Synoptic and dynamical analysis of subtropical cyclone Anita (2010) and its potential for tropical transition over the South Atlantic Ocean
João Rafael Dias Pinto,
Institute of Astronomy, Geophysics and Atmospheric Sciences, University of São Paulo, São Paulo, Brazil
Corresponding author: J. R. Dias Pinto, Institute of Astronomy, Geophysics and Atmospheric Sciences, University of São Paulo, Rua do Matão, 1226, Cidade Universitária, 05508-090 São Paulo, Brazil. (email@example.com)
 Subtropical cyclogenesis and tropical transitions (TT) over the South Atlantic Ocean only received attention after the first documented Hurricane Catarina occurred close to the southern Brazilian coast in March 2004. However, due to the lack of studies in this part of the Atlantic Ocean, it is still unclear what the main environmental conditions and dynamical processes associated with TT or even subtropical cyclogenesis are over the region. This study presents a synoptic and dynamical analysis of the subtropical cyclone Anita which occurred in March 2010 near the Brazilian coast. This system started as a pure subtropical cyclone, evolved to a condition favorable to TT, later developed into a cold-core structure, and decayed as an extratropical cyclone. During the period favorable for TT, the turbulent heat fluxes (latent plus sensible) from the ocean decreased, and Anita started interacting with another extratropical disturbance, preventing the TT to happen. This interaction, in turn, increased the vertical wind shear, allowed the extratropical transition to occur, and promoted the westward displacement of Anita to colder waters, thus decreasing the turbulent heat fluxes. The results suggest that the combination of a dipole blocking pattern aloft, with contribution from barotropic energy conversions, and strong turbulent fluxes is an important ingredient for tropical storm development. Hybrid storms in such environmental conditions can be one form of precursors of hurricanes over the South Atlantic.
 In recent times, the meteorological community has increased attention to the complexity of cyclonic systems. Among the developments has been the identification and diagnosis of hybrid systems, which can exhibit tropical and extratropical characteristics at different stages in their life cycle.
 According to Hart , subtropical cyclones present such a hybrid structure which is characterized by a warm core in the lower levels and a cold core aloft. These systems occur in several places of the world as documented in the Mediterranean Sea [Emanuel, 2005; Fita et al., 2007], near Australia [Garde et al., 2009], in the North Atlantic basin [Guishard et al., 2009], and in the South Atlantic basin [Evans and Braun, 2012]. Guishard et al.  presented a climatology of subtropical cyclones over the North Atlantic Ocean. They found a mean frequency of four systems per year (with a larger occurrence in October and June). Moreover, these authors and Guishard  listed the necessary (but not sufficient) environmental features for subtropical cyclogenesis as follows: (a) they are initiated by baroclinic processes associated with an approaching upper level trough or a cutoff low; (b) the environment should present cyclonic low-level relative vorticity (which can be probably related to a shear zone associated with an old frontal boundary linked to the upper level trough); (c) a sufficiently larger Coriolis parameter; (d) low-tropospheric to mid-tropospheric high relative humidity; (e) sea surface temperature (SST) higher than 25°C; and (f) vertical wind shear (between 200 and 925 hPa) less than 10 m s−1. The last two conditions are not rigid since Guishard et al.  also verified that many subtropical cyclones occur in environments with SST ranging from 16°C to 30°C and/or vertical wind shear ranging from 1 to 40 m s−1. Many of these environmental conditions for subtropical cyclogenesis were also identified by Davis  through idealized simulations of subtropical cyclones in a limited-area baroclinic model.
 Over the South Atlantic Ocean (SAO), until now, the unique climatology of subtropical cyclones was carried out by Evans and Braun . They showed that in the SAO basin, there is a mean of 1.2 subtropical cyclones per year, which is lower than that over the North Atlantic [Guishard et al., 2009]. Another difference between both basins is that the higher percentage of the systems occurs on SST lower than 22°C over the SAO [Evans and Braun, 2012]. Moreover, subtropical cyclones in the SAO are usually formed in an environment with higher values of vertical wind shear, which differs from those of the North Atlantic that occur with shear lower than 10 m s−1. Although the climatology of Evans and Braun  has provided valuable information on the frequency and area of occurrence of subtropical cyclones in the SAO, there is yet little knowledge referring to their vertical structure, dynamics, and similarities/differences in relation to the systems in the Northern Hemisphere.
 Cyclones undergoing either extratropical or tropical transtions (ET and TT, respectively) can also present a hybrid vertical structure during this stage. ET is a gradual process by which an initially warm-core tropical cyclone transforms into a cold-core extratropical cyclone [Jones et al., 2003]. The inverse process characterizes the TT. Davis and Bosart  described two kinds of environment that favor TT (i.e., transition from an extratropical to a tropical cyclone). In the first one, there is a strong extratropical cyclone in low levels, which favors wind-induced surface heat exchanges (WISHE) [Emanuel, 1987] and a trough in the upper troposphere westward of the surface low just prior of the transition. TT would occur along with the reduction of the vertical wind shear due to the convection, which induces upper tropospheric outflow and diabatic redistribution of potential vorticity (PV). In this process, an environment resembling an occlusion and a tropical cyclone may be generated. In the second kind of TT, a weak extratropical cyclone develops due to the combination of a mid-tropospheric mesoscale cyclonic vortex and near-surface baroclinicity. Without the mesoscale vortex, the weak extratropical cyclone cannot amplify by WISHE. TT in this situation is not fully understood [Davis and Bosart, 2004].
 Hurricane Catarina (March 2004) was the first documented example over the SAO [Pezza and Simmonds, 2005] of how a hybrid structure can acquire full properties of a hurricane. It originated through a TT of an extratropical cyclone within a predominately barotropic environment provided by a strong dipole-blocking-like structure [McTaggart-Cowan et al., 2006; Veiga et al., 2008]. The occurrence of this structure in the middle and upper troposphere reduced the vertical wind shear, which is one of the main ingredients that favors tropical cyclogenesis [Zehr, 1992; DeMaria et al., 1993]. In such a configuration of flow, Catarina presented easterly propagation due to the resulting flow in the blocking region [McTaggart-Cowan et al., 2006, Figure 2]. In other words, in the Southern Hemisphere, an upper level low is located equatorward from a high pressure within a dipole blocking. While this low produces clockwise circulation, the high pressure produces anticlockwise circulation. In the east-west direction, the junction between these flows weakens the westerly winds and the vertical wind shear.
 Near the southern coast of Brazil, Miky Funatsu et al. , Reboita et al. , Iwabe and da Rocha , and Dias Pinto and da Rocha  have already indicated the presence of cyclone systems with different characteristics from a pure extratropical cyclone. They described some surface cyclogenesis occurring below an upper level cutoff low, which is also an important feature during subtropical cyclogenesis [Guishard et al., 2009; Evans and Braun, 2012] and TT (see the Catarina case) [McTaggart-Cowan et al., 2006]. The fast and strong deepening of some observed cyclones in this area of the Atlantic was pointed out by Dal Piva et al.  and Gozzo and da Rocha  as a result of the intense latent and sensible turbulent heat fluxes provided by the warm waters of the Brazilian Current. In the absence of such heating, Reboita et al.  suggested through regional climate modeling experiments for the present climate reduction of the number of cyclogenesis over the SAO. These studies indicate a potential environment for evolution of different types of cyclogenesis either tropical or subtropical, which is in line with the occurrence of the first documented Hurricane Catarina during March 2004 near the southern coast of Brazil.
 From 6 to 11 March 2010, an unusual cyclone occurred near the southern-southeastern coasts of Brazil, being later denominated as subtropical cyclone Anita by the local meteorological centers [Dutra, 2012]. It initiated as a purely subtropical system and afterward acquired similar characteristics to those presented during Catarina's development: a dipole blocking in the upper levels acting to reduce the vertical wind shear and the development of a symmetric eye-like cloudiness structure. Anita had a long life cycle, and instead of evolving into a tropical storm as it approached toward the southern Brazilian coast, the system weakened and became an extratropical cyclone, differently from Catarina. During all its life cycle, Anita produced intense winds and precipitation near the coast [Dutra, 2012]. Since the development of subtropical and tropical cyclones has not been fully documented over the SAO, meteorological centers of weather forecasting were surprised when Anita developed along the Brazilian coast. Thus, subtropical cyclone Anita is an important event for investigation because it can help to understand mechanisms that favor this kind of cyclones over the SAO and their potentials for transitions. Therefore, this paper aims to describe the synoptic and dynamic environmental conditions favoring Anita development and analyze its potential for TT. The study is organized as follows: section 2 describes the methodology and the data utilized, section 3 presents the main results, and section 4 brings the discussions and conclusions.
2 Data and Methodology
 The study was divided in two main parts. First, the evolution of the system's vertical structure is represented by the cyclone phase space (CPS), whereas its development is accomplished through a synoptic description. Besides, the main environmental characteristics were also investigated through turbulent heat fluxes, vertical wind shear and by a blocking index (BI). Then, the system's energetics are presented through the limited-area Lorenz energy cycle.
 Data from different sources were employed in this study. Fields of sea level pressure, geopotential height, air temperature, latent plus sensible heat fluxes, and winds were obtained from the Global Forecast System (GFS, http://rda.ucar.edu/datasets/ds083.2/) operational analysis. These data have a horizontal resolution of 1.0° × 1.0°, distributed in 26 pressure levels (from 1000 to 10 hPa), and are available every 6 h (0000, 0600, 1200, and 1800 UTC). Precipitation data were obtained from version 3B42 Tropical Rainfall Measuring Mission (TRMM-3B42) which has a horizontal resolution of 0.25° × 0.25° and a temporal resolution of 3 h (http://mirador.gsfc.nasa.gov/). Wind components at surface with the same horizontal resolution of TRMM-3B42 were obtained from the Cross-Calibrated Multi-Platform (CCMP) Ocean Surface Wind Components (ftp://podaac.jpl.nasa.gov/OceanWinds/ccmp/L3.0/flk/). These data have 6 h of temporal resolution [Atlas et al., 2008, 2009]. Potential temperature in the dynamic tropopause [surface at −2 PVU (potential vorticity unit)] from the ERA-Interim reanalysis [Dee et al., 2011] was used to compute a blocking index. These data have a horizontal resolution of 1.5° × 1.5°, and they are also available every 6 h. The SST analysis is from the Operational Sea Surface Temperature and Sea Ice Analysis (OSTIA) which has been developed at the Met Office. These data are based on a combination of infrared and microwave satellites, as well as in situ data. OSTIA has a horizontal resolution of roughly 5 km and a daily time frequency [Stark et al., 2007]. Besides, qualitative information about cloudiness during the cyclone's life cycle was depicted through GOES-12 infrared (IR) satellite imagery obtained from the Weather Forecast and Climate Studies Center/National Institute of Spatial Research (CPTEC/INPE) (available at http://www.cptec.inpe.br/satelite).
2.2 The Cyclone Phase Space
 The cyclone phase space (CPS) [Hart, 2003] is an objective method that describes the three-dimensional cyclone structure evolution and gives a classification of cyclonic systems (extratropical, tropical and subtropical, for example). According to Hart , the method is based on three parameters: storm-relative motion thermal asymmetry (parameter B), lower tropospheric thermal wind (parameter ), and upper tropospheric thermal wind (parameter ).
2.2.1 Parameter B: Thermal Symmetry
 Parameter B, which indicates the cyclone symmetry, is obtained from
where Z is the geopotential height (meters); subscripts R and L indicate, respectively, the right and left sides of each position of the cyclone; and h represents a positive signal (+) for analysis over the Southern Hemisphere and negative (−) over the Northern Hemisphere. A circle with a radius of 5° is defined to determine the cyclone size. As in other studies [Hart, 2003; Guishard et al., 2009], a cyclone with −10 m < B < 10 m is symmetric (which is the case of the tropical systems), while a cyclone with B ≥ 10 m is considered asymmetric (which is the case of the extratropical systems).
2.2.2 Parameters and : Thermal Wind
 The thermal wind parameters are defined as a vertical change in the height gradient (ΔZ) between the bounding pressure levels and are calculated as
where superscripts L and U indicate the layers (lower and upper) of the thermal wind parameter (900–600 hPa and 600–300 hPa, respectively). Values < 0 and < 0 indicate a cold-core structure with the wind speed increasing with height (which is common in extratropical cyclones). On the other hand, > 0 and > 0 represent the decrease of the wind speed with height and indicate a warm-core structure. It is worthy to mention that all the algorithms related to the CPS term computation are available on Dr. Hart's home page (http://moe.met.fsu.edu/∼rhart/software.php).
 The occurrence of blockings in the atmosphere is a factor that can contribute to decrease the vertical wind shear, which, in turn, could be an important “ingredient” for the vertical organization of the cloudiness in subtropical storms.
 The possibility of blocking occurrence during Anita's life cycle was investigated using a blocking index BI defined by Pelly and Hoskins . According to the authors, in a blocking situation, there is an inversion of the potential temperature meridional gradient at the dynamic tropopause (−2 PVU). Thus, the blocking index BI at longitude λ0 can be defined as the difference in the average potential temperatures (at the dynamic tropopause) in the northern and southern boxes of a determined region (for more details, see Pelly and Hoskins [2003, Figure 2]), i.e.,
where BI is calculated along the longitude λ0 for a strip of width Δλ centered at latitude ϕ0 and meridional dimension Δϕ. The values used in equation (4) were: λ0 = 50°W, Δλ=5°, φ=37.5°S, and Δφ=20°. The index was computed from 5 to 15 March 2010 every 6 h, with blocking occurring when BI reached positive values.
 Besides the mean vertical wind shear between 200 and 850 hPa, the total heat fluxes (latent plus sensible heat fluxes) were computed within an area limited by 50°W–40°W and 37.5°S–27.5°S from 5 to 15 March 2010. This domain was used because it encloses the region where Anita remained semistationary during most of its life cycle.
2.4 Limited-Area Lorenz Energetics
 The energetics were determined with the Lorenz approach [Lorenz, 1955, 1967] in which the available potential and kinetic energies are divided in the zonal and eddy forms. Since the analyses were focused on an individual system rather than for the entire globe, all the terms of the energy balance were computed within a limited portion of the atmosphere. Following Muench , the budget equations for limited-area energetics are given by
 In these equations, A and K are the available potential and kinetic energies, respectively, where subscripts Z and E denote the zonal and eddy forms of energy, respectively; CZ indicates the conversion between AZ and KZ; CA indicates the conversion between AZ and AE; CK indicates the conversion between KE and KZ (barotropic conversion); and CE indicates the conversion between AE and KE (baroclinic conversion). GZ and GE are the generation terms of the zonal and eddy available potential energies, respectively. DZ and DE denote the dissipation of the zonal and eddy kinetic energies, respectively. Besides, due to the inclusion of the boundary energy transport, there are four more terms which represent the exchange of energy between the domain and the encircling regions (BKZ, BKE, BΦZ, and BΦE). Muench  included the appearance of kinetic energies within the volume of a limited region through the work produced at its boundaries (terms BΦZ and BΦE). The approximate mathematical expressions for all terms and symbols utilized here are detailed in Brennan and Vincent .
 Figure 1 depicts the limited-area energy cycle given by equations (5)-(8) by showing the positive sense of the conversions, generation, and transports among the zonal and eddy kinetic and available potential energies. According to Lorenz  and Asnani , positive correlations between temperature and diabatic heating generate available potential energy by enhancing thermal contrasts in either latitudinal direction (GZ) or in the same latitude circle (GE). Axisymmetric latitudinal heat transports (which is given by CA) induced by developing baroclinic waves [Holton, 2004] will generate temperature contrasts in the same latitude circle, thus generating AE at the expense of AZ. Upward motion of relatively warm air and downward motion of relatively cold air promote releasing of available potential energy into kinetic energy. Direct thermally induced circulation in a latitudinal sense is represented by CZ, whereas an induced vertical circulation at the same latitude circle is represented by CE [Lorenz, 1967], which is commonly defined as the baroclinic term. Synoptic-scale eddy motions tend to maintain the zonal mean flow by transferring their kinetic energy to the midlatitude jets [James, 1994]. In episodes of barotropic instability, however, the disturbance extracts its kinetic energy through momentum transports from the mean flow, which is expressed by CK. In the literature [see, for example, Wiin-Nielsen and Chen, 1993], the senses of energy flow AZ → AE → KE and AZ → KZ → KE are called, respectively, baroclinic and barotropic chains. Finally, the dissipation of both the zonal and eddy kinetic energies is represented by the DZ and DE terms, respectively.
 Although Anita was a nonstationary system, a space- and time-fixed reference domain was chosen for the computation of the energy cycle (a deeper discussion on the use of a moving or fixed domain can be found in Dias Pinto and da Rocha ). The domain chosen was based on a 25° × 25° latitude-longitude area (40°S–15°S and 60°W–35°W), where the Anita was the unique synoptic feature on it most of the time. Despite the limitations that arise by adopting such a big domain, the results presented here are reliable enough to reflect in a straightforward way the energetics of the system and the exchanges among it, the surrounding environment, and the regions outside the domain.
 As usual in the literature [e.g., Brennan and Vincent, 1980; Michaelides, 1987; Wahab et al., 2002], the terms of generation, dissipation, and boundary work pressure (B ΦZ and B ΦE) were obtained as residual terms from equations (5)-(8) as follows:
where symbol ϵ indicates numerical accumulated errors from the calculus included in the residual value. It is important to emphasize that equations (11) and (12) do not only contain the dissipative effects but also include the effects of the boundary pressure work within the volume.
 Due to the unusual formation, the study of the Anita life cycle was divided in four distinct phases, denominated, initial, hybrid, potential, and transitioned. We chose four periods, beginning at 0000 UTC 6 March for the initial phase (i.e., formation), at 1200 UTC 9 March for the hybrid one (related to the “mature” stage of the subtropical phase), at 0600 UTC 10 March for the potential (for TT), and, finally, at 1800 UTC 11 March 2010 for the transition (from a subtropical to an extratropical system). The selection of these periods considered combined analysis among the CPS, the shape and evolution of the cloudiness patterns in the satellite imagery, and the temporal evolution of the eddy kinetic energy. Therefore, the phases related to the onset of the system, the hybrid stage, the potential for tropical transition, and the transition to an extratropical system are all covered.
3.1 Vertical Structure Evolution Through the CPS
 The time evolution of the vertical structure of the subtropical cyclone Anita is described by computing the terms of thermal symmetry and thermal wind and plotting them on the CPS diagram. The A and Z points in Figure 2 denote, respectively, the beginning and ending points from the tracking of the system and therefore indicate its life cycle. The trajectory lines on the CPS indicate that Anita presented at the initial phase a symmetric structure in the vertical and a shallow low-level warm core, indicating that it had already started as a subtropical system. This vertical thermal structure was maintained during the cyclone development until 10 March, when the system became neutral in both diagrams; that is, it reached the threshold between the two regimes (asymmetric and symmetric, shallow warm core to deep cold core). During this period, the mean radius of 925 hPa gale force winds decreased, indicating that the radius of the system also decreased as it evolved from the subtropical phase to the potential one (favorable for TT). This decreasing radius can be graphically visualized in both diagrams by the size change of the circles between 9 and 10 March (Figure 2).
 During the transition phase from subtropical to extratropical (around 11 March), the system achieved another structure of moderate low- and high-level cold core and an asymmetric vertical structure (Figure 2). The low-level warm core weakened during this period, and the system acquired a frontal nature with an increase in the radius of gale force winds. Therefore, the trajectory lines throughout the CPS diagrams showed that the cyclone started as a warm-core system in low levels with a symmetric nature, that is, a subtropical system, and evolved in a transition to an asymmetric and cold-core system.
 Although the CPS cannot show the main underlying physical processes behind the system evolution, the analysis of the vertical structure highlighted the fact that this kind of cyclone underwent a great structural change during its life cycle, and thus, the synoptic and energetic analyses will help to understand this evolution.
3.2 Synoptic Evolution and Main Environmental Conditions
 The lower level precursor environment that led to the system was characterized by a weak low pressure area around 20°S–37.5°W, near the southeastern Brazilian coast, which was associated with a large area of convection and moisture convergence (figure not shown). At 0000 UTC 6 March, this large cloudy region was still over the genesis area, presenting in 925 hPa a preexistent cyclonic circulation, and little baroclinicity, that is, weak horizontal temperature gradients (Figures 3a and 4e). Accompanying this near-surface configuration, a midtropospheric trough over southeastern South America amplified and evolved to a cutoff low system at 1200 UTC 6 March (figure not shown), extending toward southern Brazil. This association between a cutoff low and a surface subtropical cyclone was also identified through the climatology of South Atlantic subtropical cyclones carried out by Evans and Braun .
 According to Ndarana and Waugh , the development of cutoff low systems is associated with Rossby wave breaking which, in turn, is defined as the rapid and irreversible deformation of potential vorticity (PV) material contours [McIntyre and Palmer, 1983]. Our analysis showed that the cutoff low at 1200 UTC 6 March was associated with a horizontal incursion of stratospheric PV from latitudes higher than 45°S to the south of Brazil (figures not shown). Although weak, this incursion of stratospheric potential vorticity (PV) contributed to the cutoff low's growth. During the evolution of Anita, the intensification of the cyclonic circulation at midlevels was related to a horizontal PV incursion from midlatitudes (Figure 5a). At 0600 UTC 9 March, a PV core in 300 hPa (Figure 5b) with values between −2.0 and −2.5 PVU (1 PVU = 1 × 10−6 m2 K kg−1 s−1) was observed. Indeed, at 1200 UTC 9 March, during the hybrid stage, Figure 4b shows a large cutoff low area in geopotential height at 500 hPa. This synoptic pattern resembles closely those observed in the dipole-blocking events [Rex, 1950a, 1950b; Coughlan, 1983; McTaggart-Cowan et al., 2006] where the flow is diffluent, presenting a cyclonic circulation branch northward and an anticyclonic one southward. Positive values of BI confirmed the presence of the blocking pattern (Figure 8a) between 8 and 9 March, indicating an obstruction of the zonal flow, i.e., intense westerlies are in the north and south sectors of the blocking as indicated by the wind fields at 500 and 200 hPa (figure not shown). It is interesting to mention that Hurricane Catarina, during its tropical transition, was also inserted in a dipole-blocking environment that remained for 4 days [see McTaggart-Cowan et al., 2006, Figure 9].
 From 6 to 9 March, Anita displaced southwestward in low levels and acquired a barotropic equivalent structure extending from the surface to the upper levels (Figures 4b and 4f). The movement toward the southern Brazilian coast was favored by the easterly flow at middle and high levels within the blocking-like area (figures not shown). At this time, the clouds presented an inverted comma-shaped pattern split from the main cloud band, extending from the center of Brazil to the Atlantic Ocean (Figure 3b).
 The development continues on 10 March when the cyclone center reached the nearest position of the coast at 0600 UTC (Figure 4g). At this time, it was classified as a neutral system in both CPS diagrams (Figure 2), although the near-circular pattern of the cloudiness (Figure 3c) was similar to that observed during Catarina's transition [see McTaggart-Cowan et al., 2006, Figure 4f]. In low levels, it was characterized by a small area of strong cyclonic vorticity below a cutoff low in 500 hPa, which indicates maintenance of the barotropic vertical structure (Figure 4c). As shown in the previous period, the dipole-blocking-like pattern allowed the system to propagate westward, and therefore, the associated cloudiness approached toward the southern Brazilian coast. As the system evolved, the cloudiness pattern acquired a more symmetrical structure (Figure 3c); Figure 6 suggests that Anita sustained an eye-like feature for at least 6 h, with higher clouds and heavier precipitation around a region of significantly less convective activity around 30°S and 48°W. The characteristics of a cloudy rounded structure showing a free cloud area in the center were reported by Emanuel  and Fita et al.  during the life cycle of tropical-like storms over the Mediterranean Sea, which are commonly named Medicanes (Mediterranean + Hurricanes).
 Figures 7c and 7d show that the small values of vertical shear (<10 m s−1) in the Anita region were maintained nearly constant between the hybrid and potential stages. Therefore, the lack of intense vertical shear of the westerly winds contributed to the organization of the convective activity in a more symmetrical structure, as shown in Figures 3c and 6. However, the sensible and latent heat fluxes from the ocean decreased in the vicinity of the cyclone (Figures 7a and 7b). The tracking of Anita in Figure 8 suggests that the system remained semistationary over waters with temperature around 24–26°C (here, SST were averaged over the period 6–11 March 2010) for about 2 days.
 Since turbulent heat fluxes depend on the wind intensity at surface and vertical gradients of temperature and humidity, some factors could contribute to the decrease of these fluxes. As the convective activity became more organized between the hybrid and potential stages (Figures 3b and 3c, respectively), moistening of the surrounding atmosphere (figure not shown) was observed, contributing to decrease the latent heat fluxes. According to several authors [Brand, 1971; Chang and Anthes, 1978; Wu et al., 2007; Sriver and Huber, 2007; Dare and McBride, 2011], cyclones can decrease SST through vertical mixing within the ocean. The SST cooling, in turn, may reduce the source of thermal energy to the turbulent heat fluxes, implying weakening of the cyclones. A decrease of SST at the center and southern parts of Anita was observed during 9 and 10 March (figure not shown). Besides, SST may locally decrease due to the reduction of solar energy reaching the surface due to the cloud coverage.
 Eventually, Anita underwent an extratropical transition as it interacted with an extratropical cyclone, becoming a cold-core and asymmetric system around 1800 UTC 11 March (Figure 2). The 500 hPa cutoff low reached its final stage as a small trough embedded in another short wave southward, and the cloudiness reorganized into an inverted comma shape, typical in extratropical cyclones (Figures 3d and 4d). The passage of the midlatitude system (and the resulting coupling with) not only contributed to the extratropical transition and eastward propagation of Anita but also imported increased shear into the region between 10 and 11 March (Figure 8a). At the surface, turbulent fluxes were drastically reduced in the cyclone area (Figure 8b), which could be due to Anita displacement from the subtropical to the extratropical latitudes and, thus, toward cooler SST (see the inset in Figure 8).
 Although relatively weak vertical wind shear was observed between the hybrid and potential stages (around 9 and 10 March), both the decrease of the turbulent heat fluxes and the coupling with an extratropical system prevented Anita to make transition to the tropical cyclone. Therefore, it is hypothesized that if there were a larger supply of energy to build up a deeper warm core and the lack of any extratropical forcing in this environment of low vertical wind shear, subtropical cyclone Anita could undergo such an organization capable of self-amplification as a tropical system. The description of a barotropic vortex genesis and the increasing importance of diabatic processes to strengthen the hybrid system—potentially leading to tropical transition—are consistent with the climatology of the South Atlantic subtropical cyclones by Evans and Braun .
3.3 Limited-Area Lorenz Energetics
 The time evolution of the vertically integrated energy amounts (Figure 9a) shows during the initial phase of the cyclone the decreasing of both zonal forms of kinetic and available potential energies (KZ and AZ, see Figure 1). KZ reduced, reaching the smallest values around 9 March during the hybrid period, and after that, it started increasing until 0000 UTC 13 March when it attained values close to 12 × 105 J m−2 (almost twice the initial value). AZ presented a similar behavior with minimum values at 9 March, growing again at the transitioned stage of the cyclone around 11 March. Differently from these terms, KE appeared with two maxima, one at 0000 UTC 8 March and another at 0000 UTC 12 March, showing the peaks of intensity at the hybrid and transitioned stages of the cyclone. However, as shown in Figure 9a, the intensity of KE during the hybrid stage is almost twice that of the transitioned one, with 8 ×105 J m−2 and 4 × 105 J m−2, respectively. The eddy form of available potential energy AE acquired the smallest variation compared with other energy terms during all its life cycle. Its temporal evolution shows just slightly increasing from 0000 UTC to 1800 UTC 7 March; after 0000 UTC 10 March, AE remained nearly constant.
 The temporal change of the conversion terms (Figure 9b) depicts two distinct stages during all the cyclone life cycle. The period until 0600 UTC 9 March was marked by intense activity of CZ and CK (AZ → KZ and KZ → KE conversions), while after that, during the potential and transitioned stages, they acquired smaller values. The term CZ was the most intense and positive during the initial phase with a maximum of ~5 W m−2. Positive values of CZ indicate conversion of AZ into KZ through upward motion of relatively warm air and sinking of cold air in the longitudinal plane supplying kinetic energy to KZ [Lorenz, 1967; Asnani, 1993]. The barotropic conversion CK was negative (minimum of ~3 W m−2) during the same period. Such a behavior indicates transfer of kinetic energy from the zonal flow to the disturbance through the momentum transports [Lorenz, 1967; Wiin-Nielsen and Chen, 1993]. One of the baroclinic terms, CA (AZ → AE conversion), presented larger values between 11 and 12 March, indicating that the cross-latitudinal heat transfer was important in producing eddy available potential energy AE through AZ during the transition period. The baroclinic term CE (AE → KE conversion) was relevant just between 8 and 9 March and later between 11 and 12 March, becoming the only source, even small, of kinetic energy during the transitioned stage of the cyclone. This indicates, therefore, that the barotropic processes were the most relevant during the formation and growing of the hybrid cyclone, whereas the baroclinic ones became important just at the transition stage.
 Based on the temporal evolution of KE (Figure 9a), the vertically integrated energy budget was averaged over three distinct periods: 0000 UTC 5 March to 1800 UTC 7 March, which is related to the positive temporal derivative of KE; 0000 UTC 8 March to 1800 UTC 10 March; for the negative one; and 0000 UTC 11 March to 0000 UTC 12 March, again, for the positive changes of KE. Thus, the complete limited-area Lorenz energy cycle was determined, and it is depicted in Figure 10.
 Figure 10a shows that during the earlier stages of the subtropical cyclone, the main energy flow was due to the barotropic chain (AZ → KZ → KE) supplying kinetic energy from the mean zonal wind to the disturbance at the rate of 2 W m−2. As observed in the temperature fields or even in the symmetry term of the CPS [Figure 4e and Figure 2 (top)], the baroclinity over the genesis region was very weak, and thus, the creation of KE due the heat transports and induced vertical movements (terms CA and CE) was not important. Even with the large input of KZ through the computational boundaries (term BKZ), the residual term RKZ acted as a sink of energy which also contributed to the large decrease of KZ. These temporal averaged values show that the main underlying process in the cyclone's initial stages was the barotropic mechanism, providing energy for the trough amplification (Figures 4a and 4b) through the eddy momentum transports. Such a behavior was pointed out by Guishard et al.  and Evans and Braun  as a manifestation of the Rossby wave breaking, leading to an incipient subtropical cyclone vortex aloft (section 3.2 highlights the existence of this breaking occurring during Anita formation).
 As the system evolved and later reached the potential stage for transition, its energy cycle had a considerable change. The eddy kinetic energy presented negative variation at the rate of 2.24 W m−2 (Figure 10b), mainly due to the residual term RKE acting as a sink of energy. The baroclinic and barotropic chains acted as a source of kinetic energy for the system, thus maintaining it against dissipation processes. During this period, the cyclone underwent a strong structural change and achieved a stage favorable to a tropical transition. As hypothesized in the synoptic description, at this stage, Anita could not evolve to a tropical system due to the reduction of the turbulent fluxes and the interaction with an extratropical cyclone. Regarding the kinetic energy evolution, the cyclone's life cycle was characterized by a mixed energy source: first, the barotropic ones extracting energy from the zonal jet (CK) and, second and third, the conversion of eddy available potential energy generated by CA andGE, in which the last one is represented by the diabatic processes (mostly latent heat release due to the convective activity) in the cyclone area (see Figure 6). However, KE had negative variation throughout the period due to the boundary terms, which transferred energy outward the computational domain.
 Figure 10c displays the terms averaged over the transitioned stage, i.e., the period in which the cyclone underwent an extratropical transition, evolving from a neutral (between 10 and 11 March, the system stayed in the neutral region of the CPS diagram) to a cold-core system in both low and high levels and presented a frontal nature (Figure 2). The energy cycle shows that KE started increasing again (at the rate of 1.06 W m−2), indicating new strengthening of the circulation. This occurred mainly due to the juxtaposition with a midlatitude extratropical cyclone as the hybrid cyclone propagated southeastward (Figures 4d and 4h). At this time, the baroclinic chain (AZ → AE → KE) became the main source of kinetic energy for the cyclone, whereas the barotropic one acted in an inverse sense, i.e., transferring energy to the zonal flow. Large values of KZ reflect the increase of the wind vertical shear in the region.
4 Discussion and Conclusions
 This study presented the environmental and energy cycle analyses of subtropical cyclone Anita. During this unusual event, both the tropical and extratropical characteristics were observed over the SAO, where tropical transitions are not common. It began as a purely subtropical system and evolved to an extratropical cyclone, presenting, during this life cycle, conditions for transition to a hurricane.
 Anita started as a surface cyclonic anomaly of relative vorticity near 20°S–37.5°W, close to the southeastern Brazilian coast, and it was associated with a large area of convection and moisture convergence at low levels. This cyclonic anomaly increased as a mid-tropospheric cutoff low approached toward the region. The system developed through Rossby wave breaking in midlevels, which is accompanied by horizontal incursion of cyclonic PV from extratropics into lower latitudes. According to Hanley et al. , in a Rossby wave breaking event, constructive interaction between a PV maximum and the trough is observed, and therefore, a dipole blocking could develop. This feature was observed in this study through the blocking index. At the middle and upper levels, this synoptic pattern was characterized by a cyclonic circulation area equatorward (cutoff low) of an anticyclonic one. This feature induces anomalous easterly flow due to a coherent reversal in the meridional height gradient, which contributes for the westward displacement of the systems. Both a cutoff low and a dipole blocking analogue pattern occurred during the development of Anita, and they were very similar to the synoptic pattern observed during the 2004 Hurricane Catarina [McTaggart-Cowan et al., 2006] and the 2009 Australian Duck Cyclone [Garde et al., 2009]. Such a synoptic configuration generally reduces the middle and upper levels zonal flow with consequent reduction of the vertical wind shear. These conditions were also observed in Anita, and during 8 and 9 March, the system displaced toward a region with negative vertical wind shear which indicated more intense wind near surface than in the middle and upper levels. Such condition of low vertical wind shear is one of the main ingredients favoring tropical system development [Zehr, 1992; DeMaria et al., 1993].
 At the potential stage, Anita reached the closest position to the southern Brazilian coast, and its cloudiness pattern resembled closely that observed during Hurricane Catarina with an eye-like feature. However, systematic decreasing of the turbulent fluxes along with an interaction with a midlatitude system moving from south prevented Anita to perform the TT. The resulting coupling with the extratropical system not only contributed to the extratropical transition but also imported increased shear into the region at the end of the potential phase. In a hypothetical scenario with stronger and persistent turbulent fluxes, providing moisture to feed the convective activity, and no extratropical external forcing, Anita would have all the ingredients to transition to a tropical storm. This fact could have led to serious impacts to the southern Brazilian coast, which is densely populated and has many important harbors.
 Regarding the energy cycle, complex dynamical processes took place during Anita development and maintenance. It had a mixed energy source, meaning that there was energy supply from both barotropic and moist baroclinic processes. At the initial phase, it had a strong contribution from the barotropic term, via upper level eddy momentum transports. Dias Pinto and da Rocha , analyzing three cases of cyclogenesis with different levels of baroclinity (a baroclinically weak cyclone, a classical extratropical one, and a strong bomb), observed a similar behavior during the baroclinically weak Reg1 cyclone development, in which the strong blocking-like pattern within a predominately barotropic environment was also present.
 Such a barotropic energy signature was pointed out by Guishard et al.  and Evans and Braun  as a manifestation of the Rossby wave breaking, leading to an incipient subtropical cyclone vortex in upper levels. Although Anita presented a similar blocking structure and the pronounced barotropic component, it developed as a multiple-phase “tropical aborted” system. In other words, the system started as a subtropical cyclone, evolved to a condition favorable to TT, later developed into a cold-core structure, and decayed as an extratropical cyclone.
 Between the hybrid and potential stages, Anita had a mixed source from both barotropic and moist baroclinic energy conversions. The generation of eddy available potential energy was due to the warming of the midtroposphere in response to convective heating processes as found by Brennan and Vincent  for Hurricane Carmen (1974). Although smaller than the barotropic term, the baroclinic one could indicate thermally driven circulations in Anita's vicinity due to the increasing of convective activity. However, during this same period, both forms of kinetic energy were exported out from the Anita domain across the boundaries, which resulted from the interaction with the preexisting extratropical feature southward of the Anita region. This kind of interaction was also observed by Palmén , Harr and Elsberry , Harr et al. , and Jones et al.  during episodes of ET. As Anita moved southeastward, the baroclinity of the environment increased, and the system underwent an ET.
 Our results show that subtropical systems over SAO could evolve to tropical cyclones under a specific synoptic environment. This transition would occur within a dipole-blocking-like environment combined with energy contribution from barotropic sources and strong latent and sensible fluxes from the ocean. The coupling of weak vertical wind shear and intense turbulent heat fluxes is an important ingredient to strengthen subtropical cyclones, eventually allowing them to organize in such a structure capable of self-amplification. These synoptic characteristics may be one form of precursors of hurricanes in the South Atlantic basin, being important features to be analyzed in the daily practices at the meteorological center in Brazil. Due to the lack of cases in the South Atlantic basin for comparison studies, numerical sensitivity experiments should be carried out in order to further understand the relative role of the upper level forcing and surface fluxes at the boundary layer during transition processes.
 The authors would like to thank GFS, TRMM, and CCMP for providing the data set used in this study; CPTEC/INPE for the satellite images; R. E. Hart for the CPS algorithms; and, finally, the anonymous reviewers for all suggestions. This research was supported by FAPESP (07/56758-8), CNPq (558121/2009-8 and 307202/2011-9), and CAPES.