Correction to MEPS 695:203-216 (2022) - DOI: https://doi.org/10.3354/meps14112
Ecosystem shifts inferred from long-term stable isotope analysis of male Antarctic fur seal Arctocephalus gazella teeth
Renan C. de Lima*, Julieta D. Cebuhar, Javier Negrete, Afonso Ferreira, Eduardo R. Secchi, Silvina Botta
*Corresponding author: renancdl@gmail.com
June 8, 2023:
The Suess effect correction in δ13C values was equivocated. After appropriately correcting, some details regarding δ13C must be reconsidered. The main outcomes of the paper were not affected by this change.
In the abstract, the decrease rate of 0.04 yr–1 should be disregarded; the long-term period considered for carbon temporal analysis should be 1979–2015 and not 1974–2015—that is because a new reassessment of data demanded the exclusion of points before 1979 to assure model robustness and avoid overdispersion. There is a δ15N increase after the late 1990s (not a decrease as previously stated).
In Table 1 on p. 207, the mean raw δ13C and mean Suess-corrected δ13C columns have been corrected, and in Table 2 on p. 210, generalized additive models (GAMs) of the relationship between environmental variables/climate anomalies and bulk dentine δ13C values have been amended. Furthermore, Fig. 3 (p. 208), Fig. 5 (p. 211) and Fig. 7 (p. 211) have been updated.
In the Results section on p. 209, The δ13C values of growth layer groups (GLGs) now range from –17.5 to –22.2‰ (mean = –20.5 ± 1.2‰). The intraindividual variability for δ13C values was 0.4 to 1.7‰.
Shapiro-Wilk test results for age class assessment should be H3 = 98.7, p < 0.01. The p-value for yearling-juvenile GLGs comparison should be 0.05.
After checking residuals’ plots, outliers from δ13C data had to be removed to assure model robustness (24 points, instead of the original 26, from the 1974–2015 dataset and 10, instead of 6, from 1998–2015).
Long-term data in the carbon models refer to 1979–2015 (not 1974–2015 as it is for nitrogen), and the p-value for the influence of year should be 0.003. The 0.04‰ yr–1 decrease rate of δ13C values through time should be disregarded.
Finally, in the Supplement, the δ13C models’ residuals plots have been amended in Figs. S1 & S3. Figs. S5 & S6 were missing a few individuals and have been updated.