ArticlePDF Available

An Advective-Reflective Conceptual Model for the Oscillatory Nature of the ENSO

Authors:

Abstract

Recent findings about zonal displacements of the Pacific warm pool required a notable modification of the delayed action oscillator theory, the current leading theory for the El Niño–Southern Oscillation (ENSO). Simulations with a linearized coupled ocean-atmosphere model resulted in 3- to 6-year ENSO-like oscillations, with many of the variable model parameters found to be very close to their observed values. This simple model suggests that ocean processes that are ignored or underestimated in the delayed action oscillator theory, such as zonal current convergence, zonal advection of sea surface temperature, and equatorial wave reflection from the eastern ocean boundary, are fundamental to the development of the ENSO, in particular to its manifestations in the central equatorial Pacific.
An
Advective-Reflective Con,ceptual Model for
the Oscillatory Nature of the ENSO
J.
Picaut,"
F.
Masia,
Y.
du
Penhoatf
Recent findings about zonal displacements of the Pacific warm
pool
required a notable
modification
of
the delayed action oscillator theory, the current leading theory for the El
Niño-Southern Oscillation (ENSO). Simulations with a linearized coupled ocean-atmo-
sphere model resulted in
3-
to 6-year ENSO-like oscillations, with many of the variable
model parameters found to be very close to their observed values. This simple model
suggests that ocean processes that are ignored
or
underestimated in the delayed action
oscillator theory, such as zonal current convergence, zonal advection
of
sea surface
temperature, and equatorial wave reflection from the eastern ocean boundary, are
fundamental to the development of the ENSO, in particular to its manifestations in the
central equatorial Pacific.
Earth climate variations on interannual
time scales are dominated
by
a coupled
ocean-atmosphere interaction in the Pa-
cific. This interaction connects a large-
scale oceanic sea surface temperature
(SST)
anomaly of the tropical Pacific
(El
Niño) to the large-scale atmospheric
Southern Oscillation, which is character-
ized
by
a sea-level pressure seesaw between
French Polynesia and north Australia [de-
fined
by
the Southern Oscillation Index
(SOI)].
This coupled phenomenon,
named the ENSO, oscillates irregularly
Groupe SURTROPAC, L'institut Français de Recherche
Scientifique pour le Développement en Coopération-
ORSTOM,
BP
A5,98848, Nouméa, New Caledonia.
To
whom correspondence should be addressed. E-mail:
picaut@noumea.orstom.nc
?Present address: Groupe de Recherche en Géodésie
Spatiale,
14
avenue Edouard Belin,
31401,
Toulouse,
France.
(roughly every
4
years) into a warm phase
and a cold phase (Fig.
1).
The warm phase,
El Niño, is characterized
by
warm
SST
and
weak easterly winds in the central and
eastern equatorial Pacific, energetic west-
erly winds in the western Pacific, and
negative
SOI;
whereas the cold phase, La
Niña, is characterized by cold
SST
and
strong easterly winds in the central and
eastern equatorial Pacific, weak westerly
winds in the western Pacific, and positive
SOI.
The gross features of the ENSO, and
some of its dramatic climatic impacts, can
be predicted
6
months to a year in advance
by dynamical coupled ocean-atmosphere
models (1-3). However, the prediction
skills of these models are still limited by
our insufficient understanding of the in-
trinsic mechanism that is responsible for
the ENSO.
Bjerknes
(4)
proposed that the ENSO
is a self-sustained system in which
SST
variations in the eastern and central equa-
torial Pacific produce wind variations,
which in turn produce
SST
changes. How-
ever, this scenario leads to a never ending
warm or cold state.
A
mechanism for the
oscillatory nature of the
ENSO
was origi-
nally proposed by McCreary
(5),
based on
the reflection of a subtropical oceanic up-
welling Rossby wave against the western
ocean boundary. Battisti, Hirst, Schopf,
and Suarez
(6-8)
proposed a concept that
was similar to McCreary's (but was better
supported
by
observations and equatorial
wave theory), known as the delayed action
oscillator, in which equatorial Rossby
waves reflected as upwelling equatorial
Kelvin waves are essential (9). Given the
9-month total travel time of the equatorial
upwelling Rossby and reflected Kelvin
waves, this concept asserts that it is the
continuous arrival of upwelling Kelvin
waves that slowly erodes the growing
SST-
wind interaction in the eastem equatorial
Pacific, finally stops it after
1
or 2 years,
and eventually turns the El Niño event
into a La Niña event.
The delayed action oscillator theory is
currently the leading theory for the
ENSO, although it has several flaws. First,
the maxima in the coupled SST-wind
stress fields, simulated
by
the different
models that led to this theory, are located
20"
to
40"
too far into the eastern equa-
torial Pacific as compared with observa-
tions (Fig.
1)
(1
O).
Second, on the basis of
mooring data all along the equatorial Pa-
cific and satellite altimetry data, several
authors have questioned the effectiveness
of the western ocean boundary as an equa-
torial wave reflector
(
1
1-1 4).
In
contrast,
it seems that equatorial Kelvin waves re-
flect quite well on the eastern boundary as
equatorial Rossby waves (14). Third, the
models that have led to this theory of the
ENSO are based on the dominant role of
thermocline displacements on
SST
in the
eastern equatorial Pacific, and they under-
estimate or misinterpret the effects of zon-
al advection (15).
As
a result, these mod-
els consider the eastern equatorial Pacific
(where ENSO-related
SST
variations are
the strongest) to be the source of the
ENSO air-sea interaction, instead of the
central equatorial Pacific (Fig.
1)
(10).
The central equatorial Pacific has been
confirmed as the source of the
ENSO
in
recent studies
(1
6-1
9), which indicated
that the central equatorial Pacific
SST
varies between 26" and
30"C,
predomi-
nantly on the
ENSO
time scale, as a result
of the strong eastward and westward dis-
placements of the eastern edge of the
western Pacific warm pool. Because the
SST
varies around the approximate 28°C
www.sciencemag.org
SCIENCE
VOL.
277
1
AUGUST
1997
663
threshold required for the maintenance of
organized atmospheric convection, the
central equatorial Pacific is at tlie origin of
the dominant mode of ENSO coupled
ocean-atmosphere variability on a global
scale (Fig.
1)
(17).
The dominance of
zonal advection in these zonal displace-
ments was demonstrated in several of
these studies. This was supported
by
the
discovery of
a
zone of convergence of wa-
ter masses into a well-defined salinity
front that moves together with the eastern
edge of the warm pool, eastward and west-
ward during the
El
Niño and La Niña
phases, respectively (Fig. 1) (19). The
oceanic convergence zone is due to the
confluence, within the equatorial wave
guide
(20),
of sporadic currents from the
west and nearly continuous currents from
the east
(19,
21
).
The zonal displacements
of the oceanic convergence zone-eastern
edge of the warm pool (hereafter referred
to as
OCEE)
are therefore predominantly
governed
by
the variations of zonal surface
A
SOI
140"E 180" 140"W
1OO"W
C
20"N
,
/
1
O"N
EQ
100s
200s
..
.
...
146"E 180'
14b"W
IOO'W
currents generated
by
the combined ef-
fects of (i) local wind forcing (Fig.
lB),
(ii) free equatorial Kelvin and first merid-
ional mode Rossby waves, and (iii) reflect-
ed equatorial waves on the western and
eastern ocean boundaries.
In
particular, it
appears from observations that the eastern
boundary reflection of a downwelling
Kelvin wave into a first meridional mode
downwelling Rossby wave may have been
responsible for the shift of the 1986-87
El
Niño into tlie
1988-89
La Niña
(18).
Therefore we propose a concept for the
oscillatory nature of the ENSO that can be
considered a significant modification of
the delayed action oscillator theory, with
an important role played
by
zonal advec-
tion and equatorial wave reflection on the
eastern ocean boundary.
This concept can be illustrated with a
simple ocean-atmosphere coupled model,
based on recent in situ and satellite obser-
vations.
A
linear wind-forced ocean mod-
el, derived from a low-frequency, long-
in
g
20 20
c
mo
O
v
-20
-20
a
-40
-40
E
130"E 180" 80"W
-
wave, approximation model
(22),
was re-
stricted to the zonal current of the first
baroclinic Kelvin and first meridional
Rossby modes
(20,
23).
The ocean model
did not simulate
SST,
and the location of
the oceanic zone of convergence was used
as a surrogate for the location of the east-
ern edge of the warm pool
(24).
As
in a
previous study (19), the location of the
OCEE was determined
by
the trajectory of
a hypothetical drifter moving with the
average zonal currents within the equato-
rial band.
In
the present study, this band
was usually set to 2"N to 2"S, and the
zonal currents were calculated as the sum
of the modeled varying currents and the
mean currents. (Fig. 2). The atmospheric
forcing and the coupling principle used in
the model were based on a simple approx-
imation of the observed
SST
and wind
A
3
$2
+
1
130"E 180" 80°W
Fig.
1.
Longitude-time distribution of the 2"N to
2"s
averaged
(A)
SST (contour
is
every
1
OC)
and
(B)
zonal pseudo wind stress (contour
is
every
20
m2
s-~).
Superimposed as a thick white line
is
the
SOI.
All
series are low-pass filtered to suppress oscillations of
periods shorter than
3
months. Dominant mode of ENSO covariability of the
(C)
SST and
(D)
pseudo
vector wind stress over the
1975-93
period. The SST contour
ís
every 0.2"C, and a sample of
5
m2
s-~
vector wind stress
Is
included for scale [adapted from
(lo)].
EQ,
equator.
zz
s-z
2
Fig.
2.
Concept of the reflective-advective cou-
pled system:
(A)
Longitude-time distribution
within
2"N to 2"s of the common OCEE of the
warm pool (thick line). Superimposed are the
schematic representations of the equatorial
Kelvin (Kup and Kd for upwelling and down-
welling, respectively) and first meridional mode
Rossby wave
(Rup
and
Rd
for upwelling and
downwelling, respectively) propagating paths
(lines with small dark arrows) and their associat-
ed zonal currents (small gray arrows), the mean
zonal converging currents, and the westerly and
easterly winds.
Also
shown
is
an example of two
drifters converging into the OCEE of the warm
pool
(79).
(6)
Longitude distribution of the mean
zonal currents
(in
centimeters per second) aver-
aged within 2"N to 2% deduced from obsetva-
tions (dashed line) and schematized for this
model (thick line).
SCIENCE
VOL.
277
*
1
AUGUST
1997 www.sciencemag.org
~ ~~
~____~
~~
interaction in the central and westem
equatorial Pacific (Fig. 1). During El
Niño, westerly winds penetrate from the
western equatorial Pacific into the central
equatorial Pacific simultaneously with the
eastward displacement of the eastern edge
of the warm pool. The pattern is reversed
during La Niña. Atmospheric model sim-
ulations (25) and Fig. 1B suggest that the
amplitude of the penetrating winds is fair-
ly
independent
of
the extent of penetra-
tion. The anomalous zonal wind stress that
forced the ocean model was therefore de-
fined as a patch of wind of constant am-
plitude whose zonal extent varied simulta-
neously with the zonal displacement of the
eastern edge of the warm pool (26). This
zonal extent was proportional to the dis-
tance between the instantaneous position
of the
OCEE
and the mean climatologic
position of the
OCEE
[hereafter called the
"midpoint" and situated around 180"
(Figs.
1
and 2) (27)].
In this model (Fig.
ZA),
at initial time
(t
=
O
month), a westerly wind patch is
applied
(28),
and it induces local zonal
currents that advect the
OCEE
toward the
east against the weaker mean zonal current.
At the same time, downwelling Kelvin
waves and upwelling Rossby waves, on each
side of the wind patch, propagate eastward
Fig.
3.
Examples
of
a 19-year simulation
with
various reflection coefficients: Ten percent reflec-
tion coefficient on the western boundary and
100% reflection coefficient on the eastern bound-
ary (curve a). Standard case,
with
100% reflection
coefficient on the western and eastern boundaries
(curve b). Ninety percent reflection coefficient on
the eastern boundary and
100%
reflection coeffi-
cient on the western boundary (curve c). The
curves are shifted
by
50"
longitude for clarity. Su-
perimposed on curve
b
as a thick gray curve
is
the
SOI
of
Fig.
1 over the 19-year period from 1975 to
1993.
and westward, respectively. The wind patch
expands eastward simultaneously with the
displacement
of
the
OCEE
and generates
local currents and equatorial waves of great-
er amplitude. As a consequence, the dis-
placement of the
OCEE
accelerates and
El
Niño enters into a growth phase. It is the
combination of two sets of zonal currents, in
opposite direction to the original wind-
forced currents, that gradually reduces the
acceleration of the eastward progression of
the
OCEE.
One set is produced
by
the
delayed arrival of the reflected equatorial
waves from both ocean boundaries (Fig.
2A); the other corresponds to the mean
zonal current that, as shown on Fig.
ZB,
increases in strength concurrently with the
eastward displacement of the
OCEE.
Even-
tually, this combination stops the
OCEE
displacement toward the east
(t
-
8
months) and finally pushes it back toward
the midpoint. Once the midpoint is crossed,
the wind shifts from westerly to easterly
(27) and
El
Niño turns into La Niña
(t
-
13
months); then La Niña turns into
El
Niño
(t
-
39
months) and the
ENSO
phas-
es repeat indefinitely.
The model includes several parameters,
and numerous sensitivity calculations were
done to determine the ranges of parameters
that result in an ENSO-like oscillation
of
Phase speed
(cm
s-1)
Midpoint
O"
Wind trapping
in
latitude Shift between wind and
SST
54
>.
/
0.30 0.35
Wind stress
am
litude
(dynes
cm-5)
Year
o
0.001
0.002
Rayleigh friction
(day")
Fig.
4.
Period (solid lines) and amplitude
(dashed lines) of the
ENSO
oscillation obtained
from the variation
of
a specific parameter,
with
all
other parameters set to the standard case.
(A)
Phase speed.
(B)
Location of the midpoint.
(C)
Wind
trapping
in
latitude.
(D)
Shift
between
wind
and SST.
(E)
Wind stress amplitude.
(F)
Rayleigh
friction.
the model. We defined a "standard case," in
which all parameters were set very close to
their mean values (29).
It
resulted (Fig. 3,
curve
b)
in
a lopsided oscillation (that is, a
longer time to go into an
El
Niño than into
a La Niña) with a +year period and an
amplitude of 25" (defined as half the dis-
tance from the crest to the trough). The
simulation resembled the observed ENSO
(Fig.
l),
which has on average a 3.8-year
period (30) and a lopsided pattern. The
sensitivity calculations are summarized on
Fig.
4,
where one parameter at a time was
changed, while the others were kept fixed
at their standard case values. The equatorial
Kelvin wave-speed range corresponded to
the first baroclinic mode and was found to
be very close to the one determined from in
situ and satellite observations (Fig.
4A)
(13,
31). The location of the midpoint
(mean
OCEE)
ranged around the observed
value of
180"
(Figs.
1
and
4B).
The latitu-
dinal trapping remained close to the ob-
served value
of
7" (Figs.
1D
and
4C).
A
close look at Fig.
1
suggests a
5"
to 15"
longitudinal shift between the eastern
edge of the warm pool and that of the
wind patch (Fig. 4D). The amplitude
of
the wind stress had a narrow range around
0.33 dynes cmd2 which fell into the
0.20
to 0.40 dynes cm-2 range of observed
ENSO wind stress anomalies (Fig.
4E)
(15).
Many ocean modelers use 2.5 year-'
for the Rayleigh friction. Comparisons
with observations (32) suggest a friction
of
6
month-' for the first baroclinic mode,
and the present sensitivity experiment im-
plied the use of friction smaller than
1.5
year-' (Fig. 4F).
A
series of tests (33)
indicated a rather limited north-south ex-
tension for the width of the equatorial
band over which the zonal currents are
averaged to displace the
OCEE
(within
1.75"N to 1.75"s and 2.25"N to 2.25"s).
A previous study
(I
8)
based
on
observa-
tions did not find a substantial change in
OCEE
displacements when the zonal cur-
rents were averaged within 2"N to
2"s
and
6"N to
6"s.
The final test considered dif-
ferent values for the mean zonal current
near the western and eastern ocean
boundaries (Fig. 2B). It appears to be dif-
ficult to get the model to oscillate for
values below
15
cm
s-'
near the westem
boundary and above
-35
cm
s-'
near the
eastern boundary. These numbers are still
too large compared with those of the ob-
servations (34).
As discussed above, an important and
controversial question about the delayed
action oscillator theory compared with the
present approach is the effectiveness of
the western and eastern ocean boundaries
as equatorial wave reflectors. Several sim-
ulations were done with our model, with
www.sciencemag.org
SCIENCE
VOL.
277
1
AUGUST
1997
665
the addition of reflection coefficients on
both boundaries.
It
is possible to get
ENSO-like oscillations with no or very
little
(<lo%)
western boundary reflec-
tion, through a reduction of the wind
stress amplitude (Fig.
3,
curve a).
In
con-
trast, it is impossible to obtain ENSO-like
oscillations with a reflection coefficient
on the eastern boundary that is smaller
than
85%
(Fig.
3,
curve c).
Despite its simplicity, the proposed cou-
pled model is adequate to illustrate the
advective-reflective concept for the oscilla-
tory nature of the
ENSO.
Many of the
parameters that yielded realistic simulations
of the ENSO were found to be surprisingly
close to their observed values. The concept
is based on the discovery of the
OCEE
that
is advected in phase with the
SOI
(I
9);
as a
consequence, the
ENSO
time scale can be
accurately determined
by
the duration of
advection of the eastem edge of the warm
pool by surface zonal currents in the equa-
torial band. This is more direct than the
estimation proposed by recent
ENSO
theo-
ries, based
on
the time taken by some neg-
ative feedback to stop the unstable growth
of the coupled system. However, our model
results in regular oscillations, and
as
seen
on
Figs.
1
and
3,
curve
b,
the ENSO is subject
to strong irregularities. This could be due to
the fact that several parameters that were
treated as constants in the model are not
really constant in nature (such as Kelvin
wave speed and wind amplitude). These
irregularities are beyond the scope of the
present study.
The present model requires stronger
than observed mean zonal converging cur-
rents (Fig.
ZB),
very likely to compensate
for the simplified model physics, such as
the exclusion of nonlinear terms and of
vertical advection for changing
SST.
Dur-
ing the fully developed El Niño, when
warm waters stretch well into the eastern
equatorial Pacific, the zonal
SST
gradient
on the eastern edge of the warm pool does
not remain nearly constant but weakens
significantly (Fig.
1A).
Zonal advection
associated with the returning displace-
ment of the
OCEE
does not bring cold
SST
into the central equatorial Pacific,
and a source of cold water is needed in the
eastern equatorial Pacific to develop a La
Niña. Heat flux contributions to
SST
variations are not very substantial within
the equatorial wave guide on an ENSO
time scale
(35);
thus, uplifting of the ther-
mocline appears to be the most likely
mechanism for this source. With the re-
turning surface westward flow, mass con-
servation along the equatorial wave guide
ensures an uplifting of the thermocline in
the eastern equatorial Pacific
(36).
In
this
case, our concept is pertinent during both
phases of the
ENSO.
On
the other hand,
the. uplifting of the thermocline can also
be accounted for
by
the original delayed
action theory through the action of re-
flected upwelling Kelvin waves issued
from the western ocean boundary.
In
this
second case, our concept is relevant during
La Niña and much of the duration of
El
Niño. Hence, as a modification
of
the
delayed action oscillator theory, we pro-
pose a concept in which equatorial wave
reflection on the eastern boundary is more
important than on the western boundary
and zonal advection is more effective
overall than vertical advection and en-
trainment for setting up the coupled
ENSO system in the right place, namely,
in the central equatorial Pacific.
REFERENCES AND NOTES
I.
D. Chen,
S.
E. febiak, A.
J.
Busalacchi, M. A. Cane,
Science
269, 1699 (1995).
2.
M. Ji,
A.
Leetmaa,V. Kousky,J. Clim.
9,3105 (1996).
3.
M.
H.
Glantz, Currents
of
Change:
El
Nino's lmpact
on Climate and Society (Cambridge Univ. Press,
Cambridge,
1996).
4.
J.
Bjerknes, Mon. Weather Rev.
97, 163
(1
969).
5.
J.
P.
McCreary, ibid.
111,
370 (1983).
6.
D.
S.
Battisti, J. Afmos. Sci. 45,
2889 (1988).
7.
-and
A.
C. Hirst, ibid.
46, 1687 (1989).
8.
P.
S.
Schopf and M. J. Suarez, ibid. 45,
549 (1988).
9.
In this concept, an initial positive SST perturbation
in the eastern equatorial Pacific produces a west-
erly wind anomaly west
of
the perturbation. This
wind anomaly locally enhances the initial SST
anomaly through a deepening
of
the thermocline
and generates equatorial upwelling Rossby waves
that propagate westward
(-0.9
m
s-l).
According
to
Bjerknes' concept, the ocean-atmosphere inter-
action gets into a growing mode and into a fully
developed EI Niño. The equatorial upwelling
Rossby waves reflect on the western ocean bound-
ary as equatorial upwelling Kelvin waves that prop-
agate eastward
(-2.7
m
s-l)
back to the region of
positive SST perturbation and counteract its
growth.
10. N.
J.
Mantua and D.
S.
Battisti,
J.
Clim.
8,
2897
(1
995).
11.
T. Delcroix, J.-P. Boulanger, F. Masia, C. Menkes, J.
12.
J.-P.
Boulanger and C. Menkes, ibid. 100,
25041
13.
W.
S.
Kessler and M.
J.
McPhaden,
J.
C/im.
8,1757
14.
J.-P.
Boulanger and
L.-L.
Fu,
J.
Geopbys. Res.
1
01,
15.
C. Perigaud and
B.
Dewitte,
J.
Clim.
9, 66
(1
996).
16.
M.
J.
McPhaden and
J.
Picaut, Science
250,
1385
17.
N.
E.
Graham and
T.
P.
Barnett, J.
Ciim.
8,
544
18.
J.
Picaut and T. Deicroix,
J.
Geopbys. Res. 100,
18393
(1
995).
19.
J.
Picaut, M. loualalen, C. Menkes, T. Deicroix, M.
J.
McPhaden, Science
274,1486 (1996).
20.
In our study, the wave guide was defined by the
equatorial band where zonal currents induced by
equatorial waves were maxima,
4"N
to
4"s
or less
(7
7).
In this band and overthe
1986-89
period, there
was almost no difference in the zonal displacements
of
the
eastern edge
of
the warm pool induced by the
total zonai surface currents and by the currents re-
stricted
to
the first baroclinic Kelvin and first meridi-
onal Rossby modes
(78).
21.
C. Frankignoul, F. Bonjean,
G.
Reverdin, J. Geo-
pbys. Res.
1
01,3629
(1
996).
22.
M. A. Cane and
R.
J.
Patton,
J.
Pbys. Oceanogr. 14,
Geopbys. Res.
99, 25093
(1
994).
(1 995).
(1995).
16361 (1996).
(1990).
(1995).
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
1853 (1 984).
The ocean model covered
130"E
to
80°W,
15"N
to
15"S, with a
0.5"
longitude by
0.125"
latitude grid.
It
was run on a 5-day time step and on an anomaly
basis. Dissipation was taken into account in the form
of Rayleigh friction. The longitudinal shape of the
mean zonal currents was approximated by a simple
analytic function, based on drifter observations
(34)
(Fig.
2s).
The observed near coincidence of the eastern edge
of
the warm pool (defined by the
29°C
isotherm) and
the oceanic zone of current convergence
(79)
was
explained by the dominance ofzonai advection in the
displacements of this edge and the presence there
of a nearly constant SST gradient (Fig. IA).
A.
E.
Gill and
E.
M. Rasmusson, Nature
305,
229
(1
983).
A
similar concept of expanding wind and zonal dis-
placement of the eastern edge of the warm pool at
the onset of an EI Niño has been previously modeled
[W.
S.
Kessler,
M.
J.
McPhaden,
K.
M. Weickmann,
J.
Geophys. Res. 100,
1061
3
(1
995)].
The size and shape of the wind stress patch were
determined from observations (Fig. ID), with merid-
ional and zonal structures approximated by two
gaussian functions. The meridional gaussian func-
tion was fixed and centered at the equator with
about
7"
of exponential decay In latitude. The band-
width of the zonal gaussian function was variable and
was set to the distance from OCEE to the midpoint.
The change from a constant westerly to a constant
easterly during the shift from EI Niño to La Niña or
vice versa (that is, when the OCEE crosses the mid-
point) was done gradually through a cosine function
over
1
O"
to
15"
longitude. Knowing that anomalous
winds due
to
SST perturbations were located west of
these perturbations (Fig.
1,
C and
D),
we shifted the
simulated eastern edge of the wind stress patch
westward by
5"
to
15"
compared with the OCEE.
The system will forget this Initiai westerly wind kiqk
once it
Is
on a perpetual ENSO oscillatory motion.
Parameters for the standard case were as foliows:
Kelvin wave speed of
2.8
m
s-l,
midpoint at
18Q",
wind trapping in latitude of
7",
westward shift be-
tween the eastern edge of the warm pool and that of
the wind patch of
IO"
longitude, wind stress ampli-
tude of
0.33
dyne cm+, Rayleigh friction of
2
year-', and total currents averaged within
2"N
to
2%
with
25
cm
s-l
and
-50
cm
s-l
for the mean
zonal currents near the western and eastern ocean
boundary, respectively (Fig.
28).
W.
H.
Quin, V. T. Neal,
S.
E. Antunez de Mayoio, J.
Geopbys. ßes.
92,14449 (1987).
T.
Delcroix,
J.
Picaut, G. Eldin, ibid.
96, 3249 (1991).
J. Picaut, C. Menkes, J.-P. Boulanger,
Y.
du Pen-
hoat, TOGA Notes
1
O,
I
1
(1993).
J. Picaut, F. Masia,
Y.
du Penhoat, data not shown.
P. Niiler et
ai.,
J.
Pbys.
Oceanogr., in press.
W.
T.
Liu, A. Zhang,
J.
K. B. Ëishop,
J.
Geopbys.
Res.
99,
12623
(1
994).
In the upper
1
O0
m of the water column and within
the wave guide, the variations of transport during
EI
Niño-La Niña events can exceed
50
sverdrups
(1
sverdrup
=
IO6
m3
s-l),
which is larger than the
mean transport of any equatorial currents
(76).
This
implies powerful discharge or recharge of water
masses in the equatorial band and therefore strong
readjustment of the equatorial thermocline together
with meridional transfer
of
water masses
[K.
Wyrtki,
ibid.
90, 7129 (1985)l.
We
thank P. Waigna for preparing ail the figures,
M.
J.
Langlade for help in the processing of several
figures, and N.
J.
Mantua and D.
S.
Battisti for their
authorization to reproduce Fig.
1,
C and
D.
The Flor-
ida State University wind stress and SST data were
provided by
J.
J.
O'Brien and
R.
W. Reynolds. Dis-
cussions and corrections on an early draft by
T.
Delcroix, P. Rual, L. M. Rothstein, G. Eldin, A.
'J.
Busalacchi, M.
J.
McPhaden, and J.
P.
McCrearyare
appreciated. Supported by ORSTOM, Programme
National d'Etudes de la Dynamique du Climat, add
Centre National d'Etudes Spatiales.
26
March
1997;
accepted
11
June
1997
666
SCIENCE
VOL.
277
1
AUGUST
1997
www.sciencemag.org
Reprint Series
1
August
1997,
Volume
277,
pp.
663-666
Pi.
SCIENCE
An Advective-Reflective Conceptual Model for
the Oscillatory Nature of the
ENS0
P
J.
Picaut,"
F.
Masia, and
Y.
du
Penhoatt
C.opyrightIO
1997
by
the American Association
for
the Advancement
of
Science
~~
... Once El Niño or La Niña events reach the mature phase, negative feedbacks work together to terminate them (Wang, 2018). Four major negative feedbacks have been proposed: (1) the reflection of Kelvin waves at the western boundary of the Pacific (Suarez and Schopf, 1988;Battisti and Hirst, 1989); (2) a discharge process resulting from Sverdrup transport (Jin, 1997); (3) wind-forced Kelvin waves at the western Pacific (Weisberg and Wang, 1997); and (4) anomalous zonal advection and wave reflection at the eastern boundary of the Pacific (Picaut et al., 1997). These four negative feedbacks are, respectively, referred to as the delayed oscillator, the recharge-discharge oscillator, the western-Pacific oscillator, and the advective-reflective oscillator (Wang, 2018). ...
... There is a consensus among the results of the most recent climate models that the ENSO itself and its global impacts will continue in the coming decades and centuries Marjani et al., 2019). Nevertheless, despite considerable progress in our understanding of physical processes and feedbacks that determine ENSO characteristics (Picaut et al., 1996;Picaut et al., 1997;Wang, 2001;Wang, 2018) and our improved understanding of the impacts of climate change on many processes contributing to ENSO variability (Collins et al., 2010), there is no clear consistency among the current climate models in terms of the response of ENSO amplitude and frequency in the future (e.g., Zheng et al., 2016;Chen et al., 2017;Beobide-Arsuaga et al., 2021). Different responses in terms of changes in the amplitude and frequency of ENSO events to global warming have been projected (e.g., Cai et al., 2014;Marjani et al., 2019;Cai et al., 2021;Callahan et al., 2021;Tang et al., 2021;Alizadeh et al., 2022), making it difficult to determine how ENSO will change under future global warming. ...
Article
The El Niño-Southern Oscillation (ENSO) significantly impacts the global climate by influencing the large-scale atmospheric circulation, underlining the great importance of understanding changes in its characteristics under future global warming. This review provides an emerging understanding of the response of ENSO to future global warming. The impact of global warming on ENSO in the future is the result of numerous complex nonlinear reinforcing or canceling influences , which together with several long-lasting biases in climate models make ENSO projections challenging and problematic. Thus, there has been little consensus on the response of ENSO to global warming among the state-of-the-art coupled general circulation models (CGCMs). The inability of climate models to simulate the mean-state changes of the tropical Pacific during the past recent decades is highlighted. This highly reduces the credibility of ENSO projections, despite the inter-model agreement on the projected trend toward more significant warming over the tropical eastern than western Pacific. Nevertheless, a few models have simulated a physically consistent response to global warming in the tropical Pacific. Realistic simulations of the mean-state changes of the tropical Pacific have also been recently achieved in several studies by correcting model biases. These recent efforts offer great promise for improving the reliability of ENSO projections. Efforts should be made to reduce biases in climate models by integrating both theory and reliable observations and through better representation of unresolved processes and coupled feedbacks.
... The Coastal (costero) El Niño type was defined in 2017 (Takahashi and Martínez, 2017), although the research that produced the definition predated the event, derived from ships' logs SST data of the 1925 and 1891 El Niños. Conceptual models of ENSO dynamics also grew in number during the period; the 'delayed oscillator' model that was prevalent during the TOGA period (McCreary Jr and Anderson, 1984;Suarez and Schopf, 1988) 10 joined by the advective-reflective model (Picaut et al., 1997), the western Pacific oscillator (Weisberg and Wang, 1997) and the unified oscillator model (Wang, 2001), although none of these models could entirely explain the observed diversity in ocean and atmosphere patterns of variability. 11 ...
Article
Full-text available
The El Niño phenomenon – and its associated phenomena El Niño Southern Oscillation (ENSO) and La Niña – have become probably the most well-known forms of natural climatic variability. El Niño forecasts underpin regional Climate Outlook Forums in many parts of the world. The declaration of El Niño conditions can unlock development aid money and El Niño events commonly receive widespread media coverage. Yet ‘El Niño’ has not always meant what it does today. The name was originally applied to an annually-occurring ocean current that affected northern Peru and Ecuador, so called because it arrived at Christmas (the Christ Child). The transition in meaning to a complex global phenomenon was related as much to commercial and geopolitical priorities as to the oceanic and atmospheric observations that underpin theories of El Niño dynamics. In this paper, I argue that scientific conceptualisations of El Niño are an example of path dependency. Badging ocean-atmosphere variability as ‘El Niño’ is unnecessary either for the advancement of science or effective disaster risk reduction; in fact, current definitions are confusing and can create problems in preparing for El Niño-related hazards, as occurred with the 2017 ‘coastal’ El Niño in Peru. This paper outlines the historical processes that led to the current conceptualisations of El Niño and suggests an alternative way of understanding ocean-atmosphere dynamics in the Pacific and beyond. It then considers the implications of this path-dependency on El Niño’s ontological politics; that is, who gets to define El Niño, and to what end.
... Indeed, the decadal variability, which results from internal dynamics, is a significant component of ENSO (Li et al., 2013). The transition between opposite phases of ENSO has been explained through conceptual models including the recharge-discharge oscillator (Jin, 1997), delayed oscillator , western Pacific oscillator (Weisberg and Wang, 1997) and advective-reflective oscillator (Picaut et al., 1997). The irregularity of ENSO can be explained if El Niño or La Niña is considered as a highly damped oscillation triggered by stochastic atmospheric/oceanic noise (Lau, 1985;Kessler, 2003;Chen et al., 2015). ...
Article
The El Niño-Southern Oscillation (ENSO) is the dominant natural climate pattern that influences the global climate on subdecadal timescales. Thus, a better understanding of the impacts of global warming on the characteristics of ENSO events offers large socioeconomic benefits. Changes in the frequency of the eastern Pacific (EP), central Pacific (CP) and total El Niño events are analyzed in the future period (2051-2100) under the two Shared Socioeconomic Pathways (SSPs) scenarios (i.e. SSP2-4.5 and SSP5-8.5) compared to the historical period (1951-2000) using outputs of eight models participating in the Coupled Model Intercomparison Project Phase 6 (CMIP6). Types of El Niño events are diagnosed based on pattern correlation coefficients (PCCs) between monthly sea surface temperature (SST) anomalies and the first two leading empirical orthogonal function (EOF) modes of SST anomalies in the tropical Pacific. Based on the ensemble of models, the number of total El Niño events decreases by 26% and 16% under the SSP2-4.5 and SSP5-8.5 scenarios, respectively, in the future compared to the historical period. A smaller decrease in the number of total El Niño events under the faster warming rate of the SSP5-8.5 scenario suggests that the response of ENSO dynamics to global warming is not linear. Analysis of the Extended Reconstructed Sea Surface Temperature, version 5 (ERSSTv5) dataset during the period 1951-2020 indicates that the ratio of the number of CP El Niño to EP El Niño has substantially increased over the past two decades. Nevertheless, under both the SSP2-4.5 and SSP5-8.5 scenarios in the future period, the ratio of the number of CP El Niño to EP El Niño does not change considerably compared to that in the historical period. This suggests that the recent increase in the frequency of CP El Niño could be a result of multidecadal variations rather than anthropogenic global warming.
... Traditional conceptual models of ENSO highlight our understanding of this phenomenon by summarizing the important mechanisms and feedbacks known at the time of their development (Suarez and Schopf 1988;Weisberg and Wang 1997;Picaut et al. 1997;Jin 1997; Summarized by Wang and Picaut 2004). One of the leading conceptual models for ENSO's apparent oscillatory behavior is Jin's (1997) recharge-discharge oscillator (RDO) theory. ...
Article
This study utilizes observations and a series of idealized experiments to explore whether eastern Pacific (EP)- and central Pacific (CP)-type El Niño–Southern Oscillation (ENSO) events produce surface wind stress responses with distinct spatial structures. We find that the meridionally broader sea surface temperatures (SSTs) during CP events lead to zonal wind stresses that are also meridionally broader than those found during EP-type events, leading to differences in the near-equatorial wind stress curl. These wind spatial structure differences create differences in the associated pre- and post-ENSO event WWV response. For instance, the meridionally narrow winds found during EP events have (i) weaker wind stresses along 5°N and 5°S, leading to weaker Ekman-induced pre-event WWV changes; and (ii) stronger near-equatorial wind stress curls that lead to a much larger post-ENSO event WWV changes than during CP events. The latter suggests that, in the framework of the recharge oscillator model, the EP events have stronger coupling between sea surface temperatures (SST) and thermocline (WWV), supporting more clearly the phase transition of ENSO events, and therefore, the oscillating nature of ENSO than CP events. The results suggest that the spatial structure of the SST pattern and the related differences in the wind stress curl, are required along with equatorial wind stress to accurately model the WWV changes during EP- and CP-type ENSO events.
... One is that the WWBs can lead to the convergence of sea water and the deepening of equatorial thermocline, thus triggering the eastward propagating equatorial Kelvin waves, resulting in the warming of the equatorial eastern Paci c Ocean (McPhaden 1999). The second is to produce strong equatorial sea surface currents, which transport warm water eastward along the equator and expand the eastern boundary of the warm pool (Picaut et al. 1997). ...
Preprint
Full-text available
Based on the criteria of oceanic Kelvin wave events in the equatorial Pacific on the pentad-monthly timescale, this study analyzed the optimal statistical relationship between westerly wind bursts (WWBs) and oceanic Kelvin waves in the equatorial Pacific through lead-lag analysis. Under the influence of sustained westerly wind anomaly (WWA) (60-day mean), it can effectively cause the positive sea temperature anomaly (60-day mean) in the upper equatorial Pacific Ocean after 20 days. However, under the influence of continuous easterly wind anomaly (EWA) (30-day mean), the negative sea temperature anomaly (30-day mean) in the upper ocean can be effectively caused after 30 days. In addition, WWBs near the international dateline (150 °E-170 °W) corresponds well to the 20°C isotherm depth anomaly in the equatorial Mid-Eastern Pacific (160 °W-100 °W). WWAs in the western Pacific (west to 170 °E) can lead to thermocline change at 150°W and stimulate the propagation of the Kelvin wave eastward to the equatorial eastern Pacific from March to August. During the mature period of El Niño–Southern Oscillation (ENSO) (September to February in the following year), Bjerknes positive feedback is formed near the international dateline with the coupling effect of WWBs, sea surface temperature (SST) and thermocline in the equatorial Eastern Pacific. The contemporaneous correlation reaches peak between the WWA in the equatorial western Pacific (near 170 °E) and the 20°C isotherm depth anomaly in the equatorial eastern Pacific (near 110 °W).
... In the past several decades, great advances have been achieved in understanding the physical nature of ENSO (e.g., Bjerknes 1969;Wyrtki 1975;Rasmusson and Carpenter 1982;Cane and Zebiak 1985;Neelin et al. 1998;Wallace et al. 1998;Timmermann et al. 2018). It is widely accepted that ENSO development and its oscillating behavior can be explained by invoking positive atmosphere-ocean feedbacks and delayed negative oceanic adjustment feedbacks (e.g., Bjerknes 1969;Wyrtki 1975;Suarez and Schopf 1988;Battisti and Hirst 1989;Jin 1997;Picaut et al. 1997;Weisberg and Wang 1997;Jin et al. 2020). ...
Article
Many previous studies have shown that El Niño exhibits a strong seasonality in its teleconnections and regional climate impacts. Aside from the seasonal synchronization of El Niño anomalous sea surface temperature (SST) itself, seasonal differences in its associated climate impacts could also stem from the local background seasonal cycle. During the El Niño developing boreal autumn (August–October) and decaying boreal spring (February–April), the El Niño–associated precipitation anomalies display remarkably different patterns over the eastern tropical Pacific despite similar SST anomaly amplitudes. This strong seasonality can be largely attributed to the seasonal cycle of the eastern tropical Pacific SST background state with the cold tongue being strongest in autumn and weakest in spring. Therefore, the El Niño–associated SST in spring is likely to exceed the convection threshold on both sides of the equator, leading to an approximately symmetric precipitation response about the equator. In contrast, this symmetric precipitation response is absent in autumn since the SST near and south of the equator remains below the threshold and pronounced eastern tropical precipitation anomalies can only be observed in the warm Northern Hemisphere. This seasonality is mainly embodied in eastern Pacific (EP) El Niño events rather than central Pacific (CP) El Niño events since the westward-shifted warm SST anomalies for the latter cannot establish an effective cooperation with the cold tongue SST annual cycle. This study has important implications for regional climate prediction by involving the different local precipitation responses over the tropical eastern Pacific.
Article
El Niño/Southern Oscillation (ENSO) mainly occurs in the tropical Pacific Ocean every a few years. But it affects the climate around the world and has a dramatic impact on the development of ecology and agriculture. The analysis and prediction of ENSO become particularly important for meteorology and disaster management. However, due to insufficient data, spring predictability barrier (SPB), and model uncertainty, traditional analysis models face challenges. To address these issues, researchers begin to apply deep learning (DL) technologies to ENSO research, exploring the impact of ENSO on the world's extreme climate changes. In recent years, deep learning-based methods have obtained impressive progress with more accurate and effective predictions of ENSO. In this paper, we summarize the attempts of DL technologies in predicting ENSO. We first introduce the properties of ENSO, followed by the architecture introduction of DL technologies and their application to ENSO. We then investigate the potential of DL technologies for ENSO prediction from various aspects, including model evaluation metrics, prediction algorithms, overcoming SPB and prediction uncertainty. Finally, we provide discussions on the future trends and challenges of using DL technologies for ENSO prediction.
Article
It is well known that the El Niño–Southern Oscillation (ENSO) is the most prominent mode of interannual climate variability in the tropics. The ENSO is a coupled, tropical ocean–atmosphere system that fluctuates on a time scale of two to seven years in the Pacific (Philander, 1990). The ENSO extremes are labeled as either a warm or cold phase, yet its amplitude varies across a continuum with essentially Gaussian statistics (Trenberth, 1997). Characterizing the warm (cold) ENSO phase is the presence of the anomalously warm (cold) sea surface temperatures (SSTs) in the eastern and/or central equatorial Pacific known as the El Niño (La Niña) event.
Preprint
Full-text available
Tropical Pacific is home to climate variability on different timescales, including El Niño Southern Oscillation (ENSO) – one of the most prominent quasi-periodic modes of variability in the Earth’s climate system. It is a coupled atmosphere-ocean mode of variability with a 2-8-year-timescale and oscillates between a warm (El Niño) and a cold (La Niña) phase. However, the dynamics of ENSO is complex, involving a variety of spatial and temporal scales as well as their interactions, which are not necessarily well understood. We use a recently developed nonlinear and nonstationary multivariate timeseries analysis tool – multivariate empirical mode decomposition (MEMD) – to revisit quasi-periodic variability within ENSO. MEMD is a powerful tool for objectively identifying (intrinsic) timescales of variability within a given system. We apply it to reanalysis and observational data as well as to climate model output (NorCPM1). Observational/reanalysis data reveal a quasi-periodic variability in the tropical Pacific on timescales ~2–4.5 years. This variability can then be related to ENSO’s recharge-discharge and simplified West-Pacific oscillator conceptual models. The latter occurs only on this timescale and is not necessarily well represented in NorCPM1. Additionally, the ~2–4.5-year variability in ENSO can be 'predicted' up to ~20 months ahead, while predicting the full ENSO amplitude remains challenging. MEMD can therefore be used for assessing climate dynamics on different timescales and for evaluating their representation in climate models.
Article
Current climate models have relatively high skills in predicting El Niño–Southern Oscillation (ENSO) phase (i.e., El Niño, neutral, and La Niña), once leaping over the spring predictability barrier. However, it is still a big challenge to realistically forecast the ENSO amplitude, for instance, whether a predicted event will be strong, moderate, or weak. Here we demonstrate that the accumulated westerly wind events (WWEs)/easterly wind surges (EWSs) and oceanic recharged/discharged states are both of importance in accurate ENSO amplitude forecasts. El Niño and La Niña events exhibit asymmetric temporal and spatial features in the atmospheric and oceanic preconditions. El Niño amplitude at the peak season is closely associated with the accumulated WWEs over the eastern equatorial Pacific from the previous December to May and the recharged state in the western equatorial Pacific during February. In contrast, the amplitude of La Niña events is sensitive to the accumulated EWSs over the equatorial western Pacific from the previous November to April and the discharged state extending from the equatorial western to central Pacific during February. Considering these asymmetric atmospheric and oceanic preconditions of El Niño and La Niña cases, a statistical model is established to accurately forecast the ENSO amplitude at its mature phase during 1982–2018, which is validated to be robust based on a 1-yr cross-validation and independent sample tests. The feasibility and the limitation of the established statistical model are also discussed by examining its practical utility.
Article
A fast, efficient numerical procedure for modeling the linear low-frequency motions on an equatorial beta plane is developed. The model is capable of simulating the seasonal and interannual variability in realistically shaped ocean basins forced by realistic winds. The timestep allowed is the order of days rather than hours as allowed by more conventional schemes. The numerical method is designed around the special characteristics of low-frequency equatorial waves. A crucial element is the formulation of proper boundary conditions, including those for a partial boundary such as the western end of the Gulf of Guinea. The response of an Atlantic-shaped basin to a periodic wind is compared with the analytic results of Cane and Sarachik for a meridionally unbounded ocean. The response along the equator is essentially the same. A narrow boundary layer forms along the Guinea coast, too narrow to be a single coastally trapped wave: it is the sum of many modes. The near lack of phase variation along the coast and the smallness of the phase difference between the equator and the coast are additional contrasts to the results for an initial value problem studied by Moore and others. The implication is that the annually recurring upwelling in the Gulf of Guinea cannot be adequately modeled as the response to an impulsively applied wind stress.
Article
Equatorial Kelvin and Rossby waves associated with the 1991-93 El Niño warm event were detected in temperature observations made by the Tropical Atmosphere-Ocean @TAO) buoy array. Intraseasonal Kelvin waves were a prominent part of equatorial thermocline depth variability and were well represented by a simple model consisting only of first- and second-mode baroclinic Kelvin waves. The second mode was essential to properly represent the observed amplitude. Thermocline depth variability at 5°N and 5°S was dominated by annual and interannual Rossby waves, which were found to have been largely wind forced in midbasin, with little if any signal associated with eastern boundary reflection. An evaluation of the Wyrtki buildup hypothesis and the delayed oscillator hypothesis in connection with the 1991-93 events showed that a long lag @about two years) occurred between the arrival of the downwelling signal in the west and the subsequent initiation of El Niño., this was considerably longer than suggested by delayed oscillator theory. No downwelling Rossby wave reflections occurred at the right time to trigger the onset of the 1991-92 warm event. Termination of the 1991-92 El Niño was due to an upwelling Rossby wave generated during the height of the warm episode, then reflected from the western boundary, consistent with delayed oscillator ideas. However, in early 1993, a second @weaker) warm episode occurred against the background of a very anomalously shallow west Pacific thermocline. This shows that the El Niño-Southern Oscillation cycle cannot be viewed simply as an oscillator mediated by the western boundary reflection of equatorial Rossby waves and that a buildup of a thick warm layer in the west is not a prerequisite to the occurrence of El Ni{o.
Article
The intraseasonal (30- to 90-day period) Kelvin waves are observed to be forced west of the date line and propagate at a speed of 2.4 m s-1, with high zonal coherence over at least 10 000 km. The oceanic intraseasonal variability and its low-frequency modulation are coherent with atmospheric intraseasonal variations (the Madden-Julian Oscillation), which are known to propagate eastward into the Pacific from the Indian Ocean as part of a planetary-scale signal. The oceanic intraseasonal variability peaks at periods near 60-75 days, while the corresponding atmospheric variations have somewhat higher frequencies (35- to 60-day periods). This period offset is potentially related to the zonal fetch of the wind compared to the frequency-dependent zonal wavelength of the Kelvin wave response. A simple model is formulated. -from Authors
Article
In this paper, the authors discuss observed climatic variability from 1982 to early 1995 and emphasize the contrasts between the period of strong interannual variability during the 1980s and the period of more persistent features beginning in 1990. Three versions of the NCEP coupled forecast model, which were developed to predict interannual sea surface temperature variability in the equatorial Pacific, are described and their performance compared for those two periods.Climatic variability during 1982-1992 in the tropical Pacific was dominated by strong low-frequency interannual variations characterized by three warm and two cold El Niño episodes. However, beginning in 1990, the climate state has been characterized by a pattern of persistent positive SST anomalies in the tropical Pacific, especially in the central Pacific near the date line, and weaker than normal trade winds. Superimposed on this were several occurrences of short-lived, generally small-amplitude warmings in the eastern equatorial Pacific. Some of the short-lived warmings amplified into mature warm episodes, such as in spring 1993 and in late 1994.The NCEP coupled models showed useful skill in predicting low-frequency SST variability associated with warm episodes in the tropical Pacific during the 1982-1992 period. However, the short-lived warmings in spring 1993 and fall/winter 1994/95 were not well predicted by the NCEP coupled models. Neither were they predicted by most of the other dynamic or statistical forecast models. If these short-lived warmings truly represent a different behavior of the coupled ocean-atmosphere system on intraseasonal timescales, the skill levels that were developed for predicting the strong low-frequency SST variability of the 1980s are probably not relevant. The lead times for skillful forecasts of short-lived episodes such as those observed in recent years will no doubt be only a few months.
Article
The Zebiak and Cane model is used in its `uncoupled mode,' meaning that the oceanic model component is driven by the Florida State University (FSU) wind stress anomalies over 1980-93 to simulate sea surface temperature anomalies, and these are derived in the atmospheric model component to generate, wind anomalies. Simulations am compared with data derived from FSU winds, International Satellite Cloud Climatology Project cloud convection, Advanced Very High Resolution Radiometer SST, Geosat sea level, 20°C isotherm depth derived from an expendable bathythermograph, and current velocities estimated from drifters or current-meter moorings.Forced by the simulated SST, the atmospheric model is fairly successful in reproducing the observed westerlies during El Niño events. The model fails to simulate the easterlies during La Niña 1988. The simulated forcing of the atmosphere is in very poor agreement with the heating derived from cloud convection data. Similarly, the model is fairly successful in reproducing the warm anomalies during El Niño events, However, it fails to simulate the observed cold anomalies. Simulated variations of thermocline depth agree reasonably well with observations. The model simulates zonal current anomalies that are reversing at a dominant 9-month frequency. Projecting altimetric observations on Kelvin and Rossby waves provides an estimate of zonal current anomalies, which is consistent with the ones derived from drifters or from current meter moorings. Unlike the simulated ones, the observed zonal current anomalies reverse from eastward during El Niño events to westward during La Niña events. The simulated 9-month oscillations correspond to a resonant mode of the basin. They can be suppressed by cancelling the wave reflection at the boundaries, or they can be attenuated by increasing the friction in the ocean model.
Article
Buoy drifts and current meter records between January 1987 and December 1993 are used to investigate the interannual variability of the equatorial Pacific currents at a depth of 15 m. The sampling is coarse until mid-1988 but more complete afterward, so that the large-scale features of the anomaly currents can be documented on the seasonal to yearly timescale. Using objective analysis, bimonthly current anomalies are mapped between 20°N and 20°S on a 1°×5° grid, and the error covariance matrix of the analyzed fields are estimated. The current anomalies are primarily zonal, with largest amplitudes within about 8° from the equator, and they are largely linked to the El Niño-Southern Oscillation phenomenon. In particular, broad, basin-wide westward anomaly currents were encountered during the 1988 La Niña, and strong eastward currents persisted from July-August 1991 to January-February 1992, followed by westward currents from May-June to July-August 1992. An empirical orthogonal function (EOF) analysis shows that the first EOF of zonal current anomaly is largely uniform in the equatorial band, while the next two EOFs describe large-scale currents of opposite sign across the equator and across 160°W, respectively. The EOFs are rather smooth and the errors on the principal component time series relatively small, which indicates that the sampling is adequate to describe the large-scale, low-frequency zonal current fluctuations. As the dominant EOFs of meridional current are noisy and the relative errors on the principal components larger, the meridional current fluctuations are not as well captured by the data set. Correlation analysis and a singular value decomposition are used to investigate the influence of advection by the large-scale, low-frequency currents on sea surface temperature (SST) anomalies during 1987-1993. Although the data set is noisy and other terms play an important role in the SST anomaly equation, the effect of zonal and, to a lesser extent, meridional advection is seen in much of the central and eastern equatorial Pacific. The dominant terms are the anomalous zonal advection of mean SST, the mean zonal, and, intermittently, meridional advection of SST anomalies.
Article
The “high index” response of the northeast Pacific westerlies to big positive anomalies of equatorial sea temperature, observed in the winter of 1957–58, has been found to repeat during the major equatorial sea temperature maxima in the winters of 1963–64 and 1965–66. The 1963 positive temperature anomaly started early enough to exert the analogous effect on the atmosphere of the south Indian Ocean during its winter season. The maxima of the sea temperature in the eastern and central equatorial Pacific occur as a result of anomalous weakening of the trade winds of the Southern Hemisphere with inherent weakening of the equatorial upwelling. These anomalies are shown to be closely tied to the “Southern Oscillation” of Sir Gilbert Walker.
Article
Disastrous climatic events, such as the Australian–Indonesian drought and the severe flooding in South America, which occurred in the equatorial Pacific in 1982/83 are part of a pattern which can be followed in unprecedented detail using satellite data. The evolution of wind, precipitation and surface temperature fields during this remarkable example of atmosphere–ocean interaction are described in this article. Models can relate the main features of the sea-surface temperature, sea level and surface current patterns to the winds, and relate wind to precipitation (atmospheric heating) patterns. The precipitation field shows a close relationship with surface temperature, but this is less well understood.
Article
Long-range sea surface temperature forecasts from two different coupled ocean-atmosphere models of the tropical Pacific are used in conjunction with statistical models relating winter Northern Hemisphere 700-mb height and tropical SST to forecast the former field at a lead time of two seasons in advance. The forecasts show considerable skill over large areas, with a regional distribution of predictive performance that is consistent with the observed contemporaneous relation between the two fields. Comparable skills for lead time of a year or more in advance seem likely.