BIOGEOCHEMICAL (BGC) ARGO IMPROVES UNDERSTANDING AND QUANTIFICATION OF THE OCEAN’S BIOLOGICAL CARBON PUMP

Abstract

The ocean’s biological carbon pump (BCP) contributes to the sequestration of organic carbon in the deep ocean and influences atmospheric CO2 levels. Despite its significance, the BCP remains poorly characterized due to insufficient observations. Recent advances of Biogeochemical-Argo (BGC-Argo) program have greatly increased the availability of biogeochemical observations including those which contain a wealth of information about the BCP. Therefore, the application and interpretation of these new observations warrants further investigation. This thesis presents applications of BGC-Argo data to improve estimates and mechanistic understanding of the BCP by exploring their synergies with biogeochemical models, and by performing comprehensive analyses of high-frequency BGC-Argo observations. Specifically, I carried out model optimizations using different combinations of satellite chlorophyll and BGC-Argo observations. Results show that the inclusion of BGC-Argo profiles of multiple biogeochemical properties in the parameter optimization greatly improved the model’s representation of subsurface biological distributions and vertical carbon flux. Since the availability of BGC-Argo profiles is so far insufficient for sequential data assimilation in most regions, multivariate data assimilation of satellite observations was applied to a coupled physical-biogeochemical model. Repeating data assimilation experiments by using an alternative light parameterization that had been a priori calibrated with BGC-Argo profiles showed that a well-calibrated model with accurate parameterizations is fundamental to data assimilation. This motivated me to investigate to what extent BGC-Argo data can help in distinguishing different parameterizations of vertical carbon flux. I set up a 1D model framework and calibrated it using BGC-Argo data to compare common parameterizations of vertical carbon flux in the same model environment. Results show great potential for BGC-Argo to inform vertical flux parameterizations and in determining the associated parameter values. Finally, I applied a new method, based on the mass balance of particulate organic carbon and oxygen, to BGC-Argo data to estimate vertical carbon flux due to different mechanisms. Results show that, in addition to the gravitational sinking flux of large particles, small particles make a significant contribution to the vertical flux at 100 m due to multiple mechanisms and at 600 m due to fragmentation of large particles.