Trophic levels of marine consumers from nitrogen stable isotope analysis: estimation and uncertainty

Abstract Estimates of trophic levels (TLs) are used to calibrate, parameterize, and validate foodweb models and to calculate metrics and indicators of foodweb structure and human impacts. We develop a method to estimate TL from nitrogen stable isotope data (δ15N) and apply it to 5535 individuals from 62 species of marine fish and squids sampled from the Celtic Sea, English Channel, Irish Sea, and North Sea. With this method, uncertainties in δ15N at the base of the foodweb (from a marine isoscape generated with environmental data) and in trophic fractionation (from existing fixed and scaled fractionation models) are propagated through the analysis to quantify uncertainty in TL. Higher values of base and consumer δ15N lead to greater uncertainty in TL estimates for individual consumers. Base and consumer δ15N are higher in coastal regions with lower salinity, such as the Irish Sea and Channel coasts, so uncertainty in individual TL estimates is relatively high in these regions. Conversely, when base and consumer δ15N are low, as in the high salinity waters of the northern North Sea, uncertainty is relatively low. Uncertainty intervals for species' predicted TL at a reference mass (based on ≥10 individuals spanning a range of body sizes) are small compared with estimates for individual consumers, as are estimates of slope of the TL body mass relationship. For ∼50% of region and species combinations showing trends in TL with body mass, there are eight times more positive relationships than negative ones. Our approach can be applied at large spatial scales. It generates estimates of uncertainty that support more rigorous and informed comparisons of the trophic ecology of size classes, species, and species-groups. A data file that includes estimates of TL and associated uncertainty for all sampled individuals accompanies this study.