You are here:
Best practices for use of stable isotope mixing models in food-web studies
Citation:
Phillips, D., R. Inger, S. Bearhop, A. Jackson, J. Moore, A. Parnell, B. Semmens, AND E. Ward. Best practices for use of stable isotope mixing models in food-web studies. CANADIAN JOURNAL OF ZOOLOGY. NRC Research Press, Ottawa, Canada, 92(10):823-835, (2014).
Impact/Purpose:
A manuscript by an EPA scientist, along with government and academic researchers from the US, Canada, UK, and Ireland, lays out “best practices” for using stable isotope data and models for studying food webs. The manuscript describes how mathematical mixing models can be used to extract information on wildlife diets from the isotopic composition of an animal and its food sources. A number of developments that have extended the usefulness and scope of these models for this application are reviewed, many of which are the result of EPA research.
Description:
Stable isotope mixing models are increasingly used to quantify contributions of resources to consumers. While potentially powerful tools, these mixing models have the potential to be misused, abused, and misinterpreted. Here we draw on our collective experiences to address the question: what are the major challenges to the effective application of such mixing models? Mixing models have increased rapidly in sophistication. Models are using a Bayesian framework to estimate probability distributions of source contributions, have been published with user-friendly interfaces, and can now incorporate important complexities such as variability in isotope signatures, diet-tissue discrimination factors, hierarchical variance structure, covariates, and concentration dependence. While these advances hopefully have provided novel insights, they still need to be properly implemented. For the effective application of mixing models, we offer the following suggestions. First, mixing models can only be as good as the study and the data. Studies should have a clear question, be informed by knowledge of the study system, and have a strong sampling design to effectively characterize isotope variability of consumers and resources on proper spatial and temporal scales. Second, studies should use a mixing model that is an appropriate tool for the question and recognize the assumptions and limitations of the tool. Decisions such as the grouping of sources or the incorporation of concentration dependence can influence results. Third, studies should be careful about the interpretation of model outputs. Mixing models generally estimate proportions of assimilated resources with substantial uncertainty distributions. Last, a heavy dose of common sense, such as graphing the data before analyzing, can go a long way to getting the most out of these tools. We hope that these suggestions for the effective implementation of stable isotope mixing models will aid the continued development and application of this field.