Comparing Surplus Production Estimates Fit With and Without Environmental and Ecological Co-Variates for 11 Large Marine Ecosystems in the Northern Hemisphere

We fit simple production models with and without covariates for several functionally analogous species using data for 11 large marine ecosystems.  The most complete evaluation of the impacts of covariates on estimates of production were for cod (Gadus morhua) and herring (Clupea harengus).  Functionally analogous species included Greenland cod (Gadus ogac) and Pacific cod (Gadus macrocephalus), and Pacific herring (Clupea pallasii). Biological covariates included the abundance of prey (herring for cod and an index of primary production for herring).  Physical covariates included common physical forcing functions such as AMO and PDO as well as water temperature. For both functional species, we fit a hierarchy of nested models that included a full model with physical and biological covariates.  Models were compared using an information-theoretic framework.  The numerical impacts of covariates differed among systems.  However, the trends among the different ecosystems were similar. The inclusion of a covariate representing prey led generally to reductions in MSY biomasses.  The impacts of physical covariates were more variable.  For example, surplus production in several ecosystems was significantly affected by broad scale oceanic indices, e.g., PDO or AMO, but differing to degrees (i.e., direction and magnitude of effect).  Importantly, there was no region grouping of which ecosystems were affected, suggesting the presence of broad scale teleconnections in ecosystem responses.