Daniel Ford and Jamie Shutler have led a new study published in Nature Scientific Data (https://doi.org/10.1038/s41597-026-07532-5), which delivers a recalculation of the widely used Surface Ocean carbon dioxide (CO2) Atlas (SOCAT), a database of measurements that is updated each year and needed to assess the amount of carbon the ocean absorbs each year. The recalculation calibrates the dataset to a satellite climate data record of sea temperatures and is needed to enable a more accurate calculation of the air to sea movement, or flux, of CO2. This work was only possible by a collaboration between the satellite data experts, data scientists and in situ field scientists combined with the European Space Agency Climate Change Initiative (ESA-CCI) sea surface temperature (SST) climate data record.
The CO2 observations within the SOCAT dataset are measured at a multitude of depths (ranging from ~10 m to ~0.5 m) but these are generally not representative of the conditions at the surface layer (~0.02 m) where CO2 exchange occurs with the atmosphere. Using the European Space Agency Climate Change Initiative sea surface temperature (ESA-CCI) climate data record the SOCAT dataset (~41 million CO2 observations) are recalculated to the conditions at 0.2 m. This process harmonises the depth of the SOCAT observations and makes these observations more optimal for studying the air-sea exchange. Additionally, within the SOCAT record an artificial warm bias has been identified which occurs due to sampling intricacies including (1) warming of seawater within a ship before the measurement is taken, or (2) mis-calibrated temperature sensors. The impact of this artificial warm bias signal is identified in Figure 1.
Figure 1: Global SOCATv2025 warm bias assessment showing an increasing warm bias since 2000 (following orange solid line). Panel B: Global latitudinal variation in the warm bias.
Using the ESA-CCI sea surface temperature record as a reference temperature we can now correct for these temperature errors, and it shows that the global ocean carbon sink is larger by ~0.4 Pg C yr-1 in 2024 (see Figure 2). This removal of a bias within the SOCAT dataset improves the ocean sink estimates that are based on SOCAT data, with direct affects when they are used to balance global carbon budgets. Consequently, the recalculated SOCAT dataset will now be provided each year to all ocean teams submitting data to the Global Carbon Budget assessments. All data are freely accessible via the FluxEngine GitHub. The methodology for recalculating the SOCAT dataset is described in Ford et al. (2026). This work was supported by the European Space Agency through the project Ocean Carbon for Climate.
Figure 2: Resulting global ocean CO2 uptake using the SOCAT dataset as provided (Original SOCAT; dark blue line) and using the Recalculated SOCAT dataset (Teal line).
Manuscript: Ford, D.J., Shutler, J.D., Holding, T. et al. Recalculating the Surface Ocean CO2 Atlas (SOCAT) to a sea surface temperature climate data record. Sci Data (2026). https://doi.org/10.1038/s41597-026-07532-5