Species-specific ontogenetic diet shifts attenuate trophic cascades and lengthen food chains in exploited ecosystems

Ontogenetic diet shifts are pervasive in food websbut rules governing their emergence and the implications for trophic cascades are only partly understood. Recent theoretical advances in multispecies size spectrum models (MSSMs) predict that the emergence of ontogenetic diet shifts are driven primarily by size-selective predation and changes in the relative abundances of suitably sized prey. Howeverthese assumptions have not yet been tested with data. Herewe developed alternative MSSMs based on different assumptions about the nature of species and size-based preferences and tested them using an extensive dietary database for the Eastern Bering Sea (EBS). MSSMs with both size and species-specific prey preferences correctly predicted approximately three-fold more of the diet links than those that assumed fixed species preferences. Importantlythese model assumptions also had a profound effect on the strength of fishing-induced trophic cascades and the emergent trophic structure of the community with and without fishing. The diet-informed models exhibited lower predation mortality ratesparticularly for small individuals (less than 1 g) whichin turnreduced the intensity and reach of fishing-induced trophic cascades up the size spectrum. If the level and size dependency of piscivory observed in EBS predators is typical of other systemsthe potential for fishing-induced trophic cascades may be over-stated in MSSMs as they are currently formulated and parameterized. Representation of species-specific ontogenetic shifts in diet can strongly influence system responses to perturbationsand the extensions we propose should accelerate adoption of MSSMs as frameworks for exploring size-based food web theory and developing modeling tools to support strategic management decisions.