When it comes to building a sustainable settlement on Mars, the technological and engineering challenges are steep. But they take a back seat to the Human Resources department. Forget sophisticated vehicles or sensitive instrumentation—the most temperamental, fragile things we send to the Red Planet will be humans.

After all, NASA’s Opportunity rover roamed Mars for 14 years, separated from Earth by a half-hour communications delay, scoured by dust storms and irradiated by cosmic rays, and never complained or got into a fight with a colleague. 

Humans, though, will be sequestered “in a confined space about the size of a small RV for three years,” James Driskell, a research psychologist at the Florida Maxima Corporation, says of most plausible NASA Mars mission scenarios. Driskell and his company have consulted with the space agency and the US military on the psychological issues of crews in isolated and stressful situations. In tight quarters, “people get angry at each other.”

Current Mars plans, such as NASA’s proposed Artemis mission, would send astronauts there and back on a three-year round trip. But you can imagine how stressful dynamics—danger, isolation, other people—might increase on a permanent base or research station, if crews stayed for a decade (or forever). Or, rather than using your imagination, you can rely on the computer simulation of a Mars settlement produced by George Mason University Computational Social Scientist Anamaria Berea and her colleagues. 

In a forthcoming study that hasn’t yet undergone full peer review, Berea and her colleagues detail how they used an “agent-based modeling” approach—a computer system not all that different from a large video game—to calculate the survivability of different population sizes of Mars settlers. They’ve incorporated personality types, too, for the long haul. They came to two main conclusions: that only a few tens of initial settlers are needed to create a sustainable colony, and that people with more agreeable social traits did better for themselves and the larger settlement. 

[Related: Rodent astronauts suggest trips to Mars will make us anxious, forgetful, and afraid]

The new study originated as a response to other papers suggesting that between 100 and 300 people would be the minimum necessary to begin a sustainable settlement on Mars. The nonprofit Blue Marble Science Institute, which studies questions of planetary science and habitability, contacted Berea to see whether her team could verify the other studies’ minimally viable population numbers. 

Berea says she had a better idea: Creating a simulation for a space habitat that included “human, social, and behavioral factors.” Berea and her team at the computational social sciences department had created simulated humans, who were assigned a set of skills necessary for running a Mars settlement, such as producing food or maintaining life support systems. 

Each faux settler had one of four aggregate personality types: There were the “agreeables,” highly social and low in scores of aggressions or competitiveness; “socials,” extroverts with a bit more of a competitive edge; “reactives,” who were more still competitive and fixated on fixed routines; and “neurotics,” highly competitive people with difficulty coping with changes in routine or boredom. Settlement members could die in accidents, or due to “health” conditions determined by the available food and life support resources, but could also be replenished by resupply shuttles every 18 months—the researchers chose not to model sex and reproduction. 

After running multiple computer models for more than 20 simulated years, the study authors found that settlements could begin with far fewer than 100 settlers and remain sustainable, despite accidents or dips in food supplies. The lowest number to kickstart a sustainable settlement was 22 people, but that is not a hard limit, according to Berea. “It’s somewhere between 10 and 50,” she says. “It’s in the tens; It’s not in the hundreds like the other papers were saying.”

[Related: NASA rover finds evidence of carbon-based chemistry in Martian crater]

They also found that agreeable personality types were the most likely to survive to the end of each simulation run. But Bera is careful to note that the agents—the algorithmic representations of humans—do not remain static through the simulation, just as people, whatever their personalities, change over time. “The neurotic that puts his or her foot down on the planet on day zero might not be the neurotic on day 100. They interact, and they adjust,” she says.

This can be seen in real-world Mars mission simulations, such as the Hawaii Space Exploration Analog and Simulation (HI-SEAS) missions, which places crews of six people in a simulated Mars habituated on the rocky lava slopes of Mauna Loa. There, it’s vital to anticipate the ways people change over time. 

“For the first few weeks, usually of people living under stressful conditions, they can still kind of have a ‘honeymoon period’ where everyone’s still very polite and patient and can kind of get along despite some challenges,” says astrobiologist Michaela Musilova, the former director of HI-SEAS from 2018 until 2022. “Usually after the first few weeks is when people really start to struggle and if they’re not prepared for it properly.” 

That struggle could take the form of depression or rudeness with other crew members or mission control. Over the 30 simulated Moon and Mars missions for which Musilova served as commander, she found the answer was to consciously forge bonds between crew members using shared meals and evening recreation, such as karaoke

“The more the crew bonded, the longer the ‘honeymoon period’ lasted and even when it wore off, the crew still behaved politely towards one another,” she says. 

Musilova also found that selecting as diverse a group of people as possible, in terms of skills, life experience and ethnicity, helped ensure a better functioning team. 

That’s one thing that Berea and her colleagues didn’t model—all of their simulations contained equal numbers of the four personality types they had defined, rather than trying to build teams composed of different proportions of different types of people. Purposefully screening for personality is something Driskell notes is important for building teams going into difficult and isolated conditions. 

“What type of trait profiles do we want in that team? That sociability and extraversion is really good, but you don’t want a team full of it, because then they’re going to really want to just interact and get along and talk,” Driskell says. At the same time, he adds, you have people who are very competent and follow the rules and keep things running, but who are just a complete pain to live with. “Everybody’s got an example of somebody who was extremely technically adept, but you just could not get along with them,” he says. “I guess Elon Musk is a good example.”

Neither human nor computer simulations of Mars missions can ever fully predict the experience of putting human boots on the Red Planet, but each approach also takes a different slice of the problem. Computer simulations such as Berea’s and her colleagues can give researchers some idea of the large-scale population dynamics and psychology of a Mars settlement over many years. A 12-month HI-SEAS Mars mission, meanwhile, helps tease out real-life psychological nuance you can’t get from a computer model. 

Berea hopes to do more to integrate both approaches in the future, noting that NASA has just launched a new Mars analog mission, the Crew Health and Performance Exploration Analog (CHAPEA) in the Mars Dune Alpha habitat. “Once they are done with that project, it would be great to get the data and compare that with our model for validation,” she says.