tight little box of emotions is unrealistic.”

The potential cracks in the relationship between crew and mission support are already showing. Last year, when Hurricane Harvey struck Texas and forced the displacement of thousands, NASA staff decided to evacuate the members of a space simulation in Houston. For several weeks, a crew of four lives and works inside a cozy fake spaceship at the Johnson Space Center, pretending they’re coasting toward an asteroid.

“When we woke them up that Sunday morning and told them to pack up, that we were terminating the mission, they were not happy with us,” said Lisa Spence, the flight-analog project manager for the program, known as the Human Exploration Research Analog (HERA).

“One of them was pretty upset and not very complimentary and [asked us], ‘Why are you doing this? There’s no problem here, we want to continue.’ It wasn’t until the vehicle came and evacuated them and took them to the hotel, and they could see cars stranded all over the place, and boats that had been washed up onto the streets, and several feet of water across the roads—then they kind of appreciated why we stopped.”

The mission support at HERA had better information than the crew did, and because they sit in the same warehouse, just 20 feet away from the “spaceship,” they could make a decision for the crew. That won’t be possible on Mars. If the crew goes rogue, the people back on Earth might have no idea. Some degree of eavesdropping on the crew might be necessary, according to Sonja Schmer-Galunder, a researcher at Smart Information Flow Technologies, whose work on HI-SEAS was aimed at developing ways to predict the behavioral health of individuals and the team.

“I’m not the person to decide where the limits of their privacy are, and clearly you have to be able to withdraw and also have your private space. Is it ethical?” Schmer-Galunder said. “I mean, if people are signing up to go to Mars, I think everything should and must be done in order to bring the crew back safe. When you’re signing up for a Mars mission, you know that you’re giving yourself away in almost every aspect of your life. You become a tool that is being sent out there.”


The HI-SEAS program is now on hold until the University of Hawaii and NASA complete their separate reviews. Fogarty, of NASA’s Human Research Program, is supportive of Binsted and the project. Fogarty said it’s possible that the university and the space agency could come to different conclusions about the incident, which could determine the future of the HI-SEAS project.

“In the future, NASA may not participate in it if we don’t feel that our safety threshold for participants is met,” Fogarty said.

HI-SEAS, which is run mostly by volunteers, could continue on its own. But NASA’s withdrawal would be detrimental to the agency, which doesn’t have any similar Mars simulations. The longest run of HERA, the asteroid analog, was just 45 days.

Stojanovski returned to Australia soon after the mission ended. She had resigned from a job on the communications team at Rocket Lab, a U.S. spaceflight company with a subsidiary based in New Zealand, after finding out she had been selected for the HI-SEAS program. When she withdrew from the mission, the job had already been filled. She worked at a fish market for a few months when she came home. She recently found another role at Rocket Lab, as an executive assistant, and moved to Auckland in May.

According to their personal websites, Musilova and Hervieu have found work at the Canada-France-Hawaii Telescope, located near the summit of Mauna Kea, less than 40 miles north of the habitat. “I’ve been saying c’est la vie quite a lot recently, and that’s really how it is,” Musilova said in the Cosmic Shed interview. “Life happens.” Han is still listed as a professor at the University of Texas at Austin.

Stojanovski has stayed in some touch with Binsted and Musilova. She hasn’t spoken with Hervieu or Han.

Several months removed from that February morning, Stojanovski said she wishes her crew’s panicked discussions had gone differently.

“I really regret that I followed orders that were not in the spirit of crew health and safety, just for the sake of keeping the mission within the simulation,” she said.

I asked Stojanovski whether she regrets withdrawing from the mission. She said it was a hard decision. But no, she doesn’t.

“In a way, I am kind of glad that this happened, because it’s been able to be a learning opportunity that exposes all the weak points in the system,” she said. “We can make the system strong so that people, when they do eventually get to Mars and they have a situation like this, they can be in a better position to tackle it. You’re increasing their chances of surviving something like this.”

Stojanovski remembers fondly the few days her crew had on the volcano, before Earth came knocking on their door. It was cozy in that little white dome. One crew member had brought ping-pong paddles, so they cleared a table and started bouncing the ball back and forth, click, clack. Another came with an electronic keyboard and played classical compositions at night. Distinctly earthly noises in a place that felt anything like home. They wafted around the habitat, pierced its thin walls, and drifted out into the silent expanse.