Niche-dependent trophic position distributions among primary, secondary and tertiary consumers

Many consumers display flexible feeding strategies that vary among individuals or populations, through life-history, or spatiotemporally. Despite the recognized influence of flexible feeding on the structure and dynamics of food webs, the consequences of these feeding strategies on the actual shape and characteristics of trophic position distributions have received less attention. We proposed and tested several a priori hypotheses to predict the likely effect of niche-dependent (e.g. herbivore, secondary consumer) foraging on the shape and statistical properties of consumer trophic position distributions using natural abundance stable isotope data from a diverse dataset of consumers. We found evidence that the structural characteristics of consumer trophic position distributions varied as a function of trophic niche. Herbivores and tertiary consumers tended to be packed' closely near their mean trophic position, with few individuals realizing trophic positions markedly higher or lower than the mean. Conversely, secondary consumers often displayed broad trophic position distributions with many individuals dispersed away from the center of the distribution. We examined the effect of applying constant versus dynamic isotope trophic fractionation models and found that both models yielded similar although not identical results. Our findings suggest that trophic level omnivory supports a larger fraction of consumer diet at intermediate trophic positions than at either the lowest or the highest positions in aquatic food webs. These results suggest that vertical trophic niche declines among higher order consumers despite general evidence that the range of potential foraging options (i.e. horizontal trophic niche) tends to increase at higher trophic positions. Although further work is needed to test the generality of these patterns in other ecosystems, proactively examining trophic position distributions and reporting appropriate measures of central tendency (e.g. arithmetic versus geometric means) will increase the accuracy of individual trophic studies as well as the applicability of results for meta-analytical food web models.