Modelling available habitat versus available energy flux: do PHABSIM applications that neglect prey abundance underestimate optimal flows for juvenile salmonids?

Abstract

Common applications of models to predict the response of fish habitat to altered stream flow (such as the Physical Habitat Simulation Model; PHABSIM) assume that fish abundance is directly related to the area of suitable habitat for limiting life stages and usually ignore flow effects on prey abundance. However, if prey availability is flow sensitive, then fish production may be more closely related to the total flux of available prey than to habitat area. We compared instream flow predictions from PHABSIM to predictions of optimal energy flux to drift-feeding juvenile coho salmon ( Oncorhynchus kisutch ) estimated using a drift-foraging bioenergetics model. Flux of available energy to juvenile coho salmon declined much more rapidly with decreasing flow than suitable habitat area estimated using PHABSIM, so that, relative to the bioenergetic model, predictions from PHABSIM systematically overestimated productive capacity at very low flows (i.e., underestimated the negative consequences of simulated water withdrawal). Applications of habitat suitability based models like PHABSIM may systematically overestimate low-flow productive capacity for species that prefer low velocities (e.g., pools) but are dependent on energy fluxes generated in higher velocity habitats (e.g., riffles).