Variation in δ¹⁵N and δ¹³C trophic fractionation: Implications for aquatic food web studies

Use of stable isotope techniques to quantify food web relationships requires a priori estimates of the enrichment or depletion in δ 15 N and δ 13 C values between prey and predator (known as trophic fractionation; hereafter Δδ 15 N and Δδ 13 C). We conducted a broad‐scale analysis of Δδ 15 N and Δδ 13 C from aquatic systems, including three new field estimates. Carnivores had significantly higher Δδ 15 N values than herbivores. Furthermore, carnivores, invertebrates, and lab‐derived estimates were significantly more variable than their counterparts ( f‐test, p < 0.00001). Δδ 13 C was higher for carnivores than for herbivores ( p = 0.001), while variances did not differ significantly. Excluding herbivores, the average Δδ 15 N and Δδ 13 C were 3.4‰ and 0.8‰, respectively. But even with unbiased fractionation estimates, there is variation in isotopic fractionation that contributes to error in quantitative isotope model outputs. We simulated the error variance in δ 15 N‐based estimates of trophic position and two‐source δ 13 C diet mixing models, explicitly considering the observed variation in Δδ 15 N and Δδ 13 C, along with the other potential error sources. The resultant error in trophic position and mixing model outputs was generally minor, provided that primary consumers were used as baseline indicators for estimating trophic position and that end member d13C values in dietary mixing models were sufficiently distinct.