Issue #3(17) 2018 Science

How many people does it take to colonise an exoplanet?

James Vaughan
James Vaughan
Frédéric Marin Observatoire astronomique de Strasbourg, France
Camille Beluffi CASC4DE, Strasbourg, France
Rhys Taylor Czech Academy of Sciences, Prague, Czech Republic

When it was discovered that a rocky planet was orbiting around our nearest star in a favourable location and that it might have liquid water on its surface, the prospect of visiting an Earth-like world moved from the realm of science fiction to a tantalising possibility. In the Summer issue of ROOM, #2 (16), Andrew Hein’s article ‘Flying to the stars’, considered the possibilities and challenges of getting beyond our solar system to an exoplanet. Here, the authors take a step further, calculating how many people would be needed for a journey that could take dozens of generations.

In 1995, astrophysicists Michel Mayor and Didier Queloz discovered the very first planet orbiting around a sun-like star beyond our own solar system: 51 Pegasi b. For the first time, speculation about worlds around other stars began to be replaced with real, definite data. Twenty-three years later, the number of confirmed exoplanets stands in excess of 3,700. The possibility of finding a truly Earth-like planet is now, many believe, simply a matter of time. But should we find such a world, what use would it be? Could we ever visit it ourselves, or will we forever remain prisoners of our own solar system with its single life-bearing world?

For those of us who want to actually visit in person, the recent discovery of Proxima Centauri b gives some cause for optimism. Orbiting the nearest star to the Sun, this is the closest exoplanet we can find. It is likely to be a rocky (telluric) planet as its mass is close to that of Earth. Most intriguing of all, its equilibrium temperature implies that water could be in a liquid state on its surface. Located at 4.2 light years (40,000 billion km), Proxima Centauri b is almost an ideal destination - as far as exoplanets go. But while this distance may be small by astronomical standards, it remains utterly vast on the human scale. The difficulties of sending even robotic probes are daunting enough, and sending living human colonists is a formidable challenge indeed.

Artist’s concept of 51 Pegasi b, which was discovered in October 1995. The giant planet is about half the size of Jupiter and orbits its star in about four days.Artist’s concept of 51 Pegasi b, which was discovered in October 1995. The giant planet is about half the size of Jupiter and orbits its star in about four days.

Cryogenics options

Proxima Centauri b is almost an ideal destination - as far as exoplanets go

It takes light a little over four years to cover the distance between Earth and Proxima Centauri b. Our current spacecraft technology cannot reach even a pitiful fraction of this speed: at the average velocity of Apollo 11 (5,500 km/h or 1.5 km/s), the journey would take over 100,000 years. More modern spacecraft have somewhat better performance but would still require extreme journey times. The NASA Parker Solar Probe, launched in 2018, will set a new speed record of 200 km/s by taking advantage of the enormous gravitational pull of the Sun. But even this translates to a travel time to Proxima Centauri of over 6,000 years.

Such a mission would be unlike anything ever attempted in human history. While faster-than-light ‘warp drives’ are a popular staple of science fiction and allow protagonists to flit from star system to star system in the time it takes for a commercial break, in the real world they defy all known physical laws. Even with drastic improvements in technology, our best understanding of physics dictates that human missions will inevitably be multi-generational endeavours.

There are, however, a few ways to avoid this without recourse to exotic physics. One option would be cryogenics, in which the crew is preserved and revived artificially once the ship reaches its destination. But this method is currently no more viable than warp drives, because freezing cells creates ice crystals which break down cell walls as they expand (a phenomenon called vitrification), reducing the body to useless mush once it is warmed up again. Various advances in the field have been made using chemical agents to prevent vitrification but the main drawback is that it leads to a dangerously high toxicity level in the human body.


Suspended animation

At the average velocity of Apollo 11 (5,500 km/h or 1.5 km/s), the journey would take over 100,000 years

Similarly, suspended animation scenarios, where the physiological functions of the crew members are slowed down until arrival to ensure energy conservation, are proving very unstable at best. Even animals which naturally hibernate face extreme risk. For instance, the hibernation survival rate for the Alpine marmot has been estimated between 51 and 92 percent, which is hardly a rate most potential interstellar colonists would be willing to accept!

Moreover, long-duration artificial sleep would require the use of neuroprotective agents to prevent brain damage due to oxygen deficiency. Unfortunately, at present, no individual neuroprotective agents have been proven safe and effective against neurological ‘sequels’. The laws of biology it seems, can be every bit as difficult to circumvent as those of physics.

A third cryogenic option seems more plausible: bypassing the need to deal with adult humans and instead concentrating only on their genetic material. Preserving a fully-grown human is extraordinarily complicated, whereas the technology to preserve human eggs, sperm and even embryos not only already exists but is used routinely. Sending a ship laden with frozen embryos could be a realistic option to slowly populate an Earth-like exoplanet, where robotic equipment would help them to mature and develop once they reach their destination. This would massively reduce the size (and speed) of the ship required to something feasible with near-future technology. However, we are still a long way from being able to take an embryo through the full development process - much less raise and educate a fully-functional human. There has never even been a population entirely reliant on using in vitro fertilization, let alone one raised completely by robots.

Currently, there is only one option we know that can successfully maintain the human species over a long period of time: a living, naturally breeding population inside a spacecraft. This is what Konstantin Tsiolkovsky, the man considered to be the father of modern astronautics, called “Noah’s Ark”, also known as “generation ship” or “multi-generational space ship”. Although such ships occasionally appear in popular culture, they remain poorly studied scientifically.

While it is important to continue studies into future technologies that could shorten the journey time, right now the only way we know to survive the trip is through a multi-generational voyage, so why not investigate this? After all, population genetics can be studied with much more accuracy than modelling future technological advances. In fact, several international projects (such as the Icarus Interstellar mission or the 100-year Starship project) have recently started to consider these ‘world ships’ as a viable means of interstellar travel.

Multi-generational spacecraft are often envisaged as huge, city-sized structures, but are such vast ships really necessary?Multi-generational spacecraft are often envisaged as huge, city-sized structures, but are such vast ships really necessary?

Multi-generational crews

If we want to start designing an interstellar mission, we have to fully accept that our designs will be modified in the future as our understanding of physics (and biology) changes and improves. Indeed, tackling a project expected to last longer than most of human civilisation isn’t something we can hope to solve in a few months or years, but there are individual aspects we can start to get a handle on. Specifically, if we can’t rely exclusively on cryogenics or preserved genetic material, we’re forced to ask the basic question: how many people do we need to maintain a genetically healthy population? Too low and inbreeding will lead to extinction; too high and the project may be logistically impossible.

Surprisingly little previous work has been done to answer this question, but there have been a few attempts. In 2003, the anthropologist John Moore used an ethnographic tool to rigorously estimate the minimum population size for a multi-generational journey. He simulated the marital and demographic situation facing small populations of humans trapped in a closed environment. The software included various demographic variables associated with birth and death rates, as well as several cultural variables such as marriage choice or polygamy. While his code was aimed towards the human colonisation of unoccupied regions, Moore also considered a space journey where immigration and emigration were not possible, and concluded that a 200-year mission needs to start with at least 150 people.

Social engineering

Even with drastic improvements in technology, our best understanding of physics dictates that human missions will inevitably be multigenerational endeavours

In addition, Moore’s scenario included a number of social engineering principles (i.e. social and mating rules) in order to avoid too much inbreeding within the crew. For example, Moore’s starting crew would be young, childless married couples who would postpone parenthood until relatively late in their reproductive life, maximising the capacity for genetic diversity in the first generation born on the ship. The spacecraft would then be populated with smaller ‘families’, albeit unconventional ones by modern Western standards, and the population would have an age-sex distribution designed to maximise the number of potentially reproductive crew. It wouldn’t look much like the usual distribution in real populations, which have larger fractions of young and old, but it would be a good way (Moore thought) to avoid dangerous inbreeding in a small population.

This rather modest initial population size was challenged in 2013 by archaeologist and anthropologist Cameron Smith. He published a critical re-evaluation of Moore’s number by including the effects of mutagenesis, genetic migration, mate selection and birth drifts in the simulations. According to his study, a crew of 150 people would always be on the verge of extinction in the case of a large-scale catastrophe. Smith decided a better option would be to use a very much larger gene pool, with a lower limit of 14,000 people. The huge difference between Moore and Smith’s numbers reveals that this is not a trivial problem, as many parameters can influence the result. For this reason, we attempted our own model to try to constrain the theoretical minimum size of an interstellar, multi-generational crew.

Concept for a Mars Colony Transfer Habitat Interior showing passengers in suspended animation pods. Artwork by Nathan Kreuzman and Mark Elwood.Concept for a Mars Colony Transfer Habitat Interior showing passengers in suspended animation pods. Artwork by Nathan Kreuzman and Mark Elwood.

Smallest possible crew

We started the development of a new simulation tool named HERITAGE in 2017. As well as the work of Moore and Smith, we incorporated the findings of other anthropologists and sociologists and developed a code based around the Monte Carlo method, commonly used when you know probabilities but can’t determine what will actually happen in advance.

The Monte Carlo method makes it possible to estimate the success rate and associated uncertainties on the results for any kind of complex situation and HERITAGE uses this approach to compute the evolution of a human crew aboard an interstellar ship. This allows us to explore whether a crew of a proposed size could survive for multiple generations without any artificial stocks of additional genetic material. For example, we can estimate the probability that the offspring of any couple will be infertile. So, in one simulation, perhaps one percent of the second generation happens to be infertile, but in another, through sheer bad luck, it might be 50 percent (this would be extreme, but possible). With a large enough population this variation won’t matter because there will always be enough fertile members of the next generation to continue to reproduce, but in a small enough group a high infertility fraction will eventually be fatal to the population.

HERITAGE works by running thousands of simulated space missions. Each contains individual crew members with unique biological, medical and demographic data, which are allowed to vary with time (for example fertility rate decreases with age for both men and women), to simulate the billions of interactions that can occur between breeding partners. This can determine with great precision whether a ship with a given initial crew can survive without inbreeding or genetic decay for several centuries.

By tracking the ‘family tree’ of each crew member, their rate of inbreeding and corresponding probability of negative effects can be estimated. We can control the initial population demographics and even breeding rules aboard ship, and then use the Monte Carlo method to account for the unavoidable element of random chance and see how our choices affect the initial outcome. At the relatively small numbers proposed by Moore, small variations in the breeding rules and initial crew demographics can have dramatic effects on the probability of success or failure.

It’s important to note that at this stage our concerns are strictly theoretical and biological. Eugenics, where breeding is regulated according to traits deemed to be desirable and undesirable, has a long and tragic history in the human species. At present our aim is only to calculate the minimum theoretical size of an initial crew permitted by biology. Only once we have this number can we even begin a discussion of the ethics of what this would require.

Managing a starship

It isn’t the size of the population that’s the limiting factor, it’s the choice of breeding rules

HERITAGE allows us to reconsider the findings of Moore and Smith. We found that Moore’s overly-strict social engineering rules drive the population to extinction for a small initial population, while conversely, it leads to unhealthy levels of inbreeding for Smith’s larger population. This is because rules which are perfectly adapted for the initial crew may no longer match needs inside the vessel after about two centuries.

In Moore’s scenario the population begins as a group of couples of the same age, making it easy for the next few generations to find breeding partners (under the rules that they must be of similar age and with no inbreeding allowed). After a few hundred years it becomes much more diverse but ultimately the number of available partners becomes ever smaller, driving the crew towards a slow extinction.

But this doesn’t mean Moore’s small population is inevitably doomed because of its size. Actually, it isn’t the size of the population that’s the limiting factor, it’s the choice of breeding rules: the age range of partners and the number of children per couple can’t remain fixed if the mission is to succeed. The social rules must be re-evaluated each year in order to allow for long-term and stable population growth, ensuring the population level becomes neither larger than the ship’s design capacity, nor so small that an individual cannot find a breeding partner. We also demonstrated that such adaptive rules can compensate for major population losses from fatal, large-scale diseases or other extreme incidents. These adaptive social engineering principles, in our model, are what create the difference between success and failure more than just the simple size of the population.

Nevertheless, from the perspective of actually designing and building the spaceship, it’s the numbers which matter. We found that only 98 initial gender-balanced crew members are necessary to ensure a 100 percent mission success rate with a genetically healthy crew. We assume a maximum ship capacity of 500, so the population is allowed to increase for the first few generations. However, the stable population level depends more on the allowed degree of inbreeding than the arbitrary ship capacity. If we forbid inbreeding completely, it reaches a stable level of about 300; if we only forbid inbreeding between partners more closely related than cousins, it reaches a little over 400. In either case this stable level remains almost constant for most of the 6,000-year journey, showing no decline or other variation after its initial increase.

This crew number of 98 is lower than the previous studies, which is due to our examination of the reproduction outcomes and the choice of adaptive social rules. These rules, though essential to success, are not particularly extreme compared to real societies where people tend to avoid marrying close relatives anyway. In fact, we found successful missions were possible (but less likely) with even lower numbers. Our final number of 98 therefore is a useful guideline but not a strict, definitive number for colonisation projects.

Artist’s concept of the surface of TRAPPIST-1f, an exoplanet about 40 light years from Earth.Artist’s concept of the surface of TRAPPIST-1f, an exoplanet about 40 light years from Earth.

Cows in space?

Sending a ship laden with frozen embryos could be a realistic option to slowly populate an Earth-like exoplanet - robotic equipment would help them to mature and develop once they reach their destination

This result, that the required crew number for such an extended journey is relatively small is encouraging. But it is of course just one small part of the whole project. Ultimately, the most important number to ascertain is the minimum total mass of the entire spaceship

Our next step towards this will be to investigate the food requirements of the crew. While a crew of 500 might have a mass of a few tens of tonnes, real astronauts on the International Space Station require approximately 1.8 kg of food and packaging per day. If we were to feed the crew of our interstellar mission entirely from stored food, the mass required would soar to two million tonnes. Farming, which recycles nutrients, could in principle dramatically reduce this. But establishing how much of a difference this could make is not straightforward. Food requirements vary according to gender, age, and physical characteristics. As with genetics, we can estimate the distribution of such parameters but we can’t know what they will actually be.

As Confucius said, even a journey of a thousand miles must begin with a single step. In our case the journey is much longer, but only by tackling small aspects of the problem can we hope to make progress. We don’t yet know if interstellar travel, much less multi-generational spaceships, will ever actually occur. The project as a whole is vast and complex, but by slowly and meticulously addressing each of the issues that we do understand, we hope that one day the dream may become a reality.

About the authors

Frederic Marin is a postdoctoral fellow at the Astronomical observatory of Strasbourg, France, where he carries out research on the formation and evolution of black holes. He has been selected to lead observations of galaxies for a forthcoming NASA satellite called IXPE (Imaging X-ray Polarimetry Explorer, to be launched in 2021). The HERITAGE project created and lead by Marin seeks to calculate precisely how humanity could endure a journey of several hundred years to Earth-like exoplanets.

Camille Beluffi completed a PhD in particle physics at the Catholic University of Louvain-la-Neuve, Belgium. She then became a Data Scientist for CASC4DE, a start-up based in Strasbourg that builds innovative software to help scientists deal with big data in their research projects.

Rhys Taylor is a postdoctoral researcher at the Astronomical Institute of the Czech Academy of Sciences in Prague. He specialises in radio astronomy and spent two years at the Arecibo radio telescope in Puerto Rico. His main scientific interest is galaxy evolution, particularly galaxies which contain large amounts of gas but little or no ongoing star formation.

If you've enjoyed reading this article, please consider subscribing to ROOM Space Journal to gain full access to current articles and receive your own print and/or digital copies of ROOM delivered direct to your door or electronically.

Popular articles

See also


PROSPECTing for lunar polar volatiles


Preventing pilot spatial disorientation


The sci-fi noir world of Blade Runner

Popular articles

The Lunar Gateway will orbit the Moon serving as a space station and hub for lunar and deep space exploration. Astronautics

Commercial opportunities beyond Earth orbit


High hopes for Brazil’s space ambitions