Nitpicker wrote:It must be very difficult to analyse such small signals within much larger signals. Does anyone know what the margins of error are in these findings? How many eclipses are typically observed per exoplanet candidate before confirmation? How accurate are the exoplanet semi-major axes and temperatures? How accurate are the exoplanet sizes, masses and surface gravity estimations? How accurate are the derived exoplanet compositions? I'm not doubting anyone, but there must be a some level of uncertainty within the numbers.
I can answer some of those.
> How many eclipses are typically observed per exoplanet candidate before confirmation?
For a typical Earth-size Kepler planet a hundred or so eclipses are analysed. The committee doesn't accept any planet into its list of planetary candidates until a minimum of three planetary eclipses are observed.
> How accurate are the exoplanet semi-major axes?
Very good. An acceptable error is about 0.01%. I personally spotted one possibility with an inconsistency of about 0.1% in semi-major axis - enough to eliminate it as a possible planet.
On the dubious assumptions of circular orbit and ignoring any effect of a planetary atmosphere, temperatures are calculated directly from the stellar luminosity and semi-major axis. The results are pretty accurate. There is probably a table accessible of the web that gives estimated errors.
> How accurate are the exoplanet sizes?
Sizes from Kepler data are based on the time that the light curve dips from maximum to minimum brightness, with a correction based on how far the planet is out of alignment with the equator, which in turn comes from the total transit time. From what I've seen, exoplanet sizes are can be in error by up to about 20%, but I'd guess 5% error is more typical.
> masses and surface gravity estimations?
For Kepler data, these aren't particularly good. To get some idea, think three times the percentage error in size for Earth- to Neptune-size planets. For Saturn- to Jupiter-sized planets the mass can easily be wrong by a factor of 100 due to the problem that adding mass to Jupiter could cause it to shrink rather than grow because of the extra gravity compressing the gas. The mass-size relationship of Jupiter-sized planets becomes very sensitive to both planet age and composition.
> How accurate are the derived exoplanet compositions?
This is all modelled rather than measured, based on an assumed temperature at the time of planet formation. Migration of a planet inwards or outwards with age could change this considerably. The exception would be for the nearest bright stars where it is sometimes possible to measure the spectrum or thickness of the planet's atmosphere.