Lipid normalization and stable isotope discrimination in Pacific walrus tissues

Abstract

Analysis of stable carbon and nitrogen isotope values (δ 13 C and δ 15 N) of animal tissues can provide important information about diet, physiology, and movements. Interpretation of δ 13 C and δ 15 N values, however, is influenced by factors such as sample lipid content, tissue-specific isotope discrimination, and tissue turnover rates, which are typically species- and tissue-specific. In this study, we generated lipid normalization models for δ 13 C and investigated the effects of chemical lipid extractions on δ 13 C and δ 15 N in Pacific walrus ( Odobenus rosmarus divergens ) muscle, liver, and skin. We also evaluated tissue-specific isotope discrimination in walrus muscle, liver, skin, and bone collagen. Mean δ 13 C lipid-free of skin and bone collagen were similar, as were mean δ 15 N of muscle and liver. All other tissues differed significantly for both isotopes. Differences in δ 13 C lipid-free and δ 15 N among tissues agreed with published estimates of marine mammal tissue-specific isotope discrimination factors, with the exception of skin. The results of this work will allow researchers to gain a clearer understanding of walrus diet and the structure of Arctic food webs, while also making it possible to directly compare the results of contemporary walrus isotope research with those of historic and paleoecological studies.

The effects of lipid extraction on δ13C and δ15N values and use of lipid‐correction models across tissues, taxa and trophic groups

“…carnivorous and herbivorous birds and mammals and carnivorous fish) as well as to specific species (dolphins and manatees). Our data are consistent with previous studies that the precision of lipid‐correction models, and thus their utility, increases with taxonomic and trophic specificity (Clark et al, 2019; Hoffman & Sutton, 2010; Logan et al, 2008). We included both aquatic and terrestrial organisms because previous work that has analysed organisms from each habitat separately produced nearly identical models (Ehrich et al, 2011; Post et al, 2007), and we found that trophic group had a larger effect on these models.…”

“…A mixture of linear and mass balance models best fit muscle data among different taxa and trophic groups for our study, suggesting, along with the work of others, that best corrective model for muscle can vary among taxonomic and trophic groups (Doucette, Wissel, & Somers, 2010; Ehrich et al, 2011; Yurkowski, Hussey, Semeniuk, Ferguson, & Fisk, 2015). Data from skin were also a good tissue for lipid correction because they were almost exclusively fit by mass balance models, which was also true with a species‐specific model on Pacific walruses (Clark, Horstmann, & Misarti, 2019). The only time skin was fit by another model in our results was the null model for manatees, likely due to the minimal effect of LE on δ 13 C values in manatee skin (and muscle).…”

“…These findings also explain the low prediction errors for the non‐null models for manatee tissues (Table ), making correction models for these manatee tissues of little utility. Liver data were frequently fit by null models and overall had high prediction errors, a pattern that occurs within species as well (Clark et al, 2019), suggesting it is not a good tissue on which to use lipid‐correction models. Finally, in general, models on herbivore tissues seem to have little utility because LE has little effect on δ 13 C values.…”

“…We included both aquatic and terrestrial organisms because previous work that has analysed organisms from each habitat separately produced nearly identical models (Ehrich et al, 2011; Post et al, 2007), and we found that trophic group had a larger effect on these models. Studies applying species‐specific models have obtained even better fits with lower predictive models (Clark et al, 2019; Hoffman & Sutton, 2010). Thus, our analyses support using the most taxonomically specific models available as the best way to decrease error from lipid corrective models.…”

Understanding marine food web dynamics using fatty acid signatures and stable isotope ratios: Improving contaminant impacts assessments across trophic levels

“…De-lipified tissue or mathematical corrections are therefore used for SI analysis as it reduces the variation associated with lipid content (Clark, Horstmann and Misarti, 2019). was used as the baseline species as they are likely to be part of the same food web as the other samples ( Figure 1).…”