The Atlantic Meridional Overturning Circulation in the North Atlantic, focussing on 47°N - variability, trends, and meridional connectivity

Abstract

The Atlantic Meridional Overturning Circulation (AMOC) plays a vital role in the climate of Europe and the North Atlantic region. Climate model studies project an AMOC decline in the 21st century. However, they disagree on the magnitude and timescales of the weakening. Thus, monitoring AMOC changes remains essential to provide benchmarks for assessing climate models and understanding the physical processes determining AMOC variability. In this thesis, basin-wide AMOC volume transports are calculated (1993-2018). Measurements from moored instruments of the NOAC array at 47°N are combined with hydrography and satellite altimetry. Variability, trends, and meridional connectivity with the RAPID array at 26°N are analyzed. The AMOC volume transport at 47°N exhibits a mean strength of 17.2 Sv and substantial variability on inter-annual and seasonal timescales but no significant long-term trend. The NOAC AMOC shows a significant correlation with the RAPID AMOC when the NOAC AMOC leads by about one year, indicating meridional connectivity. An analysis of the AMOC at the NOAC, RAPID, and OSNAP (52°N-60°N) lines in the high-resolution forced VIKING20X model simulation reveals a mean NOAC AMOC strength within the estimated error range of the NOAC observations. In disagreement with observations, the VIKING20X AMOC decreases after the mid-1990s until 2010 at all three array lines. This decrease coincides with a significant cooling and freshening in the subpolar North Atlantic. In agreement with observations, VIKING20X shows meridional connectivity between the NOAC and RAPID AMOC when the NOAC AMOC leads by about one year. This indicates a common mechanism, determining the meridional connectivity in observations and VIKING20X. An analysis of different ANHA model simulations with varying resolution underlines the importance of model resolution for accurately representing the AMOC mean strength and variability but also stresses the need for model improvements beyond resolution.