Examples of changes from HITRAN 2016 to HITRAN2020. We use two of the small discrepancies between data and model detected by Allart et al. (2022) for the H₂O lines, based on ESPRESSO spectra. The P-Cygni profiles found by the authors indicate that the deep part of the modeled line at 7184.6 Å was slightly blueshifted with respect to the data (see the double structure at the bottom in the top plot, which results in a global blueshift), while it was the opposite in the spectral regions shown in the bottom panel. The new HITRAN prediction shifts the lines in a way to correct for these effects. We also note the new weak line that is predicted by HITRAN 2020 at 7235 Å and was absent in HITRAN 2016.