Assimilation of pCO2 data into a global coupled physical-biogeochemical ocean model

Abstract

As part of the CARBOOCEAN project, measurements of the partial pressure of carbon dioxide (pCO2) in the ocean have been collected since 2001. Our aim was to use these data in a data assimilation context to improve the representation of pCO2 and its associated prognostic variables in biogeochemical models. Based upon a sequential framework a new method for the assimilation of pCO2 data was developed. This method had to account for oceanic pCO2 being a derived quantity with a functional dependence on temperature, salinity, dissolved inorganic carbon (DIC), and alkalinity. Our method calculates pCO2 increments using an analysis correction technique and then converts these into increments in model variables. Temperature and salinity are assumed to be error free, and so only DIC and alkalinity are updated. Furthermore, in DIC and alkalinity space increments are taken to be perpendicular to the local line of constant pCO2. Our method was tested by assimilating pCO2 data from 2006 into the NEMO-HadOCC biogeochemical model and then comparing the output from this model to a control with no pCO2 assimilation. While lack of data prevented the assimilation system from correcting a significant systematic error in the subpolar North Atlantic, results from this experiment showed that pCO 2 assimilation, despite limited data, reduced RMS and mean errors and also brought the model closer to climatology in the subtropical North Atlantic.