by Dave Bonan and Matt Luong, UW CICOES Postdoctoral Scholars

A challenging region

The Southern Ocean plays a central role in regulating Earth’s climate under global change. It absorbs roughly 75% of the excess heat and about half of the excess carbon taken up by the global ocean, making it one of the most important buffers against climate change. How the Southern Ocean responds to continued anthropogenic greenhouse-gas emissions will shape future global projections of surface temperature, precipitation, and sea level rise: essential information for preparing for and adapting to a warming world.

However, because the Southern Ocean is vast, remote, and sparsely observed, we must rely heavily on climate model simulations to study its behavior. Confidence in future climate projections depends on the ability of climate models to reproduce the recent historical record, but the Southern Ocean remains one of the most challenging regions to represent in climate models.

While most of the planet has warmed since the start of the satellite era (~1980), the surface waters around coastal Antarctica and throughout much of the Southern Ocean instead show a broad cooling. Satellite measurements also indicate an increase in sea ice concentration that accompanies this cooling.

Most climate models, however, show little to no cooling, with many simulating strong warming and declining sea ice. This mismatch suggests that climate models may be missing key processes that shape the Southern Ocean’s response to climate change— a knowledge gap that captured our attention and inspired our research. Answering the question “Why do nearly all climate models struggle to capture recent Southern Ocean temperature trends?” requires expertise across oceanography, atmospheric science, and cryospheric processes. This makes it well suited to the collaborative research environment at CICOES.

Present-day stratification: An overlooked factor

Persistent biases in climate models of the Southern Ocean, together with the complexity of interacting atmosphere, ice, and ocean processes, make it challenging to identify which mechanisms drive the observed changes. Many explanations have been proposed: wind changes linked to historical ozone depletion, freshening from Antarctic ice melt, shifts in precipitation, changes in the sea-ice hydrologic cycle, and natural variability in deep ocean convection that most climate models do not resolve. Although each process may contribute to surface cooling, none fully explains why climate models fail to reproduce the observed cooling trend or how to reconcile that trend with simultaneous subsurface warming and evolving salinity.

One possible explanation that has received growing attention is ocean stratification, which refers to the vertical layering of seawater density set by temperature and salinity, and can alter how the ocean and atmosphere interact with each other. The influence of changes in ocean stratification on predictions of longterm Southern Ocean climate trends is still not well understood.

As CICOES postdoctoral fellows working with colleagues from UW and the NOAA Pacific Marine Environmental Laboratory, including Wei Cheng, Gregory Johnson, David Battisti, and Kyle Armour, we are leading complementary efforts to evaluate whether a more realistic representation of Southern Ocean stratification improves climate models’ simulation of historical temperature trends.

In search of the missing cooling

One part of our work focuses on evaluating the representation of the Southern Ocean across different phases of the Coupled Model Intercomparison Project (CMIP). Climate models participating in CMIP apply the same physical laws, but differ in how they represent key processes. By comparing the simulations of historical climate change, we are able to identify systematic patterns, biases, and points of divergence from observations.

In these simulations, we are examining how heat enters, exits, and is redistributed within the upper ocean, decomposing the surface heat budget into its major components, including solar absorption, longwave emission, latent and sensible heat exchange, and heat transport from winds and mixing. Comparing these processes between climate models and observations reveals where climate models allow too much heat in, mix heat too efficiently downward, or misrepresent atmospheric controls on surface energy exchange. Each type of error produces a characteristic signature in the heat budget that helps explain why many climate models show less cooling than observed.

We are also investigating the problem using ocean-only simulations branched from a state estimate of the ocean, essentially enabling us to examine an ocean model with a density structure similar to observations. Starting our approach from this realistic stratification allows us to test whether stratification alone is sufficient to produce surface cooling when the modeled ocean is adjusted to be in line with real-world estimates of trends in surface heat, freshwater, and wind fluxes.

In this approach, we have found that much of the Southern Ocean’s muted warming and large-scale freshening can be explained by climate change forcing alone, even without additional changes in winds or freshwater. However, changes in winds and freshwater then modify this base pattern to produce temperature and salinity trends that more closely resemble observations.

Although this ocean-only approach omits coupled feedbacks, the idealized hierarchical modeling framework removes much of the complexity that shapes more sophisticated climate model responses and enables us to more precisely attribute which portions of the observed pattern arise from specific surface forcings. This allows us to explore questions such as why increased precipitation and increased ice-sheet melt influence the Southern Ocean differently, or whether the poleward shift or strengthening of the winds has played a larger role in driving Southern Ocean temperatures. These experiments reveal not only when realistic stratification is essential to reproduce observed Southern Ocean cooling, but also whether our best estimates of observed forcing affect ocean circulation in the ways inferred from complex models.

Together, our efforts provide a clear explanation for why state-of-the-art climate models struggle in a region that plays an outsized role in Earth’s climate and why the Southern Ocean is cooling in the real world. If climate models cannot capture the Southern Ocean’s historical behavior, then their projections of future change may be overlooking something important. Identifying the origin of this “missing cooling” is a crucial step toward improving climate models and strengthening our ability to anticipate the planet’s future.

“ Together, our efforts provide a clear explanation for why state-of-the-art climate models struggle in a region that plays an outsized role in Earth’s climate. ”