Erratum for: A&A 597, A113 (2017), DOI: 10.1051/0004-6361/201629699

1. Reported ephemeris

Our reported mid-transit time T₀ contains a shift when compared to other data sets. After checking our analysis we found an error in the correction of the Kepler BJD times. In the original article we wrote that our determined T₀ is given in BJD_UTC. However, this is only true for Kepler data up to release 19. For data releases after this (data release 20/21 onwards), the reported times in the Kepler data are correctly given in BJD_TDB¹. We checked the Kepler data we used and confirm that all the Kepler data has times in BJD_TDB. Thus, our reported T₀ is in BJD_TDB as well.

We have changed Table 1 and Fig. 1 of the original paper and include the new versions in this corrigendum. The only difference is that we have not applied the subtraction of Δt = 66.184 s to T_ref of Sanchis-Ojeda & Winn to convert it into BJD_UTC, and we now correctly state that our value of T₀ is given in BJD_TDB; we note that the actual numerical value of T₀ does not change. The other results in our original paper are not affected by this correction.

Fig. 1 Measurements of mid-transit times from 206Kepler transits. The (magenta) line represents a first-order polynomial fit to the error-weighted measurements. The lower panel shows the residuals with outliers removed. See caption and text of original paper for more detailed explanations.

2. Difference to T_ref

In Sect. 3.1 of the original paper we speculate on the origin of the time difference of about two minutes between T₀ and T_ref; however, we were not able to provide a convincing answer to this inconsistency. After correction for our mistake concerning the time system, this difference ΔT between T₀ and T_ref is reduced by Δt = 66.184 s to ΔT = 64.17 s. We believe the remaining difference ΔT is caused by the fact that T_ref as given by Sanchis-Ojeda & Winn (2011) is reported to be in BJD_TDB but is actually BJD_UTC. In this case we would have to correct T_ref by adding Δt to convert it into BJD_TDB, and ΔT would reduce to approximately −2 s. This assumption is reasonable because Sanchis-Ojeda & Winn published their paper in 2011, before the Kepler release notes 19 had been published; therefore, the problem of erroneous Kepler times was not yet known and the headers of the Kepler data files incorrectly stated that the times were in BJD_TDB².

With this correction, the time shift of about one minute for early transit numbers in Fig. 1 virtually vanishes. Thus, we conclude that T_ref is most likely BJD_UTC; however, we have not changed this in Table 1 to be consistent with the table in Sanchis-Ojeda & Winn (2011) from whence this value is taken.

References

© ESO, 2017

All Tables

All Figures

	Fig. 1 Measurements of mid-transit times from 206Kepler transits. The (magenta) line represents a first-order polynomial fit to the error-weighted measurements. The lower panel shows the residuals with outliers removed. See caption and text of original paper for more detailed explanations.
In the text