Methodological uncertainty in resource mixing models for generalist fishes

Area of Science:

• Ecology
• Isotope Ecology
• Food Web Analysis

Background:

• Stable isotope ratios (carbon and nitrogen) are widely used to infer predator diet composition.
• Dietary assessment relies on linear mixing models assuming predator tissue isotopes reflect prey sources after trophic discrimination (Δ(x)X).
• Trophic discrimination factor (Δ(x)X) is species- and tissue-specific and influenced by diet quality and quantity.

Purpose of the Study:

• To investigate the impact of uncertainty in trophic discrimination (Δ(x)X) on stable isotope mixing models.
• To demonstrate methods for incorporating Δ(x)X uncertainty into non-Bayesian mixing models using field data from reef fishes.
• To provide guidance on interpreting existing and analyzing new isotope mixing model results.

Main Methods:

• Application of linear mixing models based on mass balance equations.
• Incorporation of uncertainty in trophic discrimination factor (Δ(x)X) within non-Bayesian frameworks.
• Analysis of field data from omnivorous reef fishes to illustrate methodological approaches.

Main Results:

• Dietary interpretations from mixing models are significantly influenced by both the magnitude and uncertainty of Δ(x)X.
• Non-Bayesian models can effectively incorporate Δ(x)X uncertainty, similar to Bayesian approaches.
• Literature review indicates that Δ(x)X uncertainty is often underestimated, leading to model sensitivity issues.

Conclusions:

• Accurate assessment of Δ(x)X uncertainty is crucial for reliable stable isotope mixing model outputs.
• Underestimated uncertainty in Δ(x)X can lead to inaccurate dietary inferences.
• This study offers a framework for more robust analysis and interpretation of stable isotope mixing models in food web research.