Interactions between the stratospheric polar vortex and Atlantic circulation on seasonal to multi-decadal timescales

• Published in: Atmospheric Chemistry and Physics
• Main Authors: Dimdore-Miles, O, Gray, L, Osprey, S, Robson, J, Sutton, R, Sinha, B
• Format: Article in Journal/Newspaper
• Language: English
• Published: Copernicus Publications 2022
• Subjects: Pacific; North Atlantic; North Atlantic oscillation
• Online Access: https://doi.org/10.5194/acp-22-4867-2022; https://ora.ox.ac.uk/objects/uuid:24ac54b7-574a-40db-86a7-e1b151a63d19

Variations in the strength of the Northern Hemisphere winter polar stratospheric vortex can influence surface variability in the Atlantic sector. Disruptions of the vortex, known as sudden stratospheric warmings (SSWs), are associated with an equatorward shift and deceleration of the North Atlantic jet stream, negative phases of the North Atlantic Oscillation, and cold snaps over Eurasia and North America. Despite clear influences at the surface on sub-seasonal timescales, how stratospheric vortex variability interacts with ocean circulation on decadal to multi-decadal timescales is less well understood. In this study, we use a 1000 year preindustrial control simulation of the UK Earth System Model to study such interactions, using a wavelet analysis technique to examine non-stationary periodic signals in the vortex and ocean. We find that intervals which exhibit persistent anomalous vortex behaviour lead to oscillatory responses in the Atlantic Meridional Overturning Circulation (AMOC). The origin of these responses appears to be highly non-stationary, with spectral power in vortex variability at periods of 30 and 50 years. In contrast, AMOC variations on longer timescales (near 90-year periods) are found to lead to a vortex response through a pathway involving the equatorial Pacific and quasi-biennial oscillation. Using the relationship between persistent vortex behaviour and the AMOC response established in the model, we use regression analysis to estimate the potential contribution of the 8-year SSW hiatus interval in the 1990s to the recent negative trend in AMOC observations. The result suggests that approximately 30 % of the trend may have been caused by the SSW hiatus.