Anglada-Escude and Tuomi question the statistical rigor of our analysis while ignoring the stellar activity aspects that we present. Although we agree that improvements in multiparametric radial velocity (RV) modeling are necessary for the detection
of Earth-mass planets, the key physical points we raised were not challenged. We maintain that activity on Gliese 581 induces RV shifts that were interpreted as exoplanets.
Using solar spectral irradiance measurements from the SORCE spacecraft and the F/F technique, we have estimated the radial velocity (RV) scatter induced on the Sun by stellar activity as a function of wavelength. Our goal was to evaluate the potentia
l advantages of using new near-infrared (NIR) spectrographs to search for low-mass planets around bright F, G, and K stars by beating down activity effects. Unlike M dwarfs, which have higher fluxes and therefore greater RV information content in the NIR, solar-type stars are brightest at visible wavelengths, and, based solely on information content, are better suited to traditional optical RV surveys. However, we find that the F/F estimated RV noise induced by stellar activity is diminished by up to a factor of 4 in the NIR versus the visible. Observations with the upcoming future generation of NIR instruments can be a valuable addition to the search for low-mass planets around bright FGK stars in reducing the amount of stellar noise affecting RV measurements.
Gliese 667C is an M1.5V star with a multi-planet system, including planet candidates in the habitable zone (HZ). The exact number of planets in the system is unclear, because the existing radial velocity (RV) measurements are known to contain contrib
utions from stellar magnetic activity. Following our analysis of Gliese 581 (Robertson et al. 2014), we have analyzed the effect of stellar activity on the HARPS/HARPS-TERRA RVs of GJ 667C, finding significant RV-activity correlation when using the width (FWHM) of the HARPS cross-correlation function to trace magnetic activity. When we correct for this correlation, we confirm the detections of the previously-observed planets b and c in the system, while simultaneously ascribing the RV signal near 90 days (planet d) to an artifact of the stellar rotation. We are unable to confirm the existence of the additional RV periodicities described in Anglada-Escude et al. (2013) in our activity-corrected data.
The M dwarf Gliese 581 is believed to host four planets, including one (GJ 581d) near the habitable zone that could possibly support liquid water on its surface if it is a rocky planet. The detection of another habitable-zone planet--GJ 581g--is disp
uted, as its significance depends on the eccentricity assumed for d. Analyzing stellar activity using the H-alpha line, we measure a stellar rotation period of 130+/-2 days and a correlation for H-alpha modulation with radial velocity. Correcting for activity greatly diminishes the signal of GJ 581d (to 1.5 sigma), while significantly boosting the signals of the other known super-Earth planets. GJ 581d does not exist, but is an artifact of stellar activity which, when incompletely corrected, causes the false detection of planet g.
