Version 1.0 (updated 2025-09-02)

This dataset provides high-resolution monthly surface ocean fugacity of CO₂ (fCO₂) and derived sea–air CO₂ fluxes for the global ocean over the period 2003–2022.<b><i>fCO2_monthly_mean_XXXX-XXXX_4km.nc:</i></b>These files contain monthly mean surface ocean fCO₂ at 4 km resolution, stored in 5-year intervals (2003–2007, 2008–2012, 2013–2017, 2018–2022). Values are given in 10⁻¹ μatm. The estimates were generated using a machine learning model trained with in situ measurements from the SOCAT database, together with MODIS satellite images and several environmental parameters (see Methods in the associated paper). Land and data-void (no observation) grid cells are masked with zero values.<b><i>CO2_Flux_1M_1deg_2003-2022.nc:</i></b>This file provides monthly sea–air CO₂ fluxes (mol m^-2 yr^-1) at 1° spatial resolution for 2003–2022. The fluxes were computed using flux-related parameters together with resampled fCO₂ fields (<b><i>fCO2_1M_1deg_2003-2022.nc</i></b>, in μatm) to ensure consistency of spatial resolution in the input data.<b><i>fCO2_1M_1deg_2003-2022_coarse-inputs.nc:</i></b>For comparison, we also provide a coarse-resolution (1° grid) fCO₂ product generated with the same machine learning framework but using coarser input parameters, including the ESA OC-CCI v6 gap-filled dataset. This product facilitates evaluation against previous studies and coarse-resolution reconstructions.