Variability of the Southern Hemisphere Winter Split Flow—A Case of Two-Way Reinforcement between Mean Flow and Eddy Anomalies

Xiaosong Yang ITPA/Marine Sciences Research Center, State University of New York at Stony Brook, Stony Brook, New York

Search for other papers by Xiaosong Yang in
Current site
Google Scholar
PubMed
Close
and
Edmund K. M. Chang ITPA/Marine Sciences Research Center, State University of New York at Stony Brook, Stony Brook, New York

Search for other papers by Edmund K. M. Chang in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

A new split-jet index is defined in this study, and composites based on this index show that the split-flow regime is characterized by a cold–warm–cold tripolar temperature anomaly in the South Pacific that extends equatorward from the Southern Hemisphere (SH) high latitudes, while nonsplit flow occurs when the phase of the tripolar temperature anomaly is reversed. Analyses of the heat budget reveal that the temperature anomalies associated with the split/nonsplit flow are mainly forced by mean flow advection instead of local diabatic heating or convergence of eddy heat fluxes. Localized Eliassen–Palm (E–P) flux diagnostics suggest that the zonal wind anomalies are maintained by the eddy vorticity flux anomalies.

These diagnostic results are confirmed by numerical experiments conducted using a stationary wave model forced by observed eddy forcings and diabatic heating anomalies. The model results show that the effects of the vorticity flux dominates over those of the heat flux, which tend to dampen the flow anomalies, and that tropical diabatic heating anomalies are not important in maintaining the split-/nonsplit-flow anomalies.

The organization of high-frequency eddies by the low-frequency split/nonsplit jet is also studied. Two sets of experiments using a linear storm-track model initialized with random initial perturbations superposed upon the split- and nonsplit-jet basic state, respectively, have been conducted. Model results show that the storm-track anomalies that are organized by the split/nonsplit jet are consistent with observed storm-track anomalies, thus demonstrating that the low-frequency split/nonsplit jet acts to organize the high-frequency eddies.

The results of this paper directly establish that there is a two-way reinforcement between eddies and mean flow anomalies in the low-frequency variability of the SH winter split jet.

Corresponding author address: Dr. Edmund K. M. Chang, Institute for Terrestrial and Planetary Atmospheres, Marine Sciences Research Center, The University at Stony Brook, State University of New York, Stony Brook, NY 11794-5000. Email: kmchang@notes.cc.sunysb.edu

Abstract

A new split-jet index is defined in this study, and composites based on this index show that the split-flow regime is characterized by a cold–warm–cold tripolar temperature anomaly in the South Pacific that extends equatorward from the Southern Hemisphere (SH) high latitudes, while nonsplit flow occurs when the phase of the tripolar temperature anomaly is reversed. Analyses of the heat budget reveal that the temperature anomalies associated with the split/nonsplit flow are mainly forced by mean flow advection instead of local diabatic heating or convergence of eddy heat fluxes. Localized Eliassen–Palm (E–P) flux diagnostics suggest that the zonal wind anomalies are maintained by the eddy vorticity flux anomalies.

These diagnostic results are confirmed by numerical experiments conducted using a stationary wave model forced by observed eddy forcings and diabatic heating anomalies. The model results show that the effects of the vorticity flux dominates over those of the heat flux, which tend to dampen the flow anomalies, and that tropical diabatic heating anomalies are not important in maintaining the split-/nonsplit-flow anomalies.

The organization of high-frequency eddies by the low-frequency split/nonsplit jet is also studied. Two sets of experiments using a linear storm-track model initialized with random initial perturbations superposed upon the split- and nonsplit-jet basic state, respectively, have been conducted. Model results show that the storm-track anomalies that are organized by the split/nonsplit jet are consistent with observed storm-track anomalies, thus demonstrating that the low-frequency split/nonsplit jet acts to organize the high-frequency eddies.

The results of this paper directly establish that there is a two-way reinforcement between eddies and mean flow anomalies in the low-frequency variability of the SH winter split jet.

Corresponding author address: Dr. Edmund K. M. Chang, Institute for Terrestrial and Planetary Atmospheres, Marine Sciences Research Center, The University at Stony Brook, State University of New York, Stony Brook, NY 11794-5000. Email: kmchang@notes.cc.sunysb.edu

1. Introduction and motivation

A distinct feature of the climatological time-mean upper-level flow during winter in the Southern Hemisphere (SH) is the presence of a split jet at the longitudes of Australia and New Zealand. The equatorward branch of the time-mean split jet is anchored by a strong subtropical jet (STJ), which extends eastward between 25° and 30°S from the central South Indian Ocean across Australia to the east-central South Pacific Ocean near 130°W. The poleward branch of the time-mean split jet is anchored by the polar front jet (PFJ), which is concentrated along 60°S from 140°E to 150°W. A zone of weak westerlies lies between the STJ and PFJ, extending from southeastern Australia eastward across the South Island of New Zealand to east of the date line. The existence of the wintertime-mean split jet in the SH has been well documented (Taljaard 1972; van Loon 1972a, b,c; Hurrell et al. 1998; van Heerden and Taljaard 1998; Vincent and Silva Dias 1998; Bals-Elsholz et al. 2001).

Many previous studies have suggested that the location, existence, and intensity of the SH split jet in winter were strongly influenced by the outflow from the great Asiatic summer monsoon anticyclone of the Northern Hemisphere (NH) into the SH at the longitudes of Australia via the Coriolis torques (Taljaard 1972; Newton 1972; Hurrell et al. 1998). The phase of El Niño–Southern Oscillation (ENSO) has also been suggested to contribute to the SH split jet in winter (Karoly 1989; Chen et al. 1996). Based on a GCM study, Mo et al. (1987) showed that cold air outbreaks associated with high-latitude blocking activity could modulate the split jet. Mechoso et al. (1988) related the winter split jet to the breakdown of the Antarctic polar vortex in late spring.

Kidson (1988) found a low-frequency high-latitude mode, which is similar to the zonal wind (ZW) index, by an EOF analysis applied to the SH 500-hPa zonal-mean zonal wind. This leading mode has opposing peaks at 40° and 60°S with an equivalent barotropic structure. The regional pattern of height anomalies associated with this mode is zonally symmetric poleward of 60°S, but, in the midlatitudes, the largest amplitudes are over the Indian and western Pacific Oceans, where the split jet is located. Karoly (1990) confirmed that in the SH winter the leading high-latitude mode is primarily zonally symmetric, representing out-of-phase variations of geopotential height between mid- and high latitudes. Kidson and Sinclair (1995) reexamined this mode in somewhat more detail using 10-day low-pass-filtered 500-hPa geopotential height data, and revealed that the split jet is associated with the negative phase of the high-latitude mode. Later on, this so-called high-latitude mode is also referred to as the SH annular mode (SAM) or Antarctic Oscillation (AAO; e.g., Gong and Wang 1999; Thompson and Wallace 2000).

The EOF analyses discussed above emphasized that the split jet is one of the most important features of the low-frequency variability in the SH winter, and that the split-jet paradigm is not only centered on the longitudes between Australia/New Zealand and the date line, but also has hemispheric signatures. The relationship between the split/nonsplit jet and SAM in the SH is, to a certain extent, similar to the relationship between the North Atlantic Oscillation and the NH annular mode, which is argued by Wallace (2000) and Wallace and Thompson (2002) to be two paradigms–one phenomenon, since in the SH midlatitudes the western Pacific seems to be an activity center (Kidson 1988), even though the SH annular mode is more zonally symmetric compared with the NH mode. Since the split jet is associated with the variability of the whole SH circulation, or the so-called SH annular mode, a better understanding of the split jet may help us to investigate the dynamics of the SH annular mode, which is the leading mode of low-frequency variability.

Previous studies have shown that the planetary-scale flow acts as a waveguide for baroclinic eddies. For example, Chang (1999) showed that the primary waveguide during the SH winter splits into two branches around 70°E, with the primary branch deviating equatorward to join up with the subtropical waveguide, and a secondary branch spiraling poleward along the subpolar jet. As shown by Bals-Elsholz et al. (2001), during some winter seasons, there is a pronounced split jet, while a nonsplit jet dominates other winter seasons. Therefore, the waveguides in the split region and far downstream could be quite different under the split and nonsplit scenarios, and the weather and climate in the local split region and downstream (i.e., South America) could be significantly modulated by the variations between the split/nonsplit flows.

Bals-Elsholz et al. (2001) introduced a split-flow index to quantify the magnitude of the split flow based on the 200-hPa-level vorticity field, and revealed that relatively cold conditions occur in the South Pacific (40°–60°S) in association with nonsplit-flow regimes, and split-flow regimes occur when relatively warm conditions prevail. They also surmised that frequent cold surges out of Antarctica moving into the South Pacific are associated with nonsplit-flow regimes. They also suggested that the AAO and the Southern Oscillation may both contribute to the variability of the split jet.

Most previous studies have focused on documenting the time-mean split-flow signatures on intraseasonal and seasonal time scales. On the other hand, the maintenance of the time-mean split/nonsplit jet is not yet fully understood. Localized Eliassen–Palm (E–P) flux diagnostics showed that the winter SH split jet is maintained by vorticity fluxes (Trenberth 1986, 1991; Kidson and Sinclair 1995). Contributions from the heat flux (vertical E–P flux) to the split jet were argued to be unimportant due to the cancellation by the Coriolis torque associated with the induced transformed Eulerian mean (TEM) circulation (Trenberth 1986). The purpose of this study is to further clarify how the split/nonsplit anomalies are maintained. Some questions we would like to examine are the following. Is the tropical heating important in maintaining the split/nonsplit flow? Or is it just the effect of the eddy vorticity flux? Can we show directly that the eddy heat flux is indeed not important? If it is the eddy vorticity flux that maintains the split/nonsplit flow, can we explain the eddy anomalies?

The paper is organized as follows. The data and methodology are described in section 2 and the composites of the split and nonsplit monthly mean flow are discussed in section 3. A localized E–P flux diagnosis of the zonal split/nonsplit flow and a heat budget of the temperature anomalies associated with the split/nonsplit flow are discussed in section 4. In section 5, diagnostics of the split/nonsplit jet using a stationary wave model is presented, while the organization of the eddies by the split/nonsplit flow as simulated by a linear storm-track model is discussed in section 6. The paper ends with conclusions and a discussion in section 7.

2. Data and methodology

a. The data

The data used in this study are the gridded data produced by the European Centre for Medium-Range Weather Forecasts (ECMWF) 15-yr Re-Analysis (ERA-15) project. This global dataset is on a 2.5° × 2.5° grid, with 17 pressure levels (1000–10 hPa) in the vertical, with 6-hourly data coverage from 1 January 1979 to 31 December 1993. The fields u, υ, ω, z, and T are used to compute the monthly mean heat fluxes and momentum fluxes. Since the three austral winter months of July–September (JAS) have the most pronounced split jet, JAS is defined as Southern Hemisphere winter here.

b. Split-jet index

The SH time-mean flow for JAS at 300 hPa (Fig. 1) clearly displays the split-jet feature across the South Pacific. The STJ, stretching from the South Indian Ocean across the South Pacific Ocean, is paralleled in the South Pacific Ocean by a weaker jet, the PFJ, with a distinct minimum in the westerlies between the two jets.

Our ultimate goal is to understand how the split jet evolves, hence we need to examine the time series of the instantaneous split flow, not just the monthly mean flow. Figure 2a shows a 1-month (August 1984) latitude–time plot of the zonal wind at 300 hPa from the South Pole to the equator, averaged across the longitude belt 150°E–150°W, where the split jet is centered. It happens that there is a strong nonsplit-/single-jet event (1–7 August), as well as a strong split-/double-jet event (11–16 August) during this month. If the JAS climatological zonal wind is removed, the distinct characteristics of the split jet and nonsplit jet are obvious from Fig. 2b: positive anomalies over 35°–50°S and negative anomalies over 55°–70°S for the nonsplit jet (1–7 August), and anomalies with the opposite sign for the split jet (11–16 August).

After an examination of zonal wind anomalies for all of JAS 1979–93, a simple strategy for quantifying the split flow is devised. For the zonal wind anomaly field at 300 hPa, two regions are found to best define the split/nonsplit jet. These regions are 55°–70°S, 150°E–150°W and 35°–50°S, 150°E–150°W corresponding to boxes A and B, respectively, in Fig. 1. The split-jet index (SJI) is simply defined as
i1520-0469-63-2-634-e1

In (1), UAa is the area average of the 300-hPa zonal wind anomalies in box A, while UBa is the area average over box B. SJI is positive when the jet is strongly split; it is negative for the nonsplit flow.

The SJI time series for JAS of each year shows the flow pattern fluctuating between the split and nonsplit flow. For instance, the SJI time series of August 1984 in Fig. 2c capture the strong nonsplit event and the prominent split event, which have been shown in Figs. 2a,b.

To examine how the monthly mean split/nonsplit flow is maintained, it is convenient to define a monthly split-jet index (MSJI) as the time mean of SJI each month. The normalized MSJI (NMSJI) is constructed as follows:
i1520-0469-63-2-634-e2
where the overbar means climatological mean, and σ is the standard deviation of the MSJI time series.

Figure 3 shows the time series of the NMSJI for the 45 JAS months from 1979 to 1993. Positive NMSJI corresponds to enhanced split flow, while negative NMSJI represents less split or nonsplit monthly mean flow. The NMSJI is used for the composite analysis discussed in the next section.

Bals-Elsholz et al. (2001) have defined a split-flow index based on the 200-hPa relative vorticity field, and their index is consistent with the one discussed here. For instance, both indices capture the pronounced split flow in 1984 and 1989 and the nonsplit flow in 1980, 1981, and 1992 (Fig. 5 in Bals-Elsholz et al. 2001; Fig. 3 in this paper).

3. Composites of the split and nonsplit monthly mean flow

In this section we will discuss results from composite analyses based on the NMSJI. The five months with the largest positive NMSJI (August 1983, September 1984, September 1985, August 1989, and September 1989) are combined to construct the split-flow composites, while the five months with the most negative NMSJI (July 1980, September 1980, September 1981, July 1990, and August 1992) are composited to characterize the nonsplit flow.

Figure 4a shows the zonal wind at 300 hPa for the composite split flow. A strong subtropical jet extends between 20° and 35°S from the central South Indian Ocean across Australia to the east-central South Pacific Ocean; and a pronounced PFJ lies poleward of the STJ. There is an isolated minimum in the westerlies between the two jets centered on New Zealand. A vertical cross section showing the 160°E–160°W sector-averaged zonal wind for the composite split flow is shown in Fig. 4b. Two distinct wind maxima can be seen located at 27.5° and 62.5°S, respectively. The PFJ extends from the stratospheric polar night jet above 100 hPa all the way to the surface. In contrast to the PFJ, the STJ is confined to between about 800 and 100 hPa, with a maximum zonal wind speed of over 50 m s−1 located at 27.5°S near 250 hPa.

The zonal wind at 300 hPa for the composite nonsplit flow is shown in Fig. 4c. The distinct feature of the nonsplit flow is the absence of the PFJ, while the STJ shifts slightly poleward. A vertical cross section of the sector-averaged zonal wind for the composite nonsplit flow is shown in Fig. 4d. The wind maximum associated with the STJ is located at about 30°S, and there is no indication of a significant PJF in the troposphere. The polar night jet can still be seen around 60°S above 150 hPa.

Figures 5a,b show the zonal wind anomalies (split minus nonsplit) at 300 hPa and the temperature anomalies at 700 hPa, respectively. Between 30° and 50°S, a band of negative zonal wind anomalies extends from southeastern Australia across New Zealand to about 130°W. Farther south, between 50° and 70°S, a band of positive anomalies extends from 150°E across the date line to 100°W (Fig. 5a). In the South Pacific Ocean, the temperature anomalies at 700 hPa show a tripolar pattern: a weak cold anomaly centered between 20° and 30°S just east of the date line, a warm anomaly south of New Zealand centered around 50°S, and a strong cold anomaly farther south extending all the way to Antarctica (Fig. 5b). The temperature anomalies are consistent with the zonal wind anomalies through the thermal wind balance.

The vertical structure of the zonal wind anomalies and the temperature anomalies are shown in Figs. 5c,d, which reveal that these patterns hold over almost the entire depth of the troposphere. Even though the wind anomalies appear to extend above 100 hPa, the maxima are located well below the polar night jet, suggesting that the anomalies are probably not stratospheric in origin.

Figures 5e,f shows the differences in the zonal-mean zonal wind and zonal-mean temperature between the split jet and nonsplit composites. Compared to Figs. 5c,d, we see that the zonal-mean anomalies basically reflect the signals from the split-jet sector. This pattern (as well as the magnitude) is quite similar to the composites based on the high-latitude mode index of Karoly (1990). This consistency suggests that the variability of the local split jet is coherent with the high-latitude mode.

Results shown in this section are similar to those based on the daily index by Bals-Elsholz et al. (2001). We also have examined the composite analysis with the 6-hourly index and obtained similar results (figures not shown here).

4. Localized E–P flux diagnostics and heat budget

a. The impact of transient eddies on the split flow

To evaluate changes in the eddy forcing of the mean flow, the localized E–P vector is evaluated for each composite according to the formulation of Trenberth (1986):
i1520-0469-63-2-634-e3
where σ = −(∂T/∂P) + κT/P is the static stability. With this definition of the E–P flux, the effects of eddies on the zonal mean flow is given by
i1520-0469-63-2-634-e4
Therefore, the divergence of the E–P flux can be interpreted as the transient eddy forcing on the westerly mean flow. In (3), the horizontal components form the barotropic part and the vertical component is the baroclinic part.

The change in the latitude of the momentum flux reversal is clearly shown in the barotropic component of the E–P flux, in which arrows around the date line incline equatorward from near 40°S in the single-jet case (Fig. 6c) and equatorward from around 60°S in the split-jet case (Fig. 6a). Figures 6a,c also show the reversal of the strong poleward component of the E–P flux south of the Tasman Sea. Figures 6b,d show the divergence of the barotropic component of the E–P flux at 300 hPa for the two composites (contours). For the split-jet case, the westerly acceleration zone extends from south of Australia and inclines poleward along the polar front jet, which is shaded in Fig. 6b, and there is a deceleration anomaly near the southwest corner of New Zealand. For the nonsplit-jet case (Fig. 6d), the westerly acceleration zone extends almost zonally along 40°S with deceleration to the north. The difference between Figs. 6b,d are shown in Fig. 6f. The anomaly E–P flux clearly forces a zonal wind deceleration between 30° and 50°S, and an acceleration poleward of 50°S, consistent with the zonal wind anomalies (Fig. 5a). Similar diagnostic results were obtained by Kidson and Sinclair (1995) for their high-latitude EOF mode composites.

The baroclinic term, acting through the poleward heat fluxes, reduces the north–south temperature gradient and thus the thermal wind. The difference between the divergence of the baroclinic component for the split and nonsplit composites (Fig. 6e) is larger in amplitude than that of the barotropic component and is not well correlated with the zonal wind anomalies. Previous studies (Kidson and Sinclair 1995) have suggested that a contribution from the Coriolis torque from the induced TEM circulation is needed to balance the momentum budget (Trenberth 1986). In section 5, we will use stationary model simulations to show that forcing due to the barotropic term is indeed dominant over the baroclinic term.

b. The heat budget

In section 3, the composites reveal that cold anomalies occur over Antarctica while warm anomalies prevail in the South Pacific in association with the split jet, and vice versa for the nonsplit-flow case. The heat budget is examined here to understand how the temperature anomalies are maintained.

As suggested by Wang and Ting (1999), due to the availability of all data at pressure levels, the three-dimensional diabatic heating can easily be derived from the thermodynamic equation in pressure coordinates using monthly mean data as follows:
i1520-0469-63-2-634-e5
We can rewrite (5) as
i1520-0469-63-2-634-e6
where MFA = −[vh · + (∂/∂p)] is the mean flow advection term and EHF = −{ · () + [∂()/∂p]} is the eddy heat flux term. Therefore, the mean temperature tendency can be decomposed into three terms: the mean diabatic heating, the mean flow advection, and the eddy heat flux.

Figure 7 shows the differences in the heat budget at 700 hPa between the split and nonsplit composites. The diabatic heating anomalies are negatively correlated with the temperature anomalies in South Pacific with a spatial correlation coefficient of −0.4, while the mean flow advection anomalies are positively correlated with the temperature anomalies and the spatial correlation coefficient is 0.48. Anomalies due to the eddy heat fluxes are relatively noisy, and they are weakly anticorrelated with the temperature anomalies with a spatial correlation coefficient of −0.2. Thus the temperature anomalies are mainly forced/maintained by mean flow advection.

The latitude–height cross section of the 160°E–160°W longitude belt average of the heat budget is shown in Fig. 8. The diabatic heating pattern is almost the inverse of the temperature anomalies from 40° to 75°S in the troposphere, and the sign of the mean flow advection corresponds to the temperature anomalies. The eddy heat flux is again seen to be weakly anticorrelated with the temperature anomalies.

The heat budget suggests that the temperature anomalies associated with the split jet are in response to the advection by the large-scale flow, but not to the local diabatic heating or eddy heat fluxes. As shown in section 4a, the localized E–P flux diagnostics suggests that the split/nonsplit flow is maintained by the eddy vorticity flux. Putting the two together, the suggestion is that the mean flow advection, which is seen to be the term maintaining the temperature anomalies, is due to the secondary circulation driven by the anomalies in the eddy vorticity flux. To show that this is indeed the case, stationary wave model diagnostics are presented in the next section.

5. Stationary wave model diagnostics

As shown by previous studies (e.g., Wang and Ting 1999; Held et al. 2002), the stationary wave model can be a powerful diagnostic tool for understanding the maintenance of stationary planetary waves. Furthermore, this technique has been successfully applied to diagnose the maintenance of the SH climatological stationary wave (Ting et al. 2001), so it should be applicable to diagnose the stationary wave anomaly for this study. In this section, experiments using a stationary wave model are employed to more quantatively understand how the split/nonsplit flow is maintained.

a. Description of the model

The nonlinear stationary wave model used here has been adopted from the dynamical core of the Geophysical Fluid Dynamics Laboratory (GFDL) spectral model (Held and Suarez 1994). The model is based on the three-dimensional primitive equations in σ coordinates. The basic prognostic equations are those for perturbation vorticity, divergence, temperature, and surface pressure. A semi-implicit time integration scheme is employed with a time step of 30 min. This strategy is similar to that employed by Jin and Hoskins (1995) except that we use slightly stronger damping (discussed below) to damp out baroclinic instability [as suggested by Ting and Yu (1998) and Held et al. (2002)]. The stationary wave solution is obtained by integrating the model to a quasi–steady state, which is usually achieved after approximately 20 days. The model has T-42 truncation in the horizontal and 14 unevenly spaced σ levels in the vertical. The basic state employed in this study is the 3D climatological (1979–93) basic flow in JAS taken from the ERA-15 data. The forcings for the model include orography, tropical diabatic heating, and transient vorticity and heat flux convergences.

The damping used in the nonlinear model includes Rayleigh friction, Newtonian cooling, and biharmonic diffusion. The damping parameters used in this study are the same as those used by Held et al. (2002) for the NH stationary wave modeling. The Rayleigh friction damping time scales for both the vorticity and divergence equations are 0.3, 0.5, 1.0, and 8.0 days for the lowest four levels (σ values of 0.997, 0.979, 0.935, and 0.866), and 25 days throughout the rest of the model. The time scale of the Newtonian cooling is 15 days at all levels. The time scale of the biharmonic diffusion coefficient, identical for vorticity, divergence, and temperature, is chosen to be 0.1 day. These damping terms are significantly stronger than those typically used in GCMs of this resolution, and help to suppress model-generated baroclinic waves.

The diabatic heating was computed from reanalysis data as a residual using the thermodynamic equation in pressure coordinates (5) and then spatially interpolated onto the model grid. Since we have shown that the mid- and high-latitude diabatic heating forcing is negatively correlated to the temperature tendency (Figs. 7 and 8), and is clearly a response to the temperature anomaly rather than a real forcing, only the tropical part (20°S–20°N) of the diabatic heating forcing is applied here. The stationary wave forcings for the split and nonsplit cases were obtained with the same composite method as in section 4. When the anomalous forcings are applied, the nonlinear model reaches a quasi–steady state after being integrated for around 20 days. The nonlinear model solutions shown in this study are averaged over days 31–50, a period adequate to generate results that are not sensitive to the averaging period.

b. The maintenance of the split/nonsplit jet

The model solutions for the split and nonsplit cases are shown in Figs. 9a,b, respectively, and the differences between the split and nonsplit solutions are shown in Fig. 9c. The response to all forcings for the split-flow composite (Fig. 9a) clearly shows a more pronounced split jet than the climatological flow (Fig. 1), while the response for the nonsplit case (Fig. 9b) reveals a single jet. Comparing Figs. 9c and 5a, we can see that the model simulations match well with the ERA-15 data, except that the amplitude is a bit weaker. We conclude that when all forcings are applied, the stationary wave model can reproduce the patterns of the split and nonsplit flows.

The decomposition of the model solution into parts forced by the tropical heating, the eddy heat flux convergence, and the eddy vorticity flux convergence, for the difference between the split and nonsplit composites, is shown in Fig. 10. The model response to the tropical heating (Fig. 10c) is weak compared to those forced by the vorticity and heat flux convergences. The response to the eddy vorticity flux convergence (Fig. 10b) is dominant, while the response to the eddy heat flux convergence (Fig. 10d) is negatively correlated with the observed anomaly. Figures 10b,d show that the zonal wind anomaly induced by vorticity fluxes is stronger than the opposing effects due to the heat flux. Hence, the split jet is maintained by the vorticity fluxes while the heat flux tends to dissipate the split flow. The 700-hPa temperature anomalies forced by eddy vorticity fluxes alone (Fig. 10f) display a similar pattern to the observed temperature anomalies (Fig. 5b), which confirms our previous suggestion that much of the temperature anomalies can be explained by advection due to the secondary circulation. The role of nonlinear interactions between the different forcing terms can be estimated by taking the difference between the response due to all the forcings acting together (Fig. 10a) and the sum of the individual forcing (sum of Figs. 10b–d), and the result (Fig. 10e) also suggest that nonlinearity is not important.

The stationary wave model diagnostics are consistent with the E–P flux and heat budget diagnostics in section 4, and they reveal that the split/nonsplit jet is maintained by the vorticity fluxes. These results are in good agreement with the diagnostic studies of the midlatitude storm tracks by Lau and Nath (1991). In their analyses of NH storm-track anomalies, they illustrated that in the upper troposphere, the geopotential tendency induced by the vorticity fluxes is dominant over the opposing effects due to the heat fluxes.

6. Organization of eddies by the split/nonsplit jet

The diagnostics in section 5 showed that the monthly mean split/nonsplit jet could be maintained by eddy vorticity (or momentum) fluxes. However, how these eddy flux anomalies are generated is not yet clear. Branstator (1995) has shown that there is a two-way feedback between high- and low-frequency disturbances during episodes of prominent low-frequency anomalies in a GCM; that is, some large-scale patterns may be able to organize storm-track activity in such a way that the associated eddy fluxes feed back positively onto the large-scale anomalies while other patterns cannot induce such a positive feedback. In this section, the same method used by Branstator (1995) is employed to examine if the high-frequency eddies observed here are organized by the low-frequency split/nonsplit jet.

a. The random initial condition method

Here, the initial value technique employed by Branstator (1995) is briefly described. Starting from a spatially white random distribution of perturbations superimposed on a 3D basic state, a linear model is integrated for several days to determine how disturbances would tend to evolve under the influence of the background flow. With the effects of local regions of enhanced baroclinity, and the steering effect of the background winds, there is a tendency for perturbations with certain structures in certain regions to grow and be focused into preferred sectors while other perturbations decay. From a sufficient number of such integrations, what amounts to a climatology of the dominant, fast time scale disturbances of the model can be formed. Using this technique, one can find the storm tracks for a given background state. Details of this method can be found in Branstator (1995).

b. The model formulation and the simulation results

The model used here is the same model employed in section 5. However, here we are interested in the transient eddy statistics with random initial conditions, rather than the stationary solutions with steady forcings. Though this model is nonlinear, if we use small-amplitude initial perturbations, and stop the integrations before the perturbations become nonlinear, the model is equivalent to a linear model.

The Rayleigh friction damping times for both the vorticity and divergence equations are 2, 2, 2, 4, and 18 days for the lowest five levels (0.997, 0.979, 0.935, 0.866, and 0.777), respectively, and 60 days through the rest of the model. The time scale of the Newtonian cooling is 60 days at all levels. The biharmonic diffusion coefficient, identical for vorticity, divergence, and temperature, is chosen to be 1 day. These values are consistent with the damping used by Branstator (1995). Compared with the stationary wave model, damping with a longer time scale is applied here to allow baroclinic waves to grow. Background states for the model are the composite monthly mean fields from the ERA-15 data for the split and nonsplit case. The model variables are sampled every 6 h, and the time average from each integration is removed as a simple means of isolating the transient component of the perturbation fields. Two sets of 60 integrations of the model starting from different white noise initial perturbations are generated to construct the ensembles that form the statistics of perturbations associated with a split/nonsplit background state.

To ensure that the solution is linear, we set the amplitudes for the random initial perturbations to be 0.1 m s−1 for the zonal wind and the meridional wind, and 0.1 K for temperature, which are more than one order of magnitude smaller than the basic-state quantities.

The length of each integration used to collect the storm-track statistics has been discussed by Branstator (1995). Here, we use the same length—5 days. We have tested the model with the NH winter background state with 5-day integrations, and the model reproduces the storm-track statistics of the linear model outputs of Branstator’s study.

The storm tracks of the ERA-15 data are represented by the high-frequency meridional wind variance. The high-frequency meridional wind field is obtained by the planetary wave spectral spatial filter, which removes the large spatial scales by setting the spectral coefficients for total wavenumbers less than some prescribed cutoff to zero. As suggested by Anderson et al. (2003), total wavenumber 7 is used as the cutoff here. The amplitudes of the meridional wind variance from the model outputs are tuned to the high-frequency meridional wind variance of the ERA-15 data. Note that the same tuning factor is employed for both the split and nonsplit experiments.

Figures 11a,c,e show the storm tracks for the split-/nonsplit-jet composites (Figs. 11a,c) and the difference between the two (Fig. 11e), based on the ERA data. The storm track for the split-jet composite splits into two west of New Zealand, with a branch joining up with the STJ and another branch spiraling poleward with the PFJ, and there is a prominent local minimum between the two branches. For the nonsplit-jet composite, there is only one branch across the midlatitude of the whole SH. The storm-track structures of the two composites match well to the jet structures shown in Fig. 4. The difference between the two storm tracks shows that there is a significant decrease starting from the east of New Zealand cross the South Pacific Ocean to the west of South America, and there is an increase in the high latitudes across much of the SH.

The square roots of the 300-hPa meridional wind variance from the model outputs with split- and nonsplit-jet background states are shown in Figs. 11b,d,f. The model storm tracks have a good match to the ECMWF storm tracks for both cases. The model solution of the split-jet background state shows the two branches of the storm track and the local minimum between the two, though the minimum center is slightly poleward of that in the ECMWF data. The model reproduces the nearly circular storm track across the whole SH for the single jet case. The difference between the two cases from the model solutions (Fig. 11f) also shows the same structure as ERA-15, that is, a significant decrease in the midlatitudes of the South Pacific Ocean and the ring of positive anomalies in the high latitudes around Antarctica.

As shown in sections 4 and 5, the eddy momentum fluxes are instrumental in maintaining the low-frequency split/nonsplit jet, so it is of interest to see whether the model can reproduce the eddy momentum flux anomalies. As a means of quantifying these attributes, the localized E–P flux diagnostics in section 5 are used here. The same scale of tuning used in computing the eddy variances is used to calculate the momentum fluxes. The change in the latitude of the momentum flux reversal is clearly shown in the barotropic component of the E–P flux in Figs. 12a,c, where arrows incline northward from near 40°S in the single-jet case and northward from around 60°S in the split-jet case. Figures 12b,d show the divergence of the barotropic component of the E–P flux at 300 hPa for the two composites. For the split-jet case, a region of westerly acceleration can be seen along the PFJ, which is shaded in Fig. 12b, and there is a westerly deceleration center near New Zealand. For the nonsplit-jet case, the westerly acceleration zone extends almost zonally along 30°S. Again, it is apparent that high-frequency eddies organized by the split/nonsplit jet help to maintain the time-averaged split/nonsplit jet. Most of these features are similar to those computed based on spatially filtered ERA-15 data (not shown here, but plots are similar to Fig. 8 except with weaker amplitude due to the spatial filtering). However, one should note that the model shows a well-organized acceleration along the subtropical jet for the split-jet case (Fig. 12b), which is not present in the ERA-15 data (Fig. 8b). This could be due to the fact that the model generates stronger eddies along the subtropical jet than is observed (see Figs. 11a,b).

The storm-track anomalies that are organized by the low-frequency split/nonsplit anomalies match the storm-track anomalies that accompany the low-frequency split/nonsplit-jet anomalies in the ERA-15 data, thus establishing that the low-frequency split/nonsplit-jet patterns are responsible for the coincident storm-track anomalies. In section 5, we have shown that the monthly mean split/nonsplit flow is maintained by the eddy momentum fluxes. Thus, at this stage, a two-way feedback has been clearly established between the high-frequency eddies and the low-frequency split/nonsplit jet.

7. Conclusions and discussion

In this study, a simple split-jet index is defined, and the composites based on this index show that a cold–warm–cold tripolar temperature anomaly starting from near the South Pole, extending equatorward over the South Pacific, exists in association with the split-flow regimes, and the nonsplit flow exists when the phase of the tripolar temperature anomaly is reversed. The composite temperature anomalies associated with the split/nonsplit jet, based on the new split-jet index, are consistent with those of Bals-Elsholz et al. (2001). The localized E–P flux diagnostics and the stationary wave model diagnostics both illustrate that the split/nonsplit jet is largely maintained by the vorticity flux, which is dominant over the heat flux that tends to dissipate the jet anomaly. The heat budget reveals that the temperature anomaly associated with the split/nonsplit jet is in response to the large-scale advection, and this is consistent with the interpretation that the temperature anomaly is forced by the secondary circulation driven by the convergence of vorticity flux.

While previous studies using localized E–P flux diagnostics (e.g., Trenberth 1986; Kidson and Sinclair 1995) have suggested similar roles for the eddy vorticity and heat fluxes, their results are, to certain extent, incomplete, since these studies have to rely upon the assumption that the TEM circulation can largely cancel out most of the effects of the heat fluxes, owing to the fact that the divergence of the vertical component of the E–P flux was found to be larger than that due to the horizontal components (Fig. 6). In this study, this hypothesis is explicitly confirmed by our usage of the stationary wave model to simulate the responses forced by the eddy fluxes. In addition, our model results demonstrate that tropical heating anomalies are not important in the maintenance of the split-/nonsplit-flow anomalies, even though diabatic heating is important in the maintenance of the climatological split flow itself (Wang and Ting 1999).

The linear storm-track model based on the random initial condition method pioneered by Branstator (1995) has been employed here to demonstrate that the low-frequency split-/nonsplit-jet anomalies organize the coincident storm-track structures. While this result is also somewhat expected, a direct model confirmation is still required since Branstator (1995) has shown that not all low-frequency anomalies can organize eddies to maintain themselves. Moreover, the results of Chang and Fu (2003) also showed that while storm-track and mean flow changes are well correlated, not all storm-track changes can be directly attributable to changes in the large-scale monthly mean flow (see also Whitaker and Sardeshmukh 1998); a substantial amount of storm-track variability is probably due to nonlinear climate noise. Hence, without the storm-track model results, one cannot rule out the possibility that the eddy fluxes that act to maintain the split-/nonsplit-flow anomalies have not merely arisen due to random nonlinear fluctuations of the storm tracks. In this study, the stationary wave model diagnostics and the linear storm-track model simulations directly and quantitatively establish a two-way feedback between the high-frequency eddies and the low-frequency split/nonsplit jet.

Here, our focus has been on the maintenance of the monthly mean anomalies of the split jet. We have not addressed what leads to the development of split/nonsplit flow. Previous studies have shown that the physical mechanism leading to the growth/decay of low-frequency flow anomalies may not be the same as the mechanisms that help to maintain those anomalies. As an example, Feldstein (1998) showed that while the growth and decay of a low-frequency anomaly in a GCM is dominated by low-frequency linear processes, in the time mean, nonlinear fluxes due to the high-frequency transients are seen to help prolong the anomaly. Thus, it is not clear at this stage what physical mechanisms are responsible for exciting the growth and decay of split-/nonsplit-jet anomalies.

The time series of this split-jet index clearly show that the upper-troposphere zonal flow fluctuates between the split and nonsplit jet. An extension of the current work is to understand the life cycle of the SH split jet. Composites based on the 6-hourly time series of the SJI are being examined for the growing and decaying stages, respectively, of the split jet to determine the physical processes that govern the evolution of the split jet.

Acknowledgments

The authors thank Dr. Isaac Held for providing the spectral dynamical core model that was modified to conduct the stationary wave and storm-track model experiments. Comments and suggestions from three reviewers have helped to clarify this manuscript. The reanalysis data were obtained from the NCAR data archives, and assistance from the NCAR SCD is appreciated. This research is supported by NSF Grants ATM-0296076 and ATM-0354616.

REFERENCES

  • Anderson, D., K. I. Hodges, and B. J. Hoskins, 2003: Sensitivity of feature-based analysis methods of storm tracks to the form of background field removal. Mon. Wea. Rev., 131 , 565573.

    • Search Google Scholar
    • Export Citation
  • Bals-Elsholz, T. M., E. H. Atallah, L. F. Bosart, T. A. Wasula, M. J. Cempa, and A. R. Lupo, 2001: The wintertime Southern Hemisphere split jet: Structure, variability, and evolution. J. Climate, 14 , 41914215.

    • Search Google Scholar
    • Export Citation
  • Branstator, G., 1995: Organization of storm track anomalies by recurring low-frequency circulation anomalies. J. Atmos. Sci., 52 , 207226.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., 1999: Characteristics of wave packets in the upper troposphere. Part II: Seasonal and hemisphere variations. J. Atmos. Sci., 56 , 17291747.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., and Y. Fu, 2003: Using mean flow change as a proxy to infer interdecadal storm track variability. J. Climate, 16 , 21782196.

    • Search Google Scholar
    • Export Citation
  • Chen, B. C., S. R. Smith, and D. H. Bromwich, 1996: Evolution of the tropospheric split jet over the South Pacific Ocean during the 1986–89 ENSO cycle. Mon. Wea. Rev., 124 , 17111731.

    • Search Google Scholar
    • Export Citation
  • Feldstein, S. B., 1998: The growth and decay of low-frequency anomalies in a GCM. J. Atmos. Sci., 55 , 415428.

  • Gong, D., and S. Wang, 1999: Definition of Antarctic Oscillation index. Geophys. Res. Lett., 26 , 459462.

  • Held, I. M., and M. J. Suarez, 1994: A proposal for the intercomparison of the dynamical cores of atmospheric circulation models. Bull. Amer. Meteor. Soc., 75 , 18251830.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., M. Ting, and H. Wang, 2002: Northern winter stationary waves: Theory and modeling. J. Climate, 15 , 21252144.

  • Hurrell, J. W., H. van Loon, and D. J. Shea, 1998: The mean state of the troposphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 49, Amer. Meteor. Soc., 1–46.

    • Search Google Scholar
    • Export Citation
  • Jin, F-F., and B. J. Hoskins, 1995: The direct response to tropical heating in a baroclinic atmosphere. J. Atmos. Sci., 52 , 13291340.

    • Search Google Scholar
    • Export Citation
  • Karoly, D. J., 1989: Southern Hemisphere circulation features associated with El Niño–Southern Oscillation events. J. Climate, 2 , 12391252.

    • Search Google Scholar
    • Export Citation
  • Karoly, D. J., 1990: The role of transient eddies in low-frequency zonal variations of the Southern Hemisphere circulation. Tellus, 42A , 4150.

    • Search Google Scholar
    • Export Citation
  • Kidson, D. J., 1988: Indices of the Southern Hemisphere zonal wind. J. Climate, 1 , 183194.

  • Kidson, D. J., and M. R. Sinclair, 1995: The influence of persistent anomalies on Southern Hemisphere storm tracks. J. Climate, 8 , 19381950.

    • Search Google Scholar
    • Export Citation
  • Lau, N-C., and M. J. Nath, 1991: Variability of the baroclinic and barotropic transient eddy forcing associated with monthly changes in the midlatitude storm tracks. J. Atmos. Sci., 48 , 25892613.

    • Search Google Scholar
    • Export Citation
  • Mechoso, C. R., A. O’Neill, V. D. Pope, and J. D. Farrara, 1988: A study of the stratospheric final warming of 1982 in the Southern Hemisphere. Quart. J. Roy. Meteor. Soc., 114 , 13651384.

    • Search Google Scholar
    • Export Citation
  • Mo, K. C., J. Pfaendtner, and E. Kalnay, 1987: A GCM study on the maintenance of the June 1982 blocking in the Southern Hemisphere. J. Atmos. Sci., 44 , 11231142.

    • Search Google Scholar
    • Export Citation
  • Newton, C. W., 1972: Southern Hemisphere general circulation in relation to global energy and momentum balance requirements. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 215–240.

    • Search Google Scholar
    • Export Citation
  • Taljaard, J. J., 1972: Synoptic meteorology of the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 139–211.

    • Search Google Scholar
    • Export Citation
  • Thompson, D. W. J., and J. M. Wallace, 2000: Annular modes in the extratropical circulation. Part I: Month-to-month variability. J. Climate, 13 , 10001036.

    • Search Google Scholar
    • Export Citation
  • Ting, M., and L. Yu, 1998: Steady response to tropical heating in wavy linear and nonlinear baroclinic models. J. Atmos. Sci., 55 , 35653582.

    • Search Google Scholar
    • Export Citation
  • Ting, M., H. Wang, and L. Yu, 2001: Nonlinear stationary wave maintenance and seasonal cycle in the GFDL R30 GCM. J. Atmos. Sci., 58 , 23312354.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1986: An assessment of the impact of transient eddies on the zonal flow during a blocking episode using Eliassen–Palm flux diagnostics. J. Atmos. Sci., 43 , 20702087.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1991: Storm tracks in the Southern Hemisphere. J. Atmos. Sci., 48 , 21592178.

  • van Heerden, J., and J. J. Taljaard, 1998: Africa and surrounding waters. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 49, Amer. Meteor. Soc., 141–174.

    • Search Google Scholar
    • Export Citation
  • van Loon, H., 1972a: Temperature in the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 9–22.

    • Search Google Scholar
    • Export Citation
  • van Loon, H., 1972b: Pressure in the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 59–86.

    • Search Google Scholar
    • Export Citation
  • van Loon, H., 1972c: Wind in the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 87–99.

    • Search Google Scholar
    • Export Citation
  • Vincent, D. G., and P. L. Silva Dias, 1998: Pacific Ocean. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 49, Amer. Meteor. Soc., 101–117.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., 2000: North Atlantic Oscillation/Annular mode: Two paradigms—One phenomenon. Quart. J. Roy. Meteor. Soc., 126 , 791805.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., and D. W. J. Thompson, 2002: The Pacific center of action of the Northern Hemisphere annual mode: Real or artifact? J. Climate, 15 , 19871991.

    • Search Google Scholar
    • Export Citation
  • Wang, H., and M. Ting, 1999: Seasonal cycle of the climatological stationary waves in the NCEP–NCAR reanalysis. J. Atmos. Sci., 56 , 38923919.

    • Search Google Scholar
    • Export Citation
  • Whitaker, J. S., and P. D. Sardeshmukh, 1998: A linear theory of extratropical synoptic eddy statistics. J. Atmos. Sci., 55 , 237258.

Fig. 1.
Fig. 1.

Climatology of the zonal wind at 300 hPa for JAS 1979–93 from ECMWF. Isotachs are contoured every 5 m s−1. Shaded regions represents values over 20 m s−1. Long-dashed contour denotes 50°S (same for all figures).

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 2.
Fig. 2.

Hovmöller plot of (a) the zonal wind isotachs from 1 to 31 Aug 1984 averaged between 150°E and 150°W from 0° to the South Pole and (b) the same plot of the zonal wind anomaly (climatology removed). (c) The time series of the SJI of August 1984 are shown. Isotachs are contoured every 10 m s−1. Shaded regions represent values over 20 m s−1 in (a) and positive values in (b).

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 3.
Fig. 3.

The NMSJI for JAS from 1979 to 1993.

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 4.
Fig. 4.

The 300-hPa isotachs of the zonal wind for the (a) composite split and (c) composite nonsplit flows. Vertical cross section of the zonal wind isotachs from 1000 to 10 hPa averaged at the 160°E–160°W longitude belt between 90°S and 0° for the (b) composite split and (d) composite nonsplit flows. Isotachs are contoured every 10 m s−1. Shaded areas denote values over 20 m s−1.

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 5.
Fig. 5.

Differences between composite and composite nonsplit flows for (a) the 300-hPa isotachs of the zonal wind anomaly, (b) the 700-hPa isotherms of the temperature anomaly, the vertical cross section of (c) the zonal wind anomaly isotachs and (d) the temperature anomaly isotherms from 1000 to 10 hPa averaged at the 160°E–160°W longitude belt between 90°S and 0°, the vertical cross section of (e) the zonal-mean zonal wind anomaly isotachs, and (f) the zonal-mean temperature anomaly isotherms. Shaded regions represent positive values. Isotachs are contoured every 4 m s−1, and isotherms every 1 K.

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 6.
Fig. 6.

Barotropic component of the E–P flux at 300 hPa for (a) the split-jet composite and (c) the nonsplit-jet composite. The small arrow at the bottom right scales to a value of 50 m2 s−2. Divergence of barotropic component of the E–P flux at 300 hPa for the (b) split-jet and (d) nonsplit-jet composites. Difference of divergence of the E–P flux between split and nonsplit composites for the (e) vertical and (f) barotropic components. Contour intervals for (b) and (d) are 2 m s−1 day–1, and 4 s−1 day–1 for (e) and (f). The values over 20 m s−1 of the zonal winds are shaded in (a), (b), (c), and (d).

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 7.
Fig. 7.

The difference between split and nonsplit composites of the heat budget at 700 hPa between 90°E and 90°W for (a) the diabatic heating, (c) the mean flow advection, and (d) the eddy heat flux. Contour intervals for the heat budget terms are 1 K day−1. (b) The difference of the temperature anomaly at 700 hPa, and isotherms are contoured every 1 K. Shaded regions denote positive values.

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 8.
Fig. 8.

Same as in Fig. 7 but for the vertical cross section from 1000 to 100 hPa averaged at the 160°E–160°W longitude belt between 80°S and 0° for the (a) diabatic heating, (b) potential temperature anomaly, (c) mean flow advection, and (d) eddy heat flux. Contour intervals for the heat budget terms are 0.5 K day−1, and 1 K for isotherms.

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 9.
Fig. 9.

The steady model response of the 300-hPa zonal wind between 90°E and 90°W to forcing by the diabatic heating and transient eddy flux convergences for the (a) split, (b) nonsplit, and (c) the split minus the nonsplit cases. Contour intervals for (a) and (b) are 5 m s−1, and 4 m s−1 for (c). Values over 20 m s−1 are shaded in (a) and (b), and positive values are shaded in (c).

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 10.
Fig. 10.

Differences between the split and nonsplit cases for (a) the total model response and decomposition of the total model response of the 300-hPa zonal wind anomalies into parts forced by (b) vorticity flux, (c) tropical heating, and (d) heat flux. (e) Nonlinearity is computed as a residual. (f) Difference between the split and nonsplit cases for the 700-hPa temperature anomalies forced by the vorticity flux. Contour intervals for (a) and (b) are 4 m s−1; 2 m s−1 for (c), (d), and (e); and 1 K for (f). Positive values are shaded.

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 11.
Fig. 11.

The square root of the ECMWF 300-hPa high-frequency (total wavenumber-7 cutoff) meridional wind variance for the (a) composite split, (c) composite nonsplit, and (e) composite split flow minus composite nonsplit flows. The square root of the model 300-hPa meridional wind variance for the (b) composite split, (d) composite nonsplit, and (f) composite split flow minus composite nonsplit flows. Contour intervals are 2 m s−1. Values over 12 m s−1 are shaded in (a), (b), (c), and (d); positive values are shaded in (e) and (f).

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Fig. 12.
Fig. 12.

Barotropic component of the E–P flux at 300 hPa of the model for the (a) split-jet and (c) nonsplit-jet background states. The small arrow at the bottom right scales to a value of 50 m2 s−2. Divergence of barotropic component of the E–P flux at 300 hPa of the model for the (b) split-jet and (d) nonsplit-jet basic states. Contour interval is 2 m s−1 day−1. Regions with zonal wind value over 20 m s−1 are shaded in each plot.

Citation: Journal of the Atmospheric Sciences 63, 2; 10.1175/JAS3643.1

Save
  • Anderson, D., K. I. Hodges, and B. J. Hoskins, 2003: Sensitivity of feature-based analysis methods of storm tracks to the form of background field removal. Mon. Wea. Rev., 131 , 565573.

    • Search Google Scholar
    • Export Citation
  • Bals-Elsholz, T. M., E. H. Atallah, L. F. Bosart, T. A. Wasula, M. J. Cempa, and A. R. Lupo, 2001: The wintertime Southern Hemisphere split jet: Structure, variability, and evolution. J. Climate, 14 , 41914215.

    • Search Google Scholar
    • Export Citation
  • Branstator, G., 1995: Organization of storm track anomalies by recurring low-frequency circulation anomalies. J. Atmos. Sci., 52 , 207226.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., 1999: Characteristics of wave packets in the upper troposphere. Part II: Seasonal and hemisphere variations. J. Atmos. Sci., 56 , 17291747.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., and Y. Fu, 2003: Using mean flow change as a proxy to infer interdecadal storm track variability. J. Climate, 16 , 21782196.

    • Search Google Scholar
    • Export Citation
  • Chen, B. C., S. R. Smith, and D. H. Bromwich, 1996: Evolution of the tropospheric split jet over the South Pacific Ocean during the 1986–89 ENSO cycle. Mon. Wea. Rev., 124 , 17111731.

    • Search Google Scholar
    • Export Citation
  • Feldstein, S. B., 1998: The growth and decay of low-frequency anomalies in a GCM. J. Atmos. Sci., 55 , 415428.

  • Gong, D., and S. Wang, 1999: Definition of Antarctic Oscillation index. Geophys. Res. Lett., 26 , 459462.

  • Held, I. M., and M. J. Suarez, 1994: A proposal for the intercomparison of the dynamical cores of atmospheric circulation models. Bull. Amer. Meteor. Soc., 75 , 18251830.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., M. Ting, and H. Wang, 2002: Northern winter stationary waves: Theory and modeling. J. Climate, 15 , 21252144.

  • Hurrell, J. W., H. van Loon, and D. J. Shea, 1998: The mean state of the troposphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 49, Amer. Meteor. Soc., 1–46.

    • Search Google Scholar
    • Export Citation
  • Jin, F-F., and B. J. Hoskins, 1995: The direct response to tropical heating in a baroclinic atmosphere. J. Atmos. Sci., 52 , 13291340.

    • Search Google Scholar
    • Export Citation
  • Karoly, D. J., 1989: Southern Hemisphere circulation features associated with El Niño–Southern Oscillation events. J. Climate, 2 , 12391252.

    • Search Google Scholar
    • Export Citation
  • Karoly, D. J., 1990: The role of transient eddies in low-frequency zonal variations of the Southern Hemisphere circulation. Tellus, 42A , 4150.

    • Search Google Scholar
    • Export Citation
  • Kidson, D. J., 1988: Indices of the Southern Hemisphere zonal wind. J. Climate, 1 , 183194.

  • Kidson, D. J., and M. R. Sinclair, 1995: The influence of persistent anomalies on Southern Hemisphere storm tracks. J. Climate, 8 , 19381950.

    • Search Google Scholar
    • Export Citation
  • Lau, N-C., and M. J. Nath, 1991: Variability of the baroclinic and barotropic transient eddy forcing associated with monthly changes in the midlatitude storm tracks. J. Atmos. Sci., 48 , 25892613.

    • Search Google Scholar
    • Export Citation
  • Mechoso, C. R., A. O’Neill, V. D. Pope, and J. D. Farrara, 1988: A study of the stratospheric final warming of 1982 in the Southern Hemisphere. Quart. J. Roy. Meteor. Soc., 114 , 13651384.

    • Search Google Scholar
    • Export Citation
  • Mo, K. C., J. Pfaendtner, and E. Kalnay, 1987: A GCM study on the maintenance of the June 1982 blocking in the Southern Hemisphere. J. Atmos. Sci., 44 , 11231142.

    • Search Google Scholar
    • Export Citation
  • Newton, C. W., 1972: Southern Hemisphere general circulation in relation to global energy and momentum balance requirements. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 215–240.

    • Search Google Scholar
    • Export Citation
  • Taljaard, J. J., 1972: Synoptic meteorology of the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 139–211.

    • Search Google Scholar
    • Export Citation
  • Thompson, D. W. J., and J. M. Wallace, 2000: Annular modes in the extratropical circulation. Part I: Month-to-month variability. J. Climate, 13 , 10001036.

    • Search Google Scholar
    • Export Citation
  • Ting, M., and L. Yu, 1998: Steady response to tropical heating in wavy linear and nonlinear baroclinic models. J. Atmos. Sci., 55 , 35653582.

    • Search Google Scholar
    • Export Citation
  • Ting, M., H. Wang, and L. Yu, 2001: Nonlinear stationary wave maintenance and seasonal cycle in the GFDL R30 GCM. J. Atmos. Sci., 58 , 23312354.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1986: An assessment of the impact of transient eddies on the zonal flow during a blocking episode using Eliassen–Palm flux diagnostics. J. Atmos. Sci., 43 , 20702087.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1991: Storm tracks in the Southern Hemisphere. J. Atmos. Sci., 48 , 21592178.

  • van Heerden, J., and J. J. Taljaard, 1998: Africa and surrounding waters. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 49, Amer. Meteor. Soc., 141–174.

    • Search Google Scholar
    • Export Citation
  • van Loon, H., 1972a: Temperature in the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 9–22.

    • Search Google Scholar
    • Export Citation
  • van Loon, H., 1972b: Pressure in the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 59–86.

    • Search Google Scholar
    • Export Citation
  • van Loon, H., 1972c: Wind in the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 87–99.

    • Search Google Scholar
    • Export Citation
  • Vincent, D. G., and P. L. Silva Dias, 1998: Pacific Ocean. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 49, Amer. Meteor. Soc., 101–117.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., 2000: North Atlantic Oscillation/Annular mode: Two paradigms—One phenomenon. Quart. J. Roy. Meteor. Soc., 126 , 791805.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., and D. W. J. Thompson, 2002: The Pacific center of action of the Northern Hemisphere annual mode: Real or artifact? J. Climate, 15 , 19871991.

    • Search Google Scholar
    • Export Citation
  • Wang, H., and M. Ting, 1999: Seasonal cycle of the climatological stationary waves in the NCEP–NCAR reanalysis. J. Atmos. Sci., 56 , 38923919.

    • Search Google Scholar
    • Export Citation
  • Whitaker, J. S., and P. D. Sardeshmukh, 1998: A linear theory of extratropical synoptic eddy statistics. J. Atmos. Sci., 55 , 237258.

  • Fig. 1.

    Climatology of the zonal wind at 300 hPa for JAS 1979–93 from ECMWF. Isotachs are contoured every 5 m s−1. Shaded regions represents values over 20 m s−1. Long-dashed contour denotes 50°S (same for all figures).

  • Fig. 2.

    Hovmöller plot of (a) the zonal wind isotachs from 1 to 31 Aug 1984 averaged between 150°E and 150°W from 0° to the South Pole and (b) the same plot of the zonal wind anomaly (climatology removed). (c) The time series of the SJI of August 1984 are shown. Isotachs are contoured every 10 m s−1. Shaded regions represent values over 20 m s−1 in (a) and positive values in (b).

  • Fig. 3.

    The NMSJI for JAS from 1979 to 1993.

  • Fig. 4.

    The 300-hPa isotachs of the zonal wind for the (a) composite split and (c) composite nonsplit flows. Vertical cross section of the zonal wind isotachs from 1000 to 10 hPa averaged at the 160°E–160°W longitude belt between 90°S and 0° for the (b) composite split and (d) composite nonsplit flows. Isotachs are contoured every 10 m s−1. Shaded areas denote values over 20 m s−1.

  • Fig. 5.

    Differences between composite and composite nonsplit flows for (a) the 300-hPa isotachs of the zonal wind anomaly, (b) the 700-hPa isotherms of the temperature anomaly, the vertical cross section of (c) the zonal wind anomaly isotachs and (d) the temperature anomaly isotherms from 1000 to 10 hPa averaged at the 160°E–160°W longitude belt between 90°S and 0°, the vertical cross section of (e) the zonal-mean zonal wind anomaly isotachs, and (f) the zonal-mean temperature anomaly isotherms. Shaded regions represent positive values. Isotachs are contoured every 4 m s−1, and isotherms every 1 K.

  • Fig. 6.

    Barotropic component of the E–P flux at 300 hPa for (a) the split-jet composite and (c) the nonsplit-jet composite. The small arrow at the bottom right scales to a value of 50 m2 s−2. Divergence of barotropic component of the E–P flux at 300 hPa for the (b) split-jet and (d) nonsplit-jet composites. Difference of divergence of the E–P flux between split and nonsplit composites for the (e) vertical and (f) barotropic components. Contour intervals for (b) and (d) are 2 m s−1 day–1, and 4 s−1 day–1 for (e) and (f). The values over 20 m s−1 of the zonal winds are shaded in (a), (b), (c), and (d).

  • Fig. 7.

    The difference between split and nonsplit composites of the heat budget at 700 hPa between 90°E and 90°W for (a) the diabatic heating, (c) the mean flow advection, and (d) the eddy heat flux. Contour intervals for the heat budget terms are 1 K day−1. (b) The difference of the temperature anomaly at 700 hPa, and isotherms are contoured every 1 K. Shaded regions denote positive values.

  • Fig. 8.

    Same as in Fig. 7 but for the vertical cross section from 1000 to 100 hPa averaged at the 160°E–160°W longitude belt between 80°S and 0° for the (a) diabatic heating, (b) potential temperature anomaly, (c) mean flow advection, and (d) eddy heat flux. Contour intervals for the heat budget terms are 0.5 K day−1, and 1 K for isotherms.

  • Fig. 9.

    The steady model response of the 300-hPa zonal wind between 90°E and 90°W to forcing by the diabatic heating and transient eddy flux convergences for the (a) split, (b) nonsplit, and (c) the split minus the nonsplit cases. Contour intervals for (a) and (b) are 5 m s−1, and 4 m s−1 for (c). Values over 20 m s−1 are shaded in (a) and (b), and positive values are shaded in (c).

  • Fig. 10.

    Differences between the split and nonsplit cases for (a) the total model response and decomposition of the total model response of the 300-hPa zonal wind anomalies into parts forced by (b) vorticity flux, (c) tropical heating, and (d) heat flux. (e) Nonlinearity is computed as a residual. (f) Difference between the split and nonsplit cases for the 700-hPa temperature anomalies forced by the vorticity flux. Contour intervals for (a) and (b) are 4 m s−1; 2 m s−1 for (c), (d), and (e); and 1 K for (f). Positive values are shaded.

  • Fig. 11.

    The square root of the ECMWF 300-hPa high-frequency (total wavenumber-7 cutoff) meridional wind variance for the (a) composite split, (c) composite nonsplit, and (e) composite split flow minus composite nonsplit flows. The square root of the model 300-hPa meridional wind variance for the (b) composite split, (d) composite nonsplit, and (f) composite split flow minus composite nonsplit flows. Contour intervals are 2 m s−1. Values over 12 m s−1 are shaded in (a), (b), (c), and (d); positive values are shaded in (e) and (f).

  • Fig. 12.

    Barotropic component of the E–P flux at 300 hPa of the model for the (a) split-jet and (c) nonsplit-jet background states. The small arrow at the bottom right scales to a value of 50 m2 s−2. Divergence of barotropic component of the E–P flux at 300 hPa of the model for the (b) split-jet and (d) nonsplit-jet basic states. Contour interval is 2 m s−1 day−1. Regions with zonal wind value over 20 m s−1 are shaded in each plot.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 136 36 0
PDF Downloads 89 28 1