Butterfly effect of air–sea coupling in climate models: Different ocean circulation responses to atmospheric forcing

Published in Earth & Environment
Butterfly effect of air–sea coupling in climate models: Different ocean circulation responses to atmospheric forcing

Share this post

Choose a social network to share with, or copy the shortened URL to share elsewhere

This is a representation of how your post may appear on social media. The actual post will vary between social networks

♦ Atmospheric Forcing to Ocean Circulation

The large-scale ocean circulation, often referred to as the global conveyor belt, connects and stirs up the world's oceans—although it may sound irrelevant to us living on the ground, it can actually have wide impacts on surface climate, as illustrated in Dr. Marcello's recent posting. Interestingly, the primary process that powers this global-scale conveyor is highly localized: Triggered by densification at the surface due to heat loss or salt gain, the deep sinking of seawater occurs at specific sites to feed the lower-layer currents that travel across basins. These sites include the Antarctic Ocean, Nordic seas, and the subpolar North Atlantic, making these regions a 'window' where the atmospheric variability can effectively communicate with the ocean circulation. 

In the North Atlantic, there exists an atmospheric pattern that repeatedly appears during the wintertime. It is characterized by the strengthening/weakening of the contrast between the subtropical high and subpolar low systems. Here, changes in the strength of the subpolar low system are accompanied by changes in the wind intensity over the subpolar North Atlantic (the window region), which in turn transmit a signal to the ocean circulation. More specifically, an enhancement of the subpolar low leads to vigorous ocean circulation, as stronger-than-normal westerly winds actively help dense water formation.

Generally speaking, up-to-date climate models can simulate the above sequence on their own. However, the efficiency of this series of processes—that is, how strongly ocean circulation responds— widely varies among climate models, making a reliable prediction of ocean circulation and its consequent heat redistribution difficult. To reduce these uncertainties, we need to understand where the discrepancy in climate models arises from. 

♦ The Efficiency Inferred from the Oceanic Mean State 

Our analysis of 42 climate models revealed that the strength of the atmospheric impact on ocean circulation in the North Atlantic is closely related to the oceanic mean state in that region. In particular, the sea ice coverage and mixed layer depth during the February-to-April seasons (the time of year when the mixed layer is most well-established) are the key factors. 

Sea ice covering the ocean's surface hinders the transfer of atmospheric signals into the ocean, which has a similar effect to closing a window. Hence, the climate models that simulate a wide area of average sea ice wind up in relatively weak impacts of the westerly winds on the ocean circulation strength. Or, in other words, the window should be maintained open for effective communication between the two.

Meanwhile, the mean mixed layer depth is indicative of how well the seawater properties are mixed in a column (as the term suggests). Thus, a deep mixed layer means that, when the atmosphere perturbs seawater density at the surface, the dense water parcel can easily penetrate into the deep layer to join the lower branch of the global ocean circulation. For this reason, the mean depth of mixed layer in models tends to be proportional to the magnitude of the atmospheric influence on ocean circulation.

Then, what if the mean state is more homogeneous among the models? Will the atmosphere–ocean interaction appear more consistent? Previous literature implies 'yes.' In ocean-only models where the mean state in the subpolar North Atlantic is relatively similar to each other, so is the sensitivity of ocean circulation to atmospheric disruption. In these ocean-only models, the atmospheric influence is given as an external condition, instead of being simulated by the models themselves. Thus, it leads to the conclusion that mean-state discrepancies in coupled atmosphere–ocean models (which determine the oceanic sensitivity) fundamentally arise from the freely-evolving interaction between atmospheric and oceanic processes.

♦ Mean State Formation in Coupled Models

In coupled models, the atmospheric and oceanic fields constantly interact with and adjust to each other to achieve equilibrium. In the course of reaching this delicate balance, a series of chain reactions can culminate in a diverse mean state. 

The investigation into the mean fields of 42 models gives a glimpse of the processes of the atmosphere–ocean balancing as follows. First, if the mean intensity of ocean circulation is strong in a model, the North Atlantic surface tends to be warmer due to the oceanic heat transport. As it modulates the structure of the atmospheric jet, westerly wind over the deep sinking area strengthens. Accordingly, the deep water formation can be maintained strong on average, and so can the ocean circulation. In the opposite situation, weak ocean circulation is balanced with weak westerly winds, which is associated with less oceanic heat transport and cooler surface temperature. 


The reasoning so far could be put in the reversed way as follows: The two-way feedback between atmospheric and oceanic parts in coupled models can amplify inter-model differences in the oceanic mean state; It eventually shapes the ocean circulation response to atmospheric variability by regulating (i) opening/closing of the communicating window and (ii) the efficiency of the deep subsidence. Improving the mean state in climate models is thus one of the keys to better predicting changes in ocean circulation.

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in

Follow the Topic

Earth and Environmental Sciences
Physical Sciences > Earth and Environmental Sciences

Related Collections

With collections, you can get published faster and increase your visibility.

Digital Paleoclimate: Integration, Simulation, and Assimilation

Paleoclimate studies towards a digital paradigm by using paleoclimate records integration, model simulation, and data assimilation for promoting our understanding of climate dynamics and future prediction.

Publishing Model: Open Access

Deadline: Sep 30, 2024