A fundamental issue with predicting the outcome of multispecies consumer-resource interactions is whether the combined effects of species can be predicted from their pairwise interactions. However, basic predator-prey theory often fails to predict long-term dynamical outcomes of even simple pair-wise interactions. Some important and ubiquitous mechanisms reduce our predictive ability including; 1) nonlinear changes in predation rate with prey density, 2) changes in prey densities over time, and 3) nonlinear changes in predation rate with prey size. Predator imposed mortality rates are often assumed to be constant, which is only reasonable when predators have linear functional responses, or when prey are continuously replenished. However, most predators have nonlinear functional responses and prey densities can be reduced over time due to predation, resource depletion, or disturbance. Moreover, predation rate can be more sensitive to changes in prey size than to prey density. Failure to account for heterogeneity among prey can lead to bias in theoretical predictions based on assumptions of homogenous risk, because the average phenotype of prey changes as size specific predation occurs. In this study, I develop and test a model that incorporates size- and density-dependence in mortality rates to generate more accurate null predictions of aggregate predator effects.
Experiments were modeled on food webs in riverine rock pools (i.e., potholes) that form when erosional processes cut circular depressions into bedrock providing a natural, tractable system for testing ecological theory. Using response surface regression designs, we parameterize size and density dependent functional response models for flatworms and leeches—two common predators of small snails that are the dominant grazers in this system. We show that risk of predation both predators varies as a function of both snail size and density. However, the functional forms of this dependence differ for the two predators. Specifically, mortality was highest for the smallest Physa with flatworms and decreased exponentially with increasing size, whereas Physa mortality was a unimodal function of snail size for the leeches with the highest mortality was highest for intermediate sized Physa. Finally, the fully parameterized size and density dependent model of multiple predator effects better predicts the outcome of an independent experiment where both flatworms and leeches are simultaneously able to freely depredate snails. This study shows that failure to incorporate known nonlinearities in predator prey interactions, such as size and density dependence, can reduce our ability to predict the aggregate effects of multiple predators on prey.