1 MARCH 2012
LI AND WETTSTEIN
1587
Thermally Driven and Eddy-Driven Jet Variability in Reanalysis*
CAMILLE LI
Bjerknes Centre for Climate Research, and Uni Bjerknes Centre, Uni Research, and
Department of Earth Science, University of Bergen, Bergen, Norway
JUSTIN J. WETTSTEIN
Bjerknes Centre for Climate Research, Bergen, Norway, and Climate Analysis Section, Climate and
Global Dynamics Division, National Center for Atmospheric Research, Boulder, Colorado
(Manuscript received 14 March 2011, in final form 12 July 2011)
ABSTRACT
Two important dynamical processes influence the extratropical zonal wind field: angular momentum
transport by the thermally direct Hadley circulation (thermal-driving T) and momentum flux convergence by
atmospheric waves (eddies) that develop in regions of enhanced baroclinicity (eddy-driving E). The relationship between extratropical zonal wind variability and these driving processes is investigated using 40-yr
European Centre for Medium-Range Weather Forecasts Re-Analysis (ERA-40) data. Indices representing
the processes (iT and iE) are defined based on vertically integrated diabatic heating and meridional convergence of the meridional flux of zonal momentum by eddies, respectively. Zonal wind signatures associated
with these indices are identified via composite analysis. In the Atlantic sector, zonal wind variability is mainly
associated with momentum flux convergence by baroclinic eddies, supporting the established view that the
Atlantic jet is primarily eddy driven. In the Pacific sector, zonal wind variability is associated with both driving
processes, evidence that the Pacific jet is both thermally driven and eddy driven. The thermally driven Pacific
signature reflects changes in jet strength (intensity and longitudinal extent) with some resemblance to the
zonal wind anomalies of the Pacific–North America (PNA) pattern. The eddy-driven signature reflects
a latitudinal shift of the jet exit region in both sectors that resembles the zonal wind anomalies of the North
Atlantic Oscillation (NAO) or West Pacific (WP) patterns.
1. Introduction
Extratropical atmospheric jets are commonly described
as either subtropical or midlatitude (also called eddy
driven or polar front) in character (Hartmann 2007).
The existence of these jets is attributed to different
dynamical processes: the subtropical jet forms as a result of angular momentum transport by the thermally
direct Hadley circulation (Held and Hou 1980), while
the midlatitude jet forms as a result of eddy momentum
flux convergence by atmospheric waves that develop in
* Bjerknes Centre for Climate Research Publication Number
A356.
Corresponding author address: Camille Li, Bjerknes Centre for
Climate Research, Allegaten 55, 5007 Bergen, Norway.
E-mail: [email protected]
DOI: 10.1175/JCLI-D-11-00145.1
Ó 2012 American Meteorological Society
regions of enhanced baroclinicity (Held 1975; Rhines
1975; Panetta 1993).
In reality, it is difficult to ascribe an observed jet to a
particular driving process because both processes operate
and interact continuously (Lee and Kim 2003; Walker
and Schneider 2006). Lee and Kim (2003) point out that
the Northern Hemisphere configuration of the two jets
is longitude dependent. In the North Atlantic, the zone
of strongest baroclinicity is situated relatively far poleward, resulting in distinct midlatitude and subtropical jets
and, hence, a double maximum in the upper-tropospheric
zonal wind. In the North Pacific, the zone of strongest
baroclinicity is situated near the latitude of the maximum
zonal wind, resulting in collocated midlatitude and subtropical jets. The seasonal cycle also influences the relative placement of the baroclinic zone and the latitude of
the maximum zonal wind. For example, while the two
jets are well separated in the North Atlantic during
the middle of winter (see Fig. 2 in Eichelberger and
1588
JOURNAL OF CLIMATE
Hartmann 2007), they are more latitudinally coincident
in the early and late winter.
A number of studies attempt to identify the flow
characteristics of these jet configurations (‘‘combined’’ or
‘‘separated’’ subtropical and midlatitude jets) in idealized
modeling setups. Lee and Kim (2003) investigate the
evolution of baroclinic eddies in the presence of preexisting subtropical jets of various strengths. They find that
if the preexisting (thermally driven) subtropical jet is
strong, then the flow evolves to a state with a single jet
and weak eddies that remain trapped on its poleward
flank. If the preexisting subtropical jet is weak, then the
resulting flow exhibits a vigorous, well separated (eddy
driven) midlatitude jet. In the case of the strong subtropical jet, eddy evolution does not follow a typical
Thorncroft et al. (1993) LC1 or ‘‘anticyclonic’’ life cycle
in which the tilted waves propagate equatorward and
decay through strong baroclinic energy conversions. Instead, the eddies exhibit relatively little tilt, weak meridional propagation (see Lee and Kim 2003, their Fig. 8),
and a complicated and delayed decay that occurs both
barotropically and baroclinically.
Eichelberger and Hartmann (2007) further examine
the effect of different jet configurations on the leading
pattern of zonal wind variability. They show that when
the subtropical and midlatitude jets are well separated,
the leading pattern is a latitudinal shifting of the midlatitude jet; however, when the jets are collocated, the
leading pattern is a pulsing of the combined jet. Based
on barotropic quasilinear theory, Barnes and Hartmann
(2011) suggest that the difference in variability patterns
is due to differences in eddy propagation (which in turn
depends on eddy length scales) and the subsequent impacts on the nature of eddy–mean flow interactions.
These studies demonstrate that the idealized ‘‘combined jet’’ and ‘‘separated jet’’ configurations have
features that are consistent with what is observed in the
Pacific and Atlantic sectors of the Northern Hemisphere, respectively, in terms of both the mean flow
and its variability. Because of the similarities between
the observed wind field and the idealized model simulations described above, it is often stated that the
large-scale atmospheric circulation in the North Pacific
sector is more subtropical in nature, while the North
Atlantic is more eddy driven. This characterization is
intuitively reasonable because (i) variations in tropical
Pacific heating are known to force extratropical Pacific
changes (e.g., Bjerknes 1969; Horel and Wallace 1981;
Livezey et al. 1997; Straus and Shukla 1997; Matthews
and Kiladis 1999; Ren et al. 2008), and (ii) the leading
pattern of climate variability in the North Atlantic
sector, the North Atlantic Oscillation (NAO), appears
to be driven and maintained by eddy processes (Hurrell
VOLUME 25
1995; DeWeaver and Nigam 2000; Limpasuvan and
Hartmann 2000; Benedict et al. 2004; Franzke et al. 2004;
Woollings et al. 2008).
While the similarities between the idealized modeling results and observations are compelling, we aim to
establish a more direct link between the dynamical
processes and the resultant circulation features in the
real world. Specifically, we ask whether consistent zonal
wind signatures associated with the thermal- and eddydriving processes are identifiable in reanalysis data.
Note that while some of the idealized modeling studies
described above deal with the relative roles of the driving
processes in producing the mean jets, the analysis in this
study is primarily concerned with jet variability. To be
consistent with the idealized modeling studies described
above, the reanalysis data should show evidence that
North Pacific jet variability is more influenced by tropical convection (thermally driven) and North Atlantic jet
variability is more influenced by eddy fluxes.
We address this question by examining the zonal
wind anomalies associated with indices representing the
thermal- and eddy-driving processes. The purpose is to
establish a framework for assessing the importance of
each dynamical process in driving jet variability in a particular (observed or simulated) mean state, with an eye
toward eventually being able to quantify how this balance
changes if that mean state is altered (e.g., under the influence of anthropogenic greenhouse gas forcing).
2. Data
We use data from the 40-yr European Centre for
Medium-Range Weather Forecasts Re-Analysis (ERA40) over an extended winter season [November–April
(NDJFMA)] from 1957 to 2002. Monthly detrended topof-the-atmosphere net downward radiation RT, surface
fluxes (net downward radiation RS, upward sensible heat
flux HS, upward latent heat flux LS), and precipitation
rate P are used to define the thermal-driving index,
while instantaneous daily (1200 UTC) zonal u and meridional y wind are used to define the eddy-driving index. The daily data are filtered with the same high-pass
Butterworth filter (13th order with a 50% signal cutoff at
a nominal period of 6 days) used in Wettstein and Wallace
(2010) to emphasize the variability associated with baroclinic eddies. Although the entire September 1957–August
2002 dataset was filtered, using only the NDJFMA season removes undesirable impulse effects from the filter.
Zonal wind composites are based on monthly anomalies of Northern Hemisphere u calculated by subtracting the 45-yr mean of each month to remove the
seasonal cycle. The leading patterns of Northern Hemisphere climate variability (shown in Figs. 4 and 5) are
1 MARCH 2012
1589
LI AND WETTSTEIN
obtained from standard empirical orthogonal function
(EOF) analysis of monthly 1000-hPa geopotential height
anomalies; the patterns are identical to those in Wettstein
and Wallace (2010), thus allowing for direct comparison
with the results in that study.
The North Atlantic (908W–408E) and North Pacific
(1208E–1108W) sector definitions are the same as those
used in Wettstein and Wallace (2010). Several other
sector definitions were tested (not shown), and the results presented here are not very sensitive to the exact
choice of longitudes.
3. Indices of thermal (iT) and eddy (iE) driving
We construct monthly averaged hemispheric (subscript H) and sector-based (subscripts A for Atlantic, P
for Pacific) indices, iT and iE, representing the processes
of thermal driving by tropical convection and eddy driving by momentum flux convergence in the midlatitudes,
respectively. All indices are monthly, although the eddydriving indices are constructed using monthly averages
of the high-pass-filtered meridional flux of zonal momentum (equivalently, high-pass-filtered zonal and meridional wind covariance).
The iT indices are defined as the area-weighted sum of
the vertically integrated diabatic heating (Boer 1986;
Trenberth and Solomon 1994):
ð
iT [ (RT 1 FS 1 LP) dA
5 [RT 1 (2RS 1 HS 1 HL ) 1 LP] dA,
(1)
where RT is the net downward radiation at the top of the
atmosphere; FS is the net upward surface flux; RS is the
net downward radiation at the surface; HS and HL are
the upward sensible and latent heat fluxes at the surface,
respectively; P is the precipitation rate; L is the latent
heat of evaporation; and dA represents the area integral over the hemisphere/sector and spanning the latitudes of the tropical precipitation bands during Northern
Hemisphere winter (208S–108N). High values of iT are
associated with stronger convection and a stronger overturning Hadley circulation. This relationship is shown
in Fig. 1 in regressions of the zonal mean meridional
streamfunction (c) onto iT, with the streamfunction
defined as the northward mass flux across a latitude (f)
circle above a pressure level p:
c(f, p) 5
ð
2pa cosf p
[y] dp,
g
0
(2)
where a is the radius of the earth, g is the gravitational
constant, and square brackets denote the zonal mean.
The streamfunction pattern associated with iTH is
a slightly weaker version of the iTP pattern, and both
indicate a stronger, equatorially concentrated overturning cell when the index is high. That the hemispheric
picture is dominated by the Pacific reflects the global
importance of the tropical Pacific variability, including
the El Niño–Southern Oscillation (ENSO) as well as
basinwide, intraseasonal variability unrelated to ENSO
(monthly correlations between iTP and standard ENSO
indices Niño-3, Niño-4, and Niño-3.4 range between 0.32
and 0.53). The absence of hemispherically influential
tropical Atlantic heating sources is demonstrated by weak
regressions onto iTA and lower correlations between
iTA and iTH (0.45) than between iTP and iTH (0.84).
The intersector correlation between iTA and iTP is insignificant (0.09).
An alternative approach is to use an index that represents angular momentum transport by the Hadley cell, a
quantity that is directly responsible for changes in the
zonal wind. However, the Hadley cell, while largely set by
thermal driving, is itself also influenced by extratropical
eddies (Walker and Schneider 2006; Caballero 2008;
Levine and Schneider 2011). Thus, while a circulation- or
momentum-transport-based index might better characterize the actual variability of the jet, it is not as direct
a measure of the thermal-driving process as a heatingbased index. Other definitions of iT based on midtropospheric vertical velocities (to capture the rising
branch of the Hadley cell), upper-tropospheric meridional velocities (to capture the upper branch of the
Hadley cell), and outgoing longwave radiation (OLR)
were investigated and produce consistent results (see
the appendix).
The iE indices are defined as the area- and massweighted sum of the local meridional convergence of
the meridional flux of zonal momentum (i.e., the areaaveraged acceleration of westerly winds by transient
eddies):
ð ð 50
iE [ 2
›
(u9y9) dp dA,
1000 ›y
(3)
where 9 denotes 6-day high-pass-filtered daily fields, y is
the northward direction, dA represents the area integral
poleward of a sector-dependent latitude u0 over the
hemisphere/sector, and dp represents the vertical integral from 1000 to 50 hPa. High iE values are generally
associated with a more vigorous winter storm track, as
shown in eddy kinetic energy (EKE 5 u92 1 y92) regressions in Fig. 1, and stronger mean westerlies around
558N (section 4).
The u0 is ideally chosen to coincide with the maximum
in the eddy momentum flux. However, the monthly eddy
1590
JOURNAL OF CLIMATE
VOLUME 25
FIG. 1. Indices representing (left) the thermal driving (iT) and (right) eddy driving (iE) of the
mean flow. (top) Seasonality of the indices defined over the entire hemisphere (gray), the North
Atlantic sector (purple), and the North Pacific sector (green) for the NDJFMA extended
winter season. Here iEH is scaled by a factor representing the ratio of the hemispheric to sector
areas (see text for additional explanation). Subsequent panels show various regressions onto
these indices. (left) Color shading shows regressions of zonal mean meridional streamfunction
c (1 3 1010 kg s21) onto iTH, iTA, and iTP; contours show winter-mean c (4 3 1010 kg s21),
with black (gray) denoting clockwise (counterclockwise) circulations. (right) Color shading
shows regressions of zonal mean eddy kinetic energy EKE (6 m2 s22) over the hemisphere onto
iEH, over the Atlantic onto iEA, and over the Pacific onto iEP. Black contours show the timemean EKE in the corresponding hemisphere or sector (30 m2 s22). Additional blue contours
(6 m2 s22) show the regression of the hemispheric zonal mean EKE onto iEA and iEP.
momentum flux field is quite noisy, making it difficult to
determine the relevant latitude in any given month. To
avoid the difficulty, we use a fixed u0 set to 408N for the
hemispheric iE, 358N for iEP, and 458N for iEA based on
the climatological, winter-mean distribution of stormtrack indicators in each sector [Hoskins and Hodges
(2002), also seen in Fig. 1]. Note that selecting u0 is more
straightforward in the Pacific, where features such as the
jet and storm track are more zonally oriented than in
the Atlantic, where these features exhibit a southwest–
northeast tilt and large intraseasonal variability.
There is an EKE signature related to iE in both sectors,
although there is no correlation between iEA and iEP
themselves (r 5 0.04), nor between other indicators of
Atlantic and Pacific storm-track intensity (Wettstein and
Wallace 2010). Further support for the independence of
eddy driving in the sectors is the fact that the sector iEs
exhibit different seasonal cycles (Fig. 1, top panel). For
1 MARCH 2012
LI AND WETTSTEIN
1591
FIG. 2. Composites (color shading) of monthly NDJFMA zonal wind anomalies based on
strong (1; top and third row) and weak (2; second and fourth row) values of the sector thermaldriving indices. Composites are for iTP in the Pacific (1208E–1108W; top and second row) and
iTA in the Atlantic (908W–408E; third and fourth row). Contours indicate the mean zonal wind
field, and color shading indicates the difference between the composite and the mean monthly
field. (left) Zonal means over the sector (black contours: 10 m s21; gray contours: 5 m s21;
shading: 1 m s21; small vertical rectangle on the x axis is the location of the near-surface wind
maximum); (right) 200-hPa maps (contours: 10 m s21 starting at 20 m s21; shading 1.5 m s21).
All composites are based on greater than 1s departures in the index and on the number of
months indicated at the top left of each map.
this reason, sector iE regressions (Fig. 1) calculated using
the sector mean EKE are shown; regressions of hemispheric zonal mean EKE onto the sector indices produce
similar patterns, but with magnitudes about 4 times weaker
(blue contours). The hemispheric picture has contributions
from both the Atlantic and Pacific [r(iEA, iEH) 5 0.47,
r(iEP, iEH) 5 0.58], and can be interpreted as an average over sectors exhibiting little interannual or intraseasonal coherence in eddy driving.
It bears mentioning that iE as defined accounts for
only the momentum flux component of the transient
eddy forcing and not the heat flux component. Hence,
this analysis focuses on the barotropic decay portion
of the standard baroclinic wave life cycle, when eddy
momentum fluxes are generally acting to accelerate the
mean zonal wind (Simmons and Hoskins 1978, 1980;
Thorncroft et al. 1993). In the upper troposphere, there
is evidence that these momentum fluxes dominate the
eddy forcing of the atmospheric flow, while the contribution from the heat fluxes is weaker and generally of
the opposite sign (Lau and Nath 1991).
Finally, we note a subtle difference in the relationship between the hemispheric and sector indices in the
thermal-driving and eddy-driving cases, most easily
seen in the seasonal cycle plots (Fig. 1, top panels).
The iTs are based on the zonal mean diabatic heating,
so the hemispheric iTH resembles the mean of the sector
indices. The iEs are instead based on area integrals of
momentum flux convergence, so the hemispheric iEH
sums over the sector indices. In the seasonal cycle plot,
1592
JOURNAL OF CLIMATE
VOLUME 25
FIG. 3. As in Fig. 2, but for the sector eddy-driving indices iEP and iEA.
iEH has thus been scaled by a factor representing the ratio
of the hemispheric to sector areas.
The remainder of the analyses will focus on sector
indices of thermal and eddy driving based on the resemblance between the hemispheric iTH and Pacific iTP
regressions, low intersector correlations for both iT and
iE, and the fact that the sector indices exhibit qualitatively different seasonal cycles.
4. Zonal wind signatures
Composites of monthly zonal wind anomalies based
on sector thermal-driving iTs are shown in Fig. 2 and on
sector eddy-driving iEs in Fig. 3.
For iT, the clearest zonal wind signatures are associated with the strong Pacific thermal-driving [iTP(1)]
composite, where the central Pacific zonal wind maximum is strengthened and extended relative to its climatological mean, and an easterly tropical anomaly
indicates a weakened Walker circulation (Fig. 2). Overall, the pattern is reminiscent of the upper-tropospheric
signature associated with El Niño conditions, the
positive phase of the Pacific–North America (PNA)
pattern, and a latitudinal shift in the Pacific storm track
(Wallace and Gutzler 1981; Hoerling and Ting 1994;
Wettstein and Wallace 2010), and can be interpreted
as a Rossby wave response to latent heat release in
areas of tropical convection (Jin and Hoskins 1995).
The weak iTP(2) composite exhibits substantially smaller
magnitudes than the iTP(1) composite, but a hint
of the Pacific jet retraction associated with La Niña
and the negative phase of the PNA pattern can be
discerned.
The zonal wind anomalies associated with iTA are
smaller than those associated with iTP and furthermore
have as much of a Pacific signature as an Atlantic signature, a sign of the pervasive (global) influence of
tropical Pacific variability (Kidson 1975; Saravanan and
Chang 2000). The lack of a strong link between extratropical and tropical variability within the Atlantic sector
itself is a consequence of weak equatorial heat sources
there during Northern Hemisphere winter.
Composites of u based on the eddy-driving indices
produce zonal wind signatures with a more consistent
1 MARCH 2012
LI AND WETTSTEIN
1593
FIG. 4. Comparison of zonal wind signatures for the driving indices (top)–(bottom) iTP, iEP,
and iEA and for common climate variability indices. Shading (51.5 m s21) shows composites of
200-hPa zonal wind anomalies for anomalously strong (left) and weak (right) values of the iT
and iE indices. Color contours (3 m s21) show similar composites based on 1-s anomalies in
common climate variability indices (PNA, WP, and NAO patterns). Black contours show the
mean zonal wind field (10 m s21 starting at 20 m s21). All composites are based on greater than
1s departures in the relevant index and the number of months in the climate variability
composites is indicated in parentheses at the top of each map. PNA, WP, and NAO indices are
defined via EOF analysis of 1000-hPa geopotential height anomalies from the seasonal cycle in
the Pacific or Atlantic sector.
interpretation in the two sectors (Fig. 3). The signature
is a dipole that straddles (more so in the Atlantic than
the Pacific) the latitude of maximum zonal wind in
the lower troposphere (see latitude–height sections),
a feature commonly used to denote the position of the
midlatitude jet (Hartmann 2007). Thus, anomalously
strong eddy momentum flux convergence in each sector
is associated with a poleward shift of the midlatitude,
eddy-driven jet.
The importance of eddy driving has been established
in the context of eddy–mean flow feedbacks related
to the NAO or annular mode patterns of climate variability (e.g., Hurrell 1995; DeWeaver and Nigam 2000;
Limpasuvan and Hartmann 2000; Lorenz and Hartmann
FIG. 5. As in Fig. 4, but for the total zonal wind field. Contour interval for both sets of
composites is 10 m s21, but the iT and iE composites (shading) begin at 10 m s21. Climate
variability index composites (white contours) begin at 20 m s21 and include a bold contour at
40 m s21.
1594
JOURNAL OF CLIMATE
2003). The existence of a similar (albeit weaker) eddy–
mean flow relationship in the North Pacific has received
less attention. However, it is consistent with the results
of Wettstein and Wallace (2010), who identify anomalies in storm-track intensity and jet position in both the
Atlantic and Pacific, and link them to the NAO and the
West Pacific (WP) patterns of variability, respectively.
An interesting extension of this work is the potential
to solidify the link between the EOF/statistical view
of jet variability and the process/dynamical view of jet
variability. Figure 4 provides a preliminary comparison
of the 200-hPa zonal wind signatures (via composites)
associated with (i) the thermal- and eddy-driving indices, and (ii) leading patterns of Northern Hemisphere
variability derived from EOF analysis of the 1000-hPa
geopotential height field. The similarity of the zonal
wind signatures reinforces the interpretation that some
patterns (e.g., PNA) are more tropically forced, while
others (e.g., NAO–WP) are more directly associated
with midlatitude eddy processes. The ‘‘more’’ must be
emphasized, especially in the case of the PNA because
while it is driven by shifts in tropical convection, eddy
forcing is important for the development and maintenance of the pattern (Franzke et al. 2011). Some discrepancies are also expected because of differences in
the proportion of intraseasonal and interannual variance in the driving indices relative to the pattern indices.
The total wind composites are also shown because they
provide a more intuitive picture of the actual fluctuations in jet extent and structure (Fig. 5).
The reanalysis-based results shown here are compatible with the idealized modeling studies described in
the introduction (Lee and Kim 2003; Eichelberger and
Hartmann 2007; Barnes and Hartmann 2011). Strong
thermal driving produces a mean state with collocated
jets, which in turn favors variability in the strength (intensity, longitudinal extent) of the jet core (Fig. 2). Strong
eddy driving produces a mean state with well separated
jets, which in turn favors variability in the form of latitudinal jet shifts (Fig. 3).
5. Concluding remarks
We examine variability in the extratropical zonal wind
field associated with indices constructed to represent
thermally driven (via tropical convection) and eddydriven (via eddy momentum flux convergence) processes. In the Atlantic sector, zonal wind variability is
mainly associated with eddy momentum flux convergence, supporting the established view that the Atlantic
jet is primarily eddy driven. In the Pacific sector, there
is evidence for zonal wind variability associated with
VOLUME 25
both driving processes, meaning that on intraseasonal to
interannual time scales, the Pacific jet should be considered both thermally driven and eddy driven. The
thermally driven Pacific signature is associated with the
intensity and extent of the Pacific jet, and resembles
the zonal wind anomalies of the Pacific–North America
(PNA) pattern, while the eddy-driven signature in both
sectors is a latitudinal jet shifting that resembles the
zonal wind anomalies of the North Atlantic Oscillation
(NAO) or West Pacific (WP) patterns.
The idea that eddy driving dominates North Atlantic
extratropical variability and that thermal driving dominates North Pacific extratropical variability is not new.
This work has, however, demonstrated a simple, direct link between the dynamical driving processes and
zonal wind variability in reanalysis data, thereby bridging a gap between the real world and the idealized flows
produced under different ‘‘driving’’ conditions in modeling studies (Lee and Kim 2003; Walker and Schneider
2006; Eichelberger and Hartmann 2007; Barnes and
Hartmann 2011). The similarity in zonal wind anomalies
associated with the process-based perspective and those
associated with well-known patterns of variability further bolsters the physical interpretations presented here.
The possibility for measuring the sensitivity of the extratropical zonal wind field to changes in the balance of
the driving processes has the potential to help evaluate
the response of midlatitude jets in a range of climatological mean states.
Acknowledgments. We wish to thank D. Battisti,
I. Bethke, N. G. Kvamstø, and I. Seierstad for many
insightful discussions and three anonymous reviewers
for their thoughtful comments on the manuscript. We
acknowledge the European Centre for Medium-Range
Weather Forecasts for providing the ERA-40 data, which
were obtained from the ECMWF data server. The authors contributed equally to this work.
APPENDIX
Tropical Signature of Alternative iTs
Alternative definitions of the thermal-driving index iT
represent very similar changes in the large-scale tropical
features of interest. Composites (based on 1s) for a selection of these alternative definitions (Fig. A1) show
consistent signatures in indicator variables of the largescale tropical circulation, as described in the figure
caption. The iT definitions shown are the heating-based
index used in the paper (Fig. A1a); area-averaged outgoing longwave radiation over 208S–108N, with a sign
change such that high values of the index indicate more
1 MARCH 2012
LI AND WETTSTEIN
1595
FIG. A1. (a)–(d) Composites of various large-scale tropical circulation indicators vs alternative definitions of the sector iTs (see text for
details). (left and third column) Latitude–height sections: black contours show meridional streamfunction (5 3 1010 kg s21), with black
(gray) denoting clockwise (counterclockwise) circulations; green contours show zonal-mean vertical velocity in sector (0.1 m s21), with
shading denoting upward velocities .0.2 m s21. (second and fourth column) Maps: color shading shows column diabatic heating
(100 W m22); green contours show upward vertical velocity averaged over 600–200 hPa (0.2 m s21); black contours show OLR ,
220 W m22 (10 W m22). Number of months in each composite is indicated at the top left of each latitude–height section.
1596
JOURNAL OF CLIMATE
deep convection (Fig. A1b); area- and mass-weighted
average upward velocities over 308S–308N, 600–200 hPa
to capture the rising branch of the Hadley cell (Fig. A1c);
and maximum northward meridional velocity over 308S–
308N, 300–100 hPa to capture the northward flow in the
upper branch of the Hadley cell (Fig. A1d). Total fields
are shown in the composites to facilitate the display of the
multiple variables in each panel.
REFERENCES
Barnes, E. A., and D. L. Hartmann, 2011: Rossby wave scales,
propagation, and the variability of eddy-driven jets. J. Atmos.
Sci., 68, 2893–2908.
Benedict, J. J., S. Lee, and S. B. Feldstein, 2004: Synoptic view of
the North Atlantic Oscillation. J. Atmos. Sci., 61, 121–144.
Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97, 163–172.
Boer, G. J., 1986: A comparison of mass and energy budgets from
two FGGE datasets and a GCM. Mon. Wea. Rev., 114, 885–902.
Caballero, R., 2008: Hadley cell bias in climate models linked to
extratropical eddy stress. Geophys. Res. Lett., 35, L18709,
doi:10.1029/2008GL035084.
DeWeaver, E., and S. Nigam, 2000: Zonal-eddy dynamics of the
North Atlantic Oscillation. J. Climate, 13, 3893–3914.
Eichelberger, S. J., and D. L. Hartmann, 2007: Zonal jet structure
and the leading mode of variability. J. Climate, 20, 5149–5163.
Franzke, C., S. Lee, and S. B. Feldstein, 2004: Is the North Atlantic
Oscillation a breaking wave? J. Atmos. Sci., 61, 145–160.
——, S. B. Feldstein, and S. Lee, 2011: Synoptic analysis of
the Pacific–North American teleconnection pattern. Quart.
J. Roy. Meteor. Soc., 137, 329–346.
Hartmann, D. L., 2007: The atmospheric general circulation and its
variability. J. Meteor. Soc. Japan, 85B, 123–143.
Held, I. M., 1975: Momentum transport by quasi-geostrophic
eddies. J. Atmos. Sci., 32, 1494–1497.
——, and A. Y. Hou, 1980: Zonal jet structure and the leading
mode of variability. J. Atmos. Sci., 37, 515–533.
Hoerling, M. P., and M. Ting, 1994: Organization of extratropical
transients during El Niño. J. Climate, 7, 745–766.
Horel, J., and J. M. Wallace, 1981: Planetary-scale atmospheric
phenomena associated with the Southern Oscillation. Mon.
Wea. Rev., 109, 813–829.
Hoskins, B. J., and K. I. Hodges, 2002: New perspectives on the
Northern Hemisphere storm tracks. J. Atmos. Sci., 59, 1041–
1061.
Hurrell, J. W., 1995: Transient eddy forcing of the rotational flow
during northern winter. J. Atmos. Sci., 52, 2286–2301.
Jin, F., and B. J. Hoskins, 1995: The direct response to tropical
heating in a baroclinic atmosphere. J. Atmos. Sci., 52, 307–319.
Kidson, J. W., 1975: Tropical eigenvector analysis and the Southern
Oscillation. Mon. Wea. Rev., 103, 187–196.
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., 286,
2589–2613.
VOLUME 25
Lee, S., and H.-K. Kim, 2003: The dynamical relationship between
subtropical and eddy-driven jets. J. Atmos. Sci., 60, 1490–
1503.
Levine, X. J., and T. Schneider, 2011: Response of Hadley circulation to climate change in an aquaplanet GCM coupled to
a simple representation of ocean heat transport. J. Atmos. Sci.,
68, 769–783.
Limpasuvan, V., and D. L. Hartmann, 2000: Wave-maintained
annular modes of climate variability. J. Climate, 13, 4414–
4429.
Livezey, R. E., M. Masutani, A. Leetmaa, H. Rui, M. Ji, and
A. Kumar, 1997: Teleconnective response of the Pacific–North
American region atmosphere to large central equatorial Pacific SST anomalies. J. Climate, 10, 1787–1820.
Lorenz, D., and D. L. Hartmann, 2003: Eddy-zonal flow feedback in the Northern Hemisphere winter. J. Climate, 16,
1212–1227.
Matthews, A. J., and G. N. Kiladis, 1999: Interactions between
ENSO, transient circulation, and tropical convection over the
Pacific. J. Climate, 12, 3062–3086.
Panetta, R. L., 1993: Zonal jets in wide baroclinically unstable regions: Persistence and scale selection. J. Atmos. Sci., 50, 2073–
2106.
Ren, X., Y. Zhang, and Y. Xiang, 2008: Connections between
wintertime jet stream variability, oceanic surface heating, and
transient eddy activity in the North Pacific. J. Geophys. Res.,
113, D21119, doi:10.1029/2007JD009464.
Rhines, P. B., 1975: Waves and turbulence on a beta-plane. J. Fluid
Mech., 69, 417–443.
Saravanan, R., and P. Chang, 2000: Interaction between tropical
Atlantic variability and El Niño–Southern Oscillation. J. Climate, 13, 2177–2194.
Simmons, A. J., and B. J. Hoskins, 1978: The life cycles of some
nonlinear baroclinic waves. J. Atmos. Sci., 35, 414–432.
——, and ——, 1980: Barotropic influences on the growth and
decay of nonlinear baroclinic waves. J. Atmos. Sci., 37, 1679–
1684.
Straus, D. M., and J. Shukla, 1997: Variations of midlatitude
transient dynamics associated with ENSO. J. Atmos. Sci., 54,
777–790.
Thorncroft, C. D., B. J. Hoskins, and M. E. McIntyre, 1993: Two
paradigms of baroclinic-wave life-cycle behaviour. Quart. J. Roy.
Meteor. Soc., 119, 17–55.
Trenberth, K. E., and A. Solomon, 1994: The global heat balance:
Heat transports in the atmosphere and ocean. Climate Dyn.,
10, 107–134.
Walker, C. C., and T. Schneider, 2006: Eddy influences on Hadley
circulations: Simulations with an idealized GCM. J. Atmos.
Sci., 63, 3333–3350.
Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the
geopotential height field during the Northern Hemisphere
winter. Mon. Wea. Rev., 109, 784–812.
Wettstein, J. J., and J. M. Wallace, 2010: Observed patterns of
month-to-month storm-track variability and their relationship
to the background flow. J. Atmos. Sci., 67, 1420–1437.
Woollings, T., B. Hoskins, M. Blackburn, and P. Berrisford, 2008:
A new Rossby wave–breaking interpretation of the North
Atlantic Oscillation. J. Atmos. Sci., 65, 609–626.