Multi-decadal and centennial modes of AMOC variability and their dependence on mean state in a climate model
- 1Department of Earth and Planetary Sciences, Yale University, New Haven, USA (alexey.fedorov@yale.edu)
- 2LOCEAN-IPSL, Sorbonne University, Paris, France (brady.ferster@locean.ipsl.fr)
- 3NCAS-Climate, University of Reading, Reading, UK
Climate models exhibit large differences in the mean state and variability of the Atlantic meridional overturning circulation (AMOC), including in AMOC strength and the characteristic amplitude and frequency of its variability. Across different GCMs, AMOC long-term variability ranges from decadal to multi-centennial and its magnitude from a fraction of to several Sverdrups (Sv). In this study, we conduct ensemble experiments, using the latest coupled model from Institut Pierre Simon Laplace (IPSL-CM6A-LR), to investigate systematically how AMOC variability depends on the AMOC mean state. In the control simulations of this model AMOC mean volume transport is about 12Sv, while AMOC variability is dominated by two distinct modes – a multi-decadal mode with periodicity between 20-30 years and a centennial mode with periods of 100-200 years. The former mode is weaker and driven by temperature variations, while the latter is stronger and driven by salinity anomalies. To modify the mean state of the AMOC in the model we use an indirect method based on robust atmospheric teleconnections from the tropical Indian ocean (TIO) to the Atlantic as described in two recent studies (Hu and Fedorov, 2019; Ferster et al., 2021). Both studies have shown that warming the TIO results in an increased AMOC strength, while cooling the TIO results in a weakened AMOC. To change the Indian ocean temperature in our perturbation experiments we nudge TIO SST by -2°C, -1°C, +1°C, and +2°C; and the experiments last for approximately 1000 years. This allows us to go from a nearly collapsed AMOC state below 3Sv to a more realistic mean state of about 16Sv. We find that both modes of AMOC variability persist throughout the experiments while their amplitude increases almost linearly with AMOC mean strength, yielding linear relationships between the amplitude of variability (standard deviation) and AMOC mean strength of +0.04 Sv per 1 Sv and +0.07 Sv per 1 Sv, respectively. In the experiments that generate 16Sv of AMOC transport, the amplitudes of the two modes reach nearly 0.7 and 1.4Sv. Lastly, we compare the dynamical mechanisms of the two modes and their climate impacts. A corollary of this study is that in this model, a stronger AMOC would lead to stronger climate variability.
How to cite: Fedorov, A., Ferster, B., Mignot, J., and Guilyardi, E.: Multi-decadal and centennial modes of AMOC variability and their dependence on mean state in a climate model, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-5162, https://doi.org/10.5194/egusphere-egu22-5162, 2022.