An assessment of Brazil Current surface velocity and associated transport near 22°S: XBT and altimetry data

Abstract

The use of Expendable Bathythermographs (XBTs) is widespread throughout the world’s oceans. 15 XBT high-density transects are located only in the Atlantic Ocean. However, the only transect that exclusively monitors the Brazil Current (BC) is the NOAA/AOML AX97 transect, running from Rio de Janeiro to Trindade Island. The BC is the Western Boundary Current linked to the circulation of the South Atlantic subtropical gyre. Surface velocity field based on the AX97 XBT data reference transect and two different altimetry datasets (AVISO and ATOBA) are compared with the objective of assessing the impact of different horizontal resolution altimetry products on averaged BC dynamical features. Furthermore, a sensibility study is performed considering different approaches to estimate BC volume transports. Data from 43 AX97 cruises, covering the period from Jan/2004 to Dec/2013 were used. AVISO (ATOBA) data are daily (weekly) and present a spatial resolution of 1/4° (1/12°). AX97 geostrophic velocity values were calculated based on XBT data using the isopycnal of ${\sigma }_{\theta }=26.8$ kg/m³ as the reference level of no motion, whereas the altimetry surface geostrophic velocities were obtained based on maps of absolute dynamic topography. About 80.5% (74.4%) of the time, at least one altimetry dataset observed the BC core at AX97 region. ATOBA dataset presented a better performance if compared to AVISO when the BC is highly unstable. Also, in coastal areas, where the flow frequently changes direction, a high-resolution altimetry product (ATOBA) tends to represent the horizontal structure of the BC better than AVISO, where BC horizontal pattern is smoother. Finally, the merge between AX97 data and altimetry datasets presented some inconsistencies when surface currents are in opposite directions. It is recommended to combine AX97 data with another type of data acquired in deeper regions, where near-zero velocities will propagate a smaller error along the water column.

Introduction

The use of Expendable Bathythermographs (XBTs) for temperature data acquisition is vital for the understanding of the upper ocean. An overall of 18,000–20,000 XBTs are deployed per year around the globe (Bringas and Goni, 2015, Cheng et al., 2016). Most of these launches are performed by local or regional research groups. On the other hand, the coordination and support of international programs are indispensable for successful data acquisition, quality control and application in different studies (Abraham et al., 2013). In the Atlantic Ocean, the XBT Network is a project created by the National Oceanic and Atmospheric Administration/Atlantic Oceanographic and Meteorological Laboratory (NOAA/AOML) and oversight by Ship-Of-Opportunity Program Implementation Panel. NOAA/AOML XBT Network coordinates 15 high-density transects on the Atlantic Ocean. A high-density transect is characterized by XBT deployments every 10–15 nm and realization of approximately four times per year (Cheng et al., 2016). NOAA/AOML XBT Network data are used in a series of studies regarding ocean surface currents (e.g. Goes et al., 2013a), meridional heat transport (e.g. Garzoli and Baringer, 2007) and Atlantic Meridional Overturning Circulation (AMOC) (e.g. Dong et al., 2009, Dong et al., 2015, Goes et al., 2015a, Goes et al., 2015b). Due to the importance of XBT data on oceanographic studies, several studies focused on the understanding of their uncertainties, such as inferred depth based on Fall Rate Equation (FRE) and XBT temperature biases (e.g. Reverdin et al., 2009, DiNezio and Goni, 2010, DiNezio and Goni, 2011, Goes et al., 2013b, Hutchinson et al., 2013, Bringas and Goni, 2015, Ribeiro et al., 2018). In addition, Cheng et al. (2016) compared 10 available correction schemes that consider global and historical XBT datasets and considered the correction scheme proposed by Cheng et al. (2014) as the most appropriated for XBT data. However, Cheng et al. (2016) also stated that XBT data with no correction scheme can provide a good estimate for geostrophic current. Finally, Goes et al. (2015a) analyzed the impact of XBT biases on AMOC estimations and found that XBT biases after 2010 could result on errors of up to 3% (0.38 Sv) on the AMOC.

The use of XBT is widespread throughout the world’s oceans. From the 15 XBT high-density transects located in the Atlantic Ocean, almost one third is fully or partially positioned in the Southern Hemisphere. The main dynamic feature of the South Atlantic Ocean is the South Atlantic subtropical gyre. The South Atlantic Current, the Benguela Current, the South Equatorial Current (SEC) and the Brazil Current (BC), that completes the gyre, compose the subtropical gyre. The BC, a Western Boundary Current (WBC), is originated at the bifurcation of SEC, however, additional flow inputs are observed at different depths and latitudes (Stramma and England, 1999). Pereira et al. (2014) performed a series of model runs and detailed the SEC bifurcation process for the three main water masses: between 13° and 15$°$S for the level of the Tropical Water (TW); about 22$°$S for the South Atlantic Central Water (SACW) level, and between 28° and 30$°$S at the Antarctic Intermediate Water level. On the other hand, Calado et al. (2008) stated that the SACW bifurcation was north of 22$°$S and also that the BC transports TW and SACW between 20 and 28$°$S. Finally, the only NOAA/AOML XBT Network transect that exclusively monitors the BC is the AX97 (Fig. 1).

Between 18$°$S and 24$°$S, the BC pattern is strongly influenced by regional bathymetry and mesoscale features (e.g. Schmid et al., 1995, Stramma and England, 1999, Soutelino et al., 2011, Mill et al., 2015). Abrupt changes in flow direction are frequently observed around 22$°$S. Eddies and meanders are the main sources of variability of BC flow (Campos et al., 1995, Campos, 2006, Soutelino et al., 2011). Lima et al. (2016) compared the BC baroclinic velocity (−0.17 ± 0.30 m/s) and transport (−2.7 ± 3.6 Sv) with previous studies (e.g. Campos et al., 1995, da Silveira et al., 2008, Mata et al., 2012, Rocha et al., 2014) and confirmed the development of eddies and meanders at 22$°$S. The meandering pattern of BC corroborates for the development of eddies near 22$°$S. As a consequence, several authors observed the occurrence of opposite velocities on the shelf break region (e.g. Campos et al., 1995, da Silveira et al., 2004, da Silveira et al., 2008, Rocha et al., 2014). The intense variability of BC is linked to weather condition at the vicinity of AX97 transect (Mill et al., 2015).

Centered at 20$°$S, the Vitória-Trindade Ridge (VTR) is the main physical obstacle for the BC along the Brazilian coast (Fig. 1). VTR is a sequence of seamounts displaced zonally at 20$°$S from the shelf break to the meridian of 29$°$W. The contrast of depth up to 2000 m and seamounts that can reach tens of meters from the ocean surface creates a complex environment for the continuation of the flow of TW and SACW (Fu, 1981, Evans et al., 1983, Evans and Signorini, 1985). A schematic representation of BC main path and water masses is presented in Calado et al. (2008). In order to pass these obstacles, BC divides its flow along the different channels among the seamounts. Evans et al. (1983) used hydrographic data to describe the BC flow between 19$°$S and 22$°$S, where the main flow, located between Besnard Bank and Vitória Seamount (Fig. 1), was responsible for a transport of 3.8 Sv up to 500 m with velocities up to 0.50 m/s. Afterward, the bifurcated current is then, reorganized due to the conservation of potential vorticity into a unique branch close to the continental slope (Evans et al., 1983, Legeais et al., 2013). This mechanism is also responsible for the development of eddies and meanders south of 20$°$S (Mascarenhas et al., 1971, Signorini, 1978, Schmid et al., 1995, da Silveira et al., 2008, Mill et al., 2015). Some of these eddies are able to translate northward and cross the VTR (Arruda et al., 2013). In addition, frequent anticyclonic eddies are also observed north of the VTR (Soutelino et al., 2011).

Most of the regional studies focused on BC at the vicinity of AX97 transect use hydrographic data on its analysis (Signorini, 1978, Evans et al., 1983, Campos et al., 1995, Müller et al., 1998, da Silveira et al., 2008, van Caspel et al., 2010, Soutelino et al., 2011, Mata et al., 2012, Biló et al., 2014, Palóczy et al., 2014, Rocha et al., 2014, Mill et al., 2015, Lima et al., 2016). Therefore, the reference level adopted is a crucial step for the work and one of the main uncertainties. Even though the studies mentioned above have been based on different reference levels, they converge to the fact that most of the transport of BC is confined in the upper 200 m of the water column.

Satellites are useful tools for the understanding of ocean currents, however, the analysis of WBC dynamics could be challenging for altimetry because the main flow could deviate from isobath-parallel flow toward coastal, and shallower, areas. Coastal areas provide a great source of error for altimetry products. The proximity with land generates problems related with correction of tides, high-frequency atmospheric signal and wet tropospheric components (Dufau et al., 2011).

The main goal of the present study is to assess the performance of two altimetry datasets with different spatial resolutions on capturing the horizontal structure of BC along the AX97 transect, against observations from the NOAA/AOML XBT Network. In order to reach this objective, two secondary goals are addressed: (i) an analysis of the BC core location along the coastal and the AX97 transects and (ii) the discussion of the (dis)agreement between different altimetric products and velocity field based on XBT measurements. In addition, a sensitivity study will be performed considering the same volume transport approach for different velocity fields.