Global Surface Ocean CO2 Partial Pressure Estimates (CSIR-ML6 v2021)

12652100 data points provide surface ocean partial pressure of CO2 (pCO2) estimates using the updated CSIR-ML6 method. The dataset integrates SOCAT observations with ERA5 meteorological data, created by Pedro M S Monteiro and updated in 2021.

Use Cases

• Calibrate and validate ocean carbon flux models using the pCO2 estimates derived from SOCAT and ERA5 data.
• Analyze spatial and temporal trends in marine carbonate chemistry using the machine learning-derived pCO2 fields.
• Assess the impact of updated ERA5 wind and MSLP data on pCO2 predictions compared to previous method versions.
• Incorporate the pCO2 estimates into Earth system models to improve projections of ocean carbon uptake.

Strengths

• Contains 12652100 individual data points for analysis.
• Updated methodology uses higher-quality ERA5 wind and mean sea level pressure data.
• Explicitly extends temporal coverage through the end of 2020 for SOCATv2021.
• Released under a permissive CC BY 4.0 license for wide reuse.

Limitations

• The exact column structure and variable definitions are not provided in the input.
• The dataset size is reported as 0.0 B, indicating a potential metadata issue or extremely compact format.
• As a model-derived product, it contains uncertainties inherent to the CSIR-ML6 method and its input data.

Provenance

Source

Derived from SOCAT (Surface Ocean CO2 Atlas) v2020 and v2021 data.

Collection Method

Generated using the CSIR-ML6 machine learning method, updated with ERA5 meteorological reanalysis.

Time Range

Extends through the end of 2020 for the SOCATv2021 component.

Freshness

Methodology updated to v2021; metadata record last updated in March 2026.

Geography

Global surface ocean coverage.