Air-sea heat exchanges characteristic to a prominent

Air-sea heat exchanges characteristic to a prominent midlatitude oceanic front in
the South Indian Ocean as simulated in a high-resolution coupled GCM
Masami Nonaka
Research Institute for Global Change, JAMSTEC, Yokohama, Japan
Hisashi Nakamura
Research Institute for Global Change, JAMSTEC, Yokohama, Japan
also Graduate school of Science, the University of Tokyo, Tokyo, Japan
Bunmei Taguchi
Nobumasa Komori
Akira Kuwano-Yoshida
Earth Simulator Center, JAMSTEC, Yokohama, Japan
Koutarou Takaya
Research Institute for Global Change, JAMSTEC, Yokohama, Japan
Submitted to the Journal of Climate
December 5, 2008
revised, June 23, 2009
Corresponding author address:
Masami Nonaka
Research Institute for Global Change,
Japan Agency for Marine-Earth Science and Technology (JAMSTEC),
3173-25 Showa-machi, Kanazawa-ku, Yokohama, 236-0001 Japan
[email protected]
-1 -
Abstract
An integration of a high-resolution coupled general circulation model whose ocean
component is eddy-permitting and thus able to reproduce sharp gradient in sea surface
temperature (SST) is analyzed to investigate air-sea heat exchanges characteristic to
midlatitude oceanic frontal zone. Our focus is placed on a prominent SST front in the
South Indian Ocean, which is collocated with the core of the Southern Hemisphere
storm-track. Time-mean distribution of sensible heat flux is characterized by a distinct
cross-frontal contrast. It is upward and downward on the warmer and cooler flanks,
respectively, of the SST front, acting to maintain the sharp gradient of surface air
temperature (SAT) that is important for preconditioning the environment for the
recurrent development of storms and thereby anchoring the storm-track. Induced by
cross-frontal advection of cold (warm) air associated with migratory atmospheric
disturbances, the surface flux is highly variable with intermittent enhancement of the
upward (downward) flux predominantly on the warmer (cooler) flank of the front.
Indeed, several intermittent events of cold (warm) air advection, whose total duration
accounts for only 21% (19%) of the entire analysis period, contribute to as much as
60% (44%) of the total amount of sensible heat flux during the analysis period on the
warmer (cooler) flank. This anti-symmetric behavior yields the sharp cross-frontal
gradient in the time-mean flux. Since the flux intensity is strongly influenced by local
magnitude of the SST–SAT difference that tends to increase with the SST gradient, the
concentration of the flux variance to the frontal zone and cross-frontal contrasts in the
mean and skewness of the flux all become stronger during the spin-up of the SST front.
Synoptically, the enhanced sensible heat flux near the SST front can restore SAT toward
the underlying SST effectively with a time scale of a day, to maintain a frontal SAT
gradient against the relaxing effect of atmospheric disturbances. The restoration effect
of the differential surface heating at the SST front, augmented by the surface latent
heating concentrated on the warm side of the front, represents a key process through
-2 -
which the atmospheric baroclinicity and ultimately the storm-track are linked to the
underlying ocean.
-3 -
1. Introduction
Impacts of midlatitude air-sea interaction on the climate system and its long-term
variability are not fully understood. Some of the numerical experiments (e.g., Latif and
Barnett 1994) and observational and/or statistical data analysis (Schneider and
Cornuelle 2005; Qiu et al. 2007) have suggested that midlatitude air-sea interaction can
amplify decadal climate variability over the North Pacific. Nevertheless, it is still under
debate how significantly the midlatitude ocean can impact the overlying atmosphere.
Summarizing the results of numerical experiments carried out previously with
atmospheric general circulation models (GCMs), Kushnir et al. (2002) concluded that
no coherent large-scale atmospheric response had been obtained to prescribed
midlatitude sea surface temperature (SST) anomalies. Rather, midlatitude SST
anomalies in general have been shown to be forced primarily by atmospheric anomalies
(Frankignoul 1985) through changes in surface heat fluxes (Cayan 1992; Alexander
1992), Ekman heat transport (Yasuda and Hanawa 1997) and entrainment at the bottom
of the oceanic mixed layer (Miller et al. 1994). Thus the midlatitude oceans are believed
to be passive to atmospheric anomalies, including those generated under the remote
influence of tropical variability (Lau 1997; Alexander et al. 2002; Deser et al. 2004),
except that heat exchanges with the adjusted underlying ocean can reduce thermal
dumping of the atmospheric anomalies (Alexander 1992; Barsugli and Battisti 1998).
In recent years, however, impacts of midlatitude oceanic frontal zones on the
atmosphere have been gaining increasing attention, as their fine structures and
variability have become apparent by recent satellite observations and high resolution
numerical models (Xie 2004; Chelton et al. 2004; Nakamura and Kazmin 2003; Nonaka
et al. 2006; Minobe et al. 2008; Nakamura et al. 2008). Satellite observations have
-4 -
shown that SST and surface wind speed are correlated positively in such midlatitude
oceanic frontal zones as those along the Kuroshio Extension Current (Nonaka and Xie
2003), the Gulf Stream (Sweet et al. 1981; Park and Cornillon 2002; Chelton et al.
2004; Xie 2004; Minobe et al. 2008), the Antarctic Circumpolar Current (ACC; O’neill
et al. 2003; Liu et al. 2007), and the Brazil Current (Tokinaga et al. 2005). The same
SST-wind relationship is also found in historical data and in situ observations (Tokinaga
et al. 2005, 2006). Unlike the situation where SST anomalies are generated in response
to basin-scale wind anomalies, the positive correlation is an indication of oceanic
thermal forcing on the planetary boundary layer (PBL), especially in regions of strong
ocean currents including midlatitude oceanic frontal zones. This is also supported by
observed positive correlation between surface heat flux and SST anomalies (i.e.,
enhanced upward heat flux over positive SST anomalies) in the western North Pacific
subpolar frontal zone (Tanimoto et al. 2003; Yasuda and Kitamura 2003; Nonaka et al.
2006, 2008).
More recent studies have further suggested that the oceanic influence is not limited to
PBL but extend into the free troposphere (Liu et al. 2007; Minobe et al. 2008).
Atmospheric GCM (AGCM) experiments by Inatsu et al. (2003) and Inatsu and Hoskins
(2004) showed that a midlatitude oceanic frontal zone, such as the one over the South
Indian and Atlantic Oceans, can anchor the storm-track and polar-front jet (PFJ; or
subpolar jet) that accompanies it. An idealized AGCM experiment by Brayshaw et al.
(2008) with zonally symmetric SST distribution showed high sensitivity of the position
and intensity of a midlatitude storm-track to the meridional SST profile and the strength
of a subtropical jet (STJ). Another idealized AGCM experiment by Nakamura et al.
(2008) with zonally symmetric SST showed that a midlatitude frontal SST gradient
-5 -
anchors a storm-track by energizing migratory cyclones and anticyclones. The resultant
increase in eddy transport of westerly momentum from the subtropics strengthens a PFJ,
and the enhanced poleward heat transport strengthens surface westerlies along the
frontal zone. In a manner consistent with the numerical experiments, collocation is
observed over the South Indian Ocean among the cores of the midlatitude SST front,
storm-track and PFJ (Nakamura and Simpo 2004; Nakamura et al. 2004). It has been
further argued (Hoskins and Valdes 1990; Nakamura et al. 2004, 2008) that the
enhanced storm-track activity may in turn act to exert positive feedback on the
midlatitude ocean through strengthening the surface westerlies to drive the currents that
accompany the oceanic frontal zone.
A midlatitude SST frontal zone can enhance local storm-track activity through
moisture supply from the ocean on its warmer flank and through tight cross-frontal
gradient of surface air temperature (SAT). The latter dry process is important from a
viewpoint of baroclinic storm development via the coupling of upper-level eddies with a
surface baroclinic zone (Hoskins et al. 1985), but developing storms act to relax the
gradient by transporting heat poleward. Without effective restoration, the SAT gradient
would be lessened after the passage of a cyclone, and the recurrent development of
atmospheric disturbances would therefore be reduced. Nakamura et al. (2004, 2008)
argued that such a relaxed SAT gradient could be restored effectively through associated
enhancement of upward and downward sensible heat fluxes at the surface on the
warmer and cooler sides of the SST front, respectively.
In this present study, we investigate synoptic and statistical characteristics of the
restoration processes that act to maintain a frontal SAT gradient across a midlatitude
SST front, paying attention to surface heat flux distributions. Through our analysis of
-6 -
output of a high-resolution coupled GCM (CGCM) that reproduces SST fronts, we
show that sensible heat flux is enhanced intermittently on either side of a frontal zone in
association with migratory atmospheric disturbances, acting effectively to adjust SAT
toward the underlying SST and thereby restore the SAT gradient at the front. Our focus
is placed on a prominent SST front that forms along the warmer flank of the ACC in the
South Indian Ocean sufficiently away from landmass.
This paper is organized as follows. Section 2 outlines the CGCM and an atmospheric
reanalysis dataset used in this study, and section 3 describes the impacts of the SST
front on distributions of surface heat flux and SAT. These results are discussed in
section 4, and section 5 provides a summary and conclusion.
2. Model and reanalysis data
a. CFES
In this study we analyze the first five years of an integration of the CGCM for the
Earth Simulator (CFES; Komori et al. 2008a, 2008b), which consists of the
Atmospheric GCM for the Earth Simulator version 2 (AFES2, Enomoto et al. 2008) and
the Coupled Ocean-Sea-Ice model for the Earth Simulator (OIFES, Komori et al. 2005).
AFES2 is an improved version of AFES (Ohfuchi et al. 2004), which is based on the
Center for Climate System Research (CCSR, University of Tokyo)/ National Institute
for Environmental Studies (NIES) AGCM version 5.4.02 (Numaguti et al. 1997) but has
been modified substantially to achieve the best computational efficiency on the
vector-parallel architecture of the Earth Simulator (ES). A land-surface model called the
Minimal Advanced Treatments of Surface Interaction and Runoff (MATSIRO, Takata et
al. 2003) is incorporated in CFES. In OIFES, a sea-ice model is coupled with the Ocean
-7 -
model for the Earth Simulator (OFES, Masumoto et al. 2004; Sasaki et al. 2008), which
is based on the Modular Ocean Model version 3 (MOM3, Pakanowski and Griffies
2000) developed at the Geophysical Fluid Dynamics Laboratory (GFDL) with
substantial modifications added. The sea-ice model is based on a model developed at the
International Arctic Research Center (IARC) of University of Alaska (Zhang and Zhang
2001). It has been improved by introducing a scheme for elastic-viscous-plastic
rheology (Hunke and Dukowicz 2002). In CFES, surface sensible and latent heat fluxes
are calculated by using the bulk formulae whose coefficients are evaluated basically
with a method by Louis (1979) in taking the difference in the roughness length between
heat and momentum fluxes into account (Uno et al. 1995).
For the particular integration used in this study, the horizontal resolution of AFES2 is
set T239 (~50 km) with 48 σ-levels (including 12 layers below σ=0.8). The top is placed
at the σ=0.003 (~0.3 hPa) level. The horizontal resolution for OIFES is set 0.25° with 54
levels whose intervals are 5 m immediately below the surface. The model maximum
depth is 6065 m. The ocean GCM is thus eddy-permitting and coupled with the
atmospheric GCM in which meso-α scale phenomena can be fully resolved. The initial
conditions for AFES2 were taken from the atmospheric fields for November 1, 1982
based on the European Center for Medium-Range Weather Forecasts (ECMWF)
reanalysis (ERA40, Uppala et al. 2005). For the initial condition for the ocean model,
the climatological temperature and salinity fields for January were taken from the World
Ocean Atlas 1998 (Antonov et al. 1998a, b, c; Boyer et al. 1998a, b, c) while no motion
was assumed. Initially, sea ice was assigned with full concentration where the initial
SST was below the freezing temperature. After integrated over two months from the
initial state, AFES2 was coupled with OIFES for its spin-up.
-8 -
For comparison with the model output, we use 6-hourly fields of the ERA40 data
provided by the ECMWF data server. We also use monthly-mean SST fields based on
measurements by the Advanced Microwave Scanning Radiometer-Earth Observing
System (AMSR-E) on the Aqua satellite. The dataset has been produced by the Remote
Sensing Systems with horizontal resolution of 0.25°.
3. Results
a. Snapshots
As apparent in snapshots of surface fields in Fig. 1, CFES can well represent
atmospheric disturbances and oceanic frontal signatures owing to its high horizontal
resolution. On 30 July of Year 1 (the first year) of the model integration, an atmospheric
low-pressure system is migrating eastward around [42°S, 65°E] with a cold southerly
wind to the west of the trailing atmospheric cold front (Fig. 1a). The advection of the
cold air onto the warmer side of the SST front results in a large SST-SAT difference and
therefore strong upward fluxes of sensible and latent heat from the ocean. As obvious in
Fig. 1b, the distribution of the SST-SAT difference exhibits pronounced local maxima
on the warmer flank of the SST front between 45°E and 60°E. The associated maxima
of sensible heat flux (SHF), which is proportional to the SST-SAT difference, are a
manifestation of the influence exerted by the frontal SST structure that is governed
dominantly by oceanic processes.
Figure 2 shows latitude-time sections of atmospheric and oceanic variables for the
55°E meridian, the longitude that corresponds to the core of the strong SST front (Fig.
1b). Frequent equatorward spills of cold airmasses associated with recurrent
development of atmospheric disturbances (Figs. 2a and 2b) yield intermittent
-9 -
augmentation of the SST-SAT difference and the corresponding enhancement of SHF
and latent heat flux (LHF), which occur predominantly on the warmer flank of the SST
front (Figs. 2b-d) as in Fig. 1. On the cooler flank of the SST front, in contrast, the
poleward advection of warm air ahead of individual cyclones yields strong downward
SHF intermittently. Thus the time-mean SHF is maximized on the warmer flank of the
SST front (Fig. 2c), where the time-mean SST-SAT difference is also largest. The
time-mean SHF deceases rapidly poleward across the SST front, and the negative SHF
is strongest on the cooler flank of the front. A similar cross-frontal decline is evident in
the time-mean LHF (Fig. 2d), although it is weakly positive even on the cooler side of
the SST front.
To confirm the importance of transient atmospheric disturbances for the time-mean
SHF, a scatter plot between the surface meridional velocity and SHF is shown in Fig. 3a
at the warmer flank of the SST front, 40˚S, 55˚E (Fig. 3b). It is apparent that SHF is
greatly enhanced in association with strong cross-frontal equatorward winds
accompanied by transient atmospheric disturbances, consistent with Fig. 2. Furthermore,
a contribution from each 5 m s -1 bin of the surface meridional wind velocity to the total
SHF observed during the entire bi-monthly period of Fig. 2 is examined (Fig. 3b). Only
about 30% of total SHF during this bi-monthly period (black curve) is accounted by
events of weak meridional wind (whose intensity is less than 5 m s-1 ) whose total
duration accounts for about 70% of the entire bi-monthly period (black bars), and about
70% of the total SHF is released during the less frequent (about 30%) strong meridional
wind. More specifically, as much as 60% of the total SHF is released during several
intermittent events in which the meridional wind velocity (v10) is equatorward (v10>0)
and stronger at least by a unit standard deviations (8.3 m s -1) than the time-mean value
- 10 -
(1.5 m s-1 poleward), whose total duration accounts only 21% of the entire period.
Similarly, at the cooler flank of the SST front (45˚S, 55˚E), intermittent (18.5%
frequency) events of strong poleward winds whose intensities exceed the time-mean
value by more than a unit standard deviation accounts for 44% of the total negative SHF
(92.4% of the net total SHF, as about half of the negative SHF is canceled as suggested
in Fig. 2c) for the entire bi-monthly period. These results confirm the primary
importance of transient atmospheric disturbance for forming the time-mean SHF field.
Associated with the passage of transient atmospheric disturbances, both the SHF and
LHF are highly variable in time, especially within the oceanic frontal zone, where their
standard deviations exhibit well-defined maxima (Figs. 2c-d). The SHF variance is
distributed rather symmetrically about the frontal axis (Fig. 2c), since fluctuations in the
cross-frontal wind component can enhance the flux on both sides of the axis. In contrast,
the distribution of the LHF variance is markedly asymmetric about the frontal axis (Fig.
2d). This is because the wind fluctuations can induce large variations in LHF only on
the warmer side of the frontal zone, where the Bowen ratio is particularly low due to the
non-linearity in the Clausius-Clapeyron equation. This asymmetric distribution indicates
that moisture supply to individual storms occurs mostly from the warmer side of the
oceanic frontal zone.
It is noteworthy that bi-monthly SAT preserves its frontal gradient across the oceanic
frontal zone despite a number of atmospheric disturbances traveling through this region
have acted to relax the SAT gradient by transporting heat poleward. This implies a
restoration possibly through air-sea heat exchanges in the oceanic frontal zone.
Nakamura et al. (2004, 2008) argued that SAT gradient relaxed by atmospheric eddy
heat transport augments the SST-SAT difference in magnitude on both sides of the
- 11 -
oceanic frontal zone and the resultant enhancement of upward and downward SHF on
the warmer and cooler sides of the frontal zone, respectively, can effectively restore the
SAT gradient towards the underling SST gradient. Nakamura et al. (2008) called this
restoration “oceanic baroclinic adjustment”. Figure 2c suggests that essentially the same
restoration process may be operative in association with individual atmospheric eddies,
since upward and downward SHF intermittently enhances predominantly on the warmer
and cooler sides of the frontal zone, respectively. The restoration of the SAT gradient
should be important for preconditioning the environment for recurrent development of
the disturbances and thereby the formation of the storm-track.
b. SAT restoration
In the following, we attempt to assess quantitatively how effectively SAT gradient
relaxed by atmospheric processes can be restored toward the underlying SST gradient
through the oceanic baroclinic adjustment. The restoration time scale can be estimated
based on the approximate heat balance in the atmospheric mixed layer (ML), in which
we assume temperature to be vertically uniform and thus equal to SAT. Although
horizontal advection is critically important as a forcing of SAT fluctuations as discussed
in the next subsection, it may be instructive to assess the efficiency of the restoration
that SHF can potentially exert on SAT by considering a hypothetical case where a given
SAT anomaly is going to be modified primarily via SHF at the ocean surface; i.e.,
∂SAT
SHF
≈
∂t
ρC p H
(1)
where ρ is air density, Cp the specific heat of air at constant pressure and H the ML
depth. The bulk formula for SHF may be expressed as
SHF = ρCH CPW ( SST − SAT ) ,
- 12 -
(2)
where W denotes the surface wind speed. In (2) we assume both ρ and the exchange
coefficient CH to be constant for simplic ity. We consider a situation where the passage
of atmospheric disturbances changes wind direction and SAT without changing wind
speed substantially within the ML. In this “stormy” situation SHF can be approximated
to be proportional to SST-SAT, and then (1) may be written as
∂S A T
α (SST − SAT )
≈
∂t
ρC p H
,
(3)
where α is a constant of proportionality. In a midlatitude oceanic frontal zone, including
the one in the South Indian Ocean, temporal variations in SST tend to be much smaller
than those in SAT owing to the deep ocean ML (with its bottom well below 100 m depth
especially in winter) and advective effects by strong ocean currents. Then (3) can be
rewritten as
∂(SAT − SST ) −α (SAT − SST )
≈
,
∂t
ρC p H
(4)
and the restoration time scale of SAT toward SST due only to SHF can thus be given by
τR ≈ ρ Cp H/α.
In Fig. 4, time series of SST, SAT, SST-SAT, and SHF on the warmer flank of the SST
front (40°S) are depicted to validate the above assumptions. Compared to ∂SAT/∂t,
∂ SST/∂t is indeed negligibly small (Fig. 4a) due to the deep oceanic ML, despite the heat
loss through LHF and SHF that spontaneously exceeds 600 W m-2 in association with
the passage of atmospheric disturbances (Figs. 2c-d). This steadiness of SST validates
the assumption made for (4). The influence of the heat fluxes on SST can be estimated
with a heat budget equation similar to (1) but for the oceanic ML with Cp = 4000 J kg-1
K-1 and ρ = 10 3 kg m-3 for seawater. For a mixed layer of 100-m depth, heat loss of 600
W m-2 can induce cooling as weak as 0.13°C day-1. In reality, this cooling is
compensated by the thermal advection by mean currents, and SST is thus almost
- 13 -
constant on the time scale of months. Therefore the SST–SAT difference varies
primarily with fluctuating SAT. On time scales of years to decades, however, SST can
vary noticeably and may become important in determining the SST-SAT difference.
A comparison between the time series of W and SHF (Fig. 4b) indicates that W
remains above 5 m s–1 most of the time, and SHF fluctuations are not much dependent
on W. In fact, their correlation coefficient is r=0.12 for the bi-monthly period. In
contrast, the SHF fluctuations are highly correlated with SST-SAT with r=0.91 (Fig. 4c).
Therefore SHF is not sensitive to W but nearly proportional to SST-SAT with their
proportionality constant estimated as α ≈ 25 W m-2 K-1. We regard the top of the
atmospheric ML as the level at which virtual potential temperature becomes 1 K higher
than at the surface. In the periods of cold-air advection, the ML top thus defined is
estimated as H ≈ 1500 m based on Fig. 4d. From these estimations with Cp = 1004 J kg-1
K-1 and ρ = 1.2 kg m-3, the restoration time scale τR is estimated as about 7.2x104 s or
slightly shorter than 1 day. The time scale could become even shorter if part of the
moisture supplied from the ocean is condensed within the ML, as discussed in the next
subsection1 . On the cooler flank of the SST front (Fig. 2c), downward SHF is enhanced
when a warm airmass is advected by northerly winds, and its variability leads to an
1
Considering crucial importance of meridional advection as a forcing of SAT anomalies (Figs.
2 and 4), it may be possible to estimate a meridional length scale of the atmospheric response by
replacing the time derivative term in (1) with a meridional advection term. If the ratio of the
meridional wind velocity V to the total wind speed W is nearly constant, the dependency of SHF
on W can be neglected, and the length scale that is insensitive to W can be estimated. From the
aforementioned estimations with CH = 1.3x10-3, the length scale is estimated as about 1500 km
in a Lagrangian sense if V/W = 1. In reality, however, the ratio V/W does vary in time (not
shown), and so should the length scale estimated. Since V and cold-air advection across the SST
frontal zone both tend to intensify behind cyclones, the ratio V/W and the length scale both tend
to be maximized, acting to reduce the meridional SAT gradient.
- 14 -
estimation of α ≈ 25 W m-2 K-1. Under the enhanced stratification over the cool ocean,
the shallowness of the ML with H as small as ~200 m (Fig. 4d) renders the restoration
even more effective with τR well shorter than a day on the cooler side of the SST front.
This effective adjustment of SAT toward the underlying SST is crucial for the
effective restoration of the SAT gradient via the oceanic baroclinic adjustment against
the relaxing effect of atmospheric disturbances, which is necessary for their recurrent
development to anchor the storm-track (Nakamura et al. 2008). In our crude estimation
presented above, however, we purposely neglected the contribution from thermal
advection by the cross-frontal wind component. As discussed in the next subsection, the
thermal advection acts against the SHF contribution to retard the restoration effect
acting on the SAT gradient.
c. Composite analysis
The snapshots shown in Fig. 1 present a typical example of how the SST–SAT
difference is modified in the vicinity of the prominent SST front by thermal advection
associated with atmospheric disturbances. To substantiate this finding based on the
snapshots, we have conducted a composite analysis for strong disturbances that travel
through the frontal zone within the bi-monthly period as depicted in Fig. 2. As a
reference index for the compositing, we used the meridional wind velocity simulated at
the surface (more precisely, 10-m wind v’10 ) at [42°S, 55°E], to which 8-day high-pass
filtering had been applied to extract the signal of migratory atmospheric disturbances.
The compositing was based on the events of maxima of the high-pass filtered v’10 (i.e.,
southerlies) that are stronger than 1.5 times of its standard deviation during the
bi-monthly period, as shown in Fig. 5a.
- 15 -
Lag composite maps in Fig. 5 show that a region of large SST–SAT difference and
strong upward SHF moves eastward and extends toward lower latitudes, following a
migratory low-pressure system and a trailing cold front. The simultaneous composite
maps in Figs. 5f-i indicate that southerly winds behind the cyclone transport a cold
air-mass equatorward across the SST front, augmenting the SST–SAT difference and
upward SHF particularly on its warmer flank, as suggested in the snapshots (Fig. 1).
The cold-air advection relaxes the SAT gradient in the oceanic frontal zone between
40°S and 45°S (Figs. 5f and 5j), but the gradient is restored within 24~36 hours (Figs.
5n and 5r). The relaxation and the subsequent restoration of the SAT gradient are also
evident in the meridional profile of SAT for 55°E plotted in the right-most column of
Fig. 5. The restoration time scale is somewhat longer than our crude estimation
presented above, because of the counteracting advective effect by the cross-frontal wind
component. In the longitudinal sector of 50°~55°E, for example, a cold front passes
through the SST frontal zone (40°~45°S) ~12 hours before the peak time of the
southerly wind. For the next 24 hours the relaxation of the SAT gradient occurs in the
frontal zone, despite the strong upward SHF acting to restore the SST–SAT difference.
The recovering of the SAT gradient appears to be completed as the cold advection
ceases.
At [42°S, 55°E], on the warmer flank of the front, the sign reversal of SHF (thin solid
line in Fig. 5a) tends to immediately follow that of the meridional wind v’ 10 . Southerly
winds bring cool SAT anomalies (dashed line in Fig. 5a), augmenting upward SHF that
damps the anomalies to zero just after winds turn into northerly. Indeed, the heat
balance in the atmospheric ML around the SST frontal zone (Fig. 6) qualitatively2
2
Unfortunately, diabatic heating within the ML is not available and its accurate estimation is quite
difficult. Also, from the available outputs, it is not possible to precisely estimate the entrainment rate
- 16 -
indicates that strong cooling (solid curve in Fig. 6a) is caused by the cool advection
(dashed curve in Fig. 6b with the reversed sign), followed immediately by warming by
SHF (solid curve in Fig. 6b). This is confirmed by asymmetric lag-correlations between
the advection and SHF: r = -0.43, -0.66, and -0.70 when SHF leads the advection term
by 6 hours, 0 hour, and lags 6 hours, respectively. To understand their phase relationship,
we consider a simple model for the heat balance in the atmospheric ML with the
restoration and the advective effects in Appendix. From the simple model, the phase lag
between v’ 10 and SAT or SHF is estimated as 0.4 day, reflecting the strong restoration
effect of SHF. Though under a crude approximation, the phase lag is roughly consistent
with the phase relationships observed in Fig. 4a. It should be noted that the advection is
not effective for the actual warming, since the advective warming tends to lag behind
the SHF warming (Fig. 6b).
The sum of SHF and the advection (dashed curve in Fig. 6a) tends to underestimate
the actual warming tendency (the ratio of variances of the former to the latter in this
bi-monthly period is 0.34, although they have a high correlation r = 0.76), which is
likely because no contributions from diabatic heating and entrainment at the top of the
ML to the heat budget in the ML are included in our evaluation. Although the
contribution of diabatic heating cannot be estimated quantitatively, time series of cloud
cover and outgoing longwave radiation (ORL; Fig. 6c) imply the formation of low-level
clouds immediately after some events of the cold-air outbreaks. For such an event
around 15-20th July, for example, cloud cover increases after cold advection with
relatively high OLR, suggesting the formation of low-level clouds. Although cool
at the top of the ML, which can modify the ML height and temperature in the ML. Then, the heat
balance analysis here cannot be quantitative. The purpose of our heat balance analysis is limited to
confirming the relation between the temperature advection and SHF. The investigation of the
accurate heat balance in the ML is beyond the scope of the present study.
- 17 -
advections are not always accompanied by low-level cloud formation, when they are
formed, the associated diabatic heating can strengthen the restoration of SAT by SHF.
4. Discussions
a. High variability in air-sea heat exchanges within the oceanic frontal zone
As indicated by its meridional profile for 55°E shown in Fig. 2c, temporal variance of
SHF exhibits a pronounced maximum within the oceanic frontal zone between 40°S and
45°S. As discussed in the preceding section, a sharp cross-frontal contrast in SHF is
important for the effective restoration of SAT toward SST. The pronounced variability
of SHF is a manifestation of the SAT gradient restoration against the relaxing effect of
atmospheric disturbances. The restoration enables the recurrent development of eddies,
which is thus critical for the formation of the storm-track. To understand why the high
variability in SHF is confined to the oceanic frontal zone, meridional profiles of the
SHF variance and its components are plotted in Fig. 7.
In bulk formula for SHF (2), we assume both ρ and the exchange coefficient CH to be
constant for simplicity. Then, the variance of SHF can be approximated as
2
2
SHF '2 ≅ ρ 2C H 2C p 2 [W '2 ∆T + W ∆T '2 + 2W ∆T W ' ∆T ']
,
(5)
where ∆T = SST-SAT and the primes denote deviations from time averages (i.e.,
anomalies). It is evident in Fig. 7a that the sharp peak of the SHF variance within the
oceanic frontal zone is mainly due to the second term in (5). The term is proportional to
the variance of ∆T, which unlike W , indeed peaks within the oceanic frontal zone (Fig.
7b). Since SST is nearly constant in time (Figs. 2 and 4), the variance of ∆T mostly
represents the SAT variability that arises mainly from thermal advection associated with
fluctuating meridional wind (v’). The storm-track that forms along the oceanic frontal
- 18 -
zone causes particularly large fluctuations in v’ along the tight mean SAT gradient,
which contributes to the confinement of the SHF variance to the frontal zone.
As shown in Fig. 7c, the time-mean surface W is stronger in midlatitudes that
in the subtropics, due to downward transfer of westerly momentum by baroclinically
growing disturbances along the storm-track (Nakamura et al. 2004). Interestingly,
within its broad midlatitude peak, the time-mean W also exhibits a weak but
well-defined local minimum over the SST frontal zone and slightly to the south. This is
a manifestation of the corresponding local maximum in thermal wind shear associated
with tight SAT gradient maintained across the SST frontal zone.
b. Sensitivity to the SST gradient intensity
The results shown above suggest that a strong SST front can localize the heat release
from the ocean. This implies that if strong SST frontal structures are not well resolved
in AGCMs and/or CGCMs due to the limited resolution of the models or SST data
assigned to the AGCMs, the heat release from the ocean would not be represented well.
To confirm the suggested influence by strong SST gradient, we examine the SHF
distribution during the first few months of the CFES integration. At that time, the ocean
component of CFES has not been fully spun up from its initial state at rest, and neither
the ocean currents nor the attendant SST front in the South Indian Ocean have been
matured (Fig. 8). This gives us a good opportunity to confirm the importance of SST
gradient for the SHF distribution without adding any artificial modifications to the
model output or performing any additional model experiments. For this purpose, we
compare the surface fields on the same calendar day (1 Feb.) between Years 1, 3 and 5
of the integration plotted in Figs. 8 and 9. We focus on the month of February, because
- 19 -
frontogenesis is fairly fast in the CFES ocean surface layer and the frontal zone with
tight SST gradient has been matured by July of Year 1 (Fig. 2). In February of Year 1,
however, the SST front is still premature (Figs. 8b and 9a) with the SST gradient not as
tight (less than 3K per latitude) as in Year 3 (Figs. 8d and 9a; more than 4K per latitude
across the frontal zone). Correspondingly, localized enhancement of the SST–SAT
difference around the SST frontal zone in Year 1 is less pronounced than in Year 3 (Figs.
8b and 8d), and so is the time-mean SHF contrast between the warmer and cooler flanks
of the frontal zone (Fig. 9b). In addition, Fig. 9c indicates that the SHF variability at
55°E in Year 1 is much less confined to the frontal zone than in Years 3 and 5, which
arises from the weaker and less confined variability in the SST–SAT difference in
February, Year 1 under the weaker SST gradient across the oceanic frontal zone around
43°S. These results suggest that frontal SST gradient is important for the localized
enhancement of SHF and thereby for the restoration of SAT gradient. In fact, SAT
gradient in February across the SST frontal zone between 20°E and 70°E in Year 3 (Fig.
8c) is apparently stronger than in Year 1 (Fig. 8a).
Figure 9d shows the corresponding meridional profiles of the SHF skewness along the
55°E meridian. In a general meteorological circumstance, skewness is significant for a
variable whose distribution is characterized by a zone of its tight meridional gradient
(Nakamura and Wallace 1991), and SAT in the vicinity of the SST front is such a
variable. Indeed, large skewness is simulated in CFES on both sides of the SST front
(Fig. 9d), and the skewness changes its sign across the front from positive on its warmer
side to negative on its cooler side. As found in Fig. 2c, this cross-frontal sign reversal is
a statistical manifestation of the tendency for the cold (warm) air advection associated
with atmospheric disturbances to induce strong upward (downward) SHF, in an
- 20 -
intermittent rather than continuous manner, mostly on the equatorward (poleward) flank
of the SST front. In February of Year 1, in contrast, the weaker SST gradient associated
with the premature oceanic front gives rise to more gradual reversal in the sign of SHF
skewness than in the later years.
c. ERA40 data
As shown in section 3b, SST fluctuates much less than SAT on time scales of months,
and the SST-SAT difference, with which surface turbulent heat fluxes are correlated
positively (Figs. 2 and 4), varies primarily with fluctuating SAT. The steadiness of SST
in oceanic frontal zones means that local air-sea heat exchanges may be reproduced
qualitatively in atmospheric models or reanalysis data where frontal signatures are
resolved to a certain degree in the SST field prescribed. As an example, snapshots on
July 18, 2002 based on the ERA-40 data are presented in Fig. 10. In association with a
low-pressure system (Fig. 10a), the SST–SAT difference is strongly positive on the
warmer side of the SST front (Fig. 10b). Compared with the CFES simulation, however,
its meridional enhancement is less pronounced due to weaker SST gradient given to the
ERA40 reanalysis (Fig. 10b; meridional SST gradient is generally less than 2K per
latitude). In fact, the corresponding AMSR-E data measured in July 2002 (Fig. 10d)
indicate that SST gradient in the frontal zone is as tight (with maximum exceeding 5K
per latitude) as in the CFES simulation and much stronger than in the ERA40 data.
One may expect that the SST–SAT difference would be stronger if the SST field used
for the reanalysis were higher in spatial resolution. Indeed, as indicated with shading in
Fig. 10d, the corresponding field of the SST–SAT difference obtained hypothetically by
replacing the SST field used in the ERA-40 reanalysis with the corresponding AMSR-E
- 21 -
data can recover the localized enhancement of the SST–SAT difference on both flanks
of the frontal zone, especially along the 55°E meridian (Fig. 10c). As shown in Fig. 11,
latitude-time sections of SHF and LHF along the meridian for the period from July to
August 2002 based on the ERA-40 data can also recover the characteristic features of
SHF and LHF found in the CFES integration with respective to their bi-monthly means
and standard deviations (Fig. 2), but again their local enhancement and meridional
contrast across the SST frontal zone are less pronounced than in the CFES simulation.
These results demonstrate that the impact of the SST front on distributions of surface
heat fluxes and their variances, which are crucial for maintaining meridional SAT
gradient via the restoration process, is found both in CFES and in the atmospheric
reanalysis fields of ERA40. In fact, time-mean meridional SAT profile in the ERA40
data exhibits a frontal structure over the SST front (Fig. 11b).
5. Summary and conclusion
In the present study, an integration of a high-resolution CGCM is analyzed to examine
an impact of a midlatitude SST front on the spatial distribution and temporal variability
of turbulent surface heat fluxes (i.e., SHF and LHF) and SAT. For this purpose, the
CGCM must be able to resolve not only atmospheric synoptic-scale disturbances but
also frontal SST gradients realistically. This requirement is met by CFES, whose
atmospheric and oceanic components have (spectral) horizontal resolution of T239 and
0.25°, respectively. The South Indian Ocean is highlighted, where a prominent SST
front forms away from landmass. This situation is ideal for extracting the impact of an
oceanic frontal zone without any noticeable influence of land-sea thermal contrasts that
may mask the impact.
- 22 -
We have shown in snapshots and latitude-time sections based on a CFES integration
that cold-air advection behind a cyclone across the SST frontal zone augments the
SST–SAT difference mainly on the warmer flank of the front to induce local
enhancement of upward SHF and LHF at the sea surface. Likewise, warm air advected
ahead of a cyclone suppresses surface evaporation and yields downward SHF on the
cooler flank of the frontal zone. Indeed, several intermittent events of cold (warm) air
advection, whose total duration accounts for only 21% (19%) of the entire analysis
period, contribute to as much as 60% (44%) of the total amount of SHF during the
analysis period on the warmer (cooler) flank. We argue that the concentration of the
time-mean and temporal variance of SHF and LHF onto the frontal zone is attributable
to the tight SST gradient that acts to maintain the SAT gradient. Induced by atmospheric
disturbances, both SHF and LHF are highly variable, and the anti-symmetric behavior
of SHF about the SST front is manifested statistically as its positive and negative
skewness on the warmer and cooler flanks of the front, respectively. With this
contrasting fluctuations, SHF after averaged in time is still upward and downward on
the respective flanks of the SST front, acting to maintain the mean SAT gradient and
thereby anchor the surface baroclinic zone along the front and the associated
storm-track.
We have shown that strong upward (downward) SHF induced on the warmer (cooler)
flank of the SST front by cold (warm) air advection due to atmospheric disturbances can
restore SAT towards the underlying SST within 1 day. With this effective restoration,
925-hPa temperature averaged over July through August of Year 1 is characterized by a
baroclinic zone with strong meridional gradient across the oceanic frontal zone over the
South Indian Ocean (Fig. 12b). As actually observed (Nakamura and Shimpo 2004), the
- 23 -
near-surface baroclinic zone that forms along the oceanic frontal zone acts to anchor a
low-level storm-track. In fact, high-pass filtered fluctuations in surface meridional wind
velocity (Fig. 12a)3 and 850-hPa poleward heat flux both associated with subweekly
disturbances (Fig. 12c) simulated in CFES are strongest along the oceanic frontal zone.
Moisture supply from the ocean that occurs predominantly on the warmer flank of the
frontal zone also contributes to the development of cyclones along it. In the model
upper troposphere (250-hPa level), the storm-track marked as variance maxima of
high-pass filtered meridional wind velocity forms along the oceanic frontal zone with its
core region shifted slightly downstream of the core of the SST front, as actually
observed (Nakamura and Shimpo 2004). A well-defined PFJ is simulated slightly
poleward of the oceanic frontal zone, again consistent with the observations. As argued
by Nakamura et al. (2004), however, the actual position of a storm-track and its activity
depend not only on “oceanic baroclinic adjustment” and moisture supply from the ocean
but also on atmospheric processes including the intensity of a subtropical jet.
The present study suggests that maintenance of SAT frontal structure by the effective
restoration via SHF can be a key process through which a storm-track is anchored along
an oceanic frontal zone whose structure and variations are determined primarily by
oceanic processes (Nonaka et al. 2006, 2008; Taguchi et al. 2007; Qiu 2000; Xie et al.
2000; Schneider and Miller 2001; Seager et al. 2001; Tomita et al. 2002; Kelly and
Dong 2004). The present study highlights the specific restoration process, what may be
called “oceanic baroclinic adjustment” (Nakamura et al. 2008; Taguchi et al. 2009). The
importance of the tight SST gradient in the frontal zone for the tight SHF gradient and
thereby for anchoring a storm-track implies that resolving SST frontal zones is
3
The high variance in surface winds can be partly due to enhanced vertical mixing over warm SST
with reduced vertical stability.
- 24 -
necessary to simulate storm-tracks realistically in AGCMs or CGCMs. The particular
importance has been addressed in a suite of AGCM experiments by Nakamura et al.
(2008). It may be interesting to examine how differently the storm-tracks and associated
air-sea heat exchanges are simulated between such a high-resolution CGCM as CFES
and its counterpart where the resolution of its ocean component is purposely lowered.
It has been established that extratropical atmospheric variability arises mainly from
internal dynamics. Although the particular CFES simulation used in the present study is
too short to investigate climate variability, our results imply that oceanic variations
might also influence the large-scale atmospheric circulation through changes in SST
fronts. The atmospheric circulation anomalies forced by the oceanic influence could
drive such oceanic variations as to induce changes in the SST fronts, forming an
extratropical air-sea coupling system. It is interesting to investigate whether interannual
and longer-term variations in oceanic frontal zones can exert any significant
modulations in the nearby storm-tracks and air-sea interactions. For this purpose, a
longer integration of CFES is now under way, which will be utilized in our future study.
Although the present study has focused on the South Indian Ocean, there are other
major oceanic frontal zones in the western portions of the Northern Pacific and Atlantic
basins, along which major storm-tracks are collocated (Nakamura et al. 2004). It is
interesting to investigate whether similar relationship can be found between the SST
front and the surface heat flux distribution in each of those regions where air-sea
interaction has been believed to exert important influence on (inter-) decadal variability
over the basin. The corresponding relationship in the frontal zone over the western
North Pacific as revealed in regional atmospheric model simulations is discussed by
Taguchi et al. (2009). Since 2007, surface buoy monitoring of SST and surface
- 25 -
meteorological variables has been conducted to the north and south of the Kuroshio
Extension front jointly by the National Ocean and Atmosphere Administration, Pacific
Marine Environmental Laboratory (Cronin et al. 2007; Bond and Cronin 2008) and the
Japan Agency for Marine-Earth Science and Technology (JAMSTEC). This monitoring
and future extension of the buoy array system in that region, if realized, will contribute
to our deeper understanding of midlatitude air-sea interactions in which highly variable
surface heat fluxes are involved.
Acknowledgments. We thank Drs. S.-P. Xie and S. Minobe for valuable discussions. The
CFES integration was carried out on the Earth Simulator in support of JAMSTEC.
AMSR-E data are produced by Remote Sensing Systems and sponsored by the NASA
Earth Science REASoN DISCOVER Project and the AMSR-E Science Team. Data are
available at www.remss.com. HN is supported in part by the Grant-in-Aid #18204044
by the Japanese Ministry of Education, Culture, Sports, Science and Technology
(MEXT) and the Global Environment Research Fund (S-5) of the Japanese Ministry of
Environment. NK and AKY are supported in part by the Grant-in-Aid #19340130 by
MEXT.
- 26 -
Appendix
To understand the phase relationship between the advective forcing and SHF, we
consider a simple model for the heat balance in the atmospheric ML with the restoration
and the advective effects. With the advective effect added to (4), the heat balance can be
expressed in a more realistic manner:
∂ (SAT − SST )
≈ −τ R−1 ( SAT − SST ) + F ,
∂t
(A1)
where F signifies the advective effect that may be approximated as –v’10 ∂ SAT /∂y,
where the overbar denotes time averaging. In recognition of the thermal advection is
associated with migratory atmospheric eddies, we assume F to be periodic with
frequency of ω . Then, (A1) may be rewritten as
∂ (SAT − SST )
≈ −τ R−1 ( SAT − SST ) + ∆F sin ωt ,
∂t
(A2)
where ∆F represents the amplitude of F. Then, as temporal variations in SST tend to be
much smaller than those in SAT (see section 3b), (A2) can be solved for SAT anomaly
(SAT’):
SAT ' ≈
∆F
Asin(ω t − θ ) ,
ω
(A3)
with A = (1 + (ω τR)–2)–1/2 and θ = tan–1(ω τR). In the lack of the restoration effect by
SHF (i.e., τR → ∞), SAT anomaly is merely forced by the periodic advection. Thus SAT
also fluctuates periodically with the amplitude of ∆F/ω (A → 1) and the phase lag of 90°.
By contrast, under the extremely effective restoration (ω τR « 1), SAT anomaly is
approximated as
SAT’ ≈ τR ∆F sin [ω (t – τR)].
(A4)
Thus, the anomaly is strongly damped and fluctuating with small phase lag behind the
advective forcing.
- 27 -
In our case, SHF exerts strong restoration effect on SAT, corresponding to τ R ≈ 0.8
day as discussed above. For the bi-monthly period from July to August, an
auto-correlation of v’10 indicates the period of the advection (2π /ω ) is about 2.5 days.
From (A3) the phase lag θ can then be estimated as 62.0° or 0.4 day, reflecting the
strong restoration effect of SHF. Though under a crude approximation, the phase lag
thus estimated is roughly consistent with the phase relationship observed in Fig. 5a
among v’10 , SHF and SAT.
- 28 -
References:
Alexander, M. A., 1992: Midlatitude Atmosphere-Ocean interaction during El Nino.
Part I: The North Pacific Ocean. J. Climate, 5, 944-958.
Antonov, J. I., S. Levitus, T. P. Boyer, M. E. Conkright, T. O’Brien, and C. Stephens,
1998a: World Ocean Atlas 1998 Vol. 1: Temperature of the Atlantic Ocean. NOAA
Atlas NESDIS 27, U.S. Government Printing Office, Washington, D.C.
Antonov, J. I., S. Levitus, T. P. Boyer, M. E. Conkright, T. O’Brien, and C. Stephens,
1998b: World Ocean Atlas 1998 Vol. 2: Temperature of the Pacific Ocean. NOAA
Atlas NESDIS 28, U.S. Government Printing Office, Washington, D.C.
Antonov, J. I., S. Levitus, T. P. Boyer, M. E. Conkright, T. O’Brien, C. Stephens, and B.
Trotsenko, 1998c: World Ocean Atlas 1998 Vol. 3: Temperature of the Indian Ocean.
NOAA Atlas NESDIS 29, U.S. Government Printing Office, Washington, D.C.
Barsugli, J. J., and D. S. Battisti, 1998: The basic effects of atmosphere-ocean thermal
coupling on midlatitude variability. J. Atmos. Sci., 55, 477-493.
Bond, N. A., and M. F. Cronin, 2008: Regional weather patterns during anomalous
air-sea fluxes at the Kuroshio Extension Observatory (KEO). J. Climate, 21,
1680-1697.
Boyer, T. P., S. Levitus, J. I. Antonov, M. E. Conkright, T. O’Brien, and C. Stephens,
1998a: World Ocean Atlas 1998 Vol. 4: Salinity of the Atlantic Ocean, NOAA Atlas
NESDIS 30. US Government Printing Office, Washington, D.C.
Boyer, T. P., S. Levitus, J. I. Antonov, M. E. Conkright, T. O’Brien, and C. Stephens,
1998b: World Ocean Atlas 1998 Vol. 5: Salinity of the Pacific Ocean, NOAA Atlas
NESDIS 31. US Government Printing Office, Washington, D.C.
Boyer, T. P., S. Levitus, J. I. Antonov, M. E. Conkright, T. O’Brien, C. Stephens, and B.
Trotsenko, 1998c: World Ocean Atlas 1998 Vol. 6: Salinity of the Indian Ocean,
NOAA Atlas NESDIS 30. US Government Printing Office, Washington, D.C.
Brayshow, D., B. J. Hoskins, and M. J. Blackburn, 2008: The stormtrack response to
idealized SST perturbations in an aquaplanet GCM. J. Atmos. Sci., in press.
Cayan, D. R., 1992: Latent and sensible heat flux anomalies over the northern oceans:
Driving the sea surface temperature. J. Phys. Oceanogr., 22, 859-881.
Chelton, D. B., M. G. Schlax, M. H. Freilich, and R. F. Milliff, 2004: Satellite radar
measurements reveal short-scale features in the wind stress field over the world
ocean. Science, 303, 978-983.
Cronin, M. F., C. Meinig, C. L. Sabine, H. Ichikawa, H. Tomita. Surface mooring
network in the Kuroshio Extension. Submitted to IEEE Systems, 2007.
Enomoto, T., A. Kuwano-Yoshida, N. Komori, and W. Ohfuchi, 2008: Description of
- 29 -
AFES 2: Improvements for high-resolution and coupled simulations, in High
Resolution Numerical Modelling of the Atmosphere and Ocean, K. Hamilton and W.
Ohfuchi (eds.), chapter 5, pp. 77-97, Springer, New York.
Hoskins, B. J., and P. J. Valdes, 1990: On the existence of storm-tracks. J. Atmos. Sci.,
47, 1854-1864.
Hoskins, B. J., M. E. McIntyre, and A. W. Robertson, 1985: On the use and significance
of isentropic potential vorticity maps. Q. J. Roy. Met. Soc., 111, 877-946.
Hunke, E. C. and J. K. Dukowicz, 2002: The elastic–viscous–plastic sea ice dynamics
model in general orthogonal curvilinear coordinates on a sphere—incorporation of
metric terms. Mon. Wea. Rev., 130, 1848–1865.
Inatsu, M., and B. J. Hoskins, 2004: The zonal asymmetry of the Southern Hemisphere
winter storm-track. J. Climate, 17, 4882-4891.
Inatsu, M., H. Mukougawa, and S.-P. Xie, 2003: Atmospheric response to zonal
variations in midlatitude SST: Transient and stationary eddies and their feedback. J.
Climate, 16, 3314-3329.
Kelly, K. A., and S. Dong, 2004: The relation of western boundary current heat transport
and storage to midlatitude ocean-atmosphere interaction. In Earth Climate: The
Ocean-Atmosphere Interaction, C. Wang, S.-P. Xie, and J.A. Carton (eds.),
Geophysical Monograph, 147, AGU, Washington D.C., 347-363.
Komori, N., W. Ohfuchi, B. Taguchi, H. Sasaki, and P. Klein, 2008a: Deep ocean
intertia-gravity waves simulated in a high-resolution global coupled
atmosphere-ocean
GCM.
Geophys.
Res.
Lett.,
35,
L04610,
doi:10.1029/2007GL032807.
Komori, N., K. Takahashi, K. Komine, T. Motoi, X. Zhang, and G. Sagawa, 2005:
Description of sea-ice component of Coupled Ocean–Sea-Ice Model for the Earth
Simulator (OIFES). J. Earth Simulator, 4, 31–45.
Komori, N., A. Kuwano-Yoshida, T. Enomoto, H. Sasaki, and W. Ohfuchi, 2008b:
High-resolution simulation of the global coupled atmospheric-ocean system:
Description and preliminary outcomes of CFES (CGCM for the Earth Simulator),
in High Resolution Numerical Modelling of the Atmosphere and Ocean, K.
Hamilton and W. Ohfuchi (eds.), chapter 14, pp. 241-260, Springer, New York.
Kushnir, Y., W. A. Robinson, I. Blade, N. M. J. Hall, S. Peng, and R. Sutton, 2002:
Atmospheric GCM response to extratropical SST anomalies: Systhesis and
evaluation. J. Climate, 15, 2233-2256.
Latif, M., and T. P. Barnett, 1994: Causes of decadal climate variability over the North
Pacific and North America. Science, 266, 634-637.
- 30 -
Liu, W. T., X. Xie, P. P. Niiler, 2007: Ocean-atmosphere interaction over Agulhas
Extension meanders. J. Climate, 20, 5784-5797.
Louis, J.-F., 1979: A parametric model of vertical eddy fluxes in the atmosphere. Bound.
Layer Meteor., 17, 187–202.
Masumoto, Y., H. Sasaki, T. Kagimoto, N. Komori, A. Ishida, Y. Sasai, T. Miyama, T.
Motoi, H. Mitsudera, K. Takahashi, H. Sakuma, and T. Yamagata, 2004: A
Fifty-Year Eddy-Resolving Simulation of the World Ocean -Preliminary Outcomes
of OFES (OGCM for the Earth Simulator)-, The Journal of the Earth Simulator, 1,
31-52.
Miller, A. J., D. R. Cayan, T. P. Barnett, N. E. Graham, and J. M. Oberhuber, 1994:
Interdecadal variability of the Pacific Ocean: model response to observed heat flux
and wind stress anomalies. Climate Dyn., 9, 287-302.
Minobe, S., A. Kuwano-Yoshida, N. Komori, S.-P. Xie, and R. J. Small, 2008: Influence
of the Gulf Stream on the troposphere. Nature, 452, 206-209,
doi:10.1038/nature06690.
Nakamura, H., and J. M. Wallace, 1991: Skewness of low-frequency fluctuations in the
tropospheric circulation during the Northern Hemisphere winter. J. Atmos. Sci., 48,
1441-1448.
Nakamura, H., and T. Sampe, 2002: Trapping of synoptic-scale disturbances into the
North-Pacific subtropical jet core in midwinter. Geophys. Res. Lett., 29,
doi:10.1029/2002GL015535.
Nakamura, H., and A. S. Kazmin, 2003: Decadal changes in the North Pacific oceanic
frontal zones as revealed in ship and satellite observations. J. Geophys. Res., 108,
3078-3094.
Nakamura, H., and A. Shimpo, 2004: Seasonal variations in the Southern Hemisphere
storm tracks and jet streams as revealed in a reanalysis dataset. J. Climate, 17,
1828-1844.
Nakamura, H., T. Sampe, Y. Tanimoto, and A. Shimpo, 2004: Observed associations
among storm tracks, jet streams and midlatitude oceanic fronts. In Earth Climate:
The Ocean-Atmosphere Interaction, C. Wang, S.-P. Xie, and J.A. Carton (eds.),
Geophysical Monograph, 147, AGU, Washington D.C., 329-345.
Nakamura, H., T. Sampe, A. Goto, W. Ohfuchi, and S.-P. Xie, 2008: On the importance
of mid-latitude oceanic frontal zones for the mean state and dominant variability in
the
tropospheric
circulation.
Geophys.
Res.
Lett., 35,
L15709,
doi:10.1029/2008GL03410.
Nonaka, M., and S.-P. Xie, 2003: Covariations of sea surface temperature and wind over
- 31 -
the Kuroshio and its extension: Evidence for ocean-to-atmosphere feedback. J.
Climate, 16, 1404-1413.
Nonaka, M., H. Nakamura, Y. Tanimoto, T. Kagimoto, and H. Sasaki, 2006: Decadal
variability in the Kuroshio-Oyashio Extension simulated in an eddy-resolving
OGCM. J. Climate, 19, 1970–1989.
Nonaka, M., H. Nakamura, Y. Tanimoto, T. Kagimoto, and H. Sasaki, 2008:
Interannual-to-decadal variability in the Oyashio Current and its influence on
temperature in the subarctic frontal zone: An eddy-resolving OGCM simulation. J.
Climate, 21, 6283-6303.
Numaguti, A., M. Takahashi, T. Nakajima, and A. Sumi, 1997: Description of
CCSR/NIES atmospheric general circulation model. Study on the Climate System
and Mass Transport by a Climate Model, A. Numaguti, S. Sugata, M. Takahashi, T.
Nakajima, and A. Sumi, eds., Center for Global Environmental Research, National
Institute for Environmental Studies, CGER’s Supercomputer Monograph Report 3,
1–48.
Ohfuchi, W., H. Nakamura, M. K. Yoshioka, T. Enomoto, K. Takaya, X. Peng, S.
Yamane, T. Nishimura, Y. Kurihara, and K. Ninomiya, 2004: 10-km mesh
meso-scale resolving simulations of the global atmosphere on the Earth Simulator:
Preliminary outcomes of AFES (AGCM for the Earth Simulator). J. Earth
Simulator, 1, 8–34.
O'Neill, L.W., D.B. Chelton, and S.K. Esbensen, 2003: Observations of SST-Induced
Perturbations of the Wind Stress Field over the Southern Ocean on Seasonal
Timescales. J. Climate, 16, 2340–2354.
Onogi, K., J. Tsutsui, H. Koide, M. Sakamoto, S. Kobayashi, H. Hatsushika, T.
Matsumoto, N. Yamazaki, H. Kamahori, K. Takahashi, S. Kadokura, K. Wada, K.
Kato, R. Oyama, T. Ose, N. Mannoji and R. Taira (2007) : The JRA-25 Reanalysis.
J. Meteor. Soc. Japan, 85, 369-432.
Pacanowski R. C., and S. M. Griffies, 2000: MOM 3.0 Manual, Geophysical Fluid
Dynamics Laboratory/National Oceanic and Atmospheric Administration, 680pp.
Park, K.-A., and P. Cornillon, 2002: Stability-induced modification of sea surface winds
over Gulf Stream rings. Geophys. Res. Lett., 29, 2211, doi:10.1029/2001GL014236.
Pierce, D. W., T. P. Barnett, N. Schneider, R. Saravanan, D. Dommenget, and M. Latif,
2001: The role of ocean dynamics in producing decadal climate variability in the
North Pacific. Clim. Dyn. 18, 51-70.
Qiu, B., 2000: Interannual variability of the Kuroshio Extension system and its impact
- 32 -
on the wintertime SST field. J. Phys. Oceanogr., 30, 1486-1502.
Qiu, B., N. Schneider, and S. Chen, 2007: Coupled decadal variability in the North
Pacific: An observationally constrained idealized model. J. Climate, 20, 3602-3620.
Sasaki, H., M. Nonaka, Y. Masumoto, Y. Sasai, H. Uehara, and H. Sakuma, 2008: An
eddy-resolving hindcast simulation of the quasi-global ocean from 1950 to 2003 on
the Earth Simulator, in High resolution numerical modelling of the atmosphere and
ocean, edited by K. Hamilton and W. Ohfuchi, chapter 10, pp. 157-185, Springer,
New York.
Schneider, N., and A. J. Miller, 2001: Predicting western North Pacific oceano climate.
J. Climate, 14, 3997-4002.
Schneider, N., and B.D. Cornuelle, 2005: The forcing of the Pacific Decadal Oscillation.
J. Climate, 18, 4355-4373.
Seager R., Y. Kushnir, N. H. Naik, M. A. Cane, and J. Miller, 2001: Wind-driven shifts
in the latitude of the Kuroshio-Oyashio extension and generation of SST anomalies
on decadal timescales. J. Climate, 14, 4149-4165.
Sweet W. R., R. Fett, J. Kerling, and P. LaViolette, 1981: Air–sea interaction effects in
the lower troposphere across the north wall of the Gulf Stream. Mon. Wea. Rev., 109,
1042–1052.
Taguchi, B., S.-P. Xie, N. Schneider, M. Nonaka, H. Sasaki, and Y. Sasai, 2007: Decadal
variability of the Kuroshio Extension: Observations and an eddy-resolving model
hindcast. J. Climate, 20, 2357-2377.
Taguchi, B., H. Nakamura, M. Nonaka, and S.-P. Xie, 2009: Influences of the
Kuroshio/Oyashio Extensions on air-sea heat exchanges and storm track activity as
revealed in regional atmospheric model simulations for the 2003/4 cold season. J.
Climate, under revision (To be updated).
Takata, K., S. Emori, and T. Watanabe, 2003: Development of the minimal advanced
treatments of surface interaction and runoff. Global and Planetary Change, 38,
209– 222.
Tanimoto, Y., H. Nakamura, T. Kagimoto, and S. Yamane, 2003: An active role of
extratropical sea surface temperature anomalies in determining anomalous turbulent
heat flux. J. Geophys. Res., 108 (C10), 3304, doi:10.1029/2002JC001750.
Tokinaga, H., Y. Tanimoto, and S.-P. Xie, 2005: SST-induced surface wind variations
over the Brazil-Malvinas confluence: Satellite and in situ observations. J. Climate,
18, 3470-3482.
Tokinaga, H., Y. Tanimoto, M. Nonaka, B. Taguchi, T. Fukamachi, S.-P. Xie, H.
Nakamura, T. Watanabe, and I. Yasuda, 2006: Atmospheric sounding over the
- 33 -
winter Kuroshio Extension: Effect of surface stability on atmospheric boundary
layer structure. Geophys. Res. Lett., 33, L04703, doi:10.1029/2005GL025102.
Tomita, T., S.-P. Xie, and M. Nonaka, 2002: Estimates of Surface and Subsurface
Forcing for Decadal Sea Surface Temperature Variability in the Mid-Latitude North
Pacific. J. Meteor. Soc. Jpn, 80, 1289-1300.
Uno, I., X-M. Cai, D.G. Steyn, and S. Emori, 1995: A simple extension of the Louis
method for rough surface layer modelling. Bound.-Layer Meteor., 76, 395–409.
Uppala, S. M., P. W. K˚allberg, A. J. Simmons, U. Andrae, V. da Costa Bechtold, M.
Fiorino, J. K. Gibson, H. J., A. Hernandez, G. A. Kelly, X. Li, K. Onogi, S. Saarinen,
N. Sokka, R. P. Allan, E. Andersson, K. Arpe, M. A. Balmaseda, A. C. M. Beljaars,
L. van de Berg, J. Bidlot, N. Bormann, S. Caires, F. Chevallier, A. Dethof, M.
Dragosavac, M. Fisher, M. Fuentes, S. Hagemann, E. H´olm, B. J. Hoskins, L.
Isaksen, P. A. E. M. Janssen, R. Jenne, A. P. McNally, J.-F. Mahfouf, J.-J. Morcrette,
N. A. Rayner, R. W. Saunders, P. Simon, A. Sterl, K. E. Trenberth, A. Untch, D.
Vasiljevic, P. Viterbo, and J. Woollen, 2005: The ERA-40 re-analysis. Q. J. R.
Meteorol. Soc., 131, 2961–3012.
Xie, S.-P., 2004: Satellite observations of cool ocean-atmosphere interaction. Bull. Amer.
Meteor. Soc., 85, 195-208.
Xie, S.-P., T. Kunutani, A. Kubokawa, M. Nonaka, and S. Hosoda, 2000: Intedecadal
thermocline variability in the North Pacific for 1958-1997: A GCM simulation. J.
Phys. Oceanogr., 30, 2798-2813.
Yasuda, T., and K. Hanawa, 1997: Decadal changes in the mode waters in the
midlatitude North Pacific. J. Phys. Oceanogr., 27, 858-870.
Yasuda, T., and Y. Kitamura, 2003: Long-term variability of North Pacific subtropical
mode water in response to spin-up of the subtropical gyre. J. Oceanogr., 59,
279-290.
Zhang, X. and J. Zhang, 2001: Heat and freshwater budgets and pathway in the Arctic
Mediterranean in a coupled ocean/sea-ice model. J. Oceanogr., 57, 207–234.
- 34 -
Figure captions:
Figure 1. (a) CFES-simulated snapshot fields (30 July, Year 1) of SAT (contoured), SLP
(hPa; shaded as indicated at the bottom), and surface (10-m) wind (arrows with scaling
(m s -1) at the bottom). (b) As in (a) but for SST (contoured) and SST-SAT (°C; shaded as
indicated at the bottom. White (black) contours are added for highlighting the positive
(negative) values. In (a-b), contour intervals for SAT and SST are 1°C.
Figure 2. (left) Latitude-time sections of CFES simulated fields for 55°E in July through
August of Year 1. (a) SLP (hPa; colored), surface wind vectors (with scaling (m s-1) at
the top), (b) SAT (color contours with intervals of 1°C) and SST-SAT (°C; color
shading), (c) SHF and (d) LHF (W m-2 ; colored). Black contours in panels (b), (c), and
(d) indicate SST with intervals of 2°C. Color scales are indicated on the right of each
panel. (right) Meridional profiles for 55°E. (a) Time-mean SLP (red dotted; upper axis)
and standard deviation of surface meridional wind velocity (blue solid; lower axis),
superimposed on mean surface wind vectors (arrows). (b)Time-mean SST (black dotted;
lower axis) and SAT (red dotted; lower axis), superimposed on standard deviation of
SST-SAT (blue solid; lower axis). (c) Time mean (red dotted; upper axis) and standard
deviation (red solid; upper axis) of SHF, superimposed on time mean SST (black dotted;
lower axis). (d) As in (c) but for LHF (red lines; upper axis). Note that the standard
deviations have been doubled for rescaling for the plots.
Figure 3. (a) Scatter plot between the northward (equatorward) surface (10-m) wind
velocity, v10 (horizontal axis) and SHF (vertical axis). (b) Histograms for v10 frequency
(black bars; left axis) and the total SHF during the bi-monthly period (W m-2 ; white
bars; right axis) for each bin of v10 with 5 m s-1 intervals. The accumulative fraction (%)
- 35 -
to the total SHF is superimposed (black curve; left axis). All plots are based on 6-hourly
CFES output from July through August in Year 1 at [40˚S, 55˚E] on the warm flank of
the SST front.
Figure 4. Six-hourly time series for (a) SAT (°C; solid curve) and SST (°C; dashed
curve), (b) SHF (W m-2 ; solid curve; left axis) and surface (10-m) wind speed (m s-1 ;
dashed curve; right axis), and (c) SHF (W m-2; solid curve; left axis) and SST-SAT (°C;
dashed curve; right axis). (d) Pressure (hPa)-time section of the atmospheric ML depth
(black curve). For reference, thin grey lines are superimposed for altitudes of 200, 1000,
and 2000 (m). The top of ML is defined instantaneously as the level at which virtual
potential temperature is higher than at the surface by 1°C. All variables are simulated in
CFES from July through August in Year 1 at [40°S, 55°E] on the warm flank of the SST
front. In each of (b) and (c), correlation coefficient between the plotted time series is
indicated at the lower-left corner.
Figure 5. (a) 8-day high-pass filtered time series of surface meridional wind velocity
(v’10 ) standardized by its standard deviation (thick solid curve, left axis), SAT (dashed
curve, right axis in °C) and SHF (thin solid curve, right most axis in W m-2). All
sampled at [42°S, 55°E], just around the axis of the SST front (indicated with dots in the
left column). (b-u) Composite maps for (b, f, j, n, r) SAT, (c, g, k, o, s) SST and
SST-SAT (°C; shaded as indicated at the bottom), (d, h, l, p, t) SST and SHF (W m-2 ;
shaded as indicated at the bottom), and (e, i, m, q, u) meridional profiles of SAT (°C;
solid curve) and SST (°C; dashed curve) at 55°E. Vectors indicate surface wind. Lags of
-12 (b-e), 0 (simultaneous, f-i), 12 (j-m), 24 (n-q), 36 (r-u) hours have been imposed for
the compositing on the basis of wind maxima v’10 (> 1.5) as marked with open circles in
- 36 -
(a). Contours intervals for SST and SAT are 1°C. For shaded variables, white (black)
contours are added for highlighting their positive (negative) values.
Figure 6. Time series of (a) local tendency of vertically integrated temperature within
ML (solid curve), (b) SHF (solid curve) and horizontal temperature advection within
ML (dashed curve, sign reversed), (c) OLR (W m-2 ; thin black curve; left axis) and
cloud cover (thick grey curve; right axis; omitted for OLR<180 W m-2 ). Units are m °C
s -1for (a) and (b). In (a) the sum of the two contributions plotted in (b) is superimposed
with dashed curve. In (c) the horizontal temperature advection within ML plotted in (b)
is added for easy comparison (m °C s-1; dashed curve; right-most axis). The top of ML
is defined instantaneously as the level at which virtual potential temperature is higher
than at the surface by 1°C. All variables are simulated in CFES from July through
August in Year 1 at [40°S, 55°E] on the warm flank of the SST front. Correlation
coefficient between the plotted time series is indicated at the lower-left corner in (a).
Figure 7. (a) Meridional profiles of the (total) variance of SHF (thick black curve), the
first (grey dashed), second (thick grey), third (thin grey) and fourth (black dashed) terms
on the r.h.s. of (5), and the sum of these four terms (thin black). Units: W2 m-4. (b) As in
(a) but for the variance of SST-SAT ( ∆T ' 2 ; °C2). (c) As in (a) but for the square of the
2
mean wind speed ( W ; m2 s-2).
Figure 8. As in Fig. 1, but for 1st February in (a-b) Year 1 and (c-d) Year 3.
Figure 9. Meridional profiles for 55°E of (a) time- mean SST (°C), (b) time-mean SHF
- 37 -
(W m-2), (c) standard deviation of SHF (W m-2) and (d) skewness of SHF, all evaluated
for the 60-day periods centered on February 15 of Year 1 (thick), Year 3 (dashed), and
Year 5 (thin solid).
Figure 10. (a-b) As in Fig. 1, but for 18 July 2002 on the basis of ERA40. (c)
Meridional profile for 55°E of SST-SAT for 18 July 2002 both based on ERA40 (thick
line) and the corresponding profile based on ERA40 SAT and AMSR-E-measured SST
(thin line). (d) As in panel (b), but for the corresponding field where SAT is based on
ERA40 but SST has been replaced with measurements by AMSR-E. Units for SST and
SAT are °C.
Figure 11. As in Fig. 2, but for July through August in 2002 on the basis of ERA40.
Figure 12. Horizontal distributions (shaded with conventions on the right of the
individual panels) based on the CFES integration for July through August in Year 1.
(a) Standard deviation of 8-day high-pass filtered surface meridional wind velocity (m
s -1), (b) time-mean meridional gradient of 925-hPa potential temperature (0.1°C/100km),
(c) 850-hPa poleward flux of potential temperature (m s -1 K) associated with 8-day
high-pass filtered fluctuations, and (d) standard deviation of 8-day high-pass filtered
250-hPa meridional wind velocity (m s-1). In each panel, time-mean SST (°C)
distribution is superimposed with contour lines (every 2°C; heavy lines for every 10°C).
In (d) time-mean 250-hPa zonal wind velocity is also superimposed with smoothed
contour lines (every 5 m s-1; heavy lines for every 10 m s-1). In (b), shading is omitted
where the 925-hPa surface is placed within the surface ML.
- 38 -
- 39 -
Figure 1. (a) CFES-simulated snapshot fields (30 July, Year 1) of SAT (contoured),
SLP (hPa; shaded as indicated at the bottom), and surface (10-m) wind (arrows with
scaling (m s-1) at the bottom). (b) As in (a) but for SST (contoured) and SST-SAT
(°C; shaded as indicated at the bottom. White (black) contours are added for
highlighting the positive (negative) values. In (a-b), contour intervals for SAT and
SST are 1°C.
-1 -
Figure 2. (left) Latitude-time sections of CFES simulated fields for 55°E in July
through August of Year 1. (a) SLP (hPa; colored), surface wind vectors (with scaling
(m s -1) at the top), (b) SAT (color contours with intervals of 1°C) and SST-SAT (°C;
color shading), (c) SHF and (d) LHF (W m-2 ; colored). Black contours in panels (b),
(c), and (d) indicate SST with intervals of 2°C. Color scales are indicated on the right
of each panel. (right) Meridional profiles for 55°E. (a) Time-mean SLP (red dotted;
upper axis) and standard deviation of surface meridional wind velocity (blue solid;
lower axis), superimposed on mean surface wind vectors (arrows). (b)Time-mean SST
(black dotted; lower axis) and SAT (red dotted; lower axis), superimposed on standard
deviation of SST-SAT (blue solid; lower axis). (c) Time mean (red dotted; upper axis)
and standard deviation (red solid; upper axis) of SHF, superimposed on time mean SST
(black dotted; lower axis). (d) As in (c) but for LHF (red lines; upper axis). Note that
the standard deviations have been doubled for rescaling for the plots.
-2 -
Figure 3. (a) Scatter plot between the northward (equatorward) surface (10-m) wind
velocity, v10 (horizontal axis) and SHF (vertical axis). (b) Histograms for v10
frequency (black bars; left axis) and the total SHF during the bi-monthly period (W
m-2; white bars; right axis) for each bin of v 10 with 5 m s-1 intervals. The
accumulative fraction (%) to the total SHF is superimposed (black curve; left axis).
All plots are based on 6-hourly CFES output from July through August in Year 1 at
[40˚S, 55˚E] on the warm flank of the SST front.
-3 -
Figure 4. Six-hourly time series for (a) SAT (°C; solid curve) and SST (°C; dashed
curve), (b) SHF (W m-2 ; solid curve; left axis) and surface (10-m) wind speed (m s-1;
dashed curve; right axis), and (c) SHF (W m-2 ; solid curve; left axis) and SST-SAT (°C;
dashed curve; right axis). (d) Pressure (hPa)-time section of the atmospheric ML depth
(black curve). For reference, thin grey lines are superimposed for altitudes of 200,
1000, and 2000 (m). The top of ML is defined instantaneously as the level at which
virtual potential temperature is higher than at the surface by 1°C. All variables are
simulated in CFES from July through August in Year 1 at [40°S, 55°E] on the warm
flank of the SST front. In each of (b) and (c), correlation coefficient between the
plotted time series is indicated at the lower-left corner.
-4 -
Figure 5. (a) 8-day high-pass filtered time series of surface meridional wind velocity
(v’10) standardized by its standard deviation (thick solid curve, left axis), SAT (dashed
curve, right axis in °C) and SHF (thin solid curve, right most axis in W m-2). All
sampled at [42°S, 55°E], just around the axis of the SST front (indicated with dots in
the left column). (b-u) Composite maps for (b, f, j, n, r) SAT, (c, g, k, o, s) SST and
SST-SAT (°C; shaded as indicated at the bottom), (d, h, l, p, t) SST and SHF (W m-2 ;
shaded as indicated at the bottom), and (e, i, m, q, u) meridional profiles of SAT (°C;
solid curve) and SST (°C; dashed curve) at 55°E. Vectors indicate surface wind. Lags
of -12 (b-e), 0 (simultaneous, f-i), 12 (j-m), 24 (n-q), 36 (r-u) hours have been imposed
for the compositing on the basis of wind maxima v’10 (> 1.5) as marked with open
circles in (a). Contours intervals for SST and SAT are 1°C. For shaded variables, white
(black) contours are added for highlighting their positive (negative) values.
-5 -
Figure 6. Time series of (a) local tendency of vertically integrated temperature within
ML (solid curve), (b) SHF (solid curve) and horizontal temperature advection within
ML (dashed curve, sign reversed), (c) OLR (W m-2 ; thin black curve; left axis) and
cloud cover (thick grey curve; right axis; omitted for OLR<180 W m-2). Units are m°C
s -1 for (a) and (b). In (a) the sum of the two contributions plotted in (b) is superimposed
with dashed curve. In (c) the horizontal temperature advection within ML plotted in (b)
is added for easy comparison (m°C s -1 ; dashed curve; right-most axis). The top of ML
is defined instantaneously as the level at which virtual potential temperature is higher
than at the surface by 1°C. All variables are simulated in CFES from July through
August in Year 1 at [40°S, 55°E] on the warm flank of the SST front. Correlation
coefficient between the plotted time series is indicated at the lower-left corner in (a).
-6 -
Figure 7. (a) Meridional profiles of the (total) variance of SHF (thick black curve),
the first (grey dashed), second (thick grey), third (thin grey) and fourth (black
dashed) terms on the r.h.s. of (5), and the sum of these four terms (thin black). Units:
W2 m-4. (b) As in (a) but for the variance of SST-SAT ( ∆T ' 2 ;°C2 ). (c) As in (a) but
2
for the square of the mean wind speed ( W ; m2 s-2).
-7 -
Figure 8. As in Fig. 1, but for 1st February in (a-b) Year 1 and (c-d) Year 3.
-8 -
Figure 9. Meridional profiles for 55°E of (a) time- mean SST (°C), (b) time-mean
SHF (W m-2 ), (c) standard deviation of SHF (W m-2) and (d) skewness of SHF, all
evaluated for the 60-day periods centered on February 15 of Year 1 (thick), Year 3
(dashed), and Year 5 (thin solid).
-9 -
Figure 10. (a-b) As in Fig. 1, but for 18 July 2002 on the basis of ERA40. (c)
Meridional profile for 55°E of SST-SAT for 18 July 2002 both based on ERA40
(thick line) and the corresponding profile based on ERA40 SAT and
AMSR-E-measured SST (thin line). (d) As in panel (b), but for the corresponding
field where SAT is based on ERA40 but SST has been replaced with measurements
by AMSR-E. Units for SST and SAT are °C.
- 10 -
Figure 11. As in Fig. 2, but for July through August in 2002 on the basis of ERA40.
- 11 -
Figure 12. Horizontal distributions (shaded with conventions on the right of the
individual panels) based on the CFES integration for July through August in Year 1.
(a) Standard deviation of 8-day high-pass filtered surface meridional wind velocity
(m s-1), (b) time-mean meridional gradient of 925-hPa potential temperature
(0.1°C/100km), (c) 850-hPa poleward flux of potential temperature (m s-1 K)
associated with 8-day high-pass filtered fluctuations, and (d) standard deviation of
8-day high-pass filtered 250-hPa meridional wind velocity (m s-1). In each panel,
time-mean SST (°C) distribution is superimposed with contour lines (every 2°C;
heavy lines for every 10°C). In (d) time-mean 250-hPa zonal wind velocity is also
superimposed with smoothed contour lines (every 5 m s -1 ; heavy lines for every 10
m s-1). In (b), shading is omitted where the 925-hPa surface is placed within the
surface ML.
- 12 -