Switch branches/tags
Nothing to show
Find file
Fetching contributors…
Cannot retrieve contributors at this time
175 lines (110 sloc) 28.4 KB
description: MAD will not work in outer space; pre-emptive strikes are nigh-guaranteed.
[Orson Scott Card](!Wikipedia)'s _[Ender's Game](!Wikipedia)_ is famous, and its ending well-known: the protagonist [Ender Wiggin](!Wikipedia) is tricked into commanding a pre-emptive invasion fleet which destroys all the [enemy alien](!Wikipedia "Formics") fleets, their homeplanet, and by a quirk of their biology, the entire alien race. The reasoning given is that they couldn't risk waiting for a (third) invasion fleet.
# First-strike
This is [Cold War](!Wikipedia) reasoning, of course; version 1.0. The strategy is similar to early Cold War views, such as those who advocated that the USA strike Russia first, before it had developed A-bombs/lots-of-bombs/H-bombs.
In retrospect, such advocacy seems foolish to us who survived the Cold War; we are not quite so impressed by the argument that Communism is inimitable to the [American Way of Life](!Wikipedia) and that it is bound by its internal logic to subjugate the world and war on us. It did instigate various wars, prop us various regimes, and so on; but the USA has done quite as much of that in its history (and often for reasons only ostensibly related to ideals such as countering communism). Nor are we impressed by the arguments that such and such a achievement must be prevented by all-out war; the world did not end because Russia developed an atomic bomb, nor when it had constructed useful quantities of them, nor when it cracked the H-bomb.
## Why strike first?
The crucial characteristics of the early Cold War, that made a first strike so compelling, is that:
1. attacks can be launched simultaneously on all fronts from multiple locations
2. each attack is utterly devastating
3. the attacks cannot be guarded against^[One thinks of the quip that [NORAD](!Wikipedia)'s [Cheyenne Mountain](!Wikipedia) base was obsolete the day it was built.]
4. the attacks are not predictable or observable sufficiently in advance to take action like retaliation
If a fleet of [ballistic missile submarine](!Wikipedia)s have anchored off one's capital, the attack can begin and end within half an hour, and the missiles - which boomers carry in substantial quantities - can target anywhere in the nation. There is no practical [missile defense](!Wikipedia) more than half a century after nuclear warfare became possible, and little prospect of one that could withstand any more than a few missiles. Surviving nuclear explosions is an extremely difficult task, and cannot be guaranteed. A government attacked would crumble instantly, yielding immediate victory to the foe. Such [decapitation attack](!Wikipedia)s are strategically attractive.
In each respect, nuclear warfare differs from conventional warfare.
1. It is difficult in conventional warfare to have troops on multiple frontiers prepared for an invasion. Planes do not much affect the linear movements of mechanized troops, as they must provide air cover and cannot range freely over enemy territory (due to anti-aircraft defenses and the enemy air force); even were total air dominance achieved, the Air Force could not parachute large formations and their equipment wherever they pleased.
2. Conventional wars using small arms and low-yield explosives rarely utterly destroy a city or opposing forces; decisive clashes are rarities, and not nearly as decisive as a mushroom cloud.
3. Fortifications and normal military forces can put up a stalwart resistance to an invasion. In some periods, defense was even stronger than the offense - for example, in World War I. With nuclear bombs, the only defense is to dig deep into the earth as possible, and then even deeper than that. From the defensive perspective, this is inferior in every way to the balance of power in WWI.
4. Mass mobilizations can be observed for weeks or months in advance; with modern telecommunications, the message of invasion can work its way up the chain of command in mere hours. (This is quite fast compared to historical empires, which might not know that they've been invaded for weeks or months, but still not fast enough to cope with [ICBM](!Wikipedia)s, much like [SLBM](!Wikipedia)s.)
A conventional war is intrinsically *limited*. A country can be invaded and lose territory without too much concern. Even the tiniest countries like Israel have space to fall back and regroup. The action happens on a human time-scale, with human levels of casualties. There can be ebb and flow; strikes can be probing, small-scale.
If one wanted an analogy, conventional warfare is like martial arts sparring: multiple moves, each strike hurting but not fatal, with a back-and-forth, and winding up after a while; while nuclear warfare is like dueling with [Desert Eagle](!Wikipedia)s. And a curtain between the duelists.
# Fun space warfare
Science fiction often deals with space warfare, but usually in an extremely unscientific or romanticized way.
Even when the author throws many details into it, with hard numbers and equations and rules he must follow, the overall picture is sometimes still absurd. ([Robin Hanson](!Wikipedia) warns us to be wary of being seduced by Near/Far thinking: it is not true that the more detail a projection or fiction has, the more true or likely it is![^hanson])
Consider the [Honorverse](!Wikipedia) of [David Weber](!Wikipedia). His ships are described in loving detail; one can't go a battle without figures like '10,000 KPS' being thrown out; the history is described in tremendous detail in appendixes and secondary works; the planets number in the hundreds, and the complexities of their interactions feel authentic. Plot events often hinge on accelerations and vectors; books are powered by technical innovations. It feels awfully realistic.
But is it really? The Honorverse is a naked retelling of the old naval genre like [Horatio Hornblower](!Wikipedia), right down to the names of the wicked French antagonists (eg. [Oscar Saint-Just](!Wikipedia)). The ship designs and the universe is constructed to make '[crossing the T](!Wikipedia)', ambushes, etc. plausible in outer space.
It may be fun to read about Warshawski sails or alpha and beta nodes, or wedges and sidewalls; but these are all _ad hoc_ bits of technology and arbitrary, most obviously in the case of the wedges^[After all, how plausible is it, really, that the wedge would provide perfect protection in 2 directions but not the other 4, and that no better substitute than sidewalls would have been devised?]. The mass of details may fool us, but they ought to make us suspicious - in the real world, the principles are simple & elegant and the applications complex; both principles and applications shouldn't be complex and detailed. We enjoy _Honor Harrington_ only by suspending a great deal of disbelief.
# Grim space warfare
After all, the humdrum alternative is something like [Larry Niven](!Wikipedia)'s [Known Space](!Wikipedia) universe, where travel between the stars takes years or millennia^[In many of his better short stories, at least; the novels are usually far enough in the future that multiple FTL drives are available.], where [solar sail](!Wikipedia)s and [Bussard ramjet](!Wikipedia)s are the best (and also nonfictional) methods of travel, where [time dilation](!Wikipedia) works its terrible magic, where trips are one-way and solar systems do not see incoming vessels (traveling at fractional _[c](!Wikipedia "Speed of light")_) until it is too late.
A more fact-based view of space warfare would look something like this analysis by the character Bean from Card's _[Ender's Shadow](!Wikipedia)_ (a companion novel to _Ender's Game_), Chapter 12:
> "Let me read you something," said Dimak. "'There are no fortifications, no magazines, no strong points ... In the enemy solar system, there can be no living off the land, since access to habitable planets will be possible only after complete victory ... Supply lines are not a problem, since there are none to protect, but the cost of that is that all supplies and ordnance must be carried with the invading fleet ... In effect, all interstellar invasion fleets are suicide attacks, because time dilation means that even if a fleet returns intact, almost no one they knew will still be alive. They can never return, and so must be sure that their force is sufficient to be decisive and therefore is worth the sacrifice.... Mixed-sex forces allow the possibility of the army becoming a permanent colony and/or occupying force on the captured enemy planet."
This is a summary of a lengthier analysis in Chapter 8:
> "Well of course fortifications are impossible in space," said Bean. "In the traditional sense, that is. But there are things you can do. Like his mini-fortresses, where you leave a sallying force outside the main fortification. You can station squads of ships to intercept raiders. And there are barriers you can put up. Mines. Fields of flotsam to cause collisions with fast-moving ships, holing them. That sort of thing."
> Dimak nodded, but said nothing.
> Bean was beginning to warm to the discussion. "The real problem is that unlike [Vauban](!Wikipedia), we have only one strong point worth defending -- Earth. And the enemy is not limited to a primary direction of approach. He could come from anywhere. From anywhere all at once. So we run into the classic problem of defense, cubed. The farther out you deploy your defenses, the more of them you have to have, and if your resources are limited, you soon have more fortifications than you can man. What good are bases on moons on Jupiter or Saturn or Neptune, when the enemy doesn't even have to come in on the [plane of the ecliptic](!Wikipedia)? He can bypass all our fortifications. The way [Nimitz](!Wikipedia "Chester W. Nimitz") and [MacArthur](!Wikipedia "Douglas MacArthur") used two-dimensional [island-hopping](!Wikipedia "Leapfrogging (strategy)") against the [defense in depth](!Wikipedia "Defence in depth") of the Japanese in World War II. Only our enemy can work in three dimensions. Therefore we cannot possibly maintain defense in depth. Our only defense is early detection and a single massed force."
> Dimak nodded slowly. His face showed no expression. "Go on."
> Go on? That wasn't enough to explain two hours of reading? "Well, so I thought that even that was a recipe for disaster, because the enemy is free to divide his forces. So even if we intercept and defeat ninety-nine of a hundred attacking squadrons, he only has to get one squadron through to cause terrible devastation on Earth. We saw how much territory a single ship could scour when they first showed up and started burning over China. Get ten ships to Earth for a single day -- and if they spread us out enough, they'd have a lot more than a day! -- and they could wipe out most of our main population centers. All our eggs are in that one basket."
> "And all this you got from Vauban," said Dimak.
> Finally. That was apparently enough to satisfy him. "From thinking about Vauban, and how much harder our defensive problem is."
> "So," said Dimak, "what's your solution?"
> Solution? What did Dimak think Bean was? I'm thinking about how to get control of the situation here in [Battle School](!Wikipedia), not how to save the world! "I don't think there is a solution," said Bean, buying time again. But then, having said it, he began to believe it. "There's no point in trying to defend Earth at all. In fact, unless they have some defensive device we don't know about, like some way of putting an invisible shield around a planet or something, the enemy is just as vulnerable. So the only strategy that makes any sense at all is an all-out attack. To send our fleet against their home world and destroy it."
> "What if our fleets pass in the night?" asked Dimak. "We destroy each other's worlds and all we have left are ships?"
> "No," said Bean, his mind racing. "Not if we sent out a fleet immediately after the Second Bugger War. After [Mazer Rackham](!Wikipedia)'s strike force defeated them, it would take time for word of their defeat to come back to them. So we build a fleet as quickly as possible and launch it against their home world immediately. That way the news of their defeat reaches them at the same time as our devastating counterattack."
# Easy warfare
Card's analysis, while good as far as it goes, doesn't go far enough. The situation is actually more unbalanced in favor of the attacker. Card seems to assume that combat will be conducted with ships, and that these ships while appropriately devastating are nevertheless rather short-ranged and must get close to a planet to attack.
But ships on a scale comparable to existing naval ships or spaceships have tremendous problems in plausible space warfare. (Space fighters analogous to airplanes have even [more problems]( They can be [seen coming]( from very far away (already private amateurs [routinely spot]( spy satellites; academic telescopes can spot something like the Space Shuttle doing a little bit of maneuvering as far away as the asteroid belt, and spot its launch from past Pluto's orbit - and the telescopes will only get better[^Kuiper]); they can't carry very much shielding^[Shielding against *what*? Slugs of metal? Every possible frequency of laser? Beams of plasma or charged particles? Nukes and their x-rays?] and metal shielding can be an outright liability as far as defense goes^[An enemy's laser can melt through metal like anything else, and if the laser is powerful, the metal might just evaporate/*explode* - which isn't very helpful at all.]; being in a vacuum, they have great difficulty [dumping heat]( generated by lasers; and require implausibly efficient engines just to carry any weapons at all! Nor do the SF stories we all think of when we think about space combat do justice to just [how effective]( planetary defenses can be against any sort of space fleet. (The fleet can't hide, can't see its targets very well, can't carry an occupation force, and the planet can build much bigger and longer-ranged weapons.)
[^Kuiper]: Another comparison I've seen; from ["City Lights Could Reveal Civilisations On Other Planets"]( and ["Detection Technique for Artificially-Illuminated Objects in the Outer Solar System and Beyond"]( respectively:
> "There is another search that could be done closer to home. With the help of some back-of-the-envelope calculations, Loeb and Turner say that today's best telescopes ought to be able to see the light generated by a Tokyo-sized metropolis at a distance of about 50 AU [Pluto is 30-49 AU], that's roughly the distance to the [Kuiper belt](!Wikipedia)."
> "This implies that an f⊕-illuminated surface would provide the same observed flux _F_ as a sunlight-illuminated object at that distance, if it is ∼√3.6×10^2^ = 19 times smaller in size. In other words, an _f_⊕-illuminated surface of size 53 km (comparable to the scale of a major city) would appear as bright as a 10^3^ km object which reflects sunlight with A = 7%. Since ∼10^3^ km objects were already found at distances beyond ∼50 AU, *we conclude that existing telescopes and surveys could detect the artificial light from a reasonably brightly illuminated region, roughly the size of a terrestrial city, located on a KBO* [Kuiper Belt Object]."
What would one do if one wanted to destroy an opposing world's civilian population, on the cheap, and also with great secrecy? One could send a fleet of powered warships straight in to smash the other fleet and then scour the world. If one can even do that, a premise I hope I cast a great deal of doubt upon with the preceding paragraph.
But there are many better alternatives. Geology teaches us that the 2 most devastating & life-destroying events are [supervolcano](!Wikipedia) eruptions, and [asteroid impacts](!Wikipedia "Impact event"). Supervolcanos are rather hard to trigger, but asteroids? (And mass and velocity are interchangeable; an asteroid is heavy and slow, a [relativistic kill vehicle](!Wikipedia) light and fast. The power demands are extreme in either case, ruling out known feasible designs - we [can't build]( a [Project Orion](!Wikipedia) rocket which would kill a planet.)
If one is a interstellar power, asteroids are quite easy to come by. One could stealthily creep into the vast [Kuiper Belt](!Wikipedia) or [Oort Cloud](!Wikipedia) and send rocks flying into the inner system. The Kuiper and Oort are composed largely of lightweight bodies (composed of things like water or methane), but they are pretty sizable objects - [Halley's Comet](!Wikipedia) is a typical Kuiper Belt inhabitant, and masses 2 to $3 \times 10^{14}$ kg. That could still hurt, to put it lightly. Neptune's [Triton](!Wikipedia "Triton (moon)") moon, which weighs a meaty $2.14 \times 10^{22}$ kg, is a former Kuiper Belt object.
And such an attack would be impossible to prevent. The Kuiper Belt alone is a shell around the Solar System from 30 to 55 AU; or an area of ~1.95^30^ kilometers[^volume]. This is a large volume of space to patrol. More feasible would be monitoring each and every body: after all, there are only ~70,000 bodies. Over 100 kilometers in diameter. Estimated.
And let's not even talk about the Oort Cloud (5-50,000 AU)!
If that's not bad enough, one could envision flinging appropriate small moons or large asteroids from *other* solar systems. Why not? Boosting a large body to fractional _c_ velocities and aiming it at a far away planet isn't inherently any more absurd than building a ship and boosting it to fractional _c_. The motion of stellar bodies is famously predictable out for many centuries. The body could guide itself: add some small motors, and it could even correct for small errors in prediction. It would be the ultimate fire-and-forget weapon of mass destruction.
## Defense
Stop it? How could one stop a small moon? If it is traveling at a good fraction of lightspeed, then its kinetic energy will have reached gargantuan proportions: no frantic last minute attacks will alter its path very much. Suppose one *did* shatter it with explosives; conservation of energy and mass will still apply - if a billion tons of ice are flying towards New York at 0.5 _c_, and a bomb shatters it, what's flying towards New York at 0.5 _c_? A billion tons of ice.
And remember the early points about the vastness of space and how enemies can attack from any degree - not just along the ecliptic. (Attacking from a Kuiper Belt, traveling through or just above or below the ecliptic may well be the most efficient path. But from another solar system located at some odd angle?) For every measure the defenders take, like extensive telescope arrays looking for fast-moving bodies, there are countermeasures like reducing [albedo](!Wikipedia) to pitch-blackness by applying soot.
## Counter-point: nowhere to hide
There is one advantage a defender may or may not have in space warfare. We already discussed, [in passing](#easy-warfare), some of the disadvantages of attacking spaceships, which must obviously be *advantages* for defenders:
1. If attackers can be seen coming from very far away, then that implies defenders can *attack* them since anything you can see, you can target with lasers
2. If attackers have difficulty carrying shielding because they are attacking, defenders can have arbitrary amounts of armor if they are, say, on a planet and can dig hundreds of meters deep underground
3. If attackers have difficulty dumping heat generated by lasers into empty space, then defenders have heat sinks handy (eg. oceans)
4. If useful attackers are difficult to engineer, then any breakthroughs for them may just increase the advantage for defenders who have fewer tradeoffs to deal with
The linked Project Rho pages investigated these factors in more numerical detail and, I think, demonstrated convincingly that a network of surveillance telescopes linked to a massive laser is superior to any even half-plausible attacking space fleet. But do those considerations apply to a large guided asteroid or other such body?
Laser ablation *is* one of the standard proposed [asteroid defenses](!Wikipedia "Asteroid-impact avoidance"), but that and other [strategies](!Wikipedia "Planetary defense#Collision avoidance strategies") all seem to fail if there is any active intelligent resistance. It does no good to shoot some rock from the Moon at an incoming asteroid if there is a laser on the asteroid ablating the Moon rock itself!
Which brings up the obvious point: the enumerated advantages for a planetary defender also apply to an asteroid! A asteroid laser can be dug into the asteroid to some degree, can use the asteroid for cooling to some degree, can see the planetary defenders (and their projectiles or laser defenses) from far away to some degree, etc. The asteroid defenses would be limited in armor or size by the smaller size of the asteroid compared to the planet, yes, but the defenses can be on the losing side of each exchange as long as the asteroid survives to impact! The planetary defenders need to win and neutralize the asteroid entirely, while the asteroid defenses just need to delay the war of attrition sufficiently.
Which has the *relative* advantage? This is not clear. It may be that the planetary defenders still have an advantage and all attackers, both spaceships and asteroids/comets, are doomed. Perhaps they are doomed but one can overcome local defenses by going past asteroids to small *moons*, but moons are still ineffective because to accelerate a moon is so energetically difficult that it can be observed from the target system and appropriate defenses set in motion to neutralize the moon; and knowing this, no one will bother with moon-strikes. I am not optimistic about either scenario, so let us continue to assume that things will not work out as one would hope (in favor of defenders).
# Nuclear & space first-strikes
To return to our 4-point nuclear warfare list.
1. Space is vast & 3D^[In the same way conventional warfare is 2D; nuclear warfare is modestly 3D.]. Attacks can be begun from any axis. There are no space equivalents of mountain passes or valleys.
So this point holds: nuclear warfare can attack from anywhere on the globe to anywhere on the globe, and space warfare is even more free in its approach.
2. Each attack could be even more devastating than nuclear attacks[^gigaton]. Even small comets or asteroids exceed normal atomic bombs, and it's about as easy to go for large bodies as small.
3. Nuclear attacks have no defenses besides being buried deep. They cannot be shot down or intercepted.
This is true of moons or asteroids as well. Remember the ultimate scenario: a moon at near-lightspeed. Such an impact is best described as 'world-shattering'. There's an old military joke: the best way to survive a nuclear attack is to not be there.
4. A SLBM attack is first observable when the missiles break the water's surface. An asteroid gone ballistic is rather difficult to spot at any point on its journey. In the nightmare scenario, even if the moon is observed immediately & reported on, the alert will reach its destination only shortly before the moon does!
In every point, space warfare is even more favorable to the attacker. There is no percentage in trying to defend.
# Second-strike
Of course, everyone knows that the first strike doctrines eventually yielded to '[second strike](!Wikipedia)' doctrines: [Mutually Assured Destruction](!Wikipedia). If point #2 is weakened to being devastating to only the nation and conventional military, sparing the nuclear elements, then retaliation is possible. There is no point in striking first if it will only provoke a retaliatory strike that does you as much damage as their first strike would have.
Can we hope for MAD? Perhaps. It's clear enough that there are all sorts of techniques: each star system could have a fleet quiescent in deep space listening to omnidirectional broadcasts. The heartbeat stops, the fleet sets off for revenge. (They would be space submarines in effect; this would also be an example of "[fail deadly](!Wikipedia)".) Or a system could be constantly accelerating bodies at a rival system, but off-course; then on an attack, the bodies could be re-aimed at their target. And so on.
## Accountability
There is one problem with MAD as applied to space. Nuclear attacks in the real world inherently involve accountability. If an ICBM lands in New York City, the satellites tracking it from launch will tell us exactly what country it came from. If a nuke is quietly shipped into San Francisco & explodes, the fallout will be immediately sampled and traced back to its manufacturing country. Within days or weeks, the USA and the world will know if it was an American nuke, or Russian, or North Korean etc.
The [isotope ratios](!Wikipedia "Isotopic signature") vary subtly, and the USA has for decades collected fallout from tests by the nuclear powers to determine the characteristic isotope ratios of those nations' uranium and plutonium. Fundamentally, there's no way to nuke a nation without being blamed for it. Even if Pakistan gave Al-Qaeda a bomb and they used it in a way that was not traceable to Pakistan, the isotopes would still finger Pakistan as the culprit and fire would rain down upon their head by the aggrieved party. Every nuclear power knows this: they will be held responsible for any use of their weapons, authorized or unauthorized - period.
## Deceiving MAD
The MAD equilibrium depends on always having someone to blame and attack. But this is not necessarily true of ballistic warfare. Suppose the attacked solar system is attacked from its own Kuiper Belt? What other system does it blame? Short of capturing the crew or discovering incriminating forensic evidence on a body, there's no way of figuring out who the attacker is. They can analysis the body all they want; all they'll find out is the obvious fact that its from their own system.
Or more sinister, what if the attackers are *stealing* a body? That is, ships from system A steal a Kuiper object from system B and accelerate it to system C? It's a poor MAD indeed whose second-strike attacks the wrong party!
# Coda
A Cold War replayed on an interstellar scale is a frightening prospect. One could only hope for as childish a reality as the Honorverse, but the sober reality seems to be a chilling escalation of the risks of the Cold War. Outer space is cold indeed.
[^volume]: The area of a sphere is given by the equation: $\frac{4}{3} \times \pi \times r^3$\
1 AU = $149.60 \times 10^6$ kilometers\
30 AU = $30 \times 149.60 \times 10^6$, or $4.488 \times 10^9$ km\
55 AU = $55 \times 149.60 \times 10^6$, or $8.228 \times 10^9$ km\
So the shell is the volume of the outer sphere minus the inner sphere:\
$(\frac{4}{3} \times \pi \times (8.228 \times 10^9)^3) - (\frac{4}{3} \times \pi \times (4.488 \times 10^9)^3)$, or $1.9546466984296578 \times 10^{30}$.
[^gigaton]: While in theory H-bombs can be scaled to indefinitely large gigatonnage, this is exceedingly impractical and large bombs are wasteful.
As the bomb becomes larger, ever more of the explosion's energy is dissipated away by the atmosphere or space, and less is directed downward into the earth. The reason the Russians went as large as the [Tsar Bomba](!Wikipedia) was that they needed overkill to compensate for poor targeting - the US with more trust in its accuracy preferred putting multiple smaller bombs onto a target.
[^hanson]: Hanson, ["Near Far In Science Fiction"](
> "Like many others I enjoyed the new _[Star Trek](!Wikipedia "Star Trek (film)")_ movie, even if I don’t especially respect myself for that, and recently just rewatched _Star Wars_ [episodes II, III](!Wikipedia "Star Wars#Prequel trilogy"). And the most compelling visuals and scenes in those movies were similar, in that they combined familiar and emotionally-true foregrounds with dramatic symbolically-meaningful backgrounds which often made little sense if you thought much about them. For example, in _Star Trek_ isolated crowded shipyards are shown scattered in simple farmland, wildly violating economies of agglomeration...
> This all supports my [detached detail]( warning: don’t assume that because the character lives described are compelling, the historical arcs are as plausible."