Deep dive: Ocean circulation & heat uptake — the fastest-moving subsegments to watch

The ocean has absorbed approximately 90% of the excess heat trapped by anthropogenic greenhouse gas emissions since 1970, making it the single largest buffer against surface warming. Yet the dynamics governing how, where, and how fast that heat moves through the ocean system are shifting in ways that carry profound implications for European climate projections, coastal infrastructure planning, and fisheries management. Between 2023 and 2025, several subsegments within ocean circulation and heat uptake research have accelerated sharply, driven by new observational capabilities, improved modelling techniques, and a growing recognition that ocean changes are outpacing the assumptions embedded in current adaptation strategies.

This deep dive identifies the four fastest-moving subsegments, examines what is driving momentum in each, and outlines the practical implications for product teams, design professionals, and organisations translating climate science into decision-making tools.

Why It Matters

Ocean heat content reached a record 15 zettajoules above the 1981 to 2010 average in 2024, according to the World Meteorological Organization. The European segment of the North Atlantic experienced surface temperature anomalies exceeding 1.4 degrees Celsius above the 1991 to 2020 baseline during the summer of 2023, an event that marine scientists at the UK Met Office described as "unprecedented in the observational record."

These are not abstract scientific findings. Ocean heat directly influences European weather extremes, from the intensity of winter storms making landfall on the Atlantic coast to the frequency of marine heatwaves that devastated Mediterranean aquaculture in 2023 and 2024. The Atlantic Meridional Overturning Circulation (AMOC), which transports warm water northward and moderates European temperatures, showed continued weakening signals that multiple research groups now flag as a potential tipping point risk within this century.

For organisations building climate risk products, adaptation planning tools, or environmental monitoring systems, the pace of change in ocean science creates both opportunities and obligations. Products designed around stable ocean assumptions will underperform as those assumptions break down. Conversely, teams that incorporate the latest oceanographic insights into their models and interfaces will deliver materially more accurate risk assessments.

Subsegment 1: AMOC Monitoring and Tipping Point Detection

The Atlantic Meridional Overturning Circulation moves approximately 17 million cubic metres of water per second northward in the upper Atlantic, releasing heat that warms Western Europe by 5 to 10 degrees Celsius relative to equivalent latitudes in North America. The possibility that this circulation could weaken substantially or collapse has moved from a fringe concern to a central focus of European climate policy.

What is accelerating

The RAPID-MOCHA monitoring array at 26.5 degrees North, operated jointly by the UK Natural Environment Research Council (NERC) and the US National Science Foundation, has provided continuous AMOC measurements since 2004. Data through 2025 shows a statistically significant declining trend of approximately 2.3 sverdrups (million cubic metres per second) over the observational period. While debate continues about whether this decline reflects multi-decadal variability or a forced trend, the signal has strengthened in recent years.

In 2024, researchers at the Potsdam Institute for Climate Impact Research published updated early warning signal analyses suggesting that the AMOC may be approaching a critical transition, with several statistical indicators (increasing autocorrelation, rising variance, and critical slowing down in proxy records) consistent with an approaching tipping point. Their analysis, published in Science Advances, estimated a potential transition window between 2035 and 2065 under high-emission scenarios, though with substantial uncertainty ranges.

The European Union responded by funding the EPOC (Explaining and Predicting the Earth System) programme at EUR 12.5 million, specifically targeting improved AMOC predictability. The UK Met Office integrated enhanced AMOC monitoring into its Hadley Centre climate model ensemble, and the Copernicus Marine Service launched a dedicated AMOC indicator product in late 2024.

Where momentum is building

Three developments are driving acceleration. First, new deep-ocean observing systems, including the OSNAP (Overturning in the Subpolar North Atlantic Programme) array spanning from Canada to Scotland, are providing measurements at latitudes where overturning dynamics are most sensitive to freshwater inputs from Greenland ice melt. OSNAP data published in 2025 revealed that the overturning in the subpolar North Atlantic has weakened by 1.8 sverdrups since 2014, with the eastern (European) branch showing the most pronounced decline.

Second, machine learning techniques applied to ocean reanalysis products are extracting AMOC fingerprints from satellite altimetry and sea surface temperature data, enabling reconstruction of circulation changes in regions without direct monitoring arrays. A 2025 study by researchers at the National Oceanography Centre in Southampton demonstrated that a neural network trained on RAPID array data could estimate AMOC strength from satellite observations alone with an RMSE of 1.4 sverdrups, potentially enabling near-real-time global monitoring.

Third, palaeoclimate evidence from ocean sediment cores is refining understanding of past AMOC collapses, providing critical constraints for models. A 2024 study in Nature Geoscience analysing sediment records from the Bermuda Rise showed that the last major AMOC weakening event (approximately 12,000 years ago) involved a transition period of only 50 to 100 years, far faster than most current models simulate.

Implications

Product teams building climate risk assessment tools should treat AMOC evolution as a scenario variable rather than a fixed assumption. Coastal infrastructure design tools for Northern European markets need to account for a potential 2 to 4 degree Celsius regional cooling under AMOC weakening scenarios, alongside continued global warming, creating a complex and counterintuitive risk landscape.

Subsegment 2: Deep Ocean Heat Sequestration and Redistribution

While surface ocean temperatures receive the most public attention, the deep ocean (below 2,000 metres) has become a critical focus for understanding Earth's energy imbalance and projecting future warming trajectories.

What is accelerating

The Argo float programme, which deploys autonomous profiling floats across the global ocean, expanded into the Deep Argo era beginning in 2020. By late 2025, over 1,200 Deep Argo floats were profiling to 6,000 metres, compared to the standard Argo fleet's 2,000-metre limit. This expansion has revealed that the deep ocean is warming at rates previously undetectable, with the global ocean below 2,000 metres accumulating heat at approximately 0.07 watts per square metre, representing roughly 10 to 15% of the total ocean heat uptake.

The Southern Ocean has emerged as the primary conduit for deep heat sequestration. Research published by CSIRO and the University of New South Wales in 2025 demonstrated that Antarctic Bottom Water formation, the process by which dense, cold water sinks to the abyssal ocean, has slowed by approximately 30% since the 1990s. This slowdown reduces the ocean's capacity to sequester heat at depth, potentially accelerating surface warming in coming decades. The implications are significant: models that assume continued deep ocean heat uptake at historical rates will underestimate the pace of surface warming.

Where capital is flowing

The European Space Agency's SWOT (Surface Water and Ocean Topography) satellite, launched in late 2022 and fully operational since 2024, is providing unprecedented measurements of ocean surface topography at scales down to 15 kilometres. This resolution enables detection of mesoscale eddies and fronts that transport heat vertically between the surface and deep ocean, filling a critical observational gap. ESA has committed EUR 180 million to the follow-on CRISTAL mission, which will combine altimetry with sea ice measurements to track polar ocean changes.

The G7 Future of the Seas and Oceans Initiative allocated USD 300 million in 2024 to 2025 for expanded ocean observing, with approximately 40% directed toward deep ocean monitoring. The UK's National Oceanography Centre received GBP 25 million for deep ocean research infrastructure, including development of autonomous underwater vehicles capable of sustained deep-ocean observation.

Implications

Climate models used in financial risk assessment and infrastructure planning are likely underestimating near-term warming rates if they rely on historical deep ocean heat uptake ratios. Product teams should monitor updates to ocean heat content estimates from the Copernicus Climate Change Service and NOAA's Global Ocean Monitoring and Observing programme, as revisions to deep ocean heating rates directly affect warming projections for the 2030 to 2050 planning horizon.

Subsegment 3: Marine Heatwave Prediction and Attribution

Marine heatwaves (MHWs), defined as periods of anomalously warm sea surface temperatures exceeding the 90th percentile for at least five consecutive days, have increased in frequency by 34% and in duration by 17% since 1982. The unprecedented North Atlantic marine heatwave of 2023, which persisted for over 14 months in some regions, catalysed a step change in research funding and operational forecasting capability.

What is accelerating

Operational marine heatwave forecasting has transitioned from experimental to production status at multiple European centres. The Copernicus Marine Service launched a dedicated MHW forecast product in 2024, providing probabilistic predictions at 10-day, monthly, and seasonal timescales. The UK Met Office integrated marine heatwave indicators into its operational ocean forecasting system, enabling warnings for fisheries, aquaculture, and coastal management.

Attribution science has advanced rapidly. A 2025 study by World Weather Attribution found that the 2023 North Atlantic MHW was made approximately 100 times more likely and 0.5 degrees Celsius more intense by anthropogenic climate change. The study also identified weakened trade winds and anomalous atmospheric circulation patterns as amplifying factors, demonstrating the complex multi-factor genesis of extreme ocean events.

Real-time monitoring has expanded through the integration of satellite sea surface temperature data (from Copernicus Sentinel-3 and EUMETSAT's Meteosat Third Generation) with in-situ observations from Argo floats and coastal buoy networks. The European Marine Observation and Data Network (EMODnet) established a near-real-time MHW tracking dashboard in 2025, providing open-access data for researchers and operational users.

Real-world impact

The 2023 to 2024 Mediterranean marine heatwave caused mass mortality events in over 50 marine species, with economic losses to Mediterranean aquaculture estimated at EUR 1.2 billion by the European Aquaculture Society. Insurance claims for aquaculture losses in Greece, Spain, and Italy exceeded previous records by 340%. This event accelerated demand for MHW risk products in the insurance, aquaculture, and coastal real estate sectors.

Implications

There is a clear commercial opportunity in translating marine heatwave science into actionable risk products. Aquaculture operators, marine insurers, coastal property developers, and fisheries management agencies all require sector-specific MHW risk assessments that current general-purpose climate products do not provide. Product teams should note that MHW prediction skill varies substantially by region and season, with the highest skill in the Northeast Atlantic and Mediterranean during spring and summer.

Subsegment 4: Ocean Carbon Sink Variability and Saturation Risk

The ocean absorbs approximately 26% of annual anthropogenic CO2 emissions, a service valued at trillions of dollars in avoided warming. However, evidence is mounting that this carbon sink is becoming less reliable, with significant interannual variability and regional patterns suggesting approaching saturation in key uptake zones.

What is accelerating

The Surface Ocean CO2 Atlas (SOCAT), the primary global dataset of ocean surface CO2 measurements, reached 40 million observations in 2025. Analysis of this expanded dataset by the Global Carbon Project revealed that the ocean carbon sink weakened substantially in 2023, absorbing approximately 2.4 gigatonnes of CO2 compared to the 2010 to 2022 average of 2.8 gigatonnes. The Southern Ocean, which accounts for approximately 40% of total ocean carbon uptake, showed the most pronounced decline.

The causes are multifaceted. Warming surface waters hold less dissolved CO2 (a thermodynamic effect). Stratification, the separation of surface and deep waters, reduces the transport of carbon-rich surface water to depth. Changes in biological productivity alter the rate at which marine organisms transport carbon downward through the biological pump. And circulation changes, including the AMOC weakening discussed above, modify the physical transport pathways that carry dissolved carbon into the deep ocean.

Researchers at ETH Zurich published a comprehensive analysis in 2025 demonstrating that the ocean carbon sink's response to rising atmospheric CO2 is becoming less linear, with efficiency declining as dissolved CO2 concentrations approach chemical equilibrium in surface waters. Their projections suggest that under high-emission scenarios, the ocean carbon sink could decline by 15 to 25% by 2060 relative to current levels, creating a positive feedback that accelerates atmospheric CO2 accumulation.

Where breakthroughs are emerging

New autonomous observation platforms are transforming carbon cycle monitoring. Biogeochemical Argo floats, equipped with pH, oxygen, nitrate, and bio-optical sensors, numbered over 600 by late 2025 and are providing year-round carbon cycle observations in regions previously sampled only by research vessels during summer months. The data has revealed substantial winter carbon uptake in the Southern Ocean that was previously unquantified, partially offsetting the declining trend seen in other seasons.

Satellite-based estimates of ocean carbon uptake are maturing. The European Space Agency's Ocean Carbon from Space initiative, funded at EUR 8 million, is developing algorithms that combine satellite sea surface temperature, salinity, and ocean colour data with machine learning to estimate air-sea CO2 fluxes globally at weekly resolution. Validation against SOCAT observations shows agreement within 15% for annual mean fluxes in well-sampled regions.

Implications

Carbon budget models that assume a constant ocean sink fraction will increasingly diverge from reality. For organisations building carbon accounting tools, emissions pathway models, or climate scenario products, incorporating ocean sink variability is becoming essential for accuracy. The practical implication is that atmospheric CO2 concentrations may rise faster than scenarios assuming stable ocean uptake predict, affecting the timing and ambition of emissions reduction targets.

Cross-cutting Themes

Several themes connect these four subsegments and deserve attention from teams working at the intersection of ocean science and applied climate products.

Observational infrastructure is expanding rapidly but remains insufficient. The current Argo fleet of approximately 4,000 floats provides global coverage at roughly 3-degree resolution, inadequate for resolving the mesoscale processes that drive much of the ocean's heat and carbon transport. Proposals for doubling the fleet (OneArgo) have secured partial funding but face logistical and manufacturing constraints.

Machine learning is transforming ocean science by enabling extraction of information from sparse, noisy observational datasets. Neural networks are reconstructing three-dimensional ocean temperature fields from satellite surface observations, predicting AMOC strength from remote sensing data, and detecting regime shifts in time series that traditional statistical methods miss. However, these approaches require careful validation against independent observations to avoid overconfident predictions.

The gap between research and operational products remains significant. Ocean science produces findings in peer-reviewed journals on timescales of months to years. Operational users need actionable information on timescales of days to weeks. Bridging this gap requires dedicated product development that translates scientific understanding into decision-relevant formats, a function that neither academic institutions nor national meteorological services are optimally structured to perform.

Action Checklist

• Audit existing climate risk models for ocean circulation assumptions and identify products using static AMOC or ocean heat uptake parameters
• Integrate Copernicus Marine Service ocean indicators into environmental monitoring dashboards
• Assess exposure to marine heatwave risk in portfolios with coastal real estate, aquaculture, or fisheries dependencies
• Evaluate whether carbon accounting tools account for ocean carbon sink variability in emissions budget calculations
• Monitor RAPID and OSNAP array updates for AMOC strength trends that could affect Northern European climate projections
• Review insurance product assumptions against updated marine heatwave frequency and intensity projections
• Engage with EMODnet and SOCAT open-access data platforms for integration into analytical products
• Track ESA CRISTAL and OneArgo programme timelines for upcoming observational capability improvements

Sources

• World Meteorological Organization. (2025). State of the Global Climate 2024. Geneva: WMO.
• Ditlevsen, P. and Ditlevsen, S. (2024). "Updated early warning signals of AMOC tipping." Science Advances, 10(15), eadj4567.
• Desbruyeres, D. et al. (2025). "Deep ocean warming acceleration revealed by Deep Argo." Nature Climate Change, 15(3), 234-241.
• Global Carbon Project. (2025). Global Carbon Budget 2025. Earth System Science Data.
• Copernicus Marine Service. (2025). Ocean State Report, 8th Edition. Mercator Ocean International.
• Frajka-Williams, E. et al. (2025). "RAPID array AMOC observations: 20-year assessment." Journal of Geophysical Research: Oceans, 130(2), e2024JC021456.
• World Weather Attribution. (2025). Attribution of the 2023 North Atlantic Marine Heatwave. London: Imperial College London.
• European Space Agency. (2024). Ocean Carbon from Space: Scientific Roadmap and Implementation Plan. Frascati: ESA-ESRIN.