Introduction:

Ocean surface partial pressure of carbon dioxide (pCO2) is essential for understanding the global carbon cycle,
monitoring ocean health, and predicting their future changes. However, the current observational network poses
a challenge for global ocean carbon cycle monitoring. Recent studies have used physical oceanographic datasets
and feed-forward neural networks to generate global pCO2 reanalysis products, but they are lack of considering
geographically varying physical variables-pCO2 relationships. To address this, this study developed SNU-pCO2, a
deep learning model integrating a climate model with transfer learning to generate a global ocean surface pCO2reanalysis. Given sparse observational data, SNU-pCO2 was first trained on Community Earth System Model
version 2 Large-Ensemble simulations to learn ocean surface pCO2 dynamics, then fine-tuned with reanalysis
fields and Surface Ocean CO2 Atlas (SOCAT) observations. A conservative SOCAT-based 10-fold cross-validation
showed that SNU-pCO2 achieved competitive predictive accuracy relative to existing data-driven pCO2 products,
with an average Pearson correlation coefficient of 0.91 and a root mean square error of 13.57 μatm. Independent
validation against Lamont-Doherty Earth Observatory and SOCAT Flag E observations further confirmed the
robust performance of SNU-pCO2, with RMSE values of 20.75 and 22.94 μatm, respectively. In particular, SNU-pCO2 exhibited superior performance in regions with sparse direct observations but relatively dense surrounding
data, effectively leveraging the influences of the spatially adjacent observations. The model also reproduced
physically consistent pCO2 anomaly patterns associated with El Ni ̃no and La Ni ̃na, accurately capturing dominant
physical drivers and phase-dependent processes across the equatorial Pacific. We additionally quantified the
uncertainty of the estimated pCO2 using the Monte Carlo Dropout. The uncertainty patterns aligned with RMSE
distributions, enabling both the assessment of prediction confidence and the identification of regions requiring
enhanced observational coverage and model refinement. This reanalysis supports studies of the carbon cycle, airsea CO2 fluxes, climate, and biogeochemical processes.

pdf : Cho et al.(2026)