CELLO: Climate Exploration in Lively Liaison with the Ocean
From an auditory perspective, the rain descending on the roof of the tent, where the poster exhibition is taking place, and the applause echoing through the wood-paneled lecture hall of Bucerius Law School bear some resemblance: a sonic surge that briefly intensifies before gradually dissipating. The name of the conference organized by the Max Planck Institute for Meteorology (MPI-M) has a ring to it as well: “Climate Exploration in Lively Liaison with the Ocean.” The acronym CELLO alludes to the deep tones characteristic of the string instrument—a metaphor for the ocean’s low-frequency response to the atmosphere’s high-frequency signals, which MPI-M founding director and Nobel Prize winner Klaus Hasselmann had revealed with the help of his stochastic climate models.
The institute’s first three-dimensional ocean circulation models had mainly been developed by Hasselmann’s colleague Ernst Maier-Reimer. A little over 20 years ago, scientists at the MPI-M, together with partners from the German meteorological service, created the ICON model, which has been optimized for the parallel computing structure of modern supercomputers. ICON’s ocean component is based on the work of Peter Korn and his colleagues. In the tradition of Hasselmann and Maier-Reimer, the model has been continuously developed further since then.
CELLO: The ocean conference in the anniversary year
Peter Korn and Nils Brüggemann jointly head the “Complex Modeling and Extreme Computing” working group at the MPI-M. Together with Jochem Marotzke, director of the “Climate Variability” department at the MPI-M, they have organized the CELLO conference. The institute hosted four international conferences on Earth system modeling in Hamburg in the past, the most recent one in 2017. In 2025, on the occasion of its 50th anniversary, the institute focuses the fifth edition of this event on the ocean, which was also the main interest of founding director Hasselmann.
The concept of the conference: Instead of splitting up into different parallel sessions, all 220 participants listen to the plenary talks together. This allows theorists, modelers, and researchers who work with observational data or conduct measurements to engage in lively discussions with each other. “The conference accommodates a wide range of different disciplines within oceanography,” says Raffaele Ferrari from the Massachusetts Institute of Technology, one of the keynote speakers at the conference. “Unlike very large conferences, here you have the chance to actually try and understand what the speakers present and discuss their research directly with them.”
New developments in ocean research
There are many issues to be addressed: Highly simplified concepts that ignore the ocean’s complex and dynamic nature are being called into question. Models with ever-higher spatial resolution now allow the physics of small-scale processes to be explicitly represented. This also brings the spatial scales of modeling and observations closer together. In her presentation on her latest scientific cruise, Eleanor Frajka-Williams from the University of Hamburg points out: “We are slowly getting into a region where we can actually compare data from high-resolution models to observations.”
In contrast to overturning circulations and large-scale currents, which can be described and studied on a large scale, small-scale processes are often more chaotic. Turbulence, i.e., eddies on different spatial scales, plays an important role in ocean mixing and energy transport. Raffaele Ferrari compares the ocean to a lung. In the lung, thin ducts in the alveoli facilitate an effective gas exchange. The ocean has similar fine structures: ocean fronts, where water masses with very different properties such as temperature and salinity meet. Coarse-resolution climate models, like those used for the reports published by the Intergovernmental Panel on Climate Change (IPCC), do not resolve these fine dynamics. However, the ocean’s respiration, i.e., the exchange of gases and heat between the atmosphere and the deep ocean, depends on small-scale turbulence forming at these fronts. Accordingly, a climate model that omits these small-scale processes is like a model of the lungs without alveoli. The workaround researchers have used are simplifications that attempt to capture the collective effect of these small-scale processes. But these parameterizations are imperfect. Parts of the dynamics are missing; a lively liaison with the ocean isn’t established.
High-resolution modeling, made possible by the enormous computing power of modern supercomputers, could enable important advances in this field. “With the help of high-resolution modeling, we can now finally treat the ocean as a turbulent fluid. This is a major difference to the work of Hasselmann and Maier-Reimer. We have opened a door and are excited to see what’s behind it,” says Peter Korn. Modeling ocean turbulence is, on the one hand, about turbulence within the ocean. “On the other hand, turbulence always occurs where components of the climate system interact with each other.” Polar researcher Céline Heuzé from the University of Gothenburg uses models and observational data in her research to study the boundary between sea ice, the ocean, and the atmosphere. “Sea ice formation and openings, both of which happen with polynyas, trigger a lot of turbulence. In turn, this turbulence triggers many crucial large-scale processes in the atmosphere and ocean, but also in relation to the biology and biogeochemistry of the system.”
A grown community
Participants have addressed the various aspects of ocean turbulence over three conference days from September 16 to 18: The first day focused on turbulence in the ocean. Day two explored the interactions between the atmosphere, the sea, and ice. And on the last day of CELLO, participants are exchanging ideas on new modeling approaches.
Reminiscent of the conference title, the visual presentation of some contributions is very lively. Several speakers show animations in which dots dart across the globe in fast motion, large-scale currents flow back and forth between continents, small eddies rotate across the screen, or waves wash from one side of the screen to the other. The discussions among the participants and the conversations over a cup of coffee during the breaks are lively, too. Many of the participants know each other well. “This community is like a family in the sense that we push each other in positive ways,” says David Battisti from the University of Washington, one of the keynote speakers at the event. The atmospheric scientist researches the interaction between ocean dynamics and atmospheric dynamics and their influence on natural climate variability on seasonal to decadal timescales. Understanding this variability and its determining factors, in turn, is key to improving seasonal-to-decadal climate predictability. In his presentation, Battisti shows that the atmosphere’s response to sea surface temperature anomalies in the mid-latitudes varies depending on model resolution, with a stronger response in high-resolution models. “There are observations, there is theory, and there are models that can resolve the small-scale processes. These elements are coming together, and there is some real potential here to make progress in predicting the climate on decadal timescales.”
The potential of new methods
Machine learning and artificial intelligence (AI) methods are being used for climate predictions as well. Peter Korn draws another parallel here with the work of the founding director of the MPI-M: “In the 1970s, Hasselmann developed a statistical method that can be used to distinguish the anthropogenic signal of global warming from natural climate fluctuations. Essentially, AI is also a statistical method. It remains to be seen whether it will also lead to such groundbreaking findings. But current research is definitely interested in understanding the method and its possibilities in more detail. Therefore, I am delighted that we have been able to gather several speakers on this topic here as well.”
From shipborne measurements of ocean movements to modeling tides, from theoretical calculations to AI: the conference contributions cover a wide range of topics, encompassing not only different disciplines and spatial scales, but also taking participants from the atmosphere to the deep sea and across all regions of the globe from Antarctica to the tropical Pacific to the North Atlantic. Atmospheric researcher Battisti describes what this feels like: “Attending the conference is like being a kid in a candy store. You are always going to learn something about how nature works.”
Further information
Contact
Dr. Nils Brüggemann
Max Planck Institute for Meteorology
nils.brueggemann@mpimet.mpg.de
Dr. Peter Korn
Max Planck Institute for Meteorology
peter.korn@mpimet.mpg.de
Prof. Dr. Jochem Marotzke
Max Planck Institute for Meteorology
jochem.marotzke@mpimet.mpg.de