The Kepler satellite, which will seek out those habitable planets, launches March 6. What do you think it is going to tell us?
I'm betting it's going to tell us that Earths are quite common. Kepler's going to be looking at 100,000 stars for three or four years. We expect that Earth-like planets will have their orbits aligned in such a way that basically one out of 1,000 can be seen by Kepler. So if every single star has one Earth-like planet, that means that Kepler will see roughly 100 Earths. But if it turns out that we're wrong and, say, only one out of 100 stars has an Earth-like planet, then Kepler might find one Earth—if it's lucky—or it might find zero. And if Earths only occur around one out of 1,000 stars, we'd have to be really lucky for Kepler to find anything at all.
So Kepler basically assumes that it should be able to find some Earths as long as they occur at roughly a 1 percent rate or higher. There's a very good chance that Earths are available essentially all the time, but we can be a little bit conservative and say that Kepler should find dozens of Earths.
If Kepler does find a number of Earth-like planets, then we'll know how many there are in general because Kepler will have searched such a large sample. And then we'll know how to go about planning the next phase of the search, which is to search the nearby stars. Kepler's going to be searching stars hundreds of light-years away all in one direction of the sky. It's sort of like after looking at another neighborhood in another city, we want to then look at our own block. Kepler will tell us how many houses we have to search on our block to look for life.
Given that our planet is the only Earth we know of, how can we extrapolate how many Earths there should be out there?
First, if you talk to astronomers who look at young stars, they will tell you that when stars form, they tend to have a little bit of angular momentum, which means that they can't accrete all of their matter and they end up having a disk around them. Such disks are what planetary systems form out of, basically the leftovers from the star-formation process. Essentially all young stars have these disks, so we expect that these young stars at least have the possibility of having planetary systems.
Second, those who worry about planet-formation processes find that it's very hard to stop Earth-like planets, or some sort of large, rocky object, from forming. Earths in some sense are easier to build than Jupiters, but we already know from our extrasolar planet census that Jupiters exist around at least 10 percent, and probably around 20 percent, of stars. So Earths should be even more common than that.
Finally, and even more directly, the planetary searches are already beginning to find a new class of planets called super-Earths with masses maybe five, 10 or 15 times the mass of Earth that orbit a little closer to their star than our planet does. These guys occur on roughly one third of nearby solar-type stars. And these are sort of the oddballs in some sense, which I think are very much just the tip of the iceberg of the spectrum of Earth-like planets. In any theoretical model of planet formation that people talk about, there should be a ton of Earths compared to these oddball super-Earths, so when we do a complete census we should find a lot of Earths. If these oddballs are there 30 percent of the time and the Jupiters are there 20 percent of the time, that means the ones we can't quite see should be there essentially all the time. So it's a very compelling story, and all the evidence from several different directions points toward Earths being quite common.