An Evaluation of Global Ocean Models in the North Atlantic using BGC-Argo float Observations

Abstract

Testing, evaluating, and enhancing global ocean models holds paramount significance in climate and oceanographic research, particularly for projecting climate change impacts. While the evaluation of physical model components has become increasingly comprehensive and sophisticated, an evaluation of the biogeochemical (BGC) components has so far been difficult to accomplish because of a lack of extensive global-scale BGC observations. In this thesis, North Atlantic (BGC-)Argo profile data of chlorophyll-a, nitrate, and oxygen from the surface to 2000 m depth are compared to the corresponding properties simulated by state-of-the-art global ocean models of the CMIP6 ensemble, as well as Mercator Ocean’s data assimilating models MOI-BIO4 and MOI-GLO12. The physical variables salinity and temperature are included to investigate relationships between model physics and biogeochemistry. World Ocean Atlas (WOA) climatologies serve as a secondary dataset for model assessment. They are employed to validate the methodology of utilizing (BGC-)Argo data for model evaluation and to conduct comparisons with prior studies that have emphasized on WOA analyses.

When evaluating BGC model complexity, resolution, and performance, it's evident that neither heightened BGC complexity nor increased model resolution independently guarantee improved performance.