Max Planck Institute for Meteorology: How to Make Model Experiments on the Atlantic Overturning Circulation More Realistic

Melting ice in the Arctic is causing an increasing amount of freshwater to enter the North Atlantic, which is expected to result in a weakening of the Atlantic overturning circulation. However, many modeling studies make unrealistic assumptions about how this water enters the ocean. A new study shows that the timing, location, and source of freshwater input can have a considerable impact on its eventual fate and should therefore be taken into account in future model experiments.

Map of the North Atlantic. Continents are black, and the ocean is colored with a blue-to-yellow color code. — Salinity in the North Atlantic (increasing from yellow to blue), simulated using the ICON climate model. Credit: EERIE/ICON/MPI-M

The Atlantic Meridional Overturning Circulation (AMOC) is considered the heating system of northwestern Europe. Driven by differences in water temperature and salinity, it carries warm water from the tropics northward and cold deep water back to the south. However, under global warming, the sinking of cold, salty water masses at high latitudes is a vulnerable point: Melting of ice in the Arctic is causing an increasing amount of freshwater to enter the North Atlantic, which dilutes the water masses and impedes their sinking to the deep. This is expected to weaken the AMOC, with far-reaching consequences for Europe and the global climate.

Many studies investigating this weakening are based on so-called hosing experiments: model simulations in which freshwater is added to the entire North Atlantic to simulate ice melt and investigate its effect on the AMOC. A new study by Fraser Goldsworth of the Max Planck Institute for Meteorology highlights the limitations of these experiments and ways to make them more realistic in the future.

Novel framework reveals the weaknesses of previous hosing experiments

“Many hosing experiments just dump freshwater into the North Atlantic without considering when and where exactly the freshwater input occurs, or what its sources are,” says Goldsworth. However, this can make a big difference because not all freshwater entering the North Atlantic automatically becomes part of the AMOC. To obtain a more realistic picture, Goldsworth applied a concept developed for estuarine studies to the entire Greenland coast. Unlike conventional approaches, which budget freshwater transport using an arbitrary reference value for salinity, the novel framework tracks the transformation of freshwater masses based on the conservation of their total salt content.

Goldsworth applied the framework to a simulation performed with the coupled ICON climate model, which has a horizontal resolution of five kilometers in the ocean and distinguishes between 72 depth levels therein. The oceanographer examined four different regions around Greenland. “The big surprise was how much the processes vary with season and region,” says Goldsworth. For instance, in summer, when Arctic ice melt is strong, freshwater input tends to join near-surface estuarine currents. Conversely, in winter, it is more likely to interact with the AMOC. Experiments that do not account for this seasonality tend to overestimate the proportion of freshwater entering the AMOC, and consequently the AMOC’s sensitivity to freshwater input.

The analysis revealed that wintertime mixing is particularly strong south of Greenland. “This is probably because of a lack of insulating sea ice, which allows the surface water to become very cold, sink and then mix with salty waters” says Goldsworth. The study thus hints at physical mechanisms that need to be investigated, and points out ways to make future hosing experiments more realistic.

Original publication

Goldsworth, F. (2026). 'A novel framework for studying oceanic freshwater transport and its application in discerning the modelled fate of freshwater around the coast of Greenland'. Ocean Modelling, 199: 102599. DOI: 10.1016/j.ocemod.2025.102599

Contact

Dr. Fraser Goldsworth
Max Planck Institute for Meteorology
fraser.goldsworth@we dont want spammpimet.mpg.de

Back