Climate model diversity in the North Atlantic and its impact on prediction skill on interannual-to-decadal timescales

Project Information

DivPredSkill

Grant agreement ID: 101026271

DOI

10.3030/101026271

Project closed

EC signature date 8 April 2021

Start date 19 July 2021

End date 18 July 2023

Funded under

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Total cost

€ 212 933,76

EU contribution

€ 212 933,76

Coordinated by

THE UNIVERSITY OF READING
United Kingdom

Periodic Reporting for period 1 - DivPredSkill (Climate model diversity in the North Atlantic and its impact on prediction skill on interannual-to-decadal timescales)

Reporting period: 2021-07-19 to 2023-07-18

Summary of the context and overall objectives of the project

The Atlantic Meridional Overturning Circulation (AMOC) is an important component of the climate system. It transports heat from the equatorial Atlantic to the higher latitudes of the North Atlantic, which contributes to the mild winters in Europe. Another component relevant for the weather and climate in Europe is the North Atlantic Oscillation (NAO). The NAO describes atmospheric pressure differences over the Atlantic which affect temperature and precipitation over Europe, especially in winter. Thus, understanding the interaction of the AMOC and the NAO is of interest for the political, economic, and agricultural sectors.
One problem though is that observational records for the AMOC are too short in time to learn more about its interaction with the NAO. Therefore, climate models have to be used to study this topic. Still, climate models are subject to uncertainties and systematic errors. In this project, we focus on the differences across a large number of climate models. We compare them in terms of their AMOC and NAO variability and interaction, to learn more about the effect of model uncertainties.
We conclude that the AMOC-NAO interaction of a model is very sensitive to its tendency to be, on average, rather warm and salty or cold and fresh in the subpolar region of the North Atlantic. Models that fall into the warm-salty category, show a stronger and longer-lasting response of the AMOC to the NAO compared to the cold-fresh models. This is linked to the models' sea ice cover and the stability of the water column in the Labrador Sea. Furthermore, it was found that the potential to make skillful decadal climate predictions, is higher for the models categorised as cold-fresh.
These results stress that climate model behaviour is very diverse suggesting that research findings related to North Atlantic climate should not rely only on a single model result or the multi-model mean. Moreover, this project's results identify key elements in the models that lead to uncertainty in the North Atlantic climate variability, which is highly valuable for future model improvement enabling also better climate predictions.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The first step of this project was to collect a large amount of climate model data from the Coupled Model Intercomparison Project Phase 6. The next step was to write and run programme code that reduced the size of the overall data. Then, the analysis of the data could be performed. The analysis involved various statistical methods and testing of several scientific hypotheses centred around the diversity of climate model behaviour in the North Atlantic.

Overview of results: There are large differences across the models regarding their mean state in different regions, their variability in time and space and the interaction details of different variables. Specifically, it was found that the tendency of a model to have a warm and salty (versus a cold and fresh) surface mean state in the subpolar North Atlantic was shown to be important for the response of the ocean (specifically, the AMOC) to the atmosphere (specifically, the NAO). Compared to the cold-fresh models, the warm-salty models had a lower sea ice cover over the Labrador Sea, a higher surface heat loss related to the NAO variability, as well as a weaker density structure of the water column in the Labrador Sea. These factors combined enable the warm-salty models to create a stronger and longer-lasting response of the AMOC to the NAO.

Dissemination and exploitation: The project's results are highly relevant for better physical process understanding in the North Atlantic, and for future climate model improvement. Therefore, results were shared at various occasions. This includes 4 oral and 3 poster presentations at international conferences, as well as 4 invited talks at European and US academic institutions. Furthermore, results were presented and discussed at two meetings in the UK of projects with related topics, and at the host institution's departmental and group seminars and meetings.
Making use of these occasions, results were disseminated to a large and broad scientific audience. Also, there was intense exchange with experts in the fields, for example, through participating in the corresponding conference sessions.
A scientific publication arising from this project is currently under review at the Journal of Climate. In line with the Horizon2020 targets, the publication will be made open access. Python codes used to achieve the presented results are openly available at Zenodo at https://doi.org/10.5281/zenodo.8224157 .

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

Climate models are widely used by the climate science community to investigate specific topics. One known issue with climate models are errors in the mean state. Before the start of this project, the effect of the models' mean state differences on the AMOC-NAO link had not been studied. The results of this project provide not only the effects of specific mean state tendencies on the AMOC-NAO link but also explain the mechanism behind it. Key variables leading to uncertainties were identified. Knowing these key variables during model development is crucial to improve future climate models. Better models will provide more reliable weather and climate predictions, supporting decision makers in the political, economic, and agricultural sectors.

submission-manuscript.png