Combined analyses reveal environmentally driven changes in the North Sea ecosystem and raise questions regarding what makes an ecosystem model's performance credible?

When an ecosystem model of the North Sea is calibrated to data from multiple trophic levels, the model estimated the primary production required to support the food web correlates temporally with observed changes in sea temperature and nutrient levels, supporting evidence from empirical analyses. However, a different result is given from an alternative calibration using fish stock data only. The inference taken from the emergent primary production--temperature relationship and empirical data are that, on balance, there is stronger overall evidence to support the calibration constrained at multiple trophic levels. Two important implications of the findings are (i) that the relative importance of fishing and environmental effects is likely to be interpreted differently depending on the calibration approach and (ii) the contrasting model calibrations would give different responses to fishing policies. It raises questions regarding how to judge the performance (and credibility) of an ecosystem model and the critical importance of conducting empirical and modelling analyses in parallel. Adopting a combined approach to ecosystem modelling is an important step in the pursuit of operational and defensible tools to support the ecosystem approach to management. Quand un modele d'ecosysteme de la mer du Nord est cale sur des donnees provenant de differents niveaux trophiques, l'estimation de la production primaire necessaire pour supporter le reseau trophique produite par ce modele est correlee dans le temps aux changements observes de la temperature de la mer et des teneurs en nutriments, appuyant les resultats decoulant d'analyses empiriques. Une autre approche de calage reposant uniquement sur des donnees sur les stocks de poissons donne toutefois un resultat different. La conclusion decoulant de la relation emergente entre production primaire et temperature et des donnees empiriques est que, dans l'ensemble, les donnees appuient plus fortement le calage sur des donnees provenant de differents niveaux trophiques. Deux importantes conclusions decoulent de ces resultats, a savoir que (i) l'importance relative de la peche et d'effets du milieu peut etre interpretee differemment selon l'approche de calage et (ii) differentes approches de calage de modele meneraient a differentes conclusions quant aux consequences de politiques sur la peche. Cela souleve des questions concernant l'evaluation de la performance (et de la credibilite) des modeles d'ecosysteme et l'importance capitale de realiser parallelement des analyses empiriques et de modelisation. L'adoption d'une approche combinee de modelisation de l'ecosysteme constitue un important pas vers l'elaboration d'outils efficaces et defendables pour appuyer une approche ecosystemique de gestion. [Traduit par la Redaction]
Introduction Empirical evidence shows that climate driven changes in ocean conditions have a widespread influence on biological components throughout North Atlantic ecosystems (Ottersen et al. 2001). Both direct and indirect [...]