Hatlen Heimdal et al. (2025): Quantifying errors in ocean carbon reconstructions in the context of marine carbon dioxide removal (mCDR)

Thea Hatlen Heimdal, Galen A McKinley, Abby P Shaum, Viviana Acquaviva, Amanda R Fay and Adrienne J. Sutton, IN: ESS Open Archive, DOI: 10.22541/essoar.175434006.69051349/v1

The ocean sequesters anthropogenic carbon without direct intervention through the background ocean carbon sink, a process with large variability in both time and space that complicates its quantification. A major challenge for the effective implementation of marine CDR (mCDR) technologies will be to quantify its additionality above background ocean carbon uptake. Here, the authors quantify reconstruction errors in one machine learning method for monthly mean surface ocean pCO₂ using a high-resolution (0.1°x0.1°) model as a testbed. They sub-sample this model based on the observational coverage of pCO₂ in SOCAT for years 2020-2021 and compare the pCO₂ reconstructions to the model full-field pCO₂ ‘truth’. They focus on three potential mCDR deployment areas: the Gulf, the Bering Sea and the Caribbean Sea.

LINK