The ocean now absorbs roughly a quarter of all human CO₂ emissions, making it both a critical climate buffer and a system under accelerating chemical stress.
A new peer-reviewed paper synthesis catalogues 68 existing data products that track changing ocean chemistry — spanning raw cruise observations, long-term buoy time-series, machine-learning gridded maps, and century-scale model projections.
This paper provides the most comprehensive publicly available catalogue of ocean carbonate chemistry data products, giving researchers, managers, and policymakers a single reference point for identifying which datasets are appropriate for their specific monitoring, modelling, or policy needs. In practice, teams working on ocean acidification impacts should begin with SOCAT and GLODAPv2 as foundational observation layers, select surface gridded products based on their regional focus and temporal resolution requirements, and treat ML-derived products as gap-filling tools rather than independent observations.
This is a review and catalogue paper, not an empirical study. The authors conducted a structured literature review supplemented by community consultation via the Ocean Acidification Information Exchange (OAIE) platform. Products were identified, described, and organized into six categories based on end-user needs: cruise compilations, time-series, surface gridded, interior gridded, multi-product analyses, and model/hybrid outputs. For each product, the authors document spatial and temporal coverage, methodology, key variables, highlights, and access links. No new data were generated. No quantitative comparison of product performance was conducted.
Check our interactive tools here: https://saveocean.net/interactive-tools/
Key findings:
The global ocean carbonate chemistry monitoring system now encompasses:
- 68 data products spanning observational compilations dating to 1957,
- gridded ML-based reconstructions extending back to 1982,
- interior ocean climatologies referenced to year 2000–2002,
- and forward projections to 2100 and beyond.
Two foundational datasets:
- SOCAT (40 million surface measurements)
- GLODAPv2 (rigorous interior bottle data)
Machine learning has become the dominant methodology for surface gridded products, with at least 12 distinct algorithmic approaches now applied globally.
Key gaps identified include sparse coastal coverage:
- Limited Southern Ocean and Arctic winter data,
- Near-total absence of data from the Global South.
Limitations
The paper does not quantitatively compare product accuracy, uncertainty, or fitness-for-purpose across different research applications.
It does not address data latency, pipeline fragility, or succession planning for key datasets.
The coastal ocean is structurally underrepresented.
The mCDR monitoring use case is introduced without acknowledging that none of the products were designed or validated for it.
ML methods all achieve “roughly the same uncertainty”. It is evidence that all methods are hitting the same wall (the observational gap itself) and no amount of algorithmic sophistication can substitute for data that was never collected.
The authorship and data coverage both reflect a geographic concentration in North America and Europe.
The scientific community studying ocean carbonate chemistry is not representative of the nations most vulnerable to ocean acidification impacts.
Check the products registry with the markings of the limitations here:
Full link to the paper here.
Recommendations:
Any new ocean monitoring programme should integrate with SOCAT (surface CO₂) and GLODAPv2 (interior chemistry) rather than creating parallel data streams. Compatibility with these standards dramatically increases the value of new observations.
Open-ocean products are more coverred then the coastal areas. The most significant unmet policy need — and the largest return on investment — lies in expanding coastal ocean data products, where acidification most directly threatens fisheries, aquaculture, and food security for vulnerable communities.
You can also find the live product registry at: oceanco2.github.io/co2-products
