Meta-analysis and hierarchical modeling to identify temperature and habitat effects on North Atlantic cod carrying capacity and maximum reproductive rate

Irene Mantzouni, Helle Sørensen, Bob O’Hara & Brian R. MacKenzie

This study aims to identify the effect of temperature on cod (Gadus morhua) population dynamics by meta-analyzing data across the species distributional range in the North Atlantic. Our dataset consists of spawner (S), recruitment (R) and temperature (T) time-series for 21 eastern and western cod stocks. Regarding T, we use annual spring (spawning season) estimates in the upper water layer (0-100m). We are employing a twofold approach, aiming at different but complementary issues; (i) the identification of the general pan-Atlantic pattern of T impacts, with special focus on the extreme temperature and/or recruitment events and (ii) the incorporation of these climatic effects in the spawner-recruit (SR) models in order to improve the predictive capacity and investigate the possibility that SR parameters depend on ecosystem characteristics. The first method focuses on the identification of the influence exerted on R or R survival (log(R/S)) during years of extreme low or high T. It is therefore based, in large part, on the classification of exceptional population observations according to the exceptional T events. The next approach involves the study and integration of T effects on the parameters of the SR models. To this end, we develop the models using hierarchical (mixed models and Bayesian inference) approaches which allow the parameters to be modeled across the cod stocks. It is shown that CC depends on the stock specific habitat size (40-300m) and this can explain almost half of the variability observed across the species range. More importantly, by combining data across the cod distributional range, and thus from the whole thermal range, we can make inference on the functional form of T effects on cod R survival and also predict the possible effects of ocean warming on stocks productivity.