Observing seasonal to decadal variability related to the Atlantic meridional overturning circulation at 11°S
The Atlantic Meridional Overturning Circulation (AMOC) plays a significant role in modulating the global climate by redistributing large amounts of heat, freshwater and carbon. It is projected to weaken under a continuously warming climate. Understanding the mechanisms that control its structure and variability on different timescales is, therefore, a key priority for oceanographers and climate scientists. This thesis presents the first estimate of AMOC variations for the tropical South Atlantic – based on bottom pressure observations from the TRACOS array at 11°S. Over the observed period (2013-2018), the AMOC and its components are dominated by seasonal variability, with peak-to-peak amplitudes of 12 Sv for the upper-ocean geostrophic transport, 7 Sv for the Ekman and 14 Sv for the AMOC transport. The observed seasonal variability of the geostrophic contribution to the AMOC in the upper 300 m is controlled by pressure variations at the eastern boundary, while at 500 m depth contributions from the western and eastern boundaries are similar. The respective mechanisms are investigated with the ocean general circulation model INALT01. In the model, seasonal AMOC variability at 11°S is governed, besides the Ekman transport, by the geostrophic transport variability in the eastern basin, which is modulated by oceanic adjustment to local and remote wind forcing. Seasonal transport variability in the western basin interior is instead mainly compensated by the Western Boundary Current. These analyses showed, that uncertainties in the wind forcing are particularly relevant for the resulting uncertainties of AMOC estimates at 11°S. A collection of different gridded surface wind products widely used within the oceanographic community is evaluated regarding the mean wind field and variability in the tropical Atlantic. While all of the wind products capture the mean, large-scale atmospheric circulation in this region, their ability to reproduce the wind variability observed with moored buoys depends on the considered timescale as well as wind regime. For example, the CORE II forcing dataset does not reproduce the observed intra-seasonal variability at any of the buoy locations, scatterometer winds show low correlations on interannual timescales and most of the products have problems to reproduce the wind variability in the eastern tropical South Atlantic. ....
... The differences between the wind products translate into estimates of the large-scale circulation in the tropical Atlantic. Additionally, the uncertainties from using different wind stress parametrizations can make up 15 % of the momentum flux into the ocean. For integral variables like the Ekman or Sverdrup transports across a certain latitude, the choice of wind product results in uncertainties of 15- 40 % depending on the latitude and timescale. Older, coarser products produce spurious wind stress variations, especially on decadal timescales. The small-scale information contained in scatterometer winds, but not in the coarser reanalysis products, is emphasized in indices involving spatial derivatives like the wind stress curl. Substantial inconsistencies exist regarding the seasonal variability of the wind stress curl in the eastern and western boundary regions at 11°S. Together, these results highlight the importance of testing the sensitivity of derived estimates of the wind-driven circulation to the choice of wind product as well as careful consideration or incorporation of large uncertainties in wind forcing products into ocean simulation experiments. The AMOC is closely linked to the formation, transformation and advection of water masses. Changes in the major water masses of the Atlantic Ocean are related to variations in the oceanic heat and freshwater transports and can influence basin-wide zonal pressure gradients, thus, the large-scale ocean circulation. Two full-depth transatlantic sections in the tropical South Atlantic, separated by 24 years, are analyzed regarding water mass property distributions and long-term water mass changes along 11°S. The vertical and horizontal distributions of typical property extrema in the Brazil and Angola Basins are in agreement with previous studies. Temperature differences between 1994 and 2018 confirm a warming of the central and lower intermediate waters in the western boundary region off Brazil, also extending eastward into Angola Basin after 1994. Dissolved oxygen distributions along 11°S support a vertical expansion of the tropical South Atlantic oxygen minimum zone, but also showed a reduced westward extent between 1994 and 2018. Abyssal temperature changes in the Brazil Basin are in agreement with previous studies, showing a warming of the Antarctic Bottom Water (AABW) and cooling of the lower North Atlantic Deep Water (NADW) to be mainly caused by volume changes. However, decadal temperature variations on isopycnals to also contribute to the observed signals. An oxygen increase of 2-3 µmol kg-1 dec-1 within the AABW layers of the Brazil Basin has not been observed before and could hint towards a dominance of local mixing processes over changes in the source waters. A similar oxygen increase is also observed above the western flank of the Mid-Atlantic Ridge and in the abyssal Angola Basin below 2000 m. Overall, the results from this thesis contribute to the understanding of AMOC-related variability as well as possible uncertainties in the estimates of different AMOC components derived from observations in the tropical South Atlantic. When continued into the future and combined with other programs or AMOC arrays, the observations of the TRACOS array at 11°S have great potential for further investigations of the mechanisms relevant for tropical Atlantic variability, as well as meridional coherence and long-term changes of AMOC variability.
Preview
Rights
Use and reproduction:
Please note that individual components of the publication may be subject to other licensing or copyright conditions.