Comparison Between Trophic Positions in the Barents Sea Estimated From Stable Isotope Data and a Mass Balance Model

Abstract

The trophic position concept is central in system ecology, and in this study, trophic position (TP) estimates from stable-isotopes and an Ecopath mass-balance food web model for the Barents Sea were compared. Two alternative models for estimating TP from stable isotopes, with fixed or scaled trophic fractionation were applied. The massbalance model was parametrized and balanced for year 2000, was comprised of 108 functional groups (Gs), and was based on biomass and diet data largely based on predator stomach data. Literature search for the Barents Sea Large Marine Ecosystem revealed 93 sources with stable isotope data (δ ¹⁵N values) for 83 FGs, and 25 of the publications had trophic position estimated from nitrogen stable isotopes. Trophic positions estimated from the mass-balance model ranged to 5.1 TP and were highly correlated with group mean δ ¹⁵N values, and also highly correlated with the original literature estimates of trophic positions from stable isotopes. On average, TP from the mass-balance model was 0.1 TP higher than the original literature TP estimates (TP_SIR) from stable isotopes. A trophic enrichment factor (TEF) was estimated assuming fixed fractionation and minimizing differences between trophic positions from Ecopath and TP predicted from δ ¹⁵N values assuming a baseline value for δ ¹⁵N calculated for pelagic particulate organic matter at a baseline TP of 1.0. The estimated TEF of 3.0h was lower than the most commonly used TEF of 3.4 and 3.8h in the literature. The pelagic whales and pelagic invertebrates functional groups tended to have higher trophic positions from Ecopath than from stable isotopes while benthic invertebrate functional groups tended to show an opposite pattern. Trophic positions calculated using the scaled trophic fractionation approach resulted in lower TP than from Ecopath for intermediate TPs and also a larger TP range in the BS. It is concluded that TPs estimated from δ ¹⁵N values using a linear model compared better to the Ecopath model than the TPs from scaled fractionation approach.