I stumbled across an interesting article today. Astronomers have studied the solar systems of hundreds of planets by now, and most of them don’t look like ours. It seems like the typical planetary system has large gas giants (Jupiter-sized or larger) orbiting very close to the star. Often even closer than Mercury. Only one star system found so far (HD 13931 b) appears to be similar to our own.
The implications are actually pretty significant. Astronomers & Sci-Fi types have generally assumed that there are tons of systems with planets in the “Habitable Zone,” but the reality might well be that planets like Earth are pretty rare.
Are we sure this isn’t an artifact of how we dectect other planets?
I thought our methods were biased towards finding either big planets (since they cause larger wobbles in their stars) or close planets (lots of ‘eclipses’ of their stars due to the short year). Is it really a surprise that we’re finding a lot of big, close in planets?
The comments are interesting, a little higher level than usual (not surprising since it’s an NPR story). Even they can’t avoid the cosmological battle with religionists, but oh well.
Some there point out that it’s still a fairly small and restricted sample, and what we can see is based on the limitations of our ability to observe objects at such great distances. I’m not an astronomer, but it seems like solar systems like ours (with planets spaced farther apart and farther from the sun) would be harder to spot than the kind that they are finding (with larger planets close in).
It makes me wonder how far telescopic technology can carry us in observing other star systems, without actual FTL travel.
Roddy
We studied hundreds of planets. Is that a reasonable sample size to draw a conclusion from considering there could be billions of solar systems with planets out there?
I’m not a physicist, but my understanding was the method involved examining how the gravitational pull of the planet(s) affected the sun to determine what kinds of planets were in the solar system. That kind of methodology would lead you to find planets on suns where large gas giants are rotating close to the suns surface at an incredibly fast speed. Smaller planets that are further from the sun and that move more slowly will not have as big an impact on the central sun.
Also they’ve only studied 632 systems so far. Also planets change, our solar system used to have many more smaller planets that collided into each other. I don’t know if that is relevant to the issue.
So how did our solar system start out with several small rocky planet close to the earth, many of which crashed into each other and eventually formed our first 4 planets (mercury, venus, earth mars)? What made our system different?
That is one method, but another, the Transit Method, watches for planets passing in front of the star, obscuring it’s light. This only works for planet orbits which align with our line of sight to them, but it has been successful in finding lots of planets similar in size to earth.
I read many years ago about a proposed technique that supposedly could theoretically spot continent sized objects “as far away as Andromeda”. The idea is you station a probe out at the edge of the Solar system at a specific distance, and then you can use the gravity field of the Sun as a huge lens. And having such an enormous lens grants an enormous range that objects can be spotted at. The practical problems are getting something out there with enough fuel left to stop in place, and that you’ve got only a really limited ability to change your viewpoint.
Our detection process is extremely limited. I wouldn’t put too much stock in any “conclusions” based on current sample size. After all, when sample size was one (1), they presumed that every star system was exactly like ours.
You are correct. How can anyone make an assumption based on such a small sample. If the universe is infinite, then there are infinite possibilities for all sorts of planets.
How certain conclusions one can make from a given sample size is not dependant on the size of the population that sample was taken from. A sample size of a few hundred is equally useful for estimating the properties of a group of a hundred thousand objects as it is for uncountable trillions.
The real reliability problem is not sample size, but sample bias, as The Lurker Above suggests.
Utter nonsense in this case. We have no idea what the size of the population is, or even the possible conditions where earth like planets could occur. There’s no sampling going on, the samples are the only planets we know about. It’s not sampling, it’s mindless extrapolation.
Those are problems, but they are not problems related in any way to sample size, which is what you complained about. Don’t mix up sample size problems with sample bias problems; they’re different problems, in need of different remedies.
I agree there is bias, but it’s not what Lurker mentions. Now explain how there’s no problem with sample size. I could randomly select a sample of 3 apples and determine that all apples are red. No bias there.
The difference is that three is a useless sample size, while several hundred is not. The point is that the total number of stars in the galaxy (or universe) does not determine whether or not the hundreds of stars examined is sufficiently large.
If you were to somehow randomly select 637 apples (with every apple in the world being equally likely to be selected, somehow), and all of them were red, “all or almost all apples are red” would be a perfectly valid conclusion, and it would remain valid whether the total number of apples in the world were 100 000 or 100 000 000 000 000 000 000 000 000.
The 637 is the number they found, not the sample size I believe.
If I look at a hundred stars and the vast majority have hot close Jupiters we are the freaks.
If I look at a hundred stars and find a few hot jupiters that doesn’t say much. Particularly given that detecting systems like are own is much harder and takes years of observations.