The South Atlantic Convergence Zone (SACZ) is an atmospheric system occurring in the warm season of South America from November to March (Casarin and Kousky, 1986; Kodama 1992; Kodama 1993; Kodama et al., 1997; Gan et al. 2009; Grimm, 2011; Quadro et al., 2012). It is characterized by marked low-level winds and a moisture convergence region that persists for at least four days (Carvalho et al. 2004; Rosa et al, 2020), and by an extensive northwest-southeast-oriented cloud band (NW-SE) (Kodama 1992, Kodama 1993). It extends from the southern center of the Amazon region, crossing the Central-West and Southeast regions of Brazil. Its northern portion can also reach the central-south Bahia state (near 13°S and 42°W), and its southern portion as far south as the states of Paraná (near 25°S and 50°W) and Santa Catarina (near 27°S and 50°W). The SACZ can also extend towards the Southwestern Atlantic Ocean (SWA) as shown in several previous studies (Carvalho et al., 2002, Ferreira et al., 2004, Chaves and Nobre, 2004, Brasiliense et al., 2017). Figures 1a and 1b show composites of oceanic SACZ cases. Those are composites of oceanic SACZ made with outgoing longwave radiation and moisture flux divergence, based on four cases, which are studied and further discussed here.
In the regions affected by the SACZ, one of the main consequences is the occurrence of high rainfall, particularly in late austral spring and summer months (Grimm, 2011; Quadro et al. 2012). One of the main mechanisms contributing to the SACZ configurations is the so-called South American Monsoon System (SAMS), which is responsible for moisture transport from northern Brazil and the Amazon region to central and southeast South America (Casarin and Kousky, 1986; Kodama 1992, 1993; Kodama et al., 1997; Gan et al. 2009; Grimm, 2011; Quadro et al., 2012). Jorgetti et al. (2014), however, suggest that the SACZ can occur in both active and inactive phases of SAMS, resulting in different intensities and positions of the cloudiness band. The SACZ is therefore an important large-scale feature during summertime in tropical South America region (Kodama, 1992; Nougés-Paegle and Mo, 1997; Satyamurty et al., 1998; Liebmann et al. 1999).
As a result, SAMS and SACZ modulate the seasonal cycle of precipitation over tropical South America in a region between the equator and 25°S (Silva, 2009). Gan et al. (2004) showed that 50% of the annual rainfall over tropical and subtropical South America occurs in the austral summer months (Dec/Jan/Feb) and about 90% of it in the period from October to April. Marengo (2005), analyzing the temporal and spatial variability of the moisture balance in the Amazon basin and surrounding areas, showed that the spring and summer periods present a strong convergence of moisture throughout the SACZ climatological positioning. However, atmospheric blocking events occurring in subtropical South America can prevent SACZ formation (Rodrigues and Woollings, 2017; Rodrigues et al., 2019), leading to a deficient rainy season in the Central-West and Southeast regions of Brazil.
Previous studies suggest that reduced cloudiness in late spring (end of dry season) in Central-West and Southeast Brazil would eventually increase net surface solar radiation over the Southeast Brazilian coastal region and offshore Southwest Atlantic Ocean. This condition, in turn, would favor a pressure drop, increasing convergence at low levels and an anomalous atmospheric cyclonic circulation in southeastern Brazil. These characteristics, which are associated with increased convection, tend to intensify precipitation in the Central-West region of Brazil and develop atmospheric configurations that end up favoring the establishment of the SACZ (Grimm et al., 2007).
One of the first articles that mentioned the role of the "Oceanic SACZ'' and its implications for precipitation was developed by Carvalho et al. (2002). They argue that ~65% of all extreme rainfall events occur when convective activity in the SACZ is spatially extensive and intense. In ~30% of cases where intense precipitation occurred north of São Paulo State, they were associated with an intense SACZ with deep convective activity extending towards the Atlantic Ocean, suggesting that the SACZ probably played an important role in the increase of convection in southeastern Brazil. Regarding large-scale forcings, there is an intensification of convection over the Southwest Atlantic Ocean (Carvalho et al. 2002; Carvalho et al. 2004; Ferreira et al. 2004) during El Niño years and greater convection over the continent in La Niña years. Thus, some have proposed that the oceanic SACZ is sensitive to the SST in the Southwest Atlantic and consequently to the precipitation in South America (Tascheto et al. 2008). One of the patterns of variability in SST is the South Atlantic Dipole, which is characterized by the surface thermal gradient between the Tropical and South Atlantic. This influence has been verified during neutral ENSO conditions (Bombardi et al., 2014a; Bombardi et al., 2014b).
The SACZ can also occur in association with other atmospheric and oceanic phenomena, being influenced by local or remote factors (Kodama, 1992; Kodama et al., 1997; Grimm and Silva Dias, 1995; Grimm et al. 2007; Nogués-Paegle and Mo, 1997; Jones and Horel, 1990, Marton, 2000, Chaves and Nobre, 2004, Hirata and Grimm, 2015, Pezzi et al., 2016, Pezzi et al., 2021). These issues have been addressed in numerical modeling studies that simulate the SACZ in its atmospheric and oceanic components (Chaves and Satyamurty, 2006; Chaves and Nobre, 2004), as well as in reanalysis-based studies dedicated to understanding the spatial and temporal variability of the SACZ (Carvalho et al., 2004, Ferreira et al., 2004 and Grimm and Zilli, 2009). For example, atmospheric frontal systems in the region of the SACZ may interact with cyclonic sub-synoptic scale high-level vortices (Nobre, 1988). Oscillations in the period band of 30 to 60 days may create atmospheric disturbances that trigger the convection associated with the SACZ (Casarin and Kousky, 1986) and explosive convection over central and southern Amazonia, responsible for the generation of the convergence zone at low levels (Figueroa and Nobre, 1990). Robertson and Mechoso (2000) have presented observational evidence that positive (negative) SST anomalies in the Southwest Atlantic are associated with a weakening (strengthening) of the SACZ. On the other hand, experiments with atmospheric global circulation models show that precipitation in oceanic regions tends to intensify over warm waters (Barreiro et al., 2002).
It should be noted, however, that two possible cooling mechanisms of the ocean surface can be present during SACZ events; one is here referred to as thermodynamic and the other as dynamic. Both will now be discussed and are summarized in Figure 1c. Chaves and Nobre (2004) and Almeida et al. (2007) have suggested that the occurrence of a negative SST anomaly under the SACZ position is a result of negative feedback between the SACZ, cloud cover, and SST, with the atmosphere forcing the ocean. Atmospheric convection over the SWA increases the amount of clouds. These, in turn, reduce solar radiation at the ocean surface, and promote the cooling of it. Convection over cold waters tends to diminish, dissipating the oceanic portion of the SACZ. The tendency of atmospheric general circulation models (AGCM) to underestimate precipitation over cold waters in the SACZ region, as suggested by Nobre et al. (2012), is attributed to the non-active thermodynamic coupling in this class of models, which differs from ocean-atmosphere coupled models. In this way the authors proposed that the oceanic part of the SACZ is a thermally indirect circulation cell, with intense convection occurring over waters with lower SST values. This is the thermodynamic mechanism responsible for the cooling of the surface of the ocean, Figure 1c top panel.
In addition to the aforementioned thermodynamic process, Kalnay et al. (1986) suggested that in one SACZ case the atmosphere forced negative SST anomalies in the Southwestern Atlantic Ocean via low level cyclonic vorticity, which induces upward flow in the underlying ocean (Ekman pumping), bringing colder subsurface waters to the surface layer (upwelling). At the same time, the SACZ-cloud-SST feedback intensifies the negative SST anomaly. The cyclonic circulation tends to cease over colder waters due to the Marine Atmospheric Boundary Layer (MABL) vertical mixing adjustment to SST. Over colder waters, air buoyancy and turbulence are reduced, increasing the vertical wind shear (Wallace et al., 1989; Pezzi et al. 2005). The same dynamic feedback of SACZ-Ekman pumping-SST is mentioned in Chaves and Nobre (2004), analyzing December, January, and February averaged SST fields. They found, however, that this dynamic mechanism was one order of magnitude weaker than the thermodynamic one. Up to now, from what is known, the thermodynamic mechanism has been taken as the major mechanism responsible for the oceanic surface cooling (Figure 1b, lower panel). Our study will challenge this idea by reanalyzing in more detail the dynamic mechanism as further discussed in our results and conclusions section.
An important sector of the SE Brazilian coast is the South Brazil Bight (SBB), a highly productive semi-enclosed marine ecosystem extending from Cabo Frio (23°S, Rio de Janeiro State, RJ) to Cabo de Santa Marta (28°40'S, Santa Catarina State, SC), where intense fishing activity is concentrated. Its location can be seen in the studies of Castro (2014) and Soares et al (2011), and in Figure 1. This is a region encompassing four Brazilian states and a population of about 82 million people. Despite its importance, the SACZ atmospheric effects on the cooling of this Southwest Atlantic region, and associated impact on regional climate and marine ecosystems, has not yet been given due attention. The SBB has a strong seasonal cycle in its physical oceanography - SST, vertical stratification, thermal and haline fronts, upwelling plumes, and currents (Castro, 2014; Cerda and Castro, 2014), and its biological and fishery productivity (Soares et al., 2011; Dias et al., 2014, D’Agostine et al., 2015; Endo et al., 2019). This oceanic seasonality seems to be highly associated with the atmospheric seasonality (Castro and Miranda, 1998). The most important atmospheric system affecting SBB seasonality is the South Atlantic Subtropical High (SASH), the large-scale semi-permanent pressure center that influences the wind patterns on the Brazilian coast and causes the predominantly northeast wind direction in the SBB in summer (Pezzi and Souza, 2009b). Some studies indicate a relationship between the sardine biological cycle and fisheries of SBB to the oceanic and atmospheric physical variability (Matsuura, 1998; Sunyé and Servian, 1998; Cergole et al., 2002; Gigliotti et al., 2010; Soares et al., 2011; Dias et al., 2014, D’Agostine et al., 2015; Endo et al., 2019). These results show the important connection between ocean-atmosphere interactions and the SACZ with the ecosystem dynamics of the ocean in the region.
With the above context in mind, our main objective in this study is to investigate the coupling between the atmosphere and the SWA surface layer waters in terms of air-sea interaction processes. We achieve this by using a coupled ocean-atmosphere model to simulate oceanic SACZ episodes (i.e., with strong convective activity over its oceanic portion) selected from the study made by Rosa et al. (2020). Our main interests and novel results are on verifying how the SACZ influences both dynamic and thermodynamic mechanisms in the oceanic mixed layer (OML) that contribute to the sea surface thermal balance in that region, through changes in the OML velocity field and surface net heat fluxes, respectively.
We organize the article as follows: Section 2 introduces the regional numerical coupled system used to simulate four intense oceanic SACZ episodes. In Section 3 we present a broad overview of the simulated cases and compare three different model setups: (i) atmospheric model with prescribed SST, (ii) oceanic and atmospheric coupled and (iii) oceanic, atmospheric and wave models all actively coupled. Section 4 analyses the dynamic mechanism for the whole set of cases, while Section 5 presents a mixed layer heat budget analysis for the SWA. The article finishes with conclusions and final remarks in Section 6.