Global marine biogeochemical reanalyses assimilating two different sets of merged ocean colour products

Abstract

The only observations of marine biogeochemistry with routine global coverage are satellite ocean colour, which provide measurements of ocean bio-optical properties, with coverage still incomplete and limited to the sea surface. Models are therefore required to provide full spatial coverage, and through data assimilation can be combined with observations to create a reanalysis. This can then be used to investigate both observed and non-observed variables, including those relating to the carbon cycle. As part of the Climate Modelling User Group (CMUG) within the European Space Agency's Climate Change Initiative (CCI) project, two global marine biogeochemical reanalyses have been produced by assimilating ocean colour-derived chlorophyll data into a coupled physical-biogeochemical ocean model over the period September 1997 to July 2012. One reanalysis assimilated CCI products, the other assimilated GlobColour products, with a non-assimilative hindcast run for comparison. Each has been validated against independent in situ observations of chlorophyll, nutrients and carbon cycle variables. The assimilation of either source of ocean colour data was found to improve the model's representation of chlorophyll concentration throughout the water column, including the frequency and positioning of deep chlorophyll maxima. The assimilation also resulted in a slight improvement in nutrient concentrations and surface fugacity of carbon dioxide compared with in situ observations, although the overall impact on mean fields was small. This was found to be due in part to cancelling errors within the model, with the assimilation providing information on model biases, which can be used to inform future climate model development. The reanalyses were also able to reproduce expected seasonal cycles, as well as inter-annual variability related to major climate drivers. This study concludes that both CCI and GlobColour products are suitable for assimilation purposes, and that assimilating ocean colour data is of clear benefit.