Transiting a bright star, the “Neptune desert” planet HD 219666b was one of the more important early discoveries from the TESS survey. With a depth of only 0.17 per cent, the transits would be a challenge for any ground-based transit survey.
Nevertheless, we think we’ve found them in WASP-South lightcurves dating back to 2010. Here they are:
The orange lines show times of transit, as found by the WASP transit-detection algorithms. The shallow dips seem to be real, since they align both in period and in phase with the dips seen in the TESS lightcurve. The output from the WASP search algorithm is not itself that convincing:
However, the period that it finds (6.03446 days) matches the TESS period to an accuracy of 0.03 per cent, and the WASP ephemeris then predicts the times of the TESS transits bang on (they occur at 420.99999 cycles on the WASP ephemeris), which together mean that the detection must be real. Here are the WASP data folded on a template of the TESS transit:
With a depth of 0.17 per cent, the transits of HD 219666b are the shallowest that WASP has detected.
The benefit of looking for such pre-detections of TESS planets is that we can then produce a transit ephemeris based on data spanning a baseline of 8 years, rather than the 20 days spanned by the TESS transits. This means we can predict future transits to an accuracy of minutes, instead of hours, which is highly useful for future observations. Hence this WASP-South detection of HD 219666b transits is well worth an AAS Research Note.
As TESS continues its all-sky survey it will produce high-quality data containing lots of transits for all the WASP planets. This is especially so for planets near the ecliptic poles, which TESS will observe over many sectors. With TESS Sector 4 data recently released, here are some plots borrowed from David Kipping on Twitter.
The lower plots show the variations in transit timing (O–C is the difference between the observed timing and the timing calculated from an ephemeris).
These plots seem to show something that I’ve suspected for a while, namely that there are correlated deviations in the transit timings, meaning that if one O–C value is slightly early (or late) then the next one is more likely to be the same. Such deviations can also be larger than expected given the errors (the quoted chi-squared value for WASP-100b of 44 for 32 degrees of freedom tells us that the error bars don’t fully account for the variations).
This must be the result of stellar activity, magnetic variations on the surface of the star such as star-spots and faculae. Any deviation from a smooth stellar profile can then alter the transit profile.
Properly accounting for such effects will be important for two sorts of study. The first is looking for “transit-timing variations”, changes in the transit time of a planet caused by variations in its orbit owing to the gravitational perturbations of another planet. The second is looking for long-term changes in the orbital period, such as the inward-spiral decay of the orbit predicted to be caused by tidal interactions of the planet and its host star. The literature contains marginal claims of the latter effect that might be better explained as the effect of magnetic activity of the host star.