The energetic consequences of habitat structure for forest stream salmonids

Increasing habitat availability (i.e. habitat suitable for occupancy) is often assumed to elevate the abundance or production of mobile consumers; however, this relationship is often nonlinear (threshold or unimodal). Identifying the mechanisms underlying these nonlinearities is essential for predicting the ecological impacts of habitat change, yet the functional forms and ultimate causation of consumer-habitat relationships are often poorly understood. Nonlinear effects of habitat on animal abundance may manifest through physical constraints on foraging that restrict consumers from accessing their resources. Subsequent spatial incongruence between consumers and resources should lead to unimodal or saturating effects of habitat availability on consumer production if increasing the area of habitat suitable for consumer occupancy comes at the expense of habitats that generate resources. However, the shape of this relationship could be sensitive to cross-ecosystem prey subsidies, which may be unrelated to recipient habitat structure and result in more linear habitat effects on consumer production. We investigated habitat-production relationships for juveniles of stream-rearing Pacific salmon and trout (Oncorhynchus spp.), which typically forage in low-velocity pool habitats, while their prey (drifting benthic invertebrates) are produced upstream in high-velocity riffles. However, juvenile salmonids also consume subsidies of terrestrial invertebrates that may be independent of pool-riffle structure. We measured salmonid biomass production in 13 experimental enclosures each containing a downstream pool and upstream riffle, spanning a gradient of relative pool area (14%-80% pool). Increasing pool relative to riffle habitat area decreased prey abundance, leading to a nonlinear saturating effect on fish production. We then used bioenergetics model simulations to examine how the relationship between pool area and salmonid biomass is affected by varying levels of terrestrial subsidy. Simulations indicated that increasing terrestrial prey inputs linearized the effect of habitat availability on salmonid biomass, while decreasing terrestrial inputs exaggerated a "hump-shaped" effect. Our results imply that nonlinear effects of habitat availability on consumer production can arise from trade-offs between habitat suitable for consumer occupancy and habitat that generates prey. However, cross-ecosystem prey subsidies can effectively decouple this trade-off and modify consumer-habitat relationships in recipient systems.