Multi-dimensional metrics for use with compound specific analyses of food webs

Background/Question/Methods

Stable isotope ratios are essential tools which can be employed to assess the structure and dynamics of food webs. Numerous hypothesis-testing frameworks and analytical approaches have been proposed to characterize dispersion of carbon (δ¹³C) and nitrogen (δ¹⁵N) isotope ratios in bivariate space. Of those that relate to niche metrics, the most commonly used were developed by Layman et al. (2007, 2012). Currently, however, the data for those metrics use stable isotope ratios derived from bulk tissue analyses (total organic carbon and nitrogen), which can lead to sometimes erroneous mean values with large variances (Bowes and Thorp 2015). These confounding results occur because a dietary proxy signal does not typically reflect the bulk signal of the food group but rather that of specific fractions of the food group (e.g., macronutrients, amino acids, fatty acids). This difficulty can be resolved by first using compound specific analyses, especially amino acid stable isotopes, to produce more accurate and less variable results around the mean values. However, if scientists choose to employ compound specific analyses, they are then faced with a lack of metrics to evaluate the community in niche space.

Results/Conclusions

To alleviate this problem, we developed multi-dimensional metrics for use with more than two variables (typical six or more dimensions). Using Manhattan distance, we created n-dimensional plots and metrics in which to characterize quantitatively the community-wide aspects of trophic structure, niche space, and food sources. We demonstrate the utility of these newly-developed multidimensional metrics through analysis of amino acid compound-specific stable isotopes and assess the extent to which the additional isotope data clarify niche dimensionality. The resulting metrics provide increased resolution and reveal new dimensions compared to traditional analytical frameworks. This new method can be readily applied in ecological studies, improving our understanding of food web structure and dynamics in both natural and perturbed ecosystems.