Assessing Air-sea CO2 Flux Products with Constraints from Atmospheric Inverse Analyses

Abstract

Oceanic uptake of atmospheric CO2 has been estimated to remove ~25% of  global anthropogenic emissions in recent decades (1990-2020), and this flux displays significant decadal variability (Global Carbon Budget: Friedlingstein et al. 2022). Key contributors to this ocean CO2 flux estimate are data products derived from interpolated surface ocean pCO2 measurements combined with air-sea gas-exchange parameterizations.  Although derived from a common surface ocean CO2 database (Surface Ocean CO2 Atlas (SOCAT), Bakker et al. 2016), these data products display variations on regional and global scales due to differences in their underlying construction methodologies. Here we assess three widely cited air-sea CO2 flux products, namely, Landschutzer et al. (2016), Roedenbeck et al. (2014) and Watson et al. (2020).  Our assessment uses the GEOSChem-LETKF data assimilation system (Chen et al. 2020), together with atmospheric CO2 observations from the NOAA-ESRL global network of surface sites (Obspack_CO2_Globalviewplus, 2020).  The individual air-sea flux products are implemented as alternative representations of the ocean prior flux, and we derive optimized estimates of surface CO2 fluxes in a set of atmospheric inverse analyses with the GEOSChem-LETKF system. We assess the performance of the individual ocean flux products on regional and global scales using a range of metrics derived from the atmospheric inversions including model concentration bias, CO2 flux error reduction, and comparison to independent atmospheric measurements from surface sites and aircraft. We also compare the derived posterior ocean fluxes to estimates from global ocean biogeochemistry models and discuss the implications for closure of the global carbon budget.