AI and Argo float data can support AMOC observation
A study by the GEOMAR Helmholtz Centre for Ocean Research Kiel and Kiel University introduces a more efficient and reliable way to monitor the ocean using observation networks already in place. Published in the journal Ocean Science, the research shows how artificial intelligence, combined with data from the global Argo float programme, can help track key characteristics of the Atlantic Meridional Overturning Circulation (AMOC) – a complex current system with a major influence on Europe's climate.
The international Argo programme has been quietly transforming our understanding of the ocean for more than 20 years. At its core are around 4,000 autonomous profiling floats, drifting through the water column and periodically diving to a depth of 2,000m. As each float ascends, it records temperature, salinity and pressure along the way, then surfaces to beam the data via satellite into the Argo network – a resource freely available to researchers around the world. Few observation systems have reshaped ocean science as profoundly as this one.
Combining machine learning and physical models
A new study now shows that this data can be put to use far beyond its original purpose. "We wanted to know whether we could use the scattered Argo measurements to gain insights into large-scale circulation systems that, until now, could only be recorded through very extensive measurement campaigns," says Dr Yannick Wölker, lead author of the study and until recently a PhD student in the Ocean Dynamics research unit at GEOMAR and the research group Archaeoinformatics – Data Science at Kiel University. "Artificial intelligence opens up new possibilities here. Combining machine learning with established physical models allows us to get more out of existing measurement data and better understand how key circulation systems work."
Observational data and model calculations
At the centre of the study is the Atlantic Meridional Overturning Circulation, or AMOC. Acting like a giant conveyor belt, this circulation system carries warm surface water northwards, where it cools, sinks into the depths and returns south as cold deep water. Along the way, it moves vast amounts of heat and shapes weather and climate patterns, including in Europe. Just how stable the AMOC is, and how it is responding to climate change, remains one of the most pressing questions in climate research today – yet direct observations have so far relied on only a handful of fixed measurement series in the Atlantic, all of them technically demanding and costly to maintain.
To close this gap, the researchers brought together two approaches: observational data and model calculations. Using high-resolution ocean simulations, they trained a machine learning algorithm – a subfield of artificial intelligence in which computers learn to recognise patterns in data – to identify how typical ocean current patterns relate to temperature and salinity profiles. This trained model was then applied to real Argo data from the Atlantic, making it possible to infer large-scale current strengths from individual point measurements. In particular, this allowed the team to capture the so-called geostrophic component of the circulation – governed by the distribution of temperature and salinity – which has until now been notoriously difficult to measure on a continuous basis.
New building block for ocean observation
The resulting estimates align well with established observational series and model simulations, offering an important new methodological building block for ocean observation. The authors are nonetheless clear about the method's limits: it relies on model assumptions and captures only specific time windows, meaning very short-term fluctuations and the AMOC's long-term decline can only be tracked to a limited degree. Rather than replacing existing measurement systems, the new methodology is designed to complement and strengthen them.
“Our approach is no substitute for direct measurements in the ocean,” says Yannick Wölker. “But it can help to make better use of existing data and bridge gaps in observations.”
The method could play an important role in shaping future observation networks. It shows how global programmes such as Argo can gain even greater value through modern data analysis – and how additional ocean infrastructure could be deployed more strategically.
“Particularly with regard to long-term climate monitoring, we need to consider how to design observations that are efficient, robust and internationally coordinated,” says Prof Dr Arne Biastoch, a professor at GEOMAR and co-author of the study. “Artificial intelligence methods can help to ensure regarding future ocean measurements are made in the best possible way that strategic decisions.”
The work was carried out in collaboration as part of the Helmholtz School for Marine Data Science (MarDATA), a PhD programme involving researchers in Kiel and Bremen, in which PhD students are supervised by marine scientists and computer scientists jointly. The paper can be found here.
