Stable isotope values of Marion Island myctophids using otoliths from fur seal faecal samples

Abstract

Lanternfish (Myctophidae) are abundant mesopelagic fishes in the Southern Ocean ecosystem and represent a large dietary component for sentinel marine predators, such as seals and seabirds. Their contribution to diets of predators can be quantified indirectly using stable isotope mixing models. However, bivariate isotope mixing models require nitrogen (δ15N) and carbon (δ13C) isotope values of potential prey items, which is often lacking. Published stable isotope values of potential prey from geographically dissimilar sites are often used to parameterise the isotope-based mixing models, which might not be adequate, especially when variation in the isotope values at the base of food web exists. Acknowledging the inherent, but often unavoidable spatial assumption using extrinsic data for local modelling, in this dissertation I aimed to generate isotope values for myctophids, specific to the sub- Antarctic Prince Edward Islands (PEIs; Chapter 2), which can be used to refine previous dietary models of sentinel marine predators at PEIs (Chapter 3).

Firstly (Chapter 2), muscle values were required for myctophids utilised by marine predators at PEIs. Fur seals (Arctocephalus spp.) with known foraging areas were used as bio-samplers of the myctophid community, whereby myctophid otoliths were recovered from faecal samples collected around seal colonies breeding at the PEIs. The required stable isotope difference between muscle and otoliths (otolith-to-muscle offset) was established using pairs of fresh muscle and otolith samples of Electrona antarctica (n = 6) and Gymnoscopelus opisthopterus (n = 2) from Dronning Maud Land, Antarctica. The average otolith-to-muscle offset used was -1.3 ± 0.4‰ for δ13C and 1.6 ± 0.3‰ for δ15N. Using these otolith-to-muscle offsets, the muscle δ13C and δ15N values of myctophids could be estimated where only faecal otoliths are available, provided that an additional oxidative cleaning step was used to account for faecal contamination. An oxidative cleaning oversight was corrected for and the associated uncertainty incorporated. The otolith-to-muscle offset was applied to oxidant-cleaned otoliths from fur seal faecal samples at the PEIs. The myctophid species under study were E. subaspera, G. fraseri, G. nicholsi and G. piabilis. Stable isotope values of fish from the PEIs were positively correlated to their Îles Kerguelen counterparts, illustrating similar relative stable isotope ecology of species across sites. The absolute difference between stable isotope values of PEI and Îles Kerguelen fish could be explained by a depleted baseline at the PEIs relative to Îles Kerguelen. Detecting these patterns indicate that the oxidative cleaning and offset method is a realistic way to estimate fish muscle from faecal otoliths, despite, in this case, uncertainty brought about by cleaning correction.

The novel otolith-derived muscle isotope values of the myctophids (Chapter 2) were then used to revise a previous dietary model of juvenile southern elephant seals (Mirounga leonina), in addition to construction of novel models for sub-adults, adult females and adult males, as well as killer whales (Orcinus orca; Chapter 3). Remodelling for juvenile southern elephant seals revised krill consumption from 76% to 58%, and improved modelling performance. Squid, which only included Martialia hyadesi for juveniles, went from a maximal contribution of 35% to a maximum of 28%. Older elephant seal age classes indicated an expected decreasing trend for crustacean consumption with age, but remained substantial for each age class specifically, 35% [0% - 65%] to 24% [0% - 72%] and 19% [0% - 28%] from sub-adults to adult females and adult males. Results also indicated that all classes fed on >30% squid, while fish, surprisingly, played a predominant role in adult male diets (sum of fish containing clusters’ modes 88% [23% - 100%]). The results vary from findings at the PEIs and other sites which predicted that squid predominate diets over fish. While variation may be due to differences in prey abundances or historical estimate biases due to retention of hard squid remains in seal stomachs, it is additionally evident (Chapter 2) that variation in stable isotope values at the base of food web exists between the neighbouring PEIs and Îles Kerguelen. Baseline variation needs to be considered, as it affects the source values considered and modelling outcomes. Wide credibility intervals also suggest varied niche use and fluid combinations of dietary resources within age-sex classes.

The killer whale isotope-based dietary model remained unresolved due to a lack of potential prey with high enough δ13C values to generate a mixing space around the consumer values of these top predators. In addition to highlighting the requirement for local prey isotope values to better characterise southern elephant seal diets, this dissertation reiterates the need for sampling of more northerly prey sources with higher δ13C values to parameterise killer whale models. To this end, this study’s validation of using degraded otoliths and the myctophid-specific, otolith-to-muscle isotopic offset provides means to obtain myctophid isotope values while using predators as samplers of their environments (bio-samplers).