WASP-173 and KELT-22 are the same object. The WASP and KELT teams are both trying to find transiting exoplanets around relatively bright stars, and this means that sometimes our discoveries overlap. We announced that WASP-173 hosts a hot Jupiter in a paper on arXiv on the 7th March, and then on the 21st March KELT reported an entirely independent discovery of the same planet.
Since the two teams use different facilities, techniques and software, comparing the two sets of system parameters provides an interesting check on the methods. So let’s see how similar the reports are.
The biggest difference is a somewhat different transit depth. We (WASP) report a depth of 0.0123 ± 0.0002 whereas KELT report 0.0145 ± 0.0008, where the difference is greater than the error bars quoted. Now this system is a double star, with a companion star 6 arcsecs away and 0.8 magnitudes fainter. That makes it hard to measure the depth. One either uses a much smaller photometric aperture than normal, excluding the nearby star, or one uses a much wider aperture, containing both stars, and makes a correction for the dilution of the companion. Either approach could introduce systematic errors more than normal. Then, of course, there could be red noise in the light-curves owing to observing conditions or stellar activity.
The greater depth in the KELT paper means they arrive at a slightly larger planet radius (1.29 ± 0.10 Jupiter radii) than we do (1.20 ± 0.06) but here the error ranges overlap. The planet mass (derived mostly from the radial velocity data) is comparable, 3.47 ± 0.15 Jupiter masses in the KELT paper, and 3.69 ± 0.18 in ours.
The differences in the parameters of the host star are all within the error ranges. KELT report a G2 star with an effective temperature of 5770 ± 50 K, a surface gravity (log g) of 4.39 ± 0.05, and a mass and radius of 1.09 ± 0.05 and 1.10 ± 0.08 in solar units, whereas WASP report a G3 star with effective temperature of 5700 ± 150 K, a surface gravity of 4.5 ± 0.2, and a mass and radius of 1.05 ± 0.08 and 1.11 ± 0.05.
Another comparison is the “impact factor” (how near the center-line the transit chord is), which we have as 0.40 ± 0.08 while KELT report 0.31 ± 0.18. Our higher value results from our having a higher transit width, 0.0957 ± 0.0007 days, compared to KELT’s 0.0981 ± 0.0025. Again, the differences point to red noise in the transit lightcurves, which is likely to produce uncertainties greater than the formal error bars.
Overall, the values are sufficiently similar that we can have broad confidence in the values, but the presence of systematic noise does need to be borne in mind.