The ability to predict how predators structure ecosystems has been shown to depend on identifying both consumptive effects (CEs) and nonconsumptive effects (NCEs) of predators on prey fitness. Prey populations may also be affected by interactions between multiple predators across life stages of the prey, and by environmental factors such as disturbance. However, the intersection of these multiple drivers of prey dynamics has yet to be empirically evaluated. We addressed this knowledge gap using eastern oysters (Crassostrea virginica), a species known to suffer NCEs, as the focal prey. Over four months, we manipulated orthogonally the life stage (none, juvenile, adult, or both) at which oysters experienced simulated predation (CE) and exposure to olfactory cues of a juvenile oyster predator (crab), adult predator (conch), sequentially the crab and then the conch, or none. We replicated this experiment at three sites along an environmental gradient in a Florida (USA) estuary. After obtaining field estimates of NCEs (and simulated CEs) on prey demographic rates, we used an integral projection model to simulate prey population dynamics both with and without NCEs. We used model selection to compare those simulated results to the data from oysters exposed to CEs and NCEs in the field.
Results/Conclusions
For both juvenile and adult oysters, survival was reduced solely by CEs, and variation in growth was best explained by among-site variation in water flow, with a much smaller and negative effect of predator cue. Adults exposed to conch cue exhibited reduced growth (an NCE), but this effect was outweighed by a positive CE on growth: surviving oysters grew faster at lower densities. Finally, conch cue reduced larval settlement (another NCE), but this was swamped by among-site variation in larval supply. This research highlights how strong environmental gradients and predator CEs may outweigh the influence of NCEs, even in prey known to respond to predator cues. In fact, the integral projection model without NCEs afforded a more parsimonious fit to the data than the model including NCEs, suggesting that including the NCEs was not necessary to represent prey dynamics. These findings serve as a cautionary tale for the importance of evaluating NCE processes over temporal scales and across environmental gradients relevant to prey demography.