Variability of the Meridional Overturning Circulation in the South Atlantic and Pacific Oceans

Abstract

The ocean circulation dynamics in the Southern Hemisphere have an important role in the global climate system. The system of ocean currents knwon as Meridional Overturning Circulation (MOC) connects and redistributes salt, heat, and other biogeochemical tracers across the different ocean basins. Here, we detail the South Atlantic Ocean circulation patterns by analyzing hydrographic data from several cruises using an inverse box model to adjust the reference-level velocities. The upper layer currents, such as the northwestward-flowing Benguela Current and southward-flowing Brazil Current, describe the South Atlantic anticyclonic gyre. The deep layers feature the southward transports of the deep boundary currents and a west-to-east flow between the basins near 24°S (7.5 ± 4.4 Sv) above the Mid-Atlantic Ridge. Both upper and deep layers connect the eastern and western basins; however, the abyssal waters present northward mass transports through Argentina and Cape Basins without any interbasin exchange across the Mid-Atlantic Ridge. Furthermore, the cruise data allow us to compute the upper AMOC (Atlantic MOC) strength, as well as the transports of mass, heat, and freshwater, demonstrating the characteristic northward heat transport and the dominance of evaporation over precipitation across the subtropical South Atlantic. Currently, the freshwater transport by the AMOC (Mov) across 34.5ºS in the South Atlantic has been identified as a possible indicator of AMOC stability, with a negative (southward) freshwater transport indicating a possible bistable AMOC regime and positive (northward) transport indicating a monostable regime. This Ph.D. thesis computes the Mov using observations from 49 eXpendable BathyThermograph (XBT) transects, from South America to South Africa, over nearly two decades (2002-2019), resulting a negative Mov mean of −0.15 ± 0.09 Sv which suggests a bistable AMOC regime. These results are complemented with two data sets derived from Argo float observations, four Ocean General Circulation Models (OGCMs: GLORYS, OFES, MOM6-JRA, and MOM6-MERRA), and thirty-two Coupled General Circulation Models (CGCMs: CMIP6). Both Argo and OGCMs data sets agree with the sign of the Mov computed from the XBT data. Nevertheless, more than half of the examined CGCMs, 20 out of 32, present positive Mov mean values. To investigate the causes of the differing signs of the Mov across the models, we examine the salinity vertical structure in CGCMs with positive and negative Mov. Importantly, our work highlights the different salinity structures in CMIP6 models with positive Mov means (fresher upper and saltier deep waters compared to those estimating negative Mov values), suggesting that salinity biases may be responsible for the opposite sign of Mov. As a result, this thesis highlights the importance of improving CMIP6 model representations, especially the salinity bias. In addition, we compute the South Atlantic meridional fluxes (mass, heat, and salt) at 34.5ºS, which show linear relationships, with a negative slope (positively correlated in magnitude) between Mov/MOC and Mov/MHT (Meridional Heat Transport) and a positive slope (positively correlated) between MHT/MOC. Seasonally, the South Atlantic meridional fluxes across 34.5ºS from most of the data sets considered in this thesis show a more negative Mov and a more positive MOC and MHT in the austral fall and winter, from April to August. This research further extends to the South Pacific Ocean. In the same way as the South Atlantic, we use cruise data with an inverse box model to compute the meridional circulation and transports. We use hydrographic data to compare the circulation of three decades: 1992, 2003, 2009, and 2017. This comparative analysis reveals different horizontal circulation schemes, particularly the emergence of a "bowed gyre" in 2009, which is not replicated over the entire length of any of the four OGCMs (ECCO, GLORYS, SOSE, and MOM) used. In addition, our observational and numerical model data highlight discrepancies in the representation of the East Australian Current. However, the representation of the Peru-Chile Current is consistent across the data sets. Furthermore, this thesis computes the temperature and freshwater transports from the cruise data, estimating significantly different results during the “bowed gyre” in 2009. A linear Rossby wave model is adopted to clarify the causes of these different circulation schemes, which includes the wind stress curl variability as a remote forcing and the response to sea surface height changes along 30°S in the Pacific Ocean. The present Ph.D. thesis significantly contributes to understanding ocean circulation variability in the South Atlantic and South Pacific Oceans. By combining data from observations with numerical model outputs, this research provides a comprehensive perspective on ocean dynamics and offers implications for future climate change projections.