Understanding Model Biases in Southern Ocean Carbon Seasonality

The Southern Ocean drives over 40% of global anthropogenic carbon dioxide (CO₂) uptake, yet climate models often struggle to capture its seasonality of carbon fluxes. This study investigates these biases using sensitivity experiments and Green's function-based parameter optimisation in the Biogeochemical Southern Ocean State Estimate (B-SOSE). The optimised model reduced errors in partial pressure of carbon dioxide (pCO₂) by up to 79% and improved correlation with observations by up to 92% in specific regions. A single set of parameters can adequately replicate an average annual pCO₂ cycle in the entire Southern Ocean. However, results show that phytoplankton growth and mortality parameters primarily control the seasonal timing of pCO₂ variability regionally, while alkalinity governs regional background pCO₂ magnitudes. The findings point to data assimilation as an effective way to identify key sources of model uncertainty, providing a pathway toward more accurate representation of Southern Ocean carbon cycling in Earth System Models.

Reference: Kuhn, A.M., Mazloff, M.R., Gille, S.T. & Verdy, A. (2026). Understanding regional pCO₂ model biases and uncertainties in the Biogeochemical Southern Ocean State Estimate (B‐SOSE). J. Geophys. Res.: Biogeosci., 131(5), e2025JG009491. https://doi.org/10.1029/2025JG009491