This week, we had one of the more interesting colloquia in a while. A former University of Texas student and current employee of Lawrence Livermore Laboratory, David Dearborn, gave a talk on how to use nuclear weapons to divert asteroids that are on a collision course with Earth.
This idea has been popularized in movies like Deep Impact and Armageddon. But most of us have had serious doubts that using nukes on asteroids would be a good idea. Apollo astronaut Rusty Schweickart is strongly against this idea. NASA has been considering the possibility of using nukes on asteroids, and thinks the idea may have merit. I have been (and remain) very skeptical. But David's talk did give us a lot of food for thought.
First, let me be clear. There are no known asteroids that will hit the Earth in the foreseeable future. Not one. So we have time to discuss this problem, and we don't need an answer tomorrow. Though we know we can call on Bruce Willis if we need to.
Asteroids are bits of debris left over from the formation of the Solar System, ranging from small rocks up to a thousand kilometers across. Most asteroids are between the orbits of Mars and Jupiter, but a few (maybe a few thousand or so that are bigger than a few hundred feet across) get close enough to the Earth that they may, one day, hit the Earth.
Asteroids orbit the sun due to gravity, just like planets. So, you would think that we could figure out where the asteroids will be hundreds of years from now, just like we can with planets. But it ain't that easy.
Because asteroids are small, it is fairly easy for them to get tossed about. An asteroid going past the Earth won't do much to the Earth, but the Earth's gravity can fling the asteroid about like it is a toy. The gravity of undiscovered asteroids is enough to change their orbits in ways we can't predict. The strength of the Milky Way Galaxy's gravity varies across the Solar System in ways that affect asteroid positions in the far distant future. But the biggest unknown comes from the Yarkovsky effect.
The Yarkovsky effect is not hard to understand. Think of a typical day on Earth. When the sun rises, the temperature starts to go up. At night, the temperature drops as the Earth re-radiates that heat. The same thing happens on asteroids.
Particles of light, called photons, have momentum, just like any other moving thing. So, when light hits the Earth or an asteroid, the rock gains the momentum of the photon. It's not a lot, but the sun is always putting out photons. When the Earth or the asteroid re-radiates that light (as infrared, heat photons), it has to give the photon a little momentum. In a nice Universe, everything would balance out. But it doesn't, and the Earth and asteroids tend to send out those photons in some directions more than others. And that tends to act like a little tiny rocket engine, slowly pushing the Earth or the asteroid around.
The Earth is so big that this virtual photon rocket engine doesn't affect the Earth's orbit, not even over billions of years. But asteroids are tiny compared to the Earth, and over 100 years, this rocketing affect can cause an asteroid to drift away from its predicted position by a million miles or more. Worse yet, we can't predict how far or in what direction this drift will happen, because it all depends on how fast the asteroid is rotating, what shape it is, what it is made out of, whether it is a solid rock or a flying pile of gravel, and many other effects.
Now the Earth is only 8000 miles across, so if the Yarkovsky effect can move an asteroid by up to a million miles in a century, it seems unlikely that we can really know if a given asteroid is going to hit us or not more than a hundred years into the future. We can tell which asteroids definitely will not hit us (if they're always more than a million miles away). However, less than a century away, we can rapidly become certain how dangerous any given asteroid is.
One of Dearborn's points on Tuesday was that, should we detect an asteroid that looks like it might hit the Earth over 100 years into the future, we probably shouldn't try and deflect it right away. Since we can't be sure it will hit us, there's always the chance that, by moving the thing, we actually put the rock on a collision course. Instead, we should spend some time getting to know our enemy, sending robots, radio transponders, anything that will tell us: What is the asteroid made out of (stone, iron, granite)? What is the asteroid's internal structure? Is it a pile of gravel, a solid rock, a big rock with a lot of gravel around it, a giant chunk of iron, or something else? What is the asteroid's exact orbit, including effects due to rotation and gravity and other planets?
His second point is, if we find an asteroid that is going to hit Earth in the near future, nuclear weapons may be our only option. A ten or twenty year lead time may not be enough for any other option. And, since there isn't time to learn a lot about the asteroid, we should hit it and hit it hard, launching several bombs and blowing the thing apart. There may still be some dangerous-sized pieces left, but if we can spread them out over an area tens or hundreds of times the size of the Earth, the danger goes down some. And, as much as we hate to admit it, given the choice between a chunk of rock that will cause human extinction and a few chunks of rock that will destroy city-sized areas, most of us will choose the latter.
David's other point that hit home involved the idea of using a nuclear weapon to try and "nudge" an asteroid out of the way. Suppose we have a 100 year lead time, and we need to give an asteroid a nudge. A gravity tractor would work, but requires launching several hundred tons of material on a course to rendezvous with the asteroid. We've never launched so much material. We could also put an efficient rocket engine on an asteroid, and slowly fire it to move the asteroid. Again, this requires putting 10 or 20 tons of material on an asteroid, which isn't impossible, but is hard to do. And how do you keep that rocket engine always pointing in the right direction?
But nuclear weapons are light (about a ton), and we can (and, indeed have) put robots weighing that much anywhere we want in the solar system. And, we know what happens when we blow one up, so we know how much energy we'll be giving the asteroid -- according to Dearborn, a single nuke should be able to push a small asteroid enough to move it out of Earth's way 40 years down the road. And nukes are cheap. We have a lot sitting around, and we have rockets to put them on. So, in Dearborn's opinion, nukes are a cheap, easy, and fully-understood means of trying to push an asteroid out of the way.
I'm still skeptical on this last point. I think all of us are a little leery of using a nuclear device for any reason, and there are a lot of unknowns about asteroids -- I don't think we can be certain that a single explosion would do the trick. And, if it split the asteroid into two or three chunks, now we have twice or three times the trouble to deal with. So, I'm not convinced that nukes are the answer.
But Dearborn did convince me that we shouldn't just throw away the idea of using nuclear weapons to deflect asteroids. It, along with gravity tractors and ion propulsion and laser pulsing and other, even more exotic ideas, are all worth more and careful thought and, perhaps at some point even worth testing. And Dearborn convinced many of us that, should we find a new asteroid that will hit us in a few years' time, we may be able to save ourselves with the very weapons that constantly threaten our existence. Thankfully, with no asteroids on the way that we know about, we don't have to make such a fateful decision now.
After Dearborn's talk, I made a joking comment to a colleague that we should keep the nuclear option under wraps, so that a nation like North Korea or Iran doesn't try and use asteroids as a reason why they need nuclear weapons. I was later surprise to find out that such a claim has already been made.