Correction to: Conversions between gas-phase metallicities in MaNGA

We have identified an error in the calculation of O3N2-based calibrations presented in Scudder et al. (2021), which were metallicities based on Pettini & Pagel (2004) O3N2, Marino et al. (2013) O3N2, and Curti et al. (2017) O3N2. [O iii] fluxes were mistakenly multiplied by 1.33, which is required for R|$_{23}$|-based calibrations but not for the ([O iii]|$\lambda$|5007/H |$\beta$|⁠)/([N ii]|$\lambda$|6584/H |$\alpha$|⁠) line ratio. Correcting these values systematically decreases the raw metallicity values for these three calibrations, and slightly increases the number of overall spaxels with metallicities. A corrected Table 1 with metallicity values is presented here. All three metallicity catalogues (DR15, DR7, and TYPHOON) were identically processed, and have all now been corrected. We have re-run the rest of the work, and find that the scatter around our polynomial fits is functionally unaffected. We have updated table 3 of Scudder et al. (2021) here as Table 2 for completeness. Total number of metallicity values per calibration, for both the SMC and MW dust correction models, after S/N cuts, BPT classifications, and including an H |$\alpha$| EW cut. Note. All abbreviations are defined as in Scudder et al. (2021). In ascending typical 2|$\sigma$| scatter, we present the emission-line permutations between calibrations. For each set of calibrations which match the inclusions/exclusions, we find the typical offset of the 2|$\sigma$| contour (the median absolute value of the 2|$\sigma$| residuals) from our polynomial fit. We also include the smallest and largest |$2\sigma$| residuals for each set. The median (and range) of the 2|$\sigma$| scatter is smallest for all calibrations which have full overlap in their emission-line requirements: the top row includes all of the O3N2-based calibrations. The polynomial fits themselves shift horizontally or vertically when converting from or to an O3N2-based metallicity calibration into a non-O3N2-based calibration, by somewhere between 0.026 and 0.055 dex. The median magnitude of the vertical shifts between polynomials is 0.032 dex. Conversions between O3N2-based calibrations and other O3N2-based calibrations are unaffected. We show a sample figure in Fig. 1. We have updated the polynomials presented in Appendix Table A1, and in the full tables presented in the supplementary material. Comparison of a polynomial lines of best fit as published in Scudder et al. (2021) in a solid black line, and the corrected metallicities in a pink dotted line. The median vertical offset between polynomials for the range in x values with polynomial coverage is plotted in the lower left corner. In this case, the difference between polynomials is about 0.04 dex. This figure is representative of the change in the polynomials. Fig. 7 of Scudder et al. (2021) is the most directly impacted figure; qualitatively it is virtually the same, as all three populations presented in that figure were affected by the same systematic error, and for completeness we reproduce it here in Fig. 2. The right hand panel of fig. 8 of Scudder et al. (2021) is the only figure that has a visible change with the update of these metallicities, with the reduction of offsets between polynomials for PP04 O3N2-based metallicities into any other metallicities reduced by 0.04 dex to 0.1 dex. We thus show it here as Fig. 3. The median offset across all polynomials is only reduced by 0.003 dex relative to that reported in Scudder et al. (2021). Comparison of the polynomial lines of best fit. The fifth-order polynomial fit to the MaNGA data presented here are plotted in a black solid line. We plot the third-order polynomial fits to the DR7, which are also affected by this metallicity erratum, in a blue dot-dashed line. third-order polynomial fits to the TYPHOON data, also recalculated here, are plotted in a dashed green line. Comparison of the differences between polynomial lines of best fit. The left panel is functionally unchanged, with median offsets still at 0.019 dex. The median difference between MaNGA & DR7 (right) is reduced by 0.003 dex to 0.014 dex, with the published trend of PP04 O3N2 metallicities being slightly more offset now removed. All remaining figures in Scudder et al. (2021) are affected by |$\lessapprox$| 0.003 dex, with the updated typically reducing scatter, and are not visibly different from those published. Values in text are either identical or correct within 0.003 dex. Tables not reproduced here are also completely unchanged from the original published version. Supplementary online materials (all versions of figs 2, 3, 4, 7 in Scudder et al. 2021, and the full tables of polynomials) have been fully updated. The emission-line data underlying Scudder et al. (2021) are publicly available as part of the MaNGA DR17 data release, available athttps://www.sdss.org/dr17/. Metallicity values themselves are available upon reasonable request to the corresponding author. In this Appendix, we provide a sample few rows of the tables which list conversions between all calibrations, the number of spaxels used in the fitting procedure, the range of validity, and the polynomial fits used in this work as an example of the data structure. Summary of the median offset from a fifth-order polynomial best fit in both positive and negative directions, for the contour that encloses 95.5 per cent of the data. For each metallicity calibration pairing, the number of spaxels which are present is also recorded, for the SMC dust correction curve. The full table, along with the same for all other pairings, and for the MW dust correction curve, is available as supplementary material.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.