Triple check: Observations verify structural realism of an ocean ecosystem model

Abstract

[1] Improvements in fisheries and ecosystem management could be made if the prediction of key zooplankton, such as krill, were possible using ocean ecosystem models. To examine structural realism, hence the validity of a coupled physical-biogeochemical model, we compared measured spatiotemporal dynamics of krill and seabird abundance off California to hindcasted mesozooplankton derived from an independently designed model. Observed krill and modeled mesozooplankton (Z2) displayed latitudinal coherence but distinct longitudinal offsets, possibly related to unrealistic bathymetry in the model. Temporally, Z2, Thysanoessa spinifera (a neritic krill species) and seabird density and reproductive performance were well correlated, indicating that quantitative prediction regarding marine predators in upwelling ecosystems is within reach. Despite its basin-scale framework, the ROMS-CoSiNE model captures zooplankton and top predator dynamics regionally in the central California region, suggesting its utility for management of marine ecosystems and highlighting rapid advances that can be made through collaboration between empirical scientists and ecosystem modelers.