Eventually, Kepler and Kepler-like probes will be able to locate all earth-like planets with orbits that cross the path between the star and us. However, this technique will find the larger planets with shorter orbits first.
Granted I don't know there search pattern but we are talking about less than 1,000 systems at this point. And I think we have found 2 reasonably earth like planets already which suggests there are at a minimum millions of them in this galaxy.
I'm not trying to single you out here. I'm just curious what one has to say on HN so as not to be corrected.
Is there some "don't assume I've implicitly made a bunch of unsupported remarks" initialism that I can tack on to the end of all my posts?
/DAIIMABOUR
Both methods are biased towards finding big planets (which produce clear, easy to detect spikes) that are close to their star. (so you get enough data points) Kepler tends to be a little more biased, simply because it hasn't been running long enough: http://en.wikipedia.org/wiki/File:Exoplanet_Period-Mass_Scat...
Kepler was launched in 2009. Realistically, you'd have to run it for another three or four centuries if you wanted to detect exoplanets that orbit as far out as Neptune. (orbital period: 165 years) No wonder we're seeing a lot of planets orbiting close in!
I have no idea of what the actual numbers are, but if we need to probe 100000 stars to find one that has an exo planet, it means that the radial velocity method might not be that efficient.
For instance, maybe they are able to model the effect of the bias, at least for some range of planetary masses, and the bias is not enough to explain the frequency-vs-mass curve they are getting. Or maybe their use of multiple techniques allows them to characterize the bias.
This excerpt from their recent paper on an Earth-like planet in the habitable zone of a nearby M-class star gives an indication of where they're going with this work:
"Using the relations given by Charbonneau et al. (2007), the reported candidates have non-negligible probabilities of transiting in front of the star (∼2.7%, 1.1%, and 0.6% for planets b, c, and d, respectively). [...] With the new generation of optical and infrared spectrographs, many nearby M dwarfs will be efficiently surveyed for low mass planets. If the detection rate holds, very soon now we may have a real chance of searching for spectroscopic signatures of water and life on one of these worlds."
In other words, detect a bunch of candidates, watch for transits, and use spectroscopic information to detect water.
Yes, scientists may have been quoted in this article, but I have seen how most journalists mangle, ignore, or sensationalize things that were said.
Listen carefully to the video. She says "It could only be as common as 10-20%". Just before saying that, she pauses and looks upward, to formulate the sentence correctly. The 10-20% number is an upper bound, not a direct estimate.
You have to give her credit for communicating the idea carefully. The difficulty of doing this in real time is extreme. And if you do it wrong, you really get taken to task by your colleagues.
I only had time to look at the first few minutes, but around 6:50 she does say that our solar system is "not that common", which she quantifies as "it could only be as much as 10-20% of star systems". Maybe that means <= 10-20%?
So, this does not seem to contradict anything Vogt and Brown were quoted as saying.
Incidentally, I wasn't going by the statement in the NPR article, but also by the press release from UCSC (http://news.ucsc.edu/2012/12/tau-ceti.html), Vogt's home institution. I don't think that quote is subject to journalistic mangling.
IOW, I disagree with NPR's claim that "Our Very Normal Solar System Isn't Normal Anymore". Addendum: It may indeed be abnormal. But we don't have data to conclude that yet, or even to suspect it.
All of science is based on sampling bias, by definition. That's one of the main reasons why science isn't a perfect lens for discovering truth. But just because something is probably wrong doesn't mean it isn't a 'scientific fact.'
1) Things which orbit rapidly are easy to spot as we get lots of observations from occlusion and wobble. You can identify a planet over the course of a few weeks if it has a period measured in days.
2) Things which are massive are easy to detect using wobble methods. The closer in they are, the more wobble there is. GMm/r^2 and all that.
3) Massive close-in things with short periods are the easiest planets to detect by a country mile, as you get a nice big occlusion along with a nice big wobble, all happening on nice short timescales.
This entire article is oriented around sample bias due to current instrumentation and techniques. We have little hope of spotting the equivalent of, say, Uranus, with the same orbital period and distance, as we'd need, um, a good few centuries of observations using wobble and occlusion to be certain - and the net effects measured would be tiny, so you'd need a huge pile of data to get a statistically significant measure.
Finally, earth-mass planets at an earth-like distance - again, tricky. Less tricky than the above, arguably, particularly if they have atmospheres (spectral changes are a dead giveaway), but still tricky.
JWST could serve up the goods.
Even with sampling bias, we could still be seeing something legitimately weird. We expect to detect some certain number of exo-planets with a certain range of characteristics with our current technology. We don't think this set of exo-planets is necessarily representative of the average exo-planet in the galaxy, because of the limits of our technology, but we expect to look at X stars, and see Y planets, which look about like Z (where Z is probably Jupiter-sized planets orbiting at Jupiter-like distances). We might also expect to see a few weird planets, where the solar systems are aberrations or we just got lucky and detected something better than we'd expect with our technology.
But what this article seems to be saying (maybe a little poorly) is that the set of planets observed is not the subset of planets we expected to observe. We expected (say) to look at 100,000 stars, see 100 normal Jupiters, fail to see 9,900 other normal Jupiters that were there but we didn't detect, and see 1 aberration, like an Earth-like planet we accidentally detected or a hot Jupiter representing a weirdly captured wandering planet. Instead, we saw 1000 aberrations, hot Jupiters. More than we expected to see based on how we thought solar systems formed.
Just because the unexpected thing is actually extremely easy to detect once you're looking at the universe in the right way doesn't make it any less unexpected. Even if we eventually find millions more solar systems that look like ours than ones with hot Jupiters, we might still have to figure out why there are so many more hot Jupiters than we expected.
"Hot Jupiter" type planets aren't as much of a revelation as one might expect, however, given the prevalence of binary systems, many of which have very, very short periods (J0106-1000 has one of 39 minutes) - and there's still nothing to say that those same systems don't have rocky worlds tucked away that we can't see due to the overwhelming noise from the massive inliers.
I think this case falls under the lack of journalistic integrity, regardless of whether the overall claim is right or wrong. My view is that when a science article wants introduce the idea that reality may be different than conventionally believed, the goal should be to write an (at least mildly) well rounded, informative piece, not a strictly persuasive piece. When the first response in hundreds of armchair physicists' minds around the world is surprise that selection bias wasn't even mentioned (mine included), I think it's fair to say the article falls more into the persuasive category.