Study Suggests Greenland Ice Melt is Weakening AMOC, but There is No Impending Tipping Point Seen: Using the SSP5-8.5 Scenario Makes the Results Questionable: Thermohaline Circulation Post #4

     The AMOC is part of a heat redistribution engine that moves heat along the thermohaline circulation currents in the Atlantic Ocean, which connects the deep, shallow, and surface parts of the ocean. The AMOC is the main region where warm tropical surface water sinks to depth. That water has higher salinity since warmer surface water experiences more evaporation, which makes it denser, causing it to sink, which happens mainly in a region between Greenland and Iceland.

     The current paper, published in Science Advances, explores the effects of Greenland meltwater on the AMOC now and in the future. According to the article about the study in Phys.org:

“A recent Intergovernmental Panel on Climate Change (IPCC) model found that none of the Coupled Model Intercomparison Project phase 6 (CMIP6) models show an abrupt AMOC collapse during the 21st century.”

     Other studies have suggested that Greenland meltwater does weaken the AMOC, but there is no indication in current data of abrupt, irreversible changes ahead to 2300, which is how far into the future the simulations go.

“However, it currently remains an open question whether these meltwater-induced AMOC changes are associated with tipping point characteristics such as abruptness and irreversibility. In addition, the physical mechanisms of how Greenland meltwater affects the AMOC beyond 2100 have, to our knowledge, not yet been explored," write the authors of the new study.”

     The simulations do suggest that the effects of meltwater on the AMOC will increase after 2100.

“The model indicated that AMOC weakening is roughly linear and scales smoothly with cumulative CO2 emissions when meltwater is added, arguing against a classic tipping-style jump.”

     The researchers simulated the reversibility of a strong weakening of the AMOC and concluded that reversibility can occur but will take time, likely a couple of centuries.

     They also noted that the simulation was limited and simulations with different assumption parameters should be conducted. Indeed, they should, as will be seen later in this post.

“The team notes that the results are based on a single climate model, and other models may route meltwater differently or have different AMOC stability. They say that similar tipping-focused tests should be repeated across multiple climate models to see whether the non-abrupt and reversible result is robust.”

     Indeed, in the abstract, they note that most climate modeling has not incorporated the potential impacts of Greenland meltwater, which increases the uncertainty of whether the AMOC could shift abruptly. Their study is one of the first to assess the possibility of a sudden shift, and their conclusion seems to be that it is very unlikely to lead to one before 2100, but the probability increases after that under the scenario that they use.

 

The Simulations are Using CIMP’s SSP5-8.5, Which Has Been Widely Recognized as Extremely Unlikely and Has Been Officially Abandoned by the IPCC

  The graph below from the paper shows the projections for increases in Greenland meltwater. They are on the left y-axis. The CO2 concentrations are given on the y-axis. I found those high concentrations baffling and noticed that they are using the IPCC’s abandoned CIMP SSP5-8.5 scenario. This, in my view, makes the simulations not credible as such a catastrophic scenario as SSP5-8.5 is now widely recognized as extremely unlikely. I saw an article recently calling for tracking research that uses SSP5-8.5 so that those papers could be flagged for bias. Why would they not choose the middle of the road SSP2-4.5 scenario? On the second graph below, I drew in the red line, which shows a very rough SSP2-4.5 scenario trajectory. Since the prescribed runoff meltwater follows CO2 concentrations It becomes quite clear that using SSP2-4.5 will considerably change the results of the simulation towards much less change of a catastrophic weakening before or after 2100. In the SSP 5-8.5 scenario, atmospheric CO2 concentrations would be about 1250ppm in 2100. In the SSP2-4.5 scenario, which is far more likely, the CO2 concentrations would be less than half of that at about 600ppm and max out in 2300 between 700 and 800ppm. Thus, we can see quite clearly that using SSP5-8.5 changes the results by a large factor. I would like to see these scenarios run with a much more likely scenario. It may be good to know the absolute worst scenario parameters for an endpoint reference, but if those are highly unlikely anyway, why would the researchers choose it with their limited research dollars? Perhaps the simulations were run before the IPCC officially abandoned SSP5-8.5, which happened a few months ago. However, since SSP 5-8.5 had run out of favor long ago, why are they even considering it? We could have gotten more information and more realistic information by running SSP2-4.5.  

I added the red line to show what the CO2 concentrations would be in the SP2-4.5 scenario. The purple line would also change under the SP2-4.5 scenario. The amount of meltwater in the SP5-8.5 scenario would be roughly 4 times the amount in the SP2-4.5 scenario. 

     As noted above, the amount of meltwater in the SSP5-8.5 scenario would be roughly 4 times the amount in the SSP2-4.5 scenario. That is a huge difference.

     The graph below likely carries the same bias as the previous one, since the amount of meltwater they are projecting is likely going to be much less, since they are relying on a highly unlikely scenario.

     They also simulated shifts in the AMOC source region, which is currently between Greenland and Iceland, but again, if they are predicting far more meltwater than is likely, then how could such projections be meaningful?

     Unfortunately, the use of the SP5-8.5 scenario ruins it for me. However, we should be happy that a highly unlikely absolute worst-case scenario, such as SSP5-8.5, was not catastrophic. That means that a more realistic simulation would show that fears of an AMOC collapse are not warranted, and weakening will likely be quite manageable. Is the basis of the debate about AMOC weakening a belief on one side in the SSP5-8.5 scenario? If so, then the debate is likely settled, and there is no impending threat, at least for the next several hundred years. If there is no tipping point in sight with SSP5-8.5, there will likely be no tipping point at all. That is my take.

 

    

 References:

 

Greenland meltwater adds to AMOC weakening, but updated model finds no tipping point in sight. Krystal Kasal. Phys.org. July 4, 2026. Greenland meltwater adds to AMOC weakening, but updated model finds no tipping point in sight

Limited impact of Greenland meltwater on abruptness and reversibility of future Atlantic overturning changes. Oliver Mehling, Katinka Bellomo, Federico Fabiano, Marion Devilliers, Michele Petrini, Susanna Corti, and Jost von Hardenberg. Science Advances. 19 Jun 2026. Vol 12, Issue 25. Limited impact of Greenland meltwater on abruptness and reversibility of future Atlantic overturning changes | Science Advances

Compare IPCC Scenarios Interactive. UCAR Center for Science Education. Compare IPCC Scenarios Interactive | Center for Science Education