Diet reconstruction and resource partitioning of a Caribbean marine mesopredator using stable isotope bayesian modelling

Abstract

The trophic ecology of epibenthic mesopredators is not well understood in terms of prey partitioning with sympatric elasmobranchs or their effects on prey communities, yet the importance of omnivores in community trophic dynamics is being increasingly realised. This study used stable isotope analysis of (15)N and (13)C to model diet composition of wild southern stingrays Dasyatis americana and compare trophic niche space to nurse sharks Ginglymostoma cirratum and Caribbean reef sharks Carcharhinus perezi on Glovers Reef Atoll, Belize. Bayesian stable isotope mixing models were used to investigate prey choice as well as viable Diet-Tissue Discrimination Factors for use with stingrays. Stingray δ(15)N values showed the greatest variation and a positive relationship with size, with an isotopic niche width approximately twice that of sympatric species. Shark species exhibited comparatively restricted δ(15)N values and greater δ(13)C variation, with very little overlap of stingray niche space. Mixing models suggest bivalves and annelids are proportionally more important prey in the stingray diet than crustaceans and teleosts at Glovers Reef, in contrast to all but one published diet study using stomach contents from other locations. Incorporating gut contents information from the literature, we suggest diet-tissue discrimination factors values of Δ(15)N ≈ 2.7‰ and Δ(13)C ≈ 0.9‰ for stingrays in the absence of validation experiments. The wide trophic niche and lower trophic level exhibited by stingrays compared to sympatric sharks supports their putative role as important base stabilisers in benthic systems, with the potential to absorb trophic perturbations through numerous opportunistic prey interactions.

Figure 1. Study site location.

(A) Glovers Reef Atoll within Belize, (B) Map of Glovers Reef Atoll, (C) Enlarged satellite image of the main sampling area on Glovers Reef Atoll.

Figure 2. Density contour plot of δ¹³C vs. δ¹⁵N values for C. perezi fin (blue) and muscle tissues (red), G. cirratum fin (brown) and muscle (grey) and D. americana skin (purple) and muscle (green) sampled from wild populations at Glovers Reef Atoll, Belize.

Different colour tones represent quartile volume contours constructed using kernel density estimation.

Figure 3. Bivariate plot of δ¹³C vs. δ¹⁵N values from tissue samples collected from D. americana, potential prey species, and sympatric shark species at Glovers Reef between 2008–2010.

Values for D. americana, G. cirratum and C. perezi are displayed as combined mean (± SD) skin and muscle tissues (δ¹³C −8.09±1.7, δ¹⁵N 7.36±1.5; δ¹³C −8.32±2.62, δ¹⁵N 9.03±0.46; δ¹³C −9.37±0.98, δ¹⁵N 8.7±0.32 respectively).

Figure 4. Relationship between δ¹⁵N values and individual size (disk width) for D. americana skin (A) and muscle (B).

Red shaded areas represent 95% bivariate normal confidence ellipses for correlative relationships (A) Spearman's r(8) = 0.75 P<0.034 (B) Spearman's r(11) = 0.47, P = 0.15.

Figure 5. Bivariate plots of isotopic space depicting niche areas for muscle tissue of Dasyatis americana (red), Ginglymostoma cirratum (blue) and Carcharhinus perezi (green), using (A) convex hull areas , and (B) Bayesian ellipses .