Abstract. Given the role of the ocean in mitigating climate change through CO2 absorption, it is important to improve our ability to quantify the historical ocean CO2 uptake, including its natural variability, for carbon budgeting purposes. In this study we present an exhaustive intercomparison between two ocean modelling practices that can be used to reconstruct the historical ocean CO2 uptake. By comparing the simulations to a wide array of ocean physical and biogeochemical observational datasets, we show how constraining the ocean physics towards observed temperature and salinity results in a better representation of global biogeochemistry. We identify the main driver of this improvement to be a more realistic representation of large scale meridional overturning circulation together with improvements in mixed layer depth and sea surface temperature. Nevertheless, surface chlorophyll was rather insensitive to these changes, and, in some regions, its representation worsened. We identified the causes of this response to be a combination of a lack of robust parameter optimization and limited changes in environmental conditions for phytoplankton. We conclude that although the direct validation of CO2 fluxes is challenging, the pervasive improvement observed in most aspects of biogeochemistry when applying data assimilation of observed temperature and salinity is encouraging; therefore, data assimilation should be included in multi-method international efforts aimed at reconstructing the ocean CO2 uptake.