Impacts of physical data assimilation on the Global Ocean Carbonate System

Abstract

Prognostic simulations of ocean carbon distribution are largely dependent on an adequate representation of physical dynamics. In this work we show that the assimilation of temperature and salinity in a coupled ocean-biogeochemical model significantly improves the reconstruction of the carbonate system variables over the last two decades. For this purpose, we use the NEMO ocean global circulation model, coupled to the Biogeochemical Flux Model (BFM) in the global PELAGOS configuration. The assimilation of temperature and salinity is included into the coupled ocean-biogeochemical model by using a variational assimilation method. The use of ocean physics data assimilation improves the simulation of alkalinity and dissolved organic carbon against the control run as assessed by comparing with independent time series and gridded datasets. At the global scale, the effects of the assimilation of physical variables in the simulation of pCO2 improves the seasonal cycle in all basins, getting closer to the SOCAT estimates. Biases in the partial pressure of CO2 with respect to data that are evident in the control run are reduced once the physical data assimilation is used. The root mean squared errors in the pCO2 are reduced by up to 30% depending on the ocean basin considered. In addition, we quantify the relative contribution of biological carbon uptake on surface pCO2 by performing another simulation in which biology is neglected in the assimilated run.