Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Feb 15:8:14375.
doi: 10.1038/ncomms14375.

Abrupt cooling over the North Atlantic in modern climate models

Giovanni Sgubin  1   2 , Didier Swingedouw  2 , Sybren Drijfhout  3   4 , Yannick Mary  2 , Amine Bennabi  5
Affiliations

Abrupt cooling over the North Atlantic in modern climate models

Giovanni Sgubin et al. Nat Commun. .

Erratum in

Abstract

Observations over the 20th century evidence no long-term warming in the subpolar North Atlantic (SPG). This region even experienced a rapid cooling around 1970, raising a debate over its potential reoccurrence. Here we assess the risk of future abrupt SPG cooling in 40 climate models from the fifth Coupled Model Intercomparison Project (CMIP5). Contrary to the long-term SPG warming trend evidenced by most of the models, 17.5% of the models (7/40) project a rapid SPG cooling, consistent with a collapse of the local deep-ocean convection. Uncertainty in projections is associated with the models' varying capability in simulating the present-day SPG stratification, whose realistic reproduction appears a necessary condition for the onset of a convection collapse. This event occurs in 45.5% of the 11 models best able to simulate the observed SPG stratification. Thus, due to systematic model biases, the CMIP5 ensemble as a whole underestimates the chance of future abrupt SPG cooling, entailing crucial implications for observation and adaptation policy.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing financial interests.

Figures

Figure 1
Figure 1. Sites of deep-water formation in the North Atlantic.
Map of the maximum winter mixed layer depth (m) averaged over the 1993–2012 period according to the GLORYS reanalysis. The red contour represents the reference area for our analysis (see Methods). Its total surface measures 3.61 × 106 km2 and it entirely spans the subpolar NA, including those sites in the Labrador and Irminger Seas that are regularly subject to convective activity. The yellow contour highlights the region for which the maximum MLD averaged over the 1993–2012 period exceeds 1,000 m. This area has been used for a sensitivity test of our main findings on the particular choice of the reference region (Supplementary Fig. 9). Arrows indicate the main surface currents, including the North Atlantic Current, the western subpolar gyre in the Labrador and Irminger Seas and the eastern subpolar gyre in the Nordic Seas.
Figure 2
Figure 2. Patterns of SST response in RCP scenarios.
Ensemble mean of the 21st century SST trend normalized by its own global mean (dimensionless quantity) for (a) RCP2.6 simulations, (b) RCP4.5 simulations and (c) RCP8.5 simulations. The globally averaged SST trend ensemble mean is indicated for each scenario, that is, 0.46 10−2 oC year−1 for the RCP2.6 experiments, 1.27 10−2 oC year−1 for the RCP4.5 experiments and 3.01 10−2 oC year−1 for the RCP8.5 experiments. Since the globally averaged SST trend ensemble mean is positive for all scenarios, the non-dimensional value in each grid point is >1 when characterized by amplified warming, <1 when characterized by a subdued warming and <0 when characterized by cooling. The black contour shows regions with maximum ensemble spread (see Methods).
Figure 3
Figure 3. Characterization of the different SST responses in the SPG.
Examples of the SST, MLD and AMOC evolutions over the SPG in the three model subsets (non-abrupt, SPG convection collapse and AMOC disruption) for the RCP2.6 scenario. Only one example for each sub-ensemble is shown while the Supplementary Figs 2–6 provides a more comprehensive illustration. All time series were smoothed using a 10-year running mean to remove the high-frequency variability. (a–c) SST anomaly (oC) with respect to its initial magnitude, that is, the mean over the decade 2006–2015, in (a) NorESM1-M, that is, non-abrupt model, (b) GISS-E2-R, that is, SPG convection collapse model, (c) FIO-ESM, that is, AMOC disruption model. Values in brackets indicate SST magnitudes at the beginning of the RCP2.6 experiments (2006–2015). (d–f) Relative changes (%) of AMOC (red lines) and MLD (blue lines) in d NorESM1-M, (e) GISS-E2-R, (f) FIO-ESM with respect to their initial values (2006–2015). Absolute magnitudes of AMOC (Sv) and MLD (m) averaged over the period 2006–2015 are, respectively, displayed in red and blue brackets. It is worth noticing that the strong AMOC reduction in the FIO-ESM model already takes place during the historical period (Supplementary Fig. 6), yielding a low absolute value over the 2006–2015 period.
Figure 4
Figure 4. Different climate impacts.
Patterns of the 21st century SAT trend (oC 10−2 year−1) under the RCP4.5 scenario for: (a) non-abrupt ensemble (27 members), (b) SPG convection collapse ensemble (7 members) and (c) AMOC disruption ensemble (2 members). The GMT trend is also displayed for each subset. The light grey and dark grey contours define regions where the ensemble mean precipitation trend, respectively, exceeds 300 mm per century and is lower than −300 mm per century. Results for the RCP2.6 and RCP8.5 scenarios can be found in Supplementary Fig. 7 and Supplementary Fig. 8.
Figure 5
Figure 5. The role of stratification in SPG projections.
Scatterplot of simulated SST trends (oC 10−2 year−1) over the SPG versus (a,c,e) the relevant MLD-trend (m 10−2 year−1) and (b,d,f) the present-day stratification indicator (Kg m−3). Non-abrupt models are indicated with red circles and SPG convection collapse models with blue circles, for (a,b) the RCP2.6, (c,d) the RCP4.5, (e,f) the RCP8.5 scenario. In a,c,e the value rl indicates the linear correlation between the SST and MLD trends, whose significance above the 95% confidence level was evaluated with a two-tailed Student’s t-test. The crosses indicate the linear best-fit of the SST trends against the MLD trend, that is,. the linear regression using the least squares method. In b,d,f the value rnl indicates the non-linear correlation between SST-trend and the stratification indicator, statistically significant at the 95% confidence level (see Methods). The crosses indicate the logarithmic best-fit of the SST trends against the stratification index, that is, the logarithmic regression using the least squares method. The dashed vertical black line centred on 0 indicates the observationally based stratification index, calculated as the average of GLORYS Reanalysis (1993–2012) data and EN3 analysis data (1950–2012). The arrows at the bottom indicate the areas in the panels for which the simulated SPG stratification is either more, or less stable than in the observational data.
Figure 6
Figure 6. Different SPG stratifications in the model sub-ensembles and their comparison with the observational data.
Present-day vertical profiles of (a) winter density (kg m−3), (c) temperature (oC) and (e) salinity (psu) in the SPG region for observational data (black lines), for ensemble-mean of non-abrupt models (red lines) and for ensemble-mean of the SPG convection collapse models (blue lines). Right panels show the difference between the modelled present-day winter conditions in the SPG and the observation-based data (b) for density, (d) for temperature and (f) salinity. Horizontal dashed lines are drawn at every 1,000 metre. A zoom has been made for the first 1,000 metres.
Figure 7
Figure 7. Reduction of model uncertainty over the SPG.
Model ensemble mean and spread of the 21st century SST trend (oC 10−2 year−1) over the SPG in the RCP scenarios for different subsets of models: (black) all the 40 CMIP5 models; (red) CMIP5 models possessing a skill score S>0.8; (blue) CMIP5 models possessing a skill score S>0.9. Error bars indicate the standard deviation of the SST trend ensemble mean for the different subsets of models.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Stocker T. F. et al.. Technical Summary in Climate Change 2013: the Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2013).
    1. Hansen J. et al.. Earth's energy imbalance: confirmation and implications. Science 308, 1431–1435 (2005). - PubMed
    1. Levitus S. et al.. Anthropogenic warming of Earth’s climate system. Science 292, 267–270 (2001). - PubMed
    1. Levitus S. et al.. World ocean heat content and thermosteric sea level change (0–2000m). Geophys. Res. Lett. 39, L10603 (2012).
    1. Roemmich D., Gould W. J. & Gilson J. 135 years of global ocean warming between the Challenger expedition and the Argo Programme. Nat. Clim. Change 2, 425–428 (2012).

Publication types

LinkOut - more resources