Elon Musk says he wants to create a backup of Earth on Mars, and several brilliant people have suggested this should be one of our main objectives for sending humans into space, not only Elon Musk. Perhaps most notably, Stephen Hawking. But being brilliant doesn't mean you are always right in everything you say, or that what you say can't be questioned by anyone else!

There are other good reasons for space settlements - but would any of them be of any value as a "backup" any time in the next few hundred million years? 


First, there is no known hazard that is at all likely to make all humans on Earth extinct instantaneously. A very large asteroid impact could, but none of the planets or moons in the inner solar system inside of Jupiter have been hit by anything that big for over three billion years.

Some humans would certainly have survived the Chicxulub impact at the end of the Cretacious period. Small mammals did, also birds, turtles, many creatures did, and with our technology we have capability to survive the same things they can survive.

Then, so long as some humans survive on Earth, it remains the very best place in the solar system to rebuild civilization.

It is the same for the other possible natural threats. We don't need to escape Earth to avoid them.

In any case we have found all the 10 km or larger NEOs already and are well on our way to finding all the 1 km or larger ones. Any NEO that large from the outer solar system would give us a couple of decades or so of warning. Time to deflect or destroy or prepare.

It is well within our near future capabilities to find them all right out to the Kuiper belt. For far less than any budget to set up a human colony outside of Earth, and on a far shorter timescale, we can have a reasonably complete survey of our entire solar system.

It is hard to beat having a breathable atmosphere, and being able to go outside your habitats without a spacesuit. Or having liquid water, or atmosphere and magnetic field protecting from solar storms and cosmic radiation. So, of course, we would rebuild civilization on Earth anyway, not on Mars.

Anyone on Mars would come back here, if they could, as the obvious place to try to rebuild civilization - why continue on Mars when there is an Earth to rebuild?

About the only thing that could make us extinct - would be something that is the result of our technology. But a space colony would be more dependent on technology than any colony that has ever been in the history. Like living permanently in a submarine, but far harder than that. If technology is the problem, then any issue would probably start in the space colonies.

Space colonies would be the most technological societies ever developed.

  • Conditions never encountered before where new diseases can develop.
  • Advanced technology, requiring technology to survive, such as 3D printers - if we have nanotech in a big way it will probably be used most in space.
  • Once you have millions in space, then it is as easy to get hold of a spaceship with capability far greater than any ICBM than it is to get hold of a plane on Earth now.
  • Your space habitats explode, if hit by a spaceship at full speed, never mind nuclear bombs.

It must count as one of the worse places for technology hazards to develop and in a future with fast spaceships, you'd have as much interaction between habitats and Earth as we have between countries so any diseases or technological issues that develop in space colonies would be transferred to Earth swiftly, and vice versa.


Let's look at this a bit closer.

So, first of all - it's important that you can access your backup when you need it. If someone offered to write important files to a DVD - or some longer lasting technology like the Voyager golden disk - and send it to the Moon - or even to interstellar space - what is the point in that? If anything goes wrong we have to send an expedition to recover the data before it is any use to us. That's not going to help if we have lost space technology.

This shows the Voyager golden disk. But it's not designed as a backup system for Earth.

We could send a similar disk off into interstellar space for the most long lasting backup - but how would that help if we lose technology, and somehow we need to recover the data?

So - then how does it make a difference if we have humans as part of the backup system, but send them to interplanetary space or interstellar space?

Only if somehow the disaster would kill all humans on the Earth, but as we'll see, none of the plausible disasters do this.

The first thing the colonists will do in the event of a disaster is to get back to Earth to help deal with the problems here.

But with this plan, they are on Mars, six months away - or on the Moon. And they might not even have the technology to get back - likely to be a long time before we can make rockets on Mars using native technology. They probably die pretty soon without our help from Earth. 

If they have some amazing technology to keep going on Mars with 3D printers, even print entire rockets almost automatically without big factories - well we will have that technology back on Earth as well. All bets are off, our civilization is totally transformed. I think best to plan for a situation like that if it actually happens, so let's forget that possibility for now.

Even if they can get back - why add all that extra obstacle and the dangers of the journey back to Earth, with no technological help from us - why wait for six months for them to get back, why not have them here in the first place?

That is of course - unless there is reason for supposing they can survive the disaster itself more easily on Mars than on the Earth. More on that in a minute.

I'm taking it as obvious that they would want to come back to Earth. Surely, it doesn't make any sense to try to set up home on Mars or the Moon, in a solar system with an uninhabited Earth?

Earth, Mars and the Moon.

If something did happen to make humans extinct on Earth, or nearly extinct, and you had anyone, anywhere in the solar system who survived the disaster - where do you think they would want to go to set up home and rebuild after the disaster? Earth. Mars or the Moon?

So where is the best place for the backup?

I've never heard of any plausible astronomical disaster that could make the Earth as uninhabitable as Mars. Even Mars as it might be after 1000 years of terraforming - still - taking the most optimistic projections - you might be able to grow trees there and humans might be able to go out of doors without a spacesuit using only an air breather like an aqualung. But - no disaster on Earth could make it as uninhabitable as that. 

And - that's an optimistic projection for something that, maybe, could be achieved after 1000 years. I think myself, that we shouldn't even consider a thousand year long project until we've had a few more centuries of technology - or information from ETs, or some amazing advance in  human knowledge. And if we manage to survive 1000 years long enough to terraform Mars, if it is possible - well we probably won't need a backup strategy any more, or will have come up with something far more fool proof than our Mars colony.

I suggest, that it's better to look at some of the plausible disasters that could happen - and what the best backup strategy would be. I.e. look at the problems first, and design the backup strategy proportional to the issues. In all the examples I can think of, then Earth is the obvious and best place for your backup.


So - one disaster that's often mentioned is the issue of an asteroid impact. And you often see this artist's impression, reproduced so many times in this context:

Artist's impression by Donald Davis, Artist's impression of a 1000km-diameter planetoid hitting a young Earth

The dinosaur ending impactor was so small you couldn't see it at all at this level of magnification (see Killer Asteroids for a picture of the Earth with a 10 km asteroid superimposed to scale)

If we had an impactor as big as this, a respectable size for a moon (far larger than Phobos or Deimos), 1000 kms across - that's large enough to make large areas of the Earth molten, perhaps boil the seas etc etc. It might indeed make humans extinct, or result in a devastated Earth where the few survivors can't do much.

But asteroids as large as that only hit the Earth in the first few hundred million years soon after the formation of the Moon in the late heavy bombardment, when the solar system was still settling down and full of half formed planets and embryo planets and assorted debris from their formation.

That's also when the large dark patches were created on the Moon. There is pretty close to zero chance of us getting hit by one of those now.

The biggest asteroid impact crater in the geological record, two billion years ago, is a 300 km diameter crater in South Africa. Vredefort crater which was probably created by an asteroid between 5 and 10 kms in diameter.

The impact that ended the Dinosaur era was 66 million years ago, and that's not that much smaller - a crater 180 km in diameter, the impactor was probably 5 to 8 km in diameter.

These would have a global impact for sure. They would create an "impact winter" for several years afterwards. They might cause a global firestorm for the first few hours after the impact. If they land in the sea, they would cause enormous tsunamis.

However - many humans, especially on the other side of the world, would survive, especially if given a bit of warning to shelter from the worst effects.

Then - all the way through the following "impact winter" Earth would remain by far the most habitable place in the solar system, a far better place to be than Mars.

Impacts of this size certainly are a possibility. But not as likely as you might guess.

The chance is about one in ten million, or 0.00001% (given that we have found 90% of them already - found all the NEOs larger than 10 km, and thought to be probably 10% of those impacts come from longer period comets).


Still, I agree, even a 0.00001% chance of a civilization destroying event is worth protecting against. 

I'm not convinced we need to devise a backup yet though, the Spaceguard approach seems more appropriate. Better to devote our efforts to detecting meteorites and to find ways to deflect them. 

The dinosaurs didn't have the technology that we have. 

With our technology - first - we are almost bound to see an incoming asteroid or comet this big, probably a couple of decades before.

So - if there is an imminent strike like this - and if we can't destroy or deflect it, then I'm sure that we'd devote much of the capabilities of our civilization, for those last twenty years, to preparing for it.

We'd evacuate the impact zone. Build shelters to shelter from the firestorm underground. If necessary supplied with oxygen and anything else needed during the firestorm.

We'd organize protection for important collections of seeds and plants and animals. And do what we can to create as many shelters of that kind as we can.

We wouldn't go extinct after such an impact, unlike the dinosaurs. 


An impactor of 20 meters diameter like the Russian meteorite can hit us without any warning if it comes in rapidly on a trajectory from the sun, so we can't spot it in our telescopes until it is too late.

 But a 10 km asteroid? I think that's rather unlikely. NASA have been tasked with finding 90% of all the Near Earth asteroids of 140 meters or larger in diameter by 2020. Whether they succeed in that, Pan-STARRS particularly has already found 90% of the asteroids above one km in diameter and is continuing to find them, one a month. It's expected to find 99% of the one km diameter asteroids by the 2020s. They believe they have already found all the NEOs of 10 km diameter or more.

We are talking here about an asteroid nearly as large as Phobos or Deimos.

Even if coming from the direction of the sun - well, we know there isn't any NEO that large, so it has to come from the outer solar system, now that we know there are no NEOs this large, we'd have spotted it on an incoming trajectory before it passed behind the sun, decades before. We'd see it as far away as Jupiter, easily.

And - if it is a comet - well first, they are thought to be rarer anyway, impacts by giant comets. Also, we would know about comets this size for about a couple of decades in advance.

So, it is very hard to see how it could happen and we get almost no warning. For a ten kilometer asteroid anyway.

But - just as a thought experiment - if there was no warning - still there would be some survivors. At the very least, anyone who is deep below the sea in a submarine would survive (a tsunami is a surface phenomenon, deep below the sea you won't notice it) - unless actually at the impact site or close to it. 

 If there are global firestorms, surely the bases in Antarctica would be spared (nothing to burn in the icy expanse). In the immediate aftermath there would be many other survivors, such as people in deep shelters on the other side of the Earth from the impact. 

So - I am not sure you really need a backup strategy at all. 

But if you do - well - you can do this easily on the Earth.

You need to have - maybe a couple of groups of people, either side of the world (so they can't be both hit by the same impact), and protected from a firestorm - and with whatever you judge is most useful for rebuilding society. Then that's a far better backup than a colony on Mars. 

It would also cost far less. The colonies don't need SpaceX launchers to get to their colonies. They don't need to build their houses able to hold in ten tons per square meter of outward pressure. They don't need to make oxygen from the desert sand or ice. They don't need to wear spacesuits every time they go outside their habitat. They don't need to use spaceships to get back to Earth after the disaster.

Apart from anything else - they can breath the air outside their settlement. That's such a huge plus. Why send them to Mars for your backup?


Large areas of the Earth remain uninhabited, such as deserts, large parts of Canada, Siberia etc. We certainly don't colonize everywhere, just the most easily inhabited places, or places where there are other resources or reasons to have people on site.

Or if what you need is remoteness and isolation - well there are many uninhabited islands in all the main oceans of the world, many of them previously inhabited but abandoned because of the isolation. Here on the West coast for instance, we have the Treshnish islands. A colony on any of these islands would of course be much easier to make self sufficient than a colony on Mars.

A few of the many Treshnish isles, some of them formerly inhabited but abandoned because they are too isolated for modern life.  Photo from the Geograph project.

Lunga, one of the Treshnish isles, formally inhabited, now populated only by puffins and other wildlife.

It would be far easier to set up a self sufficient colony designed to survive any disaster except a direct impact on the site by an asteroid, on one of the many remote islands in the world than on Mars. No access by road or ferry, only visited by occasional sailors and canonists.

Few people would choose to live there, but it is an easy place to live, and very easy of access, compared to Mars.

Indeed, if this idea of a backup was thought to be that important, you don't need to rely on high technology. Every country could have its own community of a few people living in isolation and with protection from firestorms and enough supplies to last out the immediate results of the most likely disasters.


Elon Musk is also concerned about some global pandemic killing us all. It is easy to devise an Earth based backup strategy there. 

The only advantage of Mars has is isolation - that it takes six months to get there so hopefully if the pandemic starts on Earth - then it has been detected before any humans manage to get from Earth to Mars. But as transportation to Mars gets easier, if we can get there in weeks or even days with fast transport in the future - then that advantage goes away - with a large colony on Mars and frequent travel, then the pandemic would spread there as well, if it can spread to every continent on the Earth.

Best way to deal with this, if you are really concerned about it, is to arrange to have some settlements (could be the same as before) with quarantine regulations. You can only go in or out if you stay in quarantine for, say, six months or whatever you think is necessary.

In any case surely you'd have some survivors from almost any pandemic.

Another thing that's mentioned -  a gamma ray burst, (very low probability but mentioning for completeness), then the Earth itself acts as a shield for the worst effects for anyone on the opposite side of the planet. 

As for black holes - if there was a black hole headed for our solar system - the sun is by far the biggest target. If Earth was in danger from a black hole, then we'd see stars blinking out in the sky. So, I don't think we need to worry about that possibility.


What about the problems of "gray goo" - self replicating von Neumann machines etc?

Well we are a bit away from this capability.

But if it ever was possible - again - then if the problem is so pervasive that no part of the Earth is spared, then why would Mars be spared? After all the nanobots have to get from one country to another - they would surely also get to Mars as well in the spaceships traveling back and forth.

For that matter, Mars would be more reliant on this kind of technology than Earth. So the problem could as easily start on Mars.

Similarly - any high technology issue could as easily start on Mars. Not sure that having extra settlements helps here.


In the long term, yes, the Earth will become uninhabitable, generally agreed. 

But - again - the popular idea of when this will happen is a bit different from the reality I think.

By long term, here we mean really long term. As in, several hundred million years - probably at least half a billion years into the future.

To put that into perspective - that's long enough for evolution to proceed all the way from the very first multi-cellular life in the fossil record to humans, a second time.

Whatever or whoever may be living on Earth then - I don't think we really need to work out their backup strategy quite yet. 


And indeed - the ideas for terraforming Mars would terraform the planet only for the near future. 

That's because of its low gravity, and its lack of a magnetic field, and lack of continental drift (which recycles carbonates  back to CO2).

Most of any atmosphere would get lost over longer geological timescales. Or it would dissolve into the oceans and form carbonates. By 500 million years from now, chances are that Mars is even more of a dead planet than it is now, with hardly any chance of terraforming it. 

Who knows, maybe it will be needed as a temporary second home from Earth in the far distant future. But if so, I think that the best thing we can do to help make sure it is available for a backup in the future is to leave well alone.

Unless we have some extraordinary advance in understanding and technology, the chances are that all we can do right now is to mess it up and make it less useful for whoever lives on Earth millions of years into the future.

This also applies to the likes of Stanford Torus colonies

Stanford Torus type colony - the original design was for 10,000 people - constructed at a cost of about 30 billion dollars in 1975 dollars (that's the maximum debt because it was going to pay for itself by making solar power satellites).

It's easier to build than you might think because all the cosmic radiation shielding comes from materials mined in space, they envisioned a small bulldozer on the Moon loading materials into a mass driver and sending it one small load at a time to the Stanford Torus colony. Lots of details to be worked out and information needed about what humans and plants need to survive (gravity, spin tolerance etc) and closed system ecosystems etc - but same applies for colonies on a planetary surface.

Main advantage - that it can be built anywhere, first ones close to Earth - and that you can customize it to any level of gravity you like, and any amount of light.

I think myself they are more likely to work as an almost self sufficient community than a colony on a planetary surface - and also - much more customizable - you can set them to any level of gravity you like and choose sunlight levels etc. See Asteroid Resources Could Create Space Habs For Trillions; Land Area Of A Thousand Earths

But still - given the choice of a Stanford Torus colony - and the Earth - as a place to live after a global disaster - you'd choose the Earth every time. You don't need cosmic radiation shielding, you don't need to build these big habitats - just find some remote island or something and everything is much easier here.

So they make sense as space settlements if there is some motivation for going into space, e.g. mining, or solar power etc. But I can't see them working as a "backup". You still have this issue, why have your backup colony in space, where construction is bound to be more difficult and expensive? And where you have the problem of returning to Earth, or Earth having to send missions to your settlement to get you back?


So - there may be many other reasons for space settlement. Scientific exploration, learning about the origins of life and nature and formation of the solar system. Mining for platinum, ice in space, and other resources. Constructing solar satellites and solar power stations on the Moon to beam power back to the Earth. 

For that matter they can also be involved in searching for asteroids likely to impact Earth - and diverting them - or even - mining them for materials (with a long enough lead time).

But let's devise our space exploration around those aims, rather than about this idea of a backup for the Earth.

And - let's use our space exploration and settlement to help fix the issues on the Earth, not attempt to find a way to get away from them!


I was motivated to write this after watching a BBC Horizon program about Mars. A lot of it was good. But it ended with this idea that it is inevitable that humans will go to Mars. And it had a clip from Stephen Hawking saying that we need to go into space to survive as a species.

I realized, this is one of those ideas that gets a lot of media coverage - but you seldom if ever get this idea criticized in the public media.

I think it is worth looking at it more closely. Would a backup in space really achieve anything at all? Would space colonization help at all with survival of humans on Earth?

If not, then this idea is diverting our attention away from the many other things we could be doing in space that have clear benefits to the Earth. 

And it could also conflict with other things - if your main aim is to create this "backup" then at some time down the line you might end up making decisions in favour of your "backup" that actually impact on other things that do benefit the Earth, or making decisions that conflict with the needs of scientific exploration. 

I've written several articles here about how human colonization of the surface of Mars could conflict with the needs to study a pristine Mars for exobiology.

That is - if it turns out that there are indeed micro-habitats for present day Earth life on the surface of Mars, as a fair number of scientists now think. How could you send humans to the surface without greatly increasing the risk of introducing Earth life in the event of a hard landing (never mind after you land there)? 

So that's one possible conflict, and there could well be others.

Anyway that's how I see, it, what do you think? Can you think of any situation where a backup in a space settlement would make sense?