It's full of Neptunes...

Among the multitude of delicious Kepler results to digest from yesterday was an estimate of the frequency of occurrence of particular planet categories.  Now the full report by Borucki et al. is available we can take a closer look.

By allowing for the known geometric effects of transit detections, sensitivity effects from stellar brightness, observation time and transit frequency and models of noise and false positive rates then the authors take a careful stab at computing the true population numbers for planets. It's not unlike being shown a single snapshot of part of a forest that also happens to be shrouded in fog and having to guess how many trees there really are. With some logic and statistics you can probably make a pretty good estimate.

The results are intriguing. For the range of stellar types in the Kepler data (mostly normal F, G, and K stars that range from a bit less massive to a bit more massive than the Sun) then it is estimated that about 6% of all such stars harbor 'Earth-sized' planets less than 1.25 times the radius of Earth within orbits of 0.5 astronomical units - or half the size of Earth's actual orbit. This orbital cut-off is simply due to the fact that Kepler has not been observing for long enough to find planets further out - yet.

Slightly larger planets, so-called 'Super-Earths' up to 2 times the size of our homeworld are similarly numerous and occur around about 7% of all such stars. Jupiter sized planets, between 6 and 15 times the girth of Earth should be present around about 4% of stars. Again, on orbits within 0.5 astronomical units.

Remarkably, the most numerous planets are those in the 'Neptune' size range, between 2 and 6 times Earth-radius. About 17% of all such stars should play host to these hefty worlds. This is to my mind a clear and excellent challenge for our theories and models of planet formation. Whatever schemes we come up with had better reproduce this kind of population distribution.

Then there is one other tantalizing feature. Borucki et al. subdivide these results into bins according to orbital radii. The trend for all planetary sizes is remarkably flat. What does that mean? It means that if one were to extrapolate these results to larger orbital radii, to the planets yet to emerge as Kepler continues its long hard stare, we might expect very similar results for all those worlds in the magical zone that is equivalent to where our own Earth orbits its G-dwarf star. In a galaxy of 200 billion stars this would imply a few million such circumstances. Of course it's not really magical, that's just our prejudice. Nonetheless whether it's scientific or not, we would perhaps all dream more interesting dreams if we knew that small rocky worlds orbited other Suns just the way we do.