Educated search for transiting habitable planets

http://arxiv.org/abs/1002.4702

interest calculation

Quote :

To estimate the observational effort required for a RV detection for GJ 436 (V=10.7), we compute the detection limit (Zechmeister&K¨urster 2009) imposed by RV data gathered with HIRES/Keck (Maness et al. 2007). Figure 1 (bottom panel) shows the semi-amplitude above which a planet would have been detected, with a 99% confidence level (for all of our 12 trial phases). The plot shows fluctuations due to sampling efficiency and an average limit of 6.3 m.s−1. Any planet that induces a RV higher than 6.3 m.s−1 are therefore excluded with a 99% confidence limit. Given the large time coverage of HIRES/Keck data, red noise is probably largely accounted in our analysis. Adding more measurements will therefore scale the limit with the square of the number of measurements. Hence, the observational effort to detect a 1-M planet at  0.3 AU from GJ 436 (or a K  0.3m.s−1 semi-amplitude) requires at least 420 times more Keck measurements (> 3000 h). Despite its northern declination, GJ 436 is visible with HARPS for few months a year. Its typical 1 m.s−1 precision for bright M dwarfs could shorten the time to 1000 h. As a result, true Earth-mass planet (and all the more so for a Mars-mass planet) located in the HZ of GJ 436 seems beyond reach of current velocimeters.

Quote :

As for GJ 436, we evaluate how the RV method competes with photometry to find Earth-mass planet in GJ 1214’s HZ. GJ 1214 is a lower-mass star than GJ 436 and an Earth-mass planet at 0.66 AU induces a reflex motion with a semi-amplitude of 0.88 m.s−1. But GJ 1214 is also fainter than GJ 436 (V=15 vs 10.7) and HARPS achieves a median uncertainty of 4.5 m.s−1 only in 40 min exposures. To balance both effects, we compute the detection limit with the 28 HARPS RVs reported in the detection paper (Charbonneau et al. 2009). The plot in Fig. 2 (bottom panel) shows that our detection limit for GJ 1214 fluctuates more than for GJ 436. This is because we have too few data points for a clean sampling, but more data will smooth the curve and it will eventually be independent of the period. The dashed line underlining the plot (K10 m/s) is therefore a better estimate for the velocity amplitude that we want to scale. To detect an Earth-mass planet orbiting at 0.66 AU – or K0.88 m.s−1 – we therefore need 130 times more data points (> 2000 h). As for GJ 436, a true Earth-mass planet (or lower mass) requires unrealistic observing time with current velocimeters.

For comparison, find analogs Earth on HARPS in 10 nearby stars similar to the Sun is estimated at approximately 2000 hours over 5 years.

In general, even confirmed habitable planets Keplera around red dwarfs will be require a very long time.

http://ssc.spitzer.caltech.edu/warmmission/scheduling/approvedprograms/go7/70049.txt
I did not think that a full search for habitable transit planets in Glize1214b happen so quickly.