Abstract

ERA-Interim and JRA-55 reanalysis products are analyzed to document the annual cycle of the South Atlantic subtropical high (SASH) and examine how its interannual variability relates to regional and large-scale climate variability. The annual cycle of the SASH is found to have two peaks in both intensity and size. The SASH is strongest and largest during the solstitial months when its center is either closest to the equator and on the western side of the South Atlantic basin during austral winter or farthest poleward and in the center of the basin in late austral summer. Although interannual variations in the SASH’s position are larger in the zonal direction, the intensity of the high decreases when it is positioned to the north. This relationship is statistically significant in every month. Seasonal composites and EOF analysis indicate that meridional changes in the position of the SASH dominate interannual variations in austral summer. In particular, the anticyclone tends to be displaced poleward in La Niña years when the southern annular mode (SAM) is in its positive phase and vice versa. Wave activity flux vectors suggest that ENSO-related convective anomalies located in the central-eastern tropical Pacific act as a remote forcing for the meridional variability of the summertime SASH. In southern winter, multiple processes operate in concert to induce interannual variability, and none of them appears to dominate like ENSO does during the summer.

1. Introduction

Located over the South Atlantic basin, the South Atlantic subtropical high (SASH) plays a significant role in the regional climate. For instance, this high pressure system is closely related to the South Atlantic subtropical dipole (SASD) mode, characterized by sea surface temperature (SST) anomalies oriented from the southwest to the northeast (Venegas et al. 1997; Sterl and Hazeleger 2003; Haarsma et al. 2005; Morioka et al. 2011). Recent studies (e.g., Richter et al. 2010; Lübbecke et al. 2010, 2014) also suggest that the strength of this anticyclone is influential in the development of the cold tongue in the eastern equatorial Atlantic, including its onset and peak amplitude during the boreal summer. Although the Indian Ocean functions as the primary moisture source for southern Africa, rainfall variability over southern Africa is found to be associated with the intensity and position of the SASH (Hermes and Reason 2009; Vigaud et al. 2009). On the western side of the South Atlantic basin, this anticyclone also affects precipitation over southeastern South America through its impact on SSTs and the South Atlantic convergence zone (Bombardi and Carvalho 2011; Bombardi et al. 2014). Understanding variations in this high pressure system is essential for understanding regional climate variations. In the past several decades, some progress has been made on the formation of subtropical highs, but investigations of the SASH are limited, and to our knowledge there have been no previous studies dedicated to directly addressing its interannual variability.

The purpose of this study is to improve our understanding of the annual cycle and interannual variability of the SASH, and how its variability is connected to the regional and large-scale circulation and coupled with the ocean. Background about subtropical anticyclones and the SASH is summarized in section 2. Section 3 provides a description of the datasets analyzed and methodologies applied. Results regarding the annual cycle and interannual variability of the SASH are reported in section 4, while the role of large-scale circulations and the underlying ocean are discussed in section 5. Conclusions are drawn in section 6.

2. Background

The historical view on the formation of subtropical anticyclones relies on the descending arm of the Hadley cell. This notion, however, is contradictory to the observation that Northern Hemisphere subtropical anticyclones are relatively strong in the summer when the Hadley cell is weak (Hoskins 1996). Using numerical models with prescribed heating and realistic topography, Rodwell and Hoskins (2001) concluded that, while nonlinear interactions between the zonal-mean flow and mountains are essential for the formation of subtropical anticyclones in austral winter, the summertime monsoonal heating to the east is fundamental for the existence of subtropical highs in both hemispheres by generating subsidence aloft. They also suggested that downward motion to the west of a monsoon circulation can be reinforced through local cooling over the eastern oceans (the so-called radiative enhancement mechanism). A later study by Seager et al. (2003) using an atmospheric general circulation model (AGCM) coupled with an ocean mixed layer model showed that monsoonal deep convective heating alone cannot produce realistically strong subtropical highs. Seager et al. (2003) suggested that air–sea interactions are also important and demonstrated that the subtropical anticyclones can be intensified in the presence of zonal asymmetry in atmospheric heating, which itself is associated with SST variations.

Another understanding of summertime subtropical highs has been developed in the framework of stationary waves. Using a linear quasigeostrophic model, Chen et al. (2001) suggested that the summertime subtropical highs in the Northern Hemisphere are maintained as a component of stationary baroclinic Rossby waves forced by upstream deep convective heating. As pointed out by Miyasaka and Nakamura (2005), however, this interpretation is limited because the meridional dependence of the observed vertical structure of summertime planetary waves (baroclinic in the subtropics but equivalent barotropic in midlatitudes) is absent in the Chen et al. (2001) study. Using the same model as Rodwell and Hoskins (2001), Miyasaka and Nakamura (2005) showed that these high pressure systems can be reproduced as a response to a local shallow cooling–heating couplet associated with land–sea contrast. Specifically, they revealed that this near-surface thermal couplet can act as a source of planetary waves, propagating upward and diverging downstream above the surface subtropical highs. The importance of shallow continental heating and maritime cooling on the formation of surface subtropical highs was highlighted in Liu et al. (2004) and Li et al. (2012), although their interpretation was based on the Sverdrup balance and focused on vertical heating profiles.

The annual cycle of subtropical highs is determined by various factors, including the Hadley circulation, radiation, and SSTs (Hastenrath 1991; Seager et al. 2003). As the transition to early summer occurs, the wintertime zonal band of high pressure underlying the subsiding branch of the Hadley cell starts to break when convection over land commences. This is accomplished through vorticity balance, inducing poleward low-level flow into the regions of deep convection on the western flank of the subtropical high, while radiatively driven subsidence and equatorward flow dominate on the eastern flank. Consequently, summertime atmosphere is relatively stable in the east and unstable in the west because descent in the east promotes subsidence drying and cools the SST while poleward flow in the west warms the subtropics. The extensive convection over the western ocean thus enhances the forcing of the subtropical anticyclones and positions them over the eastern oceans in summer. As the summer ends, the western–eastern SST asymmetry tends to be damped by vertical flux of moist static energy, and subtropical anticyclones ultimately become more zonally symmetric when precipitation in the subtropics weakens and eventually shifts to the other hemisphere in winter.

Specifically for the SASH, Richter et al. (2008) proposed that its wintertime intensity and position are linked to Asian–African summer monsoon circulations that are largely located north of the equator, while the zonal gradient of SSTs across the subtropical Atlantic is of minor importance. This interhemispheric influence was also reported in Lee et al. (2013). Their numerical investigations using an AGCM and a two-layer model suggested that the impact is realized through anomalous overturning circulations and stationary barotropic Rossby waves propagating far beyond the tropics. They also showed that this teleconnection is most dramatic in the South Pacific rather than the South Atlantic. More work is still needed to better understand the mechanisms that are controlling the SASH in the winter and summer seasons.

3. Datasets and methodology

a. Datasets

The European Centre for Medium-Range Weather Forecasts (ECMWF) interim reanalysis (ERA-Interim, hereinafter ERAI; Dee et al. 2011) and the Japanese 55-year Reanalysis (JRA-55; Kobayashi et al. 2015) are analyzed in this study. ERAI is available at spectral T255 resolution (~0.703125°) with 37 vertical levels from the Computational and Information Systems Laboratory (CISL) at the National Center for Atmospheric Research (NCAR). JRA-55 is available on a 1.25° resolution grid with 37 vertical levels as well. This dataset is obtained from the Japan Meteorological Agency (JMA).

An advantage of using ERAI and JRA-55 is that the prescribed SST forcing differs between these two reanalyses, meaning that the two reanalyses are more independent from one another, at least as far as SST forcing goes. ERAI SSTs come from multiple sources (Dee et al. 2011), including the National Centers for Environmental Prediction (NCEP) two-dimensional variational data assimilation (2D-Var) SST, NCEP Optimum Interpolation SST (OISSTv2), NCEP real-time global (RTG) SST, and operational SST and sea ice analysis (OSTIA). In contrast, JRA-55 uses the centennial in situ observation-based estimates (COBE; Ishii et al. 2005) SSTs. This estimate is based on in situ observations and does not include satellite-derived values. As demonstrated by Vizy and Cook (2016, their Fig. 1), the JRA-55 annual mean SSTs are cooler in the South Atlantic than in the ERAI. This is of particular interest here, since the SASH can be influenced by the underlying ocean.

Additionally, the 2.5° Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP; Xie and Arkin 1997) is also analyzed. This product blends gauge measurements with satellite-derived precipitation estimates from 1979 to 2015. CMAP precipitation estimates are primarily used as proxy to evaluate atmospheric latent heating fields available in JRA-55, since such fields in the reanalysis are heavily dependent upon the reanalysis modeling system. Unlike JRA-55, ERAI does not provide heating fields at each pressure level.

b. Methodology

Both reanalyses are analyzed over the 1979–2015 period. In section 4 we analyze the climatological means and interannual variability at the monthly time scale. The 850-hPa geopotential height, rather than sea level pressure, is used to depict the SASH to reduce complications from topography over the adjacent continents. The analysis of interannual variations in section 4b is based on detrended monthly data, and the position of the high is determined as the location of maximum 850-hPa geopotential height. Unlike the climatological monthly averages, months from some years are found to have multiple SASH centers (e.g., a western center closer to South America and an eastern center near Africa), indicating a positional shift of the anticyclone within the month. These cases occur mainly in the transitional month of May and extend into the southern winter. They are neglected from our analysis and amount to 67 (71) of the total 444 months for the ERAI (JRA-55).

In section 5, the analysis is conducted using seasonal means for the solstitial seasons with December, January, and February (DJF) averaged together for austral summer and June, July and August (JJA) averaged for austral winter. For instance, the average of December of 1979 and January and February of 1980 is considered as the 1979 summer here. The traditional definitions of summer and winter are preferred because the large-scale circulation is relatively consistent within these months.

To better understand how interannual SASH variability is connected to the regional and large-scale atmospheric circulation, Z-score-based composite anomalies are constructed for austral summer (DJF) and winter (JJA) after removing the long-term trend from 1979 to 2015 in ERAI, JRA-55, and CMAP precipitation datasets. Four composites are formulated for the summer and winter seasons (viz., a south, north, west, and east composite) according to their relative position to the climatological seasonal location of the SASH for each reanalysis. Using σφ (σλ) to denote the standard deviation of the latitude (longitude) relative to the climatological mean value, a year is selected for inclusion in the composite if its standard deviation in the direction being considered (i.e., either latitude or longitude) is more than 0.5 standard deviations away from the climatological position in both reanalyses, ERAI and JRA-55. For example, a west (east) composite year must have its longitude be less than −0.5 (greater than +0.5) standard deviations away from the climatological longitude of the SASH for both reanalyses. This approach guarantees a sufficient sampling of the years and confirms that selected cases are not reanalysis dependent. As listed in Table 1, the number of cases considered in the composites amounts to ~20%–35% of the years when the SASH is well defined. Because latitude and longitude are considered independently, some years included in the south or north composite are also present as west or east cases.

Table 1.

DJF and JJA seasons used to formulate the south, north, west, and east SASH composites.

DJF and JJA seasons used to formulate the south, north, west, and east SASH composites.
DJF and JJA seasons used to formulate the south, north, west, and east SASH composites.

To explore remote forcing of SASH interannual variability, the wave activity flux proposed by Takaya and Nakamura (2001) is calculated. In spherical coordinates and for stationary waves, the horizontal components of this three-dimensional (3D) quantity can be written as follows:

 
formula
 
formula

where λ (φ) is longitude (latitude), p denotes normalized pressure (divided by 1000 hPa), a is the radius of Earth, ψ stands for geostrophic streamfunction, U indicates zonal wind, and |U| is effectively the magnitude of the climatological horizontal wind (U, V) as the vertical velocity is assumed to be zero. The primes in Eqs. (1) and (2) represent deviations from the long-term mean. This formulation is an extension of the Eliassen–Palm (E–P) flux (Eliassen and Palm 1960). It indicates the propagation of large-scale quasi-stationary Rossby waves because the local group velocity is parallel to the wave activity flux. The divergence and convergence of (Fλ, Fφ) indicate where the wave packet is emitted and absorbed, respectively. Compared with other variations of the E–P flux in 3D configurations (e.g., Plumb 1985), the Takaya and Nakamura (2001) extension is more suited for diagnosing wave activity associated with anomalous circulations under zonally varying basic flows and applied in our composite analysis using detrended seasonal data at 200 hPa, where the circulation variance is greatest since this level is close to the core of the tropospheric jet stream (Ding et al. 2012).

As described later in section 5, the composite anomalous patterns are rather complex, especially for southern winter. To achieve a better understanding, empirical orthogonal function (EOF) analysis (Davis 1986) of 850-hPa geopotential height is conducted over the region from the equator to the South Pole between 70°W and 20°E. The longitudinal boundaries of this analysis domain correspond to the westward and eastward extent of the SASH as indicated by the 10-gpm geopotential height anomalies shown in Fig. 1, and we take 90° latitude for the analysis because the composite anomalies as discussed later appear to be related to the southern annular mode (SAM; Thompson and Wallace 2000). Atmospheric fields and CMAP rainfall are then projected onto the principal component (PC) time series of these EOFs to diagnose relevant physical mechanisms.

Fig. 1.

ERAI 1979–2015 climatological monthly 850-hPa scaled geopotential height (gpm). The values are scaled by removing the annual mean of the area-averaged geopotential height (1517.16 gpm) over the South Atlantic region bounded by 45°S–0° and 45°W–10°E. The 10- and 28-gpm values are marked by black contours.

Fig. 1.

ERAI 1979–2015 climatological monthly 850-hPa scaled geopotential height (gpm). The values are scaled by removing the annual mean of the area-averaged geopotential height (1517.16 gpm) over the South Atlantic region bounded by 45°S–0° and 45°W–10°E. The 10- and 28-gpm values are marked by black contours.

4. Seasonal and interannual variability of the SASH

a. Annual cycle

Figure 1 shows ERAI 1979–2015 climatological monthly 850-hPa geopotential heights over the South Atlantic. To better quantify the size of the anticyclone as discussed later, geopotential height is scaled by subtracting the annual mean geopotential height averaged from 45°S to the equator and from 45°W to 10°E. As indicated by the 10-gpm anomalies, the SASH is confined within the South Atlantic basin during the austral summer (Figs. 1a–c). During the fall (Figs. 1d–f), the anticyclone weakens progressively and extends toward the adjacent continents. After a sudden increase from May to June, the SASH remains strong through the winter (Figs. 1g–i) before weakening and contracting once again in austral spring (Figs. 1j–l). From May to September (Figs. 1f–j), when land surface temperatures over southern Africa are colder than over the adjacent South Atlantic, the SASH is connected with the Mascarene high over the Indian Ocean (Ohishi et al. 2015). The SASH behaves similarly in JRA-55 (not shown).

Figure 2a quantifies variations in the climatological SASH strength over the annual cycle. Here, the SASH intensity is defined as the maximum monthly 850-hPa geopotential height over the South Atlantic (45°S–0°, 45°W–10°E). A distinct annual cycle occurs in the reanalyses with two peaks, the first in February (~1560 gpm) and a second maximum in July (~1568 gpm). The anticyclone is weakest during April and May (~1548 gpm). Both reanalyses exhibit the same seasonal cycle, but the intensity in JRA-55 is consistently lower than in ERAI by up to ~2.8 gpm.

Fig. 2.

The 1979–2015 climatological SASH intensity (gpm) as a function of (a) month, (b) latitude, and (c) longitude. Also 1979–2015 climatological SASH size (106 km2) as a function of (d) month, (e) latitude, and (f) longitude. Orange (green) circles denote values from ERAI (JRA-55). The numbers in (b),(c),(e),(f) denote the associated calendar month of the year for each circle.

Fig. 2.

The 1979–2015 climatological SASH intensity (gpm) as a function of (a) month, (b) latitude, and (c) longitude. Also 1979–2015 climatological SASH size (106 km2) as a function of (d) month, (e) latitude, and (f) longitude. Orange (green) circles denote values from ERAI (JRA-55). The numbers in (b),(c),(e),(f) denote the associated calendar month of the year for each circle.

The sensitivity of these results to the definition of SASH intensity is evaluated by calculating two other metrics and repeating the above analysis. The first metric defines intensity as the area-averaged geopotential height within the scaled 28-gpm anomaly contours, as they are the lowest closed isoheight anomalies at 850 hPa every month in both ERAI (Fig. 1) and JRA-55 (not shown). The second metric defines the SASH intensity as the area-averaged geopotential height within ±20° longitude and ±10° latitude of the climatological SASH centers. Both alternative metrics yield results (not shown) consistent with those presented in Fig. 2a.

Figures 2b and 2c relate the SASH intensity to its position in latitude and longitude, respectively. The SASH fluctuates about 6° latitude over the climatological annual cycle in both reanalyses; it is farthest north in the austral winter and farthest south during austral summer (Fig. 2b). The longitudinal position of the SASH varies by 14° longitude (Fig. 2c); it is farthest east during the fall (March–April) and spring (October–November) and farthest west during austral winter. In austral summer the anticyclone is centrally located in the basin. These results are broadly consistent with Hastenrath (1991, his Fig. 6.3:1), except that the SASH was reported as farthest east in February.

Figure 2d shows variations in the size of the SASH over the annual cycle for the two reanalyses at 850 hPa. Again, the 28-gpm anomalies are used to define the extent of the anticyclone. The annual cycles are similar in the two reanalysis products, with differences less than ~0.5 × 106 km2. The SASH reaches its largest size in July (~10.6 × 106 km2 in ERAI and ~10.2 × 106 km2 in JRA-55), and it is smallest during April–May (~0.5–1.0 × 106 km2 in ERAI and JRA-55). The size of the anticyclone is dependent on the definition of the SASH area, including the region used for scaling and the corresponding anomalous isoheights closed within the South Atlantic area, but results for the annual cycle are robust.

Figures 2e and 2f display the SASH size relative to the latitudinal and longitudinal positions of its center, respectively. The anticyclone is largest during the solstitial months when the center of the SASH is either closest to the equator and on the western side of the South Atlantic in the winter or farthest from the equator in the center of the South Atlantic in late summer (February).

b. Interannual variability of the SASH

Figure 3 shows the position of the monthly SASH centers when the anticyclone is well defined (see section 3b for more details). In austral summer (DJF; Figs. 3a–c), the SASH center is confined within about 4° latitude and 12° longitude of its climatological position in both reanalyses. From January to February, westward excursions of the SASH center are greater than eastward excursions. During the fall (MAM; Figs. 3d–f), the center location is more variable than in DJF, ranging from 43°W to 3°E and between 20° and 35°S. The substantial longitudinal spread, especially westward, continues in the southern winter (JJA; Figs. 3g–i), but the latitudinal spread is less (20°–30°S) than in the fall. Compared with MAM, the climatological center is more representative of where one would estimate the center using the monthly positions, although a slight eastward skew exists in June. For austral spring (Figs. 3j–l), September is similar to JJA and highly variable, while October and November behave more like DJF. Overall, the location of the high varies the most from May to August when synoptic variability is relatively strong in the Southern Hemisphere. This may partially explain why the individual cases dropped from the analysis occur primarily during these months.

Fig. 3.

Monthly scatterplots of the SASH’s central position from ERAI (orange) and JRA-55 (green) for each year from 1979 to 2015 when a SASH is well defined. The 1979–2015 climatological central SASH positions in each panel for ERAI and JRA-55 are denoted by the black solid and dashed crosses, respectively. The number of monthly cases involved in the analysis from ERAI and JRA-55 is indicated in the panel titles.

Fig. 3.

Monthly scatterplots of the SASH’s central position from ERAI (orange) and JRA-55 (green) for each year from 1979 to 2015 when a SASH is well defined. The 1979–2015 climatological central SASH positions in each panel for ERAI and JRA-55 are denoted by the black solid and dashed crosses, respectively. The number of monthly cases involved in the analysis from ERAI and JRA-55 is indicated in the panel titles.

Scatterplots of the monthly SASH intensity and the latitudinal position of the anticyclone center are presented in Fig. 4. The SASH is stronger when it is located farther south in all months. The intensity of the SASH fluctuates by about 40–60 gpm with smallest variations in May (Fig. 4f) and largest variations in February (Fig. 4c) and August (Fig. 4i). Linear regression shows that the relationship between the monthly SASH intensity and its latitudinal position (solid and dashed lines in Fig. 4) is statistically significant at the 95% level of confidence after considering autocorrelation (Zwiers and von Storch 1995). The monthly correlation coefficients r2 for each reanalysis are also reported in Fig. 4, suggesting that the variance explained by linear regression is not high (up to 0.68). This may result from the large interannual variability of the anticyclone. As documented in Fig. 4, the standard deviation of the SASH intensity σ is on the order of 10 gpm and largely proportional to its climatological magnitude, with two peaks in February and June but a minimum in May. The differences of σ between ERAI and JRA-55 are less than 1 gpm.

Fig. 4.

Monthly scatterplots of SASH intensity (gpm) from ERAI (orange) and JRA-55 (green) for each year from 1979 to 2015 when a SASH is well defined. Standard deviations for ERAI and JRA-55 are shown in each panel. Solid (dashed) black lines indicate linear regressions for ERAI (JRA-55), with corresponding r2 values indicated in the panel titles. In each month, the regression is statistically significant at the 95% level of confidence after accounting for autocorrelation (Zwiers and von Storch 1995).

Fig. 4.

Monthly scatterplots of SASH intensity (gpm) from ERAI (orange) and JRA-55 (green) for each year from 1979 to 2015 when a SASH is well defined. Standard deviations for ERAI and JRA-55 are shown in each panel. Solid (dashed) black lines indicate linear regressions for ERAI (JRA-55), with corresponding r2 values indicated in the panel titles. In each month, the regression is statistically significant at the 95% level of confidence after accounting for autocorrelation (Zwiers and von Storch 1995).

Although the monthly SASH size is positively correlated with the SASH intensity during each month, its dependence on the latitudinal position of the high is statistically insignificant in December, January, March, and July (not shown). Except for September and November, the correlation coefficients between the size and center latitude of the anticyclone are below 0.3. In contrast, the relationship between the intensity/size and longitudinal position of the SASH is not uniform and generally insignificant (not shown).

5. Connections between SASH interannual variability and the large-scale climate

To better understand interannual variability of the SASH, austral summer (DJF) and winter (JJA) composites are formulated for cases when the SASH is south, north, west, and east of its climatological seasonal mean position (section 3b). Table 1 lists the years that are included in each composite.

a. Austral summer

For reference, Figs. 5a and 5b display climatological summertime 850-hPa geopotential heights from ERAI and JRA-55, respectively. North and south composite anomalies (i.e., differences from the climatological mean) are shown in Figs. 5c,d and Figs. 5e,f, respectively. There is a dipole geopotential height anomaly centered around 55°S. The hemisphere-scale symmetry between the south and north composites implies that the summertime meridional variability of the SASH is associated with large-scale processes. Although the positive (Figs. 5c,d) or negative (Figs. 5e,f) height anomalies break down south of Australia, these anomalies are uniformly distributed along the latitude band from ~50° to 30°S at upper levels (not shown). This pattern resembles the SAM (Thompson and Wallace 2000) and indicates that the SASH tends to be positioned to the south (north) when the SAM is in its positive (negative) phase in austral summer. An examination of each composite member confirms that these mean anomalies are not dominated by one or a few years.

Fig. 5.

Austral summer (DJF) 850-hPa geopotential height (gpm) climatology (1979–2015) from the (a) ERAI and (b) JRA-55. (c)–(j) As in (a),(b), but for the anomalies associated with the (c),(d) south, (e),(f) north, (g),(h) west, and (i),(j) east composites.

Fig. 5.

Austral summer (DJF) 850-hPa geopotential height (gpm) climatology (1979–2015) from the (a) ERAI and (b) JRA-55. (c)–(j) As in (a),(b), but for the anomalies associated with the (c),(d) south, (e),(f) north, (g),(h) west, and (i),(j) east composites.

West and east composites are shown in Figs. 5g–j. The 850-hPa geopotential height anomalies are generally weaker than for the north and south composites and there is less symmetry, indicating that a west–east mode of variability is not prominent during austral summer. This is consistent with EOF analysis (not shown). The first EOF, explaining ~50% of the total variations in both reanalyses, is related to the SAM and exhibits patterns associated with the meridional displacement of the SASH, while the third EOF represents the west–east displacement of the SASH and only accounts for 9.7% of the variability in ERAI and 9.3% in JRA-55. The second EOF explains 21.2% of the variability in ERAI and JRA-55, and it is linked to both meridional and zonal displacement.

The relationship between the south–north variations of the SASH and the SAM suggested by Fig. 5 is further confirmed using the station-based SAM index from Marshall (2003). This index reflects the zonal pressure difference between 40° and 65°S, with negative (positive) values corresponding to the negative (positive) phase of the SAM. As indicated by the summer (DJF means) SAM index in Figs. 6a,b, out of a total of eight (nine) south (north) members, six (seven) occur when the SAM is in its positive (negative) phase. Since tropical SSTs in austral summer can be an important forcing for the SAM (Grassi et al. 2005), Figs. 6c,d relate the multivariate El Niño–Southern Oscillation (ENSO) index (MEI; Wolter and Timlin 2011) to the latitudinal position of the SASH. Seven (eight) out of eight (nine) south (north) composite events occur during La Niña (El Niño) events, as indicated by the negative (positive) values of the MEI (Figs. 6c,d). These results are consistent with previous studies showing that the phase of the SAM is negatively correlated with ENSO in southern summer (e.g., Fogt and Bromwich 2006; Gong et al. 2010).

Fig. 6.

Austral summer (DJF) SAM index (Marshall 2003) associated with the south and north composite years from the (a) ERAI and (b) JRA-55, where a blue (red) dot indicates that the SAM is in its negative (positive) phase. (c),(d) As in (a),(b), but for the MEI (Wolter and Timlin 2011) with La Niña (El Niño) conditions in blue (red). The 1979–2015 climatological SASH center latitudes are marked by the dashed gray lines.

Fig. 6.

Austral summer (DJF) SAM index (Marshall 2003) associated with the south and north composite years from the (a) ERAI and (b) JRA-55, where a blue (red) dot indicates that the SAM is in its negative (positive) phase. (c),(d) As in (a),(b), but for the MEI (Wolter and Timlin 2011) with La Niña (El Niño) conditions in blue (red). The 1979–2015 climatological SASH center latitudes are marked by the dashed gray lines.

Given that the physics of subtropical highs is closely related to atmospheric heating (section 2), we relate north–south displacements of the SASH to latent heating variations. Climatological 500-hPa latent heating from JRA-55 and precipitation from CMAP during Southern Hemisphere summer are shown in Figs. 7a and 7b, respectively. The two fields are in good agreement in both convective regions and areas with relatively weak rainfall, suggesting that the latent heating field in JRA-55 is realistic.

Fig. 7.

Austral summer (DJF) 1979–2015 climatological (a) JRA-55 500-hPa latent heating (K day−1) and (b) CMAP precipitation (mm day−1). (c) DJF JRA-55 500-hPa latent heating anomalies (shading; K day−1), 850-hPa geopotential height anomalies (contours; gpm), and 200-hPa wave activity flux (vectors; m2 s−2) for the south composite. (d) DJF CMAP precipitation anomalies (shading; mm day−1) and JRA-55 850-hPa geopotential height anomalies (contours; gpm) for the south composite. (e),(f) As in (c),(d), but for the DJF north composite. The 200-hPa wave activity flux vectors with magnitudes less than 0.1 m2 s−2 are omitted.

Fig. 7.

Austral summer (DJF) 1979–2015 climatological (a) JRA-55 500-hPa latent heating (K day−1) and (b) CMAP precipitation (mm day−1). (c) DJF JRA-55 500-hPa latent heating anomalies (shading; K day−1), 850-hPa geopotential height anomalies (contours; gpm), and 200-hPa wave activity flux (vectors; m2 s−2) for the south composite. (d) DJF CMAP precipitation anomalies (shading; mm day−1) and JRA-55 850-hPa geopotential height anomalies (contours; gpm) for the south composite. (e),(f) As in (c),(d), but for the DJF north composite. The 200-hPa wave activity flux vectors with magnitudes less than 0.1 m2 s−2 are omitted.

The 500-hPa latent heating anomalies from JRA-55 and CMAP precipitation anomalies associated with the south and north composites are provided in Figs. 7c–e (shaded). Vectors in Figs. 7c,e denote the associated wave activity flux at 200 hPa. Note that the wave activity flux is masked out between 10°S and 10°N because quasigeostrophic theory does not hold at low latitudes. The anomalous heating and precipitation over the Pacific (Figs. 7c–e) are consistent with Figs. 6c,d, indicating again that the SASH tends to be located to the south (north) during La Niña (El Niño) events. The heating anomalies are largest in the central-eastern tropical Pacific near 105°W, from which a wave train emanates as indicated by the wave activity flux shown in Figs. 7c,e. For both the south and north composites, the wave train from the tropics propagates toward the southeast and passes over southern South America before reaching the SASH.

Another distinct feature in Figs. 7c–f is variations of the convergence zones, either shifted or changed in magnitude. For the South Pacific convergence zone, its migration is likely related to ENSO as shown by Folland et al. (2002). As suggested by the corresponding 850-hPa geopotential height anomalies in DJF (black contours in Figs. 7c–f), displacement of the south Indian Ocean (Cook 2000) and South Atlantic convergence zones may be a response to the south–north displacement of the Mascarene high in the Indian Ocean and the SASH. The changes of the intertropical convergence zone over the equatorial Atlantic are potentially related to inter-Pacific–Atlantic variability (Wang 2006) through modifications of the SST gradient between these two basins owing to ENSO. Overall, these anomalies are likely consequences of ENSO, rather than forcings of a southward or northward shifted SASH.

Figure 8 shows regression maps of the 500-hPa latent heating anomalies from JRA-55 and CMAP precipitation associated with EOF1, EOF2, and EOF3. Although the level of statistical confidence is moderate (80% and higher), indicating the relationship between 850-hPa geopotential height and latent heating or rainfall is not perfectly linear, the regression maps of latent heating (Figs. 8a,c,e) and rainfall (Figs. 8b,d,f) agree well. The EOF1–EOF3 regression map anomalous patterns are remarkably similar to each other (cf. Figs. 8a,c,e; Figs. 8b,d,f), with the strongest signals in the tropics, and the equatorial Pacific in particular. This implies that ENSO is the primary forcing for the summertime variability of SASH.

Fig. 8.

Austral summer (DJF) regression maps of 500-hPa latent heating anomalies (K day−1) associated with EOF1 from (a) JRA-55 and (b) CMAP. (c),(d) As in (a),(b), but for EOF2. (e),(f) As in (a),(b), but for EOF3. The amplitudes correspond to the anomaly values that would occur in association with one standard deviation of each principal component time series. The variance explained by each EOF is indicated in the panel titles. Hatching denotes regions where the regression is statistically significant at the 80% level of confidence after accounting for autocorrelation (Zwiers and von Storch 1995).

Fig. 8.

Austral summer (DJF) regression maps of 500-hPa latent heating anomalies (K day−1) associated with EOF1 from (a) JRA-55 and (b) CMAP. (c),(d) As in (a),(b), but for EOF2. (e),(f) As in (a),(b), but for EOF3. The amplitudes correspond to the anomaly values that would occur in association with one standard deviation of each principal component time series. The variance explained by each EOF is indicated in the panel titles. Hatching denotes regions where the regression is statistically significant at the 80% level of confidence after accounting for autocorrelation (Zwiers and von Storch 1995).

b. Austral winter

Climatological wintertime (JJA) 850-hPa geopotential heights from ERAI and JRA-55 are shown in Figs. 9a and 9b, respectively. The two datasets are similar to each other, including for the location and strength of the subtropical anticyclones over the Pacific, Atlantic, and Indian Oceans. The 850-hPa geopotential height anomalies from ERAI and JRA-55 for the south, north, west, and east composites are shown in Figs. 9c–j. Compared with the austral summer (Figs. 5c–j), west–east variations of the SASH are more pronounced during austral winter but still weaker than north–south variability. As for austral summer, geopotential height anomalies are primarily located on the southern flank of the SASH from ~30° to 55°S.

Fig. 9.

As in Fig. 5, but for austral winter (JJA).

Fig. 9.

As in Fig. 5, but for austral winter (JJA).

Figures 10a and 10b display the climatological wintertime 500-hPa latent heating from JRA-55 and CMAP precipitation, respectively. The JRA-55 heating field is similar to the observed rainfall, not only for the tropical rainfall belt but also along the Southern Hemisphere storm track (40°–50°S). The climatological 200-hPa divergent wind in Fig. 10a (vectors) shows upper-level convergence over the SASH, and in the equatorial and tropical South Atlantic the near-surface wind is on the order of 10 m s−1 (vectors in Fig. 10b).

Fig. 10.

Austral winter (JJA) 1979–2015 climatological (a) JRA-55 500-hPa latent heating (shading; K day−1) and 200-hPa divergent wind (vectors; m s−1) and (b) CMAP precipitation (shading; mm day−1) and JRA-55 10-m winds (vectors; m s−1). (c) JJA JRA-55 500-hPa latent heating anomalies (shading; K day−1) and 200-hPa divergent wind anomalies (vectors; m s−1) for the south composite. (d) DJF CMAP precipitation anomalies (shading; mm day−1) and JRA-55 10-m wind anomalies (vectors; m s−1) for the south composite. (e),(f) As in (c),(d), but for the JJA north composite. Vectors in (c)–(f) with magnitude less than 0.1 m s−1 are omitted.

Fig. 10.

Austral winter (JJA) 1979–2015 climatological (a) JRA-55 500-hPa latent heating (shading; K day−1) and 200-hPa divergent wind (vectors; m s−1) and (b) CMAP precipitation (shading; mm day−1) and JRA-55 10-m winds (vectors; m s−1). (c) JJA JRA-55 500-hPa latent heating anomalies (shading; K day−1) and 200-hPa divergent wind anomalies (vectors; m s−1) for the south composite. (d) DJF CMAP precipitation anomalies (shading; mm day−1) and JRA-55 10-m wind anomalies (vectors; m s−1) for the south composite. (e),(f) As in (c),(d), but for the JJA north composite. Vectors in (c)–(f) with magnitude less than 0.1 m s−1 are omitted.

The 500-hPa latent heating anomalies from JRA-55 and CMAP precipitation anomalies associated with the austral winter south and north composites are shown in Figs. 10c–f. Vectors in Figs. 10c,e denote the 200-hPa divergent wind anomalies from JRA-55, while the vectors in Figs. 10d,f denote 10-m wind anomalies. When the SASH is located to the south, latent heating increases over South America but slightly decreases over the eastern Pacific (Fig. 10c). If this cooling–heating couplet is sufficiently strong (e.g., during austral summer), it can induce subsidence to the southeast of South America by generating stationary waves propagating downstream (Miyasaka and Nakamura 2010, their Fig. 6d), moving the SASH to the south. The wintertime anomalies, however, are likely too weak to produce significant impacts because the corresponding rainfall anomalies up to ~0.3 mm day−1 (Figs. 10d,f) are merely equivalent to anomalous heating of ~8.5 W m−2.

Over the equatorial Atlantic, convection is suppressed in the south composite but enhanced in the north composite (Figs. 10c,e). Discrepancies between 500-hPa latent heating anomalies and CMAP rainfall anomalies exist for the north composite; rather than a uniform enhancement over the equatorial Atlantic, precipitation decreases north of the equator but increases southward over the tropical Atlantic (Figs. 10e,f). As the equatorial Atlantic SST is colder in the south composite but warmer in the north composite (not shown), the change of convective overturning is likely a response to the underlying ocean. Further, streamlines constructed from the divergent component of the zonal wind and the vertical pressure velocity averaged between the equator and 20°N show a Walker-type anomalous circulation with its rising branch located in the Pacific and subsiding branch over the Atlantic for the south composite (not shown), suppressing convection over the equatorial Atlantic. This anomalous circulation is lacking in the north composite, in agreement with its relatively weaker heating anomalies in the equatorial Pacific (Figs. 10e,f) and warmer equatorial Atlantic SST (not shown).

Variations of equatorial Atlantic SSTs, however, are not due to SASH variations because the associated surface wind anomalies in the South Atlantic basin are mainly constrained within the subtropics for both south and north composites (Figs. 10d,f), and in the tropics the changes are small compared to their climatological magnitude (Fig. 10b). Nevertheless, the upper-level circulation responds accordingly. As seen in the 200-hPa divergent wind anomalies (Figs. 10c), upper-level divergence over the equatorial Atlantic (Fig. 10a) is weaker in the south composite, consistent with a weakening of the northern portion of the anticyclone. For the north composite, upper-level divergent wind anomalies are deflected toward Chad and the Congo basin, and convection is reduced in the former area (Fig. 10e).

Over the western Pacific warm pool, latent heating anomalies are coherent and robust in the south composite, while the composite pattern over the Indian Ocean is different in the south and north composites. To the southwest of the Indian subcontinent (i.e., the Arabian Sea), convective motion is suppressed in the south composite (Figs. 10c,d) but enhanced in north composite (Figs. 10e,f). This is consistent with Richter et al. (2008), who suggest that the position of the wintertime SASH is related to convection over eastern tropical Africa and India.

Latent heating and precipitation anomalies associated with the west and east composites are presented in Fig. 11. Vectors in Figs. 11a,c denote the anomalous divergent wind at 200 hPa for each composite, while Figs. 11b,d show the anomalous near-surface wind. Compared with the south and north composites, a distinct contrast occurs over the equatorial Pacific in the heating anomaly for the west and east composites. Instead of having the same sign as in the south and north composites (cf. Figs. 10c,e; Figs. 10d,f), the signals are of opposite sign for these two composites (cf. Figs. 11a,c; Figs. 11b,d), advocating that convective anomalies over the central equatorial Pacific play a fundamental role in the zonal displacement of the SASH. In fact, SST anomalies in the eastern equatorial Pacific associated with EOF1 and EOF3 (not shown) suggest that ENSO influences the positional variations of the high in austral winter.

Fig. 11.

Austral winter (JJA) JRA-55 500-hPa latent heating anomalies (shading; K day−1) and 200-hPa divergent wind anomalies (vectors; m s−1) for the west composite. (b) CMAP precipitation anomalies (shading; mm day−1) and JRA-55 10-m wind anomalies (vectors; m s−1) for the west composite. (e),(f) As in (c),(d), but for the JJA east composite. Vectors in (a)–(d) with magnitude less than 0.1 m s−1 are omitted.

Fig. 11.

Austral winter (JJA) JRA-55 500-hPa latent heating anomalies (shading; K day−1) and 200-hPa divergent wind anomalies (vectors; m s−1) for the west composite. (b) CMAP precipitation anomalies (shading; mm day−1) and JRA-55 10-m wind anomalies (vectors; m s−1) for the west composite. (e),(f) As in (c),(d), but for the JJA east composite. Vectors in (a)–(d) with magnitude less than 0.1 m s−1 are omitted.

The upper-level divergent wind anomalies in Fig. 11a indicate enhanced convergence over the SACZ in the west composite, which is associated with a suppression of SACZ and a westward intensification of the SASH. Because climatological wintertime moisture transport toward the SACZ is weaker than in summer, we suggest that this upper-level impact dominates low-level processes or surface conditions in the formation of the SACZ. This hypothesis is supported by the minimal changes of surface wind over South America for the west and east composites (Figs. 11b,d). As shown in Figs. 10e,f, similar anomalous 500-hPa latent heating and upper-level wind anomalies over the SACZ are also present in the north composite over the tropical and subtropical South Atlantic, suggesting that the meridional and zonal displacement of the anticyclone are both highly related during southern winter. In fact, in southern winter the SASH tends to be positioned to the north when it is displaced to the west and vice versa (Figs. 3g–i).

In Fig. 12, 850-hPa geopotential height anomalies are projected onto the PC time series associated with the first three EOFs for austral winter. If the patterns associated with EOF1 (Figs. 12a,b) are interpreted as being related to the SAM (Thompson and Wallace 2000), and inducing the meridional displacement of the high, only about 38% of the total variance is explained, much lower than during the summertime when around 50% of the variance is explained by EOF1.

Fig. 12.

Austral winter (JJA) regression maps of 850-hPa geopotential height anomalies (gpm) associated with EOF1 from (a) ERAI and (b) JRA-55. (c),(d) As in (a),(b), but for EOF2. (e),(f) As in (a),(b), but for EOF3. The amplitudes correspond to the anomaly values that would occur in association with one standard deviation of each principal component time series. The variance explained by each EOF is indicated in the panel titles. Hatching denotes regions where the regression is statistically significant at the 90% level of confidence after accounting for autocorrelation (Zwiers and von Storch 1995).

Fig. 12.

Austral winter (JJA) regression maps of 850-hPa geopotential height anomalies (gpm) associated with EOF1 from (a) ERAI and (b) JRA-55. (c),(d) As in (a),(b), but for EOF2. (e),(f) As in (a),(b), but for EOF3. The amplitudes correspond to the anomaly values that would occur in association with one standard deviation of each principal component time series. The variance explained by each EOF is indicated in the panel titles. Hatching denotes regions where the regression is statistically significant at the 90% level of confidence after accounting for autocorrelation (Zwiers and von Storch 1995).

The regression maps of geopotential height anomalies for the second EOF (Figs. 12c,d) suggest a wave train emanating east of Australia, propagating to the southeast, passing over the southern tip of South America, and reaching the South Atlantic before turning northeastward. Specifically, the northern portion of the negative geopotential height anomalies centered around 65°S, 70°W overlaps with the western part of the anomalous geopotential height associated with each composite shown in Fig. 9.

This wave train hypothesis is supported by the regression maps of 500-hPa latent heating anomalies and precipitation anomalies shown in Fig. 13. In particular, EOF2 in the JRA-55 shows anomalous cooling over the Coral Sea to the northeast of Australia (Fig. 13c), and the regression map of CMAP precipitation associated with EOF2 also shows decreased rainfall therein (Fig. 13d). A similar wave train pattern was identified by Ding et al. (2012, their Fig. 16). They suggest that this wave train is important during nonsummer seasons and may induce zonal asymmetry in the SAM.

Fig. 13.

As in Fig. 8, but for austral winter (JJA).

Fig. 13.

As in Fig. 8, but for austral winter (JJA).

For EOF3, the regressed geopotential height anomalies (Figs. 12e,f) are less uniform over the polar region than those associated with EOF1 (Figs. 12a,b), but less wavy than the anomalous geopotential height related to EOF2 (Figs. 12c,d). The centers of these negative geopotential height anomalies over the South Pacific (~50°S, 165°E), South Atlantic (~50°S, 10°W), and southern Indian Ocean (~40°S, 90°E) are located at 50°–40°S and far away from the topography of South America and southern Africa, suggesting that this mode is related to the Southern Hemisphere storm track. This is consistent with the regressed 500-hPa latent heating patterns, showing positive anomalies over these regions, especially for the South Atlantic and southern Indian Ocean (Figs. 13e). Because the anomalous upper-level convergence–divergence contrast centered around 35°S, 5°W is more prominent in the south and north composites (Figs. 10c,d) than in the west and east cases (Figs. 11a,c), this EOF may explain more about meridional variations of the SASH.

6. Summary and conclusions

The South Atlantic subtropical high (SASH) is the primary circulation feature over the South Atlantic basin. Understanding variations in this high pressure system is essential to advance our knowledge about regional climate variations. In this study, two reanalysis products, the ERAI and JRA-55, are examined to detail the SASH’s climatology and how its interannual variability relates to the large-scale climate. To achieve this, seasonal composites are constructed after detrending the data and EOF analysis is conducted.

The annual cycle of the SASH is found to have two distinct peaks in both intensity and size. The SASH is strongest and largest during the solstitial months when its center is either closest to the equator and on the western side of the South Atlantic basin as in austral winter or farthest poleward in the center of the South Atlantic basin as in late austral summer. Although year-to-year variations in the SASH’s position are larger in the zonal direction (i.e., up to ~40° longitude from May to August) than in the meridional direction, the intensity of the high decreases when it is positioned to the north. This relationship is statistically significant at the 95% level of confidence in every month.

In austral summer, the spatial variations of SASH on the interannual time scale are primarily in the meridional direction and dominated by ENSO. Take the El Niño condition for instance: the SST anomalies in the tropics can render the SAM into its negative phase via the adjustment of the Hadley circulation [see Lim et al. (2013) for a summary and the references therein for detailed dynamics]. The resulting zonal wind anomalies thus shift atmospheric mass toward the South Pole (Sen Gupta and England 2006), reducing the intensity of SASH and shifting the anticyclone equatorward (Fig. 4). Wave activity flux vectors in Figs. 7c,e suggest that ENSO-related convective anomalies located in the central-eastern tropical Pacific centered at ~105°W act as a remote forcing for the meridional variability of the high. As elucidated by the EOF analysis (Fig. 8), although the west–east mode of SASH displacement is not prominent in this season, it is also related to ENSO.

Understanding the interannual variability of the SASH during the austral winter is not as straightforward as it is during the summer. Composite and EOF analysis indicate that multiple physical mechanisms appear to impact the position of the SASH during this time of the year. For example, SAM and ENSO are still influential, explaining about 38% of the total variance. This, however, is less than during austral summer when SAM–ENSO forcing explains ~50% of the total variance. Variations in the Asian–African monsoon also appear to be influential (Figs. 10 and 11) to some extent, which is consistent with the modeling work of Richter et al. (2008). Additionally, anomalous forcing from convection over the Arabian Sea and over the Coral Sea to the northeast of Australia, as well as the variations of Southern Hemisphere storm track, is also identified as related to the variability of the SASH. Unlike during the austral summer, none of these mechanisms clearly dominates over the SASH variability, resulting in the complexity when trying to understand the austral winter SASH variability. This study is intended as a first step to identify physical processes associated with SASH variability. More work is still needed to better understand the exact role(s) of each process discussed above and how they interact to influence the SASH and the South Atlantic regional climate variability.

As the source of the southeast trade winds in the Atlantic basin, the SASH plays an important role in shaping the climate over the equatorial Atlantic. The limited surface wind anomalies associated with each composite shown in Figs. 10 and 11, however, suggest that wintertime spatial variations of the SASH are inefficient in inducing SST changes in this region, although the related wind anomalies are much larger in other seasons (not shown). This is in agreement with Trzaska et al. (2007), who concluded that the equatorial Atlantic SST anomalies during austral winter predominantly act as a forcing to the atmosphere and impact the SASH via a Rossby wave train extending to the extratropics, but not vice versa. Because remote forcing mechanisms are crucial on the interannual variability of the SASH, our results may also explain the sensitivity of simulated SASH to domain size as reported in Cabos et al. (2017).

Acknowledgments

We thank two anonymous reviewers for their constructive suggestions and comments. This research was funded by National Aeronautics and Space Administration (NASA) Award NNX13AQ76G. The ERAI product was obtained from the Computational and Information Systems Laboratory (CISL) at the National Center for Atmospheric Research (NCAR). We also gratefully acknowledge the Climate Prediction Division of the Japan Meteorological Agency for the dissemination of JRA-55. The Texas Advanced Computing Center (TACC) at The University of Texas at Austin provided the high-performance computing and data storage resources.

REFERENCES

REFERENCES
Bombardi
,
R. J.
, and
L. M. V.
Carvalho
,
2011
:
The South Atlantic dipole and variations in the characteristics of the South American monsoon in the WCRP-CMIP3 multi-model simulations
.
Climate Dyn.
,
36
,
2091
2102
, doi:.
Bombardi
,
R. J.
,
L. M. V.
Carvalho
,
C.
Jones
, and
M. S.
Reboita
,
2014
:
Precipitation over eastern South America and the South Atlantic sea surface temperature during neutral ENSO periods
.
Climate Dyn.
,
42
,
1553
1568
, doi:.
Cabos
,
W.
, and Coauthors
,
2017
:
The South Atlantic anticyclone as a key player for the representation of the tropical Atlantic climate in coupled climate models
.
Climate Dyn.
, doi:, in press.
Chen
,
P.
,
M. P.
Hoerling
, and
R. M.
Dole
,
2001
:
The origin of the subtropical anticyclones
.
J. Atmos. Sci.
,
58
,
1827
1835
, doi:.
Cook
,
K. H.
,
2000
:
The south Indian convergence zone and interannual variability over southern Africa
.
J. Climate
,
13
,
3789
3804
, doi:.
Davis
,
J. C.
,
1986
: Statistics and Data Analysis in Geology. 2nd ed. John Wiley and Sons, 646 pp.
Dee
,
D.
, and Coauthors
,
2011
:
The ERA-Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
, doi:.
Ding
,
Q.
,
E. J.
Steig
,
D. S.
Battisti
, and
J. M.
Wallace
,
2012
:
Influence of the tropics on the southern annular mode
.
J. Climate
,
25
,
6330
6348
, doi:.
Eliassen
,
A.
, and
E.
Palm
,
1960
:
On the transfer of energy in stationary mountain waves
.
Geofys. Publ.
,
22
,
1
23
.
Fogt
,
R.
, and
D.
Bromwich
,
2006
:
Decadal variability of the ENSO teleconnection to the high-latitude South Pacific governed by coupling with the southern annular mode
.
J. Climate
,
19
,
979
997
, doi:.
Folland
,
C. K.
,
J. A.
Renwick
,
M. J.
Salinger
, and
A. B.
Mullan
,
2002
:
Relative influences of the interdecadal Pacific oscillation and ENSO on the South Pacific convergence zone
.
Geophys. Res. Lett.
,
29
, 1643, doi:.
Gong
,
T.
,
S. B.
Feldstein
, and
D.
Luo
,
2010
:
The impact of ENSO on wave breaking and southern annular mode events
.
J. Atmos. Sci.
,
67
,
2854
2870
, doi:.
Grassi
,
B.
,
G.
Redaelli
, and
G.
Visconti
,
2005
:
Simulation of polar Antarctic trends: Influence of tropical SST
.
Geophys. Res. Lett.
,
32
, L23806, doi:.
Haarsma
,
R. J.
,
E. J. D.
Campos
,
W.
Hazeleger
,
C.
Severijns
,
A. R.
Piola
, and
F.
Molteni
,
2005
:
Dominant modes of variability in the South Atlantic: A study with a hierarchy of ocean–atmosphere models
.
J. Climate
,
18
,
1719
1735
, doi:.
Hastenrath
,
S.
,
1991
: Climate Dynamics of the Tropics. Kluwer, 488 pp.
Hermes
,
J. C.
, and
C. J. C.
Reason
,
2009
:
Variability in sea-surface temperature and winds in the tropical south-east Atlantic Ocean and regional rainfall relationships
.
Int. J. Climatol.
,
29
,
11
21
, doi:.
Hoskins
,
B. J.
,
1996
:
On the existence and strength of the summer subtropical anticyclones
.
Bull. Amer. Meteor. Soc.
,
77
,
1287
1291
.
Ishii
,
M.
,
A.
Shouji
,
S.
Sugimoto
, and
T.
Matsumoto
,
2005
:
Objective analyses of SST and marine meteorological variables for the 20th century using ICOADS and the Kobe collection
.
Int. J. Climatol.
,
25
,
865
879
, doi:.
Kobayashi
,
S.
, and Coauthors
,
2015
:
The JRA-55 reanalysis: General specifications and basic characteristics
.
J. Meteor. Soc. Japan
,
93
,
5
48
, doi:.
Lee
,
S.-K.
,
C. R.
Mechoso
,
C.
Wang
, and
J. D.
Neelin
,
2013
:
Interhemispheric influence of the northern summer monsoons on the southern subtropical anticyclones
.
J. Climate
,
26
,
10 193
10 204
, doi:.
Li
,
W.
,
L.
Li
,
M.
Ting
, and
Y.
Liu
,
2012
:
Intensification of Northern Hemisphere subtropical highs in a warming climate
.
Nat. Geosci.
,
5
,
830
834
, doi:.
Lim
,
E.-P.
,
H. H.
Hendon
, and
H.
Rashid
,
2013
:
Seasonal predictability of the southern annular mode due to its association with ENSO
.
J. Climate
,
26
,
8037
8054
, doi:.
Liu
,
Y.
,
G.
Wu
, and
R.
Ren
,
2004
:
Relationship between the subtropical anticyclone and diabatic heating
.
J. Climate
,
17
,
682
698
, doi:.
Lübbecke
,
J. F.
,
C. W.
Böning
,
N. S.
Keenlyside
, and
S.-P.
Xie
,
2010
:
On the connection between Benguela and equatorial Atlantic Niños and the role of the South Atlantic anticyclone
.
J. Geophys. Res.
,
115
, C09015, doi:.
Lübbecke
,
J. F.
,
N. J.
Burls
,
C. J. C.
Reason
, and
M. J.
McPhaden
,
2014
:
Variability in the South Atlantic anticyclone and the Atlantic Niño mode
.
J. Climate
,
27
,
8135
8150
, doi:.
Marshall
,
G. J.
,
2003
:
Trends in the southern annular mode from observations and reanalyses
.
J. Climate
,
16
,
4134
4143
, doi:.
Miyasaka
,
T.
, and
H.
Nakamura
,
2005
:
Structure and formation mechanism of the Northern Hemisphere summertime subtropical highs
.
J. Climate
,
18
,
5046
5065
, doi:.
Miyasaka
,
T.
, and
H.
Nakamura
,
2010
:
Structure and mechanisms of the Southern Hemisphere summertime subtropical anticyclones
.
J. Climate
,
23
,
2115
2130
, doi:.
Morioka
,
Y.
,
T.
Tozuka
, and
T.
Yamagata
,
2011
:
On the growth and decay of the subtropical dipole mode in the South Atlantic
.
J. Climate
,
24
,
5538
5554
, doi:.
Ohishi
,
S.
,
S.
Sugimoto
, and
K.
Hanawa
,
2015
:
Zonal movement of the Mascarene high in austral summer
.
Climate Dyn.
,
45
,
1739
1745
, doi:.
Plumb
,
R.
,
1985
:
On the three-dimensional propagation of stationary waves
.
J. Atmos. Sci.
,
42
,
217
229
, doi:.
Richter
,
I.
,
C. R.
Mechoso
, and
A. W.
Robertson
,
2008
:
What determines the position and intensity of the South Atlantic anticyclone in austral winter?—An AGCM study
.
J. Climate
,
21
,
214
229
, doi:.
Richter
,
I.
,
S. K.
Behera
,
Y.
Masumoto
,
B.
Taguchi
,
N.
Komori
, and
T.
Yamagata
,
2010
:
On the triggering of Benguela Niños: Remote equatorial versus local influences
.
Geophys. Res. Lett.
,
37
,
L20604
, doi:.
Rodwell
,
M. J.
, and
B. J.
Hoskins
,
2001
:
Subtropical anticyclones and summer monsoons
.
J. Climate
,
14
,
3192
3211
, doi:.
Seager
,
R.
,
R.
Murtugudde
,
N.
Naik
,
A.
Clement
,
N.
Gordon
, and
J.
Miller
,
2003
:
Air–sea interaction and the seasonal cycle of the subtropical anticyclones
.
J. Climate
,
16
,
1948
1966
, doi:.
Sen Gupta
,
A.
, and
M. H.
England
,
2006
:
Coupled ocean–atmosphere–ice response to variations in the southern annular mode
.
J. Climate
,
19
,
4457
4486
, doi:.
Sterl
,
A.
, and
W.
Hazeleger
,
2003
:
Coupled variability and air–sea interaction in the South Atlantic Ocean
.
Climate Dyn.
,
21
,
559
571
, doi:.
Takaya
,
K.
, and
H.
Nakamura
,
2001
:
A formulation of a phase-independent wave-activity flux for stationary and migratory quasigeostrophic eddies on a zonally varying basic flow
.
J. Atmos. Sci.
,
58
,
608
627
, doi:.
Thompson
,
D. W. J.
, and
J. M.
Wallace
,
2000
:
Annular modes in the extratropical circulation. Part I: Month-to-month variability
.
J. Climate
,
13
,
1000
1016
, doi:.
Trzaska
,
S.
,
A. W.
Robertson
,
J. D.
Farrara
, and
C. R.
Mechoso
,
2007
:
South Atlantic variability arising from air–sea coupling: Local mechanisms and tropical–subtropical interactions
.
J. Climate
,
20
,
3345
3365
, doi:.
Venegas
,
S. A.
,
L. A.
Mysak
, and
D. N.
Straub
,
1997
:
Atmosphere–ocean coupled variability in the South Atlantic
.
J. Climate
,
10
,
2904
2920
, doi:.
Vigaud
,
N.
,
Y.
Richard
,
M.
Rouault
, and
N.
Fauchereau
,
2009
:
Moisture transport between the South Atlantic Ocean and southern Africa: Relationships with summer rainfall and associated dynamics
.
Climate Dyn.
,
32
,
113
123
, doi:.
Vizy
,
E. K.
, and
K. H.
Cook
,
2016
:
Understanding long-term (1982–2013) multi-decadal change in the equatorial and subtropical South Atlantic climate
.
Climate Dyn.
,
46
,
2087
2113
, doi:.
Wang
,
C.
,
2006
:
An overlooked feature of tropical climate: Inter-Pacific-Atlantic variability
.
Geophys. Res. Lett.
,
33
, L12702, doi:.
Wolter
,
K.
, and
M. S.
Timlin
,
2011
:
El Niño/Southern Oscillation behaviour since 1871 as diagnosed in an extended multivariate ENSO index (MEI.ext)
.
Int. J. Climatol.
,
31
,
1074
1087
, doi:.
Xie
,
P.
, and
P. A.
Arkin
,
1997
:
Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs
.
Bull. Amer. Meteor. Soc.
,
78
,
2539
2558
, doi:.
Zwiers
,
F. W.
, and
H.
von Storch
,
1995
:
Taking serial correlation into account in tests of the mean
.
J. Climate
,
8
,
336
351
, doi:.

Footnotes

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).