Space Colonization

[Dominus Atheos]

Charles Stross

I write SF for a living. Possibly because of this, folks seem to think I ought to be an enthusiastic proponent of space exploration and space colonization. Space exploration? Yep, that's a fair cop — I'm all in favour of advancing the scientific enterprise. But actual space colonisation is another matter entirely, and those of a sensitive (or optimistic) disposition might want to stop reading right now ...

I'm going to take it as read that the idea of space colonization isn't unfamiliar; domed cities on Mars, orbiting cylindrical space habitats a la J. D. Bernal or Gerard K. O'Neill, that sort of thing. Generation ships that take hundreds of years to ferry colonists out to other star systems where — as we are now discovering — there are profusions of planets to explore.

And I don't want to spend much time talking about the unspoken ideological underpinnings of the urge to space colonization, other than to point out that they're there, that the case for space colonization isn't usually presented as an economic enterprise so much as a quasi-religious one. "We can't afford to keep all our eggs in one basket" isn't so much a justification as an appeal to sentimentality, for in the hypothetical case of a planet-trashing catastrophe, we (who currently inhabit the surface of the Earth) are dead anyway. The future extinction of the human species cannot affect you if you are already dead: strictly speaking, it should be of no personal concern.

Historically, crossing oceans and setting up farmsteads on new lands conveniently stripped of indigenous inhabitants by disease has been a cost-effective proposition. But the scale factor involved in space travel is strongly counter-intuitive.

Here's a handy metaphor: let's approximate one astronomical unit — the distance between the Earth and the sun, roughly 150 million kilometres, or 600 times the distance from the Earth to the Moon — to one centimetre. Got that? 1AU = 1cm. (You may want to get hold of a ruler to follow through with this one.)

The solar system is conveniently small. Neptune, the outermost planet in our solar system, orbits the sun at a distance of almost exactly 30AU, or 30 centimetres — one foot (in imperial units). Giant Jupiter is 5.46 AU out from the sun, almost exactly two inches (in old money).

We've sent space probes to Jupiter; they take two and a half years to get there if we send them on a straight Hohmann transfer orbit, but we can get there a bit faster using some fancy orbital mechanics. Neptune is still a stretch — only one spacecraft, Voyager 2, has made it out there so far. Its journey time was 12 years, and it wasn't stopping. (It's now on its way out into interstellar space, having passed the heliopause some years ago.)

The Kuiper belt, domain of icy wandering dwarf planets like Pluto and Eris, extends perhaps another 30AU, before merging into the much more tenuous Hills cloud and Oort cloud, domain of loosely coupled long-period comets.

Now for the first scale shock: using our handy metaphor the Kuiper belt is perhaps a metre in diameter. The Oort cloud, in contrast, is as much as 50,000 AU in radius — its outer edge lies half a kilometre away.

Got that? Our planetary solar system is 30 centimetres, roughly a foot, in radius. But to get to the edge of the Oort cloud, you have to go half a kilometre, roughly a third of a mile.

Next on our tour is Proxima Centauri, our nearest star. (There might be a brown dwarf or two lurking unseen in the icy depths beyond the Oort cloud, but if we've spotted one, I'm unaware of it.) Proxima Centauri is 4.22 light years away.A light year is 63.2 x 103 AU, or 9.46 x 1012 Km. So Proxima Centauri, at 267,000 AU, is just under two and a third kilometres, or two miles (in old money) away from us.

But Proxima Centauri is a poor choice, if we're looking for habitable real estate. While exoplanets are apparently common as muck, terrestrial planets are harder to find; Gliese 581c, the first such to be detected (and it looks like a pretty weird one, at that), is roughly 20.4 light years away, or using our metaphor, about ten miles.

Try to get a handle on this: it takes us 2-5 years to travel two inches. But the proponents of interstellar travel are talking about journeys of ten miles. That's the first point I want to get across: that if the distances involved in interplanetary travel are enormous, and the travel times fit to rival the first Australian settlers, then the distances and times involved in interstellar travel are mind-numbing.

This is not to say that interstellar travel is impossible; quite the contrary. But to do so effectively you need either (a) outrageous amounts of cheap energy, or (b) highly efficient robot probes, or (c) a magic wand. And in the absence of (c) you're not going to get any news back from the other end in less than decades. Even if (a) is achievable, or by means of (b) we can send self-replicating factories and have them turn distant solar systems into hives of industry, and more speculatively find some way to transmit human beings there, they are going to have zero net economic impact on our circumstances (except insofar as sending them out costs us money).

What do I mean by outrageous amounts of cheap energy?

Let's postulate that in the future, it will be possible to wave a magic wand and construct a camping kit that encapsulates all the necessary technologies and information to rebuild a human civilization capable of eventually sending out interstellar colonization missions — a bunch of self-replicating, self-repairing robotic hardware, and a downloadable copy of the sum total of human knowledge to date. Let's also be generous and throw in a closed-circuit life support system capable of keeping a human occupant alive indefinitely, for many years at a stretch, with zero failures and losses, and capable where necessary of providing medical intervention. Let's throw in a willing astronaut (the fool!) and stick them inside this assembly. It's going to be pretty boring in there, but I think we can conceive of our minimal manned interstellar mission as being about the size and mass of a Mercury capsule. And I'm going to nail a target to the barn door and call it 2000kg in total.

(Of course we can cut corners, but I've already invoked self-replicating robotic factories and closed-cycle life support systems, and those are close enough to magic wands as it is. I'm going to deliberately ignore more speculative technologies such as starwisps, mind transfer, or AIs sufficiently powerful to operate autonomously — although I used them shamelessly in my novel Accelerando. What I'm trying to do here is come up with a useful metaphor for the energy budget realistically required for interstellar flight.)

Incidentally, a probe massing 1-2 tons with an astronaut on top is a bit implausible, but a 1-2 ton probe could conceivably carry enough robotic instrumentation to do useful research, plus a laser powerful enough to punch a signal home, and maybe even that shrink-wrapped military/industrial complex in a tin can that would allow it to build something useful at the other end. Anything much smaller, though, isn't going to be able to transmit its findings to us — at least, not without some breakthroughs in communication technology that haven't shown up so far.

Now, let's say we want to deliver our canned monkey to Proxima Centauri within its own lifetime. We're sending them on a one-way trip, so a 42 year flight time isn't unreasonable. (Their job is to supervise the machinery as it unpacks itself and begins to brew up a bunch of new colonists using an artificial uterus. Okay?) This means they need to achieve a mean cruise speed of 10% of the speed of light. They then need to decelerate at the other end. At 10% of c relativistic effects are minor — there's going to be time dilation, but it'll be on the order of hours or days over the duration of the 42-year voyage. So we need to accelerate our astronaut to 30,000,000 metres per second, and decelerate them at the other end. Cheating and using Newton's laws of motion, the kinetic energy acquired by acceleration is 9 x 1017 Joules, so we can call it 2 x 1018 Joules in round numbers for the entire trip. NB: This assumes that the propulsion system in use is 100% efficient at converting energy into momentum, that there are no losses from friction with the interstellar medium, and that the propulsion source is external — that is, there's no need to take reaction mass along en route. So this is a lower bound on the energy cost of transporting our Mercury-capsule sized expedition to Proxima Centauri in less than a lifetime.

To put this figure in perspective, the total conversion of one kilogram of mass into energy yields 9 x 1016 Joules. (Which one of my sources informs me, is about equivalent to 21.6 megatons in thermonuclear explosive yield). So we require the equivalent energy output to 400 megatons of nuclear armageddon in order to move a capsule of about the gross weight of a fully loaded Volvo V70 automobile to Proxima Centauri in less than a human lifetime. That's the same as the yield of the entire US Minuteman III ICBM force.

For a less explosive reference point, our entire planetary economy runs on roughly 4 terawatts of electricity (4 x 1012 watts). So it would take our total planetary electricity production for a period of half a million seconds — roughly 5 days — to supply the necessary va-va-voom.

But to bring this back to earth with a bump, let me just remind you that this probe is so implausibly efficient that it's veering back into "magic wand" territory. I've tap-danced past a 100% efficient power transmission system capable of operating across interstellar distances with pinpoint precision and no conversion losses, and that allows the spacecraft on the receiving end to convert power directly into momentum. This is not exactly like any power transmission system that anyone's built to this date, and I'm not sure I can see where it's coming from.

Our one astronaut, 10% of c mission approximates well to an unmanned flight, but what about longer-term expeditions? Generation ships are a staple of SF; they're slow (probably under 1% of c) and they carry a self-sufficient city-state. The crew who set off won't live to see their destination (the flight time to Proxima Centauri at 1% of c is about 420 years), but the vague hope is that someone will. Leaving aside our lack of a proven track record at building social institutions that are stable across time periods greatly in excess of a human lifespan, using a generation ship probably doesn't do much for our energy budget problem either. A society of human beings are likely to need more space and raw material to do stuff with while in flight; sticking a solitary explorer in a tin can for forty-something years is merely cruel and unusual, but doing it to an entire city for several centuries probably qualifies as a crime against humanity. We therefore need to relax the mass constraint. Assuming the same super-efficient life support as our solitary explorer, we might postulate that each colonist requires ten tons of structural mass to move around in. (About the same as a large trailer home. For life.) We've cut the peak velocity by an order of magnitude, but we've increased the payload requirement by an order of magnitude per passenger — and we need enough passengers to make a stable society fly. I'd guess a sensible lower number would be on the order of 200 people, the size of a prehistoric primate troupe. (Genetic diversity? I'm going to assume we can hand-wave around that by packing some deep-frozen sperm and ova, or frozen embryos, for later reuse.) By the time we work up to a minimal generation ship (and how minimal can we get, confining 200 human beings in an object weighing aout 2000 tons, for roughly the same period of time that has elapsed since the Plymouth colony landed in what was later to become Massachusetts?) we're actually requiring much more energy than our solitary high-speed explorer.

And remember, this is only what it takes to go to Proxima Centauri our nearest neighbour. Gliese 581c is five times as far away. Planets that are already habitable insofar as they orbit inside the habitable zone of their star, possess free oxygen in their atmosphere, and have a mass, surface gravity and escape velocity that are not too forbidding, are likely to be somewhat rarer. (And if there is free oxygen in the atmosphere on a planet, that implies something else — the presence of pre-existing photosynthetic life, a carbon cycle, and a bunch of other stuff that could well unleash a big can of whoop-ass on an unprimed human immune system. The question of how we might interact with alien biologies is an order of magnitude bigger and more complex than the question of how we might get there — and the preliminary outlook is rather forbidding.)

The long and the short of what I'm trying to get across is quite simply that, in the absence of technology indistinguishable from magic — magic tech that, furthermore, does things that from today's perspective appear to play fast and loose with the laws of physics — interstellar travel for human beings is near-as-dammit a non-starter. And while I won't rule out the possibility of such seemingly-magical technology appearing at some time in the future, the conclusion I draw as a science fiction writer is that if interstellar colonization ever happens, it will not follow the pattern of historical colonization drives that are followed by mass emigration and trade between the colonies and the old home soil.

What about our own solar system?

After contemplating the vastness of interstellar space, our own solar system looks almost comfortingly accessible at first. Exploring our own solar system is a no-brainer: we can do it, we are doing it, and interplanetary exploration is probably going to be seen as one of the great scientific undertakings of the late 20th and early 21st century, when the history books get written.

But when we start examining the prospects for interplanetary colonization things turn gloomy again.

Bluntly, we're not going to get there by rocket ship.

Optimistic projects suggest that it should be possible, with the low cost rockets currently under development, to maintain a Lunar presence for a transportation cost of roughly $15,000 per kilogram. Some extreme projections suggest that if the cost can be cut to roughly triple the cost of fuel and oxidizer (meaning, the spacecraft concerned will be both largely reusable and very cheap) then we might even get as low as $165/kilogram to the lunar surface. At that price, sending a 100Kg astronaut to Moon Base One looks as if it ought to cost not much more than a first-class return air fare from the UK to New Zealand ... except that such a price estimate is hogwash. We primates have certain failure modes, and one of them that must not be underestimated is our tendency to irreversibly malfunction when exposed to climactic extremes of temperature, pressure, and partial pressure of oxygen. While the amount of oxygen, water, and food a human consumes per day doesn't sound all that serious — it probably totals roughly ten kilograms, if you economize and recycle the washing-up water — the amount of parasitic weight you need to keep the monkey from blowing out is measured in tons. A Russian Orlan-M space suit (which, some would say, is better than anything NASA has come up with over the years — take heed of the pre-breathe time requirements!) weighs 112 kilograms, which pretty much puts a floor on our infrastructure requirements. An actual habitat would need to mass a whole lot more. Even at $165/kilogram, that's going to add up to a very hefty excess baggage charge on that notional first class air fare to New Zealand — and I think the $165/kg figure is in any case highly unrealistic; even the authors of the article I cited thought $2000/kg was a bit more reasonable.

Whichever way you cut it, sending a single tourist to the moon is going to cost not less than $50,000 — and a more realistic figure, for a mature reusable, cheap, rocket-based lunar transport cycle is more like $1M. And that's before you factor in the price of bringing them back ...

The moon is about 1.3 light seconds away. If we want to go panning the (metaphorical) rivers for gold, we'd do better to send teleoperator-controlled robots; it's close enough that we can control them directly, and far enough away that the cost of transporting food and creature comforts for human explorers is astronomical. There probably are niches for human workers on a moon base, but only until our robot technologies are somewhat more mature than they are today; Mission Control would be a lot happier with a pair of hands and a high-def camera that doesn't talk back and doesn't need to go to the toilet or take naps.

When we look at the rest of the solar system, the picture is even bleaker. Mars is ... well, the phrase "tourist resort" springs to mind, and is promptly filed in the same corner as "Gobi desert". As Bruce Sterling has puts it: "I'll believe in people settling Mars at about the same time I see people settling the Gobi Desert. The Gobi Desert is about a thousand times as hospitable as Mars and five hundred times cheaper and easier to reach. Nobody ever writes "Gobi Desert Opera" because, well, it's just kind of plonkingly obvious that there's no good reason to go there and live. It's ugly, it's inhospitable and there's no way to make it pay. Mars is just the same, really. We just romanticize it because it's so hard to reach." In other words, going there to explore is fine and dandy — our robots are all over it already. But as a desirable residential neighbourhood it has some shortcomings, starting with the slight lack of breathable air and the sub-Antarctic nighttime temperatures and the Mach 0.5 dust storms, and working down from there.

Actually, there probably is a good reason for sending human explorers to Mars. And that's the distance: at up to 30 minutes, the speed of light delay means that remote control of robots on the Martian surface is extremely tedious. Either we need autonomous roots that can be assigned tasks and carry them out without direct human supervision, or we need astronauts in orbit or on the ground to boss the robot work gangs around.

On the other hand, Mars is a good way further away than the moon, and has a deeper gravity well. All of which drive up the cost per kilogram delivered to the Martian surface. Maybe FedEx could cut it as low as $20,000 per kilogram, but I'm not holding my breath.

Let me repeat myself: we are not going there with rockets. At least, not the conventional kind — and while there may be a role for nuclear propulsion in deep space, in general there's a trade-off between instantaneous thrust and efficiency; the more efficient your motor, the lower the actual thrust it provides. Some technologies such as the variable specific impulse magnetoplasma rocket show a good degree of flexibility, but in general they're not suitable for getting us from Earth's surface into orbit — they're only useful for trucking things around from low earth orbit on out.

Again, as with interstellar colonization, there are other options. Space elevators, if we build them, will invalidate a lot of what I just said. Some analyses of the energy costs of space elevators suggest that a marginal cost of $350/kilogram to geosynchronous orbit should be achievable without waving any magic wands (other than the enormous practical materials and structural engineering problems of building the thing in the first place). So we probably can look forward to zero-gee vacations in orbit, at a price. And space elevators are attractive because they're a scalable technology; you can use one to haul into space the material to build more. So, long term, space elevators may give us not-unreasonably priced access to space, including jaunts to the lunar surface for a price equivalent to less than $100,000 in today's money. At which point, settlement would begin to look economically feasible, except ...

We're human beings. We evolved to flourish in a very specific environment that covers perhaps 10% of our home planet's surface area. (Earth is 70% ocean, and while we can survive, with assistance, in extremely inhospitable terrain, be it arctic or desert or mountain, we aren't well-adapted to thriving there.) Space itself is a very poor environment for humans to live in. A simple pressure failure can kill a spaceship crew in minutes. And that's not the only threat. Cosmic radiation poses a serious risk to long duration interplanetary missions, and unlike solar radiation and radiation from coronal mass ejections the energies of the particles responsible make shielding astronauts extremely difficult. And finally, there's the travel time. Two and a half years to Jupiter system; six months to Mars.

Now, these problems are subject to a variety of approaches — including medical ones: does it matter if cosmic radiation causes long-term cumulative radiation exposure leading to cancers if we have advanced side-effect-free cancer treatments? Better still, if hydrogen sulphide-induced hibernation turns out to be a practical technique in human beings, we may be able to sleep through the trip. But even so, when you get down to it, there's not really any economically viable activity on the horizon for people to engage in that would require them to settle on a planet or asteroid and live there for the rest of their lives. In general, when we need to extract resources from a hostile environment we tend to build infrastructure to exploit them (such as oil platforms) but we don't exactly scurry to move our families there. Rather, crews go out to work a long shift, then return home to take their leave. After all, there's no there there — just a howling wilderness of north Atlantic gales and frigid water that will kill you within five minutes of exposure. And that, I submit, is the closest metaphor we'll find for interplanetary colonization. Most of the heavy lifting more than a million kilometres from Earth will be done by robots, overseen by human supervisors who will be itching to get home and spend their hardship pay. And closer to home, the commercialization of space will be incremental and slow, driven by our increasing dependence on near-earth space for communications, positioning, weather forecasting, and (still in its embryonic stages) tourism. But the domed city on Mars is going to have to wait for a magic wand or two to do something about the climate, or reinvent a kind of human being who can thrive in an airless, inhospitable environment.

Colonize the Gobi desert, colonise the North Atlantic in winter — then get back to me about the rest of the solar system!

[GrandMasterTerwynn]

Well, duh. Space travel is enormously expensive. And save access to stupidly cheap energy and, possibly a bending of the laws of physics, interstellar colonization will be beyond our reach for millenia, if ever.

I'm not sure I agree with his assessment of developing the solar system, though. Ultimately, to colonize that, you probably don't really want to drop down into the planetary gravity wells. You could bury colonies under enough rock and dust to protect humans from most cosmic radiation, if you burrowed into asteroids. Or you could convert asteroid materials into giant habitats. You could spin them up, and they'd feel just like Earth, though those who live on them will have to get used to a tangible Coriolis force, and the odd horizon, but whatever. And yes, you'd probably want to develop space elevators so you can cheaply get people up to the point where you'd send them off to these space habitats. So really, colonizing the solar system isn't quite as hard as he makes it sound. Yes, it's going to be a real bear to do, and it won't be cheap, but you won't have to get as close to the Clarke-tech line.

[Ariphaos]

Yes, because technology never advances, we'll never discover fusion, to say, power maglev-launch something 30 km up, or perhaps a space elevator, or perhaps...

Also, wtf are we bringing air to the moon base for? It's covered in titanium dioxide. Fusion + extraction process = give the entire Moon a breathable oxygen atmosphere if you want. Bonus points if the poles really do have ice, but Stross's argument already begins to crumble. Once the initial hurdle is passed, guess what? We don't have to build stuff on Earth! It can be made in space, too!

And yes, outrageous amounts of cheap energy. Fusion aside, build a dyson swarm around the Sun, a few AU-long coilgun arrays, and you can launch on the order of a million tonnes a second at 86% of c to nearby stars. You don't carry the fuel to get you up to that speed! You've got a gigantic fusion engine behind you, use it!

The big problem though, is with his dollar figures. The sorts of energy costs he talks about are going down, not up, with the possible exception of methane if it turns out the various massive potential reserves are not accessible. Eventually, I imagine the world will move to an energy-backed, limited-scarcity economy anyway.

People do not do things solely out of economic practicality, we do not have an economy for the sake of having an economy. We have an economy so we can do cool and entertaining things, like build rockets, send them places, and bring them back. We will terraform Mars not because it is economical, but because it is far, far cooler than minting another trillion pop singer CDs.

[metavac]

First of all, nothing would be cooler than Britney Spears going a million times platinum (if only because Primo would have to produce the whole damn album). Second, nearly three quarters of the cost of getting to anywhere interesting in the Earth Lunar system is getting to low earth orbit through 150 miles of atmosphere. Third, this is the longest justification for three generations of science fiction writers almost completely ignoring that entire period between now and economical space colonization I've ever seen. Maybe they're finally getting sick and tired of getting showed up by Tom Hanks.

[Surlethe]

He also needs to realize that, in the long, long run, the amount of energy available to be consumed on the Earth is ultimately determined by a linear function, while the amount of energy humans will consume is geometric and will increase only until it is constrained by the available energy (we're about to see this on a minor scale with peak oil, by the way). Eventually, if we don't make use of the non-renewable energy available to us to launch ourselves into space and start harvesting resources from the solar system, our daily energy consumption will be limited to the amount of energy the Sun gives us.

So I don't know about colonization -- there's still plenty of space on Earth; we could fit everyone on the planet into Texas at a comfortable suburban population density -- but we sure as hell will need to get off the planet to continue growing and avoid a Malthusian population struggle.