UDC 551.513.1:551.526.6:551.465.63
Transequatorial Effects of Sea-Surface Temperature
Anomalies in a Global General Circulation Model
JEROME SPAR-Department of Meteorology and Oceanography,
New York University, Bronx, N.Y.
I
ABSTRACT-The global response of tbe atmosphere, as
simulated by the Mintz-Arakawa, two-level, general
circulation model, t o a persistent anomalous pool of warm
sea-surface temperatures (SST) in the extratropical Pacific
Ocean is examined in this descriptive study in terms of
the meridional pole-to-pole profile of the zonally averaged
600-mb surface for periods up to 90 days. Following an
initial hydrostatic inflation of the isobaric surface in the
latitude of the warm pool, effects spread poleward within
the hemisphere, then begin to appear after about 2-3
weeks in high latitudes of the opposite hemisphere, but
with little or, no response in the Tropics. The same sea
temperature anomaly field generates a stronger response
in winter than in summer and a very different reaction
when located in the Southern Hemisphere than when in
the Northern Hemisphere. After a month of thermal
forcing, the response to an SST anomaly is a t least as
large in the opposite hemisphere as in the hemisphere of
the anomaly. A winter hemisphere responds more rapidly
to an SST anomaly in the opposite hemisphere than does a
summer hemisphere. Vacillation between low and high
meridional wave number patterns is observed in the com-
puted reaction to the warm pool.
1. INTRODUCTION
Experiments with the Mintz-Arakawa, two-level, gen-
eral circulation model have recently been carried out for
the purpose of estimating the possible response of the
atmosphere to persistent anomalies of sea-surface temper-
ature (SST) over periods of the order of a season (Spar
1972, 1973). One of the conclusions derived from these
cant response in the sea-level pressures in the opposite
hemisphere after several weeks, and that these reactions
appear to occur with little or no disturbance of the
equatorial region. It appeared from these computations
as if the transequatorial propagation took place through a
form of forced standing wave with an equatorial node.
However, no attempt was made in the previous study to
examine this interhemispheric transfer of influence in any
detail. In the present paper, an effort is made to describe
the development of the model atmosphere's response to
these SST anomalies in terms of their effect on the pole-to-
pole meridional profile of the zonal mean 600-mb surface,
which represents approximately the middle level of the
model. An earlier experiment similar to the one described
here was carried out by Rowntree (1972), but with a
hemispheric Geophysical Fluid Dynamics Laboratory
(GFDL) model and an SST anomaly in the equatorial
region. Since there are, of course, no interhemispheric
effects in Rowntree's hemispheric experiment, his results
cannot be compared with ours. Ramage and Murakami
(1973) have recently argued that the equatorial wall in
the hemispheric model is responsible for some of Rown-
tree's principal results.
A complete documentation of the Mintz-Arakawa two-
level model has been provided by Gates et al. (1971)' and a
I
experiments was that SST anomalies may induce a signifi-
The research reported in this paper has been supported by the National Aeronautics
and Space AdminiWation, Goddard Space Flight Center, under Grant NGR 33-016-174.
554 / Vol. 101, No. 7 / Monthly Weather Review
brief description of the model can be found in Spar (1973).
I n the Northern Hemisphere SST anomaly experiments
(designated NHTA) , the climatological mean annual sea-
surface temperatures, which are employed as fixed bound-
ary conditions in the model computations, were altered
by adding from 2" to 6°C to the SST over a "box" in the
North Pacific Ocean bounded by latitudes 22' and 42'N
and longitudes 140"W and 180'. For the Southern Hemi-
sphere SST experiments (designated SHTA), the same
anomaly field was added to the corresponding box in the
South Pacific Ocean. In both cases, the maximum anomaly
(+6"C) isotherm corresponds to a "rectangle" defined by
longitudes 150" and 170"W and latitudes 30' and 34', so
that the center of the warm pool in the ocean is at latitude
32" and longitude 160"W. A Northern Hemisphere anomaly
experiment was conducted for both the winter (NHTA-W)
and summer (NHTAS) seasons with the same SST fields,
while the Southern Hemisphere anomaly experiment was
carried out for northern summer (southern winter) only.
Initial conditions for the three experiments were taken
from history tapes provided by the UCLA group under
Yale Mintz. (See Mintz et al. 1972 for examples of the
climate simulations generated by the model.) The effect
of the SST anomaly is measured by comparing the
anomaly'' run with a "control" run that is identical in
every respect except for the absence of the SST anomaly.
Differences between anomaly and control runs presumably
represent effects of the SST anomaly.
In each experiment, the expected initial effect of the
positive SST anomaly is an augmented heat transfer from
sea to air over the local anomaly region. It is impossible to
anticipate a priori precisely how this increased heating
effect will propagate away from the warm ocean area or
what its ultimate dynamical consequences will be. HOW-
ever, one can expect that the immediate hydrostatic
(I
,....' W I N T E R
Anomaly I day 90)
0 0 0 0
I I I I
t o
i o \ 000
\
\
\
\,90d
'0'
I I I I
LATITUDE ''IJTH N O R T H
FIGURE 1.-Meridional profile of zonal mean 600-mb height for
the winter control run, experiment NHTA-W, on days 15, 45,
and 90. (Circles represent the profile on day 90 for the anomaly
run.)
effect of the augmented heating will be an inflation of the
isobaric surfaces aloft. Since the present study is concerned
primarily with the meridional propagation of these effects,
it is desirable to examine zonally averaged quantities.
For our purpose, we have chosen the 600-mb level as a
characteristic isobaric surface and have computed for
each day of each 90-day run and for every 4' of latitude
the zonal mean geopotential height of that surface. The
initial effect of the SST anomaly on this quantity is ex-
pected to be a rise (relative to the corresponding control
case) of the zonal mean height within the latitude band
containing the warm pool. This initial reaction and the
subsequent evolution of the 600-mb meridional height
profile, including the response in the opposite hemisphere,
are illustrated and discussed in the following sections.
2. EXPERIMENT NHTA-W
To illustrate the characteristic meridional 600-mb
profiles generated by the model in the northern winter
season, we have reproduced in figure 1 the profiles for
the winter control run for days 15, 45, and 90, correspond-
ing to early, middle, and late winter.2 The large fluctua-
tions in height found in the Arctic and Antarctic (where
the maximum height ranges computed for the whole
Because of the fact that initial conditions for the first day of each control run were se-
lected at random from the history tapes for the appropriate season, whereas the model
sun was reset to its proper solstitid position for the actual control run, the model appears
to go through a transient adjustment period during the first few days. Therefore, day 15
rather than the first day (approximately the solstice) was selected as representative of
early winter.
season were 501 and 377 m respectively) may be un-
realistic. Similarly, the total range of heights at the
Equator during the season (computed to be 85 m) also
may be too large. Between latitudes 50"N and 50°S, the
profiles are nearly symmetrical. The mean zonal geo-
strophic wind speeds corresponding to the mean height
profiles at 600 mb between latitudes 30' and 50' in both
hemispheres are approximately 20 m/s and are slightly
higher in the Southern Hemisphere than in the Northern
Hemisphere, despite the fact that figure 1 represents
northern winter.3 (The small circles in figure 1 represent
the zonal mean 600-mb heights at 90 days for the anomaly
run and will be discussed later). The large decrease in
height in the Antarctic at the end of the season appears
to indicate that the effect of radiative cooling is already
becoming apparent there by the time of the equinox.
The reaction of the model atmosphere to the SST
anomaly during the first month is illustrated in figure 2
in the form of a time-latitude cross section of height
differences between the anomaly and control runs.
Positive values represent an increase in the height of the
600-mb surface relative to the control case. Difference
isopleths, including the zero isopleth, are drawn a t an
interval of f 20 m up to 100 m, and f 40 m at higher
values. Hatching is used to indicate positive effects in
excess of + 20 m. The latitude of the SST anomaly is
indicated schematically by the bar graphs on either side
of the figure, the length of the bar being proportional to
the magnitude of the anomaly at each latitude. For the
first 3 weeks, the response is largely confined to the
Northern Hemisphere. The 600-mb surface is initially
elevated within the latitude band of the SST anomaly
and lowered to the north. The crest of this inflation wave
migrates slightly poleward after about 2 weeks before
vanishing. The maximum height difference within the
latitude band of the SST anomaly, shown in figure 3 for
the entire 90-day period, rises at an almost constant rate
for the first 2-3 weeks. After that time, the inflation effect
in the warm pool latitude band disappears briefly, and is
replaced by a marked elevation of the 600-mb surface in
the polar regions. At the North Pole, this effect, which is
probably spurious, reaches a maximum value of 260 m
on day 25. Almost simultaneously a dubious response
appears in the Antarctic;4 near the end of the month,
large effects are seen in middle latitudes of the Southern
Hemisphere. At the end of the 30 days, the inflation in
the latitude band of the SST anomaly begins to reappear,
and another period of inflation lasting about 3 weeks
begins, as shown in figure 3.
During the first month, the largest height difference
computed in the equatorial region was only 14 m; during
the entire 9O-day period, the equatorial response did not
exceed 18 m. The small equatorial reaction is illustrated
3 As shown, for example, by van Loon (1964, 1965), the geostrophic zonal westerlies a t
600 mb in the Southern Hemisphere in southern summer are stronger than in southern
winter, and at least as strong as the zonal westerlies at 5GU mb in the Northern Hemisphere
in northern winter. Thus, the symmetry of the profiles in figure 1 is probably realistic.
4 A recently discovered coding error (Gates et al. 1971) in the computation of the albedo
over snow and ice has been found (Gates 1972) to produce spurious results in the simulated
climatology, especially in the Antarctic. This, together with certain compUtatiOna1 prob-
lems in the vicinity of the poles, renders suspect all results poleward of about latitude 709.
JUIY 1973 1 Spar I 555
DAY
I 2 3 4 5 6 7 8 9 IO II 12 13 14 15 16 17 18 19 2021 22 23 24 25 26 27 28 29 30 31
1 1 1 )1 1 1 1 1 1 ~1 \1 1 1 1 1 1 1 1 1 1 1 1 1 1 ~1 \\
l o *220 *180*140 1
I 2 3 4 5 6 7 8 . 9 IO I I 12 13 14 I S 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
FIGURE a.-Time-latitude cross section of anomaly-minus-control 600-mb height differences for the first month of experiment NHTA-W.
Isopleths are drawn for an interval of 20 rn up to 100 m, and 40 m above 100 m. Positive differences in excess of f 2 0 m are indi-
cated by hatching. The latitude band of the warm pool is indicated by the bar graphs.
I
I
for the whole season in figure 4, where we have plotted for
each day the largest height difference (positive or nega-
tive) in the latitude band from 6'N to 6's. Some disturb-
ance of the equatorial region does in fact appear after
about 3 weeks, suggesting a propagation of influence
across the Equator. However, there is no evidence either
in figure 2 or in the corresponding time-latitude sections
for the following 60 days (not shown) of traveling meridi-
onal waves crossing the Equator. While the small-ampli-
tude equatorial oscillation shown in figure 4 does ap-
parently represent a response to the SST anomaly, it
appears to have the character of a standing oscillation
rather than a progressive wave.
The global response of the 600-mb surface to the SST
anomaly is further illustrated in figure 5 showing daily
meridional profiles of the anomaly-control height differ-
ences drawn a t 10-day intervals beginning on day 20.
Latitude is represented on a sine scale; therefore, the
surface area of any latitude band is proportional to hori-
zontal distance in the figure. No values are plotted or
analyzed poleward of latitude 70'. The bar graphs in the
figure indicate the latitude band of the SST anomaly.
Days 20, 30, 40, and 50 are shown in figure 5A and days
60, 70, 80, and 90 appear in figure 5B. From figure 5, it is
apparent that, with the possible exception of the shallow
trough on days 20 and 30, no disturbances of significant
amplitude crossed the Equator during the season. I t can
also be seen in figure 5 that, following the initial inflation
of the 600-mb surface in the latitude band of the SST
I
I
-10-
-2 0 -
-30 -
-4 0 0 5 IO 15 20 25 30 35 4 0 45 50 55 60 65 70 75 BO 85 90
D A Y
FIGURE 3.-Maximum (positive or negative) anomaly-minus-
control 600-mb height differences, Az (m), in the latitude band
22'-42"N for experiment NHTA-W.
anomaly, all subsequent inflations took place north of the
center of the warm pool. The standing character of the
oscillation induced in the Northern Hemisphere is indi-
556 / Vol. 101, No. 7 / Monthly Weather Review
0 5 10 \5 20 25 30 35 40 45 50 55 60 65 70 75 80 85
DAY
FIGURE 4.-Same as figure 3 for latitude band 6’N-6”S.
cated in figure 5, where the amplitude of the oscillation is
seen to diminish with time in middle latitudes and increase
in higher latitudes.
The Southern Hemisphere response first appears clearly
at 30 days in figure 5A in middle and high latitudes, then
shifts toward the Antarctic. However, after 60 days, as
shown in figure 5B, the influence of the SST anomaly
becomes “locked in” in the Southern Hemisphere, with a
permanent anomaly in the slope of the 600-mb surface.
This anomalous corrugation is characterized by relatively
high geopotential in the Antarctic and low geopotential
near latitude 4OoS, as indicated also by the circles repre-
senting the anomaly profile on day 90 in figure 1. A t
latitude 50’5, the anomalous slope on day 90 corresponds
to a decrease of about 7 m/s (equivalent to more than 30
percent) in the speed of the geostrophic westerlies. [A
similar diminution of the surface Southern Hemisphere
westerlies in the latter part of the season was also noted by
Spar (1973), based on 30-day mean sea-level pressure
fields]. On the other hand, at latitude 70’N on day 90,
the anomaly profile shown in figure 1 represents a reversal
of the geostrophic zonal flow from weak easterlies in the
control case to westerlies in the anomaly run.
The response of the 600-mb surface to the SST anomaly
can also be examined in terms of the day-to-day changes
in the spectral characteristics of the height difference
profile. A simple representation of this history is shown
in figure 6, where the meridional wave number of maxi-
mum amplitude (the “dominant wave number”) for each
day is plotted against time with the relative amplitude
of each dominant harmonic indicated by the length of a
vertical spike. (In the Fourier analysis of the meridional
height difference profiles, the domain was taken as twice
30. 50° SOUTH
70.90.
180-
160-
140-
120
100-
80 -
€a-
40 -
c 20-
-
-
L
E“ 0- - Y -20-
-40 -
-60 -
-80
-100 -
-120 -
-140 -
a
-
-E -180
7
FIGURE 5.-Meridional profiles of anomaly-minus-control 600-mb
height differences, Az (m), for (A) days 20, 30, 40, 50, and (B)
days 60, 70, 80, and 90, of experiment NHTA-W. The latitude
of the SST anomaly is indicated by the bar graphs.
the distance from pole to pole, so that wave number 2
represents one wavelength from pole to pole, wave number
4 one wavelength from pole to Equator, etc.)
For the first 10 days, the profile is characterized by a
low-wave-number (3) regime of small but increasing
amplitude, corresponding to the initial hydrostatic infla-
tion of the 600-mb surface. The meridional propagation
of the anomaly effect within the Northern Hemisphere is
represented by an abrupt transition to a high-wave-
number (7-13) regime lasting 9 days. As the SST effect
crosses the Equator between days 21 and 3 1, a second low-
wave-number regime (2-4) is established. Then from
days 32 to 39, the wave numbers fluctuate indecisively
before settling into a high-wave-number (8) pattern
between days 40 and 45. This latter period was charac-
terized by the reappearance of inflation in the Northern
Hemisphere latitude band of the SST anomaly (as shown
July 1973 j Spar j 557
13
12
1 1 .
I O .
5
!-
z a
2 4 - E
0 0
3
2
I "I I
' I I I 11 I I I I
1 1
I I
I I
1 I (m e t e r s )
I
I I
I I 1
Amplitude
Stole
I I
I
in fig. 3), while a t the same time the Southern Hemisphere
reaction was increasing in magnitude.
From day 46 until the end of the season, the Southern
Hemisphere response becomes "locked in," as noted
previously, and dominates the profile. As a result, the
anomaly-control height difference pattern is characterized
by low wave numbers during all of the latter half of the
season. Thus, in terms of the meridional 600-mb anomaly
profile spectrum, the response to the northern winter SST
anomaly is a vacillation between low- and high-wave-
number regimes in the first half of the season, and a fixed
low-wave-number anomaly pattern in the second half.
3. EXPERIMENT NHTA-S
Representative meridional 600-mb profiles generated
by the model for the northein summer season are illus-
trated in figure 7 for days 15,45, and 90 from the summer
control run. (The circles in the figure represent the
anomaly profile for day 45, which will be discussed
A comparison of figure 7 with figure 1 reveals both the
small annual variation in middle latitudes of the Southern
Hemisphere and the large annual variation in middle
and high latitudes of the Northern Hemisphere, as well
as the asymmetry between the hemispheres chaI acteristic
of northern summer. The plofile for day 90 indicates that
the effect of radiative cooling at high latitudes in the
Northern Hemisphere may already be apparent in the
558 1 Vol. 101, No. 7 1 Monthly Weather Review
I I I I 1
I I I I I I
I I l l II I I l l I d I I I I I I I I
Control
....... day 90
--- day 45
day 15
3900
4100i
Anomaly (day 4 5 )
0 0 0 0 0
LATITUDE SOUTH NORTH
FIGURE 7.-Same as figure 1 for the summer control run, experiment
NHTA-S. (Circles represent the profile on day 45 for the anomaly
run.)
model in the increased slope of the 600-mb level by the
time of the equinox.
The time-latitude cross section in figure 8 shows the
response (anomaly-minus-control height differences) to
the SST anomaly in northern summer for the first month.
Compared with the northern winter experiment (fig. 2),
the initial inflation in summer in the latitude band of the
SST anomaly is both slower and smaller. This is also
apparent from figure 9, where the maximum (positive or
negative) anomaly-con trol height differences in the
latitude band of the SST anomaly have been plotted
for 87 days. During the first month, the maximum inflation
effect is less than 20 m in summer compared with more
than 60 m in winter {fig. 3). {Only at the end of the season,
as shown in figure 9, does an inflation effect of 45m, com-
parable to that of the winter experiment, appear.) Accord-
ing to the bulk aerodynamic formula used in the model
{Gates et al. 1971) to compute heat. transfer from sea to
air, the anomalous heating, and hence the inflation effect,
should be proportional to the magnitude of the SST
anomaly, which is the same in both the summer and winter
experiments. However, the heat transfer is also assumed to
be proportional to the surface wind speed, which is
generally greater in winter than in summer in the region
of the warm pool. This undoubtedly contributes to the
difference between the two seasons. (The relatively large
inflation effect at the end of the summer season noted in
figure 9 is probably also due in part to stronger surface
winds.) Other nonlinear factors, such as greater instability
and stronger convective heat transfer over the ocean
D A Y
I 2 3 4 5 6 7 8 9 IO II 12 13 14 15 16 17 18 I9 20 21 22 23 24 25 26 27 28 29 3031
- -______- ---*-
I 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 I
30° -
/--'\
\
\ 40' / -
/ \
\
I
\ \
\
.
---------a/ . . . .
50' -
60° -
700 -
\
. .. \
\
800 - \ -12
900-
I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
FIGURE &--Same as figure 2 for the first month of experiment NHTA-S.
in winter, also contribute to the seasonal difference
between the inflation effects.
In the equatorial region, the height differences in experi-
ment NHTA-S between anomaly and control runs ale
extremely small during the entire season with maximum
absolute values of less than 7 m (compared with 18 m
in NHTA-W). Despite this evidence that meridional
wave propagation does not take place across the Equator,
the response in the Southern Hemisphere to the SST
anomaly in the opposite hemisphere appears earlier and is
larger in southern winter (fig. 8) than in southern summer
(fig. 2). It appears that the winter hemisphere is more
responsive to SST anomalies in the opposite hemisphere
than is the summer hemisphere.
The remaining history of experiment NHTA-S beyond
the first month is illustrated in figure 10, which shows the
meridional height difference profiles on days 40, 60, and
80. As in figure 5, latitude is plotted on a sine scale and
no Iesults are shown poleward of latitude 70°N and 70's.
In these curves, there is no evidence of the locked-in re-
sponse noted in the Southern Hemisphere in figure 5.
Instead, the reaction in the Southern Hemisphere to the
thermal forcing in the north appears to be in the form of a
standing oscillation. The inflation effect in the Northern
Hemisphere at the end of northern summer noted in
figure 9 also appears in the curve for day 80 in figure 10
accompanied by a response of even larger amplitude in
the Southern Hemisphere.
Returning again to figure 7, we see that the anomaly
profile at 600 mb, represented for day 45 by the circles,
is not very different from the control profile on the same
day. The effects shown in figures 8 ,9 , and 10 do not re-
present radical alterations' of the zonal mean flow, but
only relatively small-amplit ude perturbations. However,
these small effects on the zonally averaged profiles may
be associated with synoptic effects of considerable magni-
tude (Spar 1973).
4. EXPERIMENT SHTA
The possibility that unobseived events in the Southern
Hemisphere may be transmitting significant signals to the
Northern Hemisphere over periods of t.he order of a month
to a season was the motivation for experiment SHTA. I n
this experiment, the SST anomaly in the South Pacific
Ocean was introduced at the beginning of southern winter
(northern summer). Hence, the control profiles for the
SHTA experiment are also represented by the curves in
figure 7.
The initial response in this experiment closely resembles
that of the previous two. Maximum inflation begins in
the latitude band of the SST anomaly and grows rapidly.
However, as shown in figure 11, where the inflation curves
for the SST anomaly belts are plotted for the first 10
days for all three experiments, the resemblance abruptIy
ends after 4 days. Up to day 4, the inflation effect is almost
the same for NHTA-W and SHTA; that is, for both
July 1973 f Spar f 559
5 0 25
20
1 5 -
- t -
L -
E
a -
-- u
10
-4 0 1
-
-
-
-
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 <
D A Y
FIQURE 9.-Same as figure 3 for experiment NHTA-S.
I O 0
80-
60
40
I
?? 2 0 -
L
I - w -2 0 -
4 -4 0
-60
-e o -
-100
-
-
- -
0-
-
-
-
- I
F-
'..
&
iATlTUDE NORTH
....
y .I
SOUTH
FIQURE 10.-Same as figure 5 for days 40, 60, and 80, of experi-
ment NHTA-S.
winter hemisphere SST anomaly experiments. (The sum-
mer hemisphere inflation rate represented by NHTA-S
is generally smaller.) After day 4, however, the initial
inflation effect rapidly disappears in SHTA, whereas in
NHTA-W the inflation continues for about three weeks
(fig. 3). Thesubsequentresponse to the SHTA SST anomaly
in the latitude band of the Southern Hemisphere warm
pool is shown in figure 12, which may be compared with
the corresponding figure 3 for NHTA-W. Both figures
exhibit positive maxima a t 40-45 days and 75-80 days,
with minima (negative maxima) at 60-65 days. Except for
these gross similarities however, the two curves bear little
resemblance to each other.
I
I
I
I
I
I
I
I
I
I
I
I
I
i
I
I
I I
I
,"HTA - w
I
I'
/
I
I
I
D A Y
FIQURE ll.-Maximum anomaly-minus-control 600-mb height
differences, Az (m), in the latitude band of the SST anomaly for
the first 10 days of experiments NHTA-W, NHTA-S, and SHTA.
D A Y
FIGURE l2.-Same as figure 3 for latitude band 22O-42OS for experi-
ment SHTA.
560 1 Vol. 101, No. 7 1 Monthly Weather Review
90.
eo*-
70-
60'
50-
40.
30°-
20'
0-
100-
IO.-
200-
I 2 3 4 5 6 7 0 9 I O I I I2 13 14 I5 16 17 10 19 2021 22 23 24 25 26 27 28 29 30 31
I 2 3 4 5 6 7 8 9 IO I I 12 I 3 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
I 1 I I I I I I I I I I I' I I I I I I I I I I I I I I I I 1 1
- 90°NORTH
*40 *2 0 - 00.
0 ,,eo -70.
/ - 60.
0 - .' / /-
/
/
:I-:",-- ' \ -14 -2 0
.
I
I
-
.-- - ------ - f -1s
- /- --. ' -----o -50.
- 40. ___--- 0--, .. \ .Ox, +I I
\ - / ,\ \
f -4 1 \\ -3 \
\
/ \ \ - 30. \
'---_
/ - 20. \ / I /
I ,-----/ W
I
-
I 0
- 100 2
a
10. -1
/ I
' -2 I
I 00 E
I I ;; I
\
\
I
I I l !f / _-,----o - I
1 c/ I
--01 \ ,---
I -1 / // \\-/'
FIGURE 13.-Same as figure 2 for the first month of experiment SHTA.
As shown in the time-latitude cross section for the first
month of SHTA (fig. 13), significant height effects do
appear in the Southern Hemisphere within a few days,
generally poleward of the latitude band of the SST
anomaly. However, the response in the opposite hemi-
sphere is minimal in the first month compared with that
found in NHTA-W (fig. 2) or NHTA-S (fig. 8). This
same lack of response in the Northern Hemisphere during
the first month of SHTA was also noted by Spar (1973)
in both the sea-level pressures and regional 600-mb cir-
culation indices. One can only speculate about the reasons
for both the cut-off of inflation and the slow transequa-
torial response in the SHTA experiment. Whatever the
reasons (different initial conditions, effects of continen-
tality, etc.), it is apparent that the response to SST
anomalies is complex and not readily anticipated.
The history of the meridional response at 600 mb in
the SHTA experiment is illustrated in figure 14 by the
height difference profiles for days 40, 60, and 80. Again
the response appears to take the form of a standing oscil-
lation in both hemispheres but with a larger amplitude
in the hemisphere opposite the SST anomaly. Relatively
large height differences appear on day 80 in the equatorial
region in this experiment compared with the previous
two experiments (figs. 5, 10). This is further illustrated in
figure 15, where the largest positive or negative height
differences in the equatorial belt, 6'N-6'S, are plotted
for each day. Here it can be seen that the equatorial
response, represented by a lowering of the isobaric surface
relative to the control, began only after two months, but
continued to the end of the season.
5. SUMMARY AND CONCLUSIONS
The global response of the model atmosphere to a posi-
tive SST anomaly located in extratropical latitudes of
the Pacific Ocean has been studied in terms of the merid-
ional profile of the 600-mb surface from pole to pole. The
initial effect of the SST anomaly is an inflation of the
isobaric surfaces, relative to a control run, within the
latitude band of the warm pool. This is apparently a
direct hydrostatic result of the augmented warming asso-
ciated with increased sea-to-air heat transfer. The initial
inflation rate is greater in winter than in summer. In
experiments with a North Pacific warm pool, the inflation
continued for about 3 weeks in both winter and summer.
However, when the same SST anomaly was placed in the
South Pacific Ocean, the initial inflation of the isobaric
surface lasted only a few days.
During the first month of each experiment, the atmos-
pheric reaction within the hemisphere of the SST anomaly
appeared mainly poleward of the latitude band of the
warm pool. This was true in all three experiments but
was more apparent in the winter hemisphere experiments
(NHTA-W and SHTA), as shown in figures 2 and 13,
than in the summer hemisphere experiment (NHTA-S),
.
July 1973 1 Spar 1 561
120-
100
80
60-
40
- i 20-
: .- - * -2 0 -
a
-40
-60
-80
-100
- -
..I .
-
LATITUDE
7
-
-
-
0-
-
-
-
-
-
FIGURE 14.-Same as figure 5 for days 40,60, and 80, of experiment
SKTA.
illustrated in figure 8. Although, as shown in figure 11,
the initial inflation persisted longer and grew ultimately
larger in the Northern Hemisphere summer experiment
(NHTAS) than in the Southern Hemisphere winter
experiment (SHTA) , the reactions poleward of the infla-
tion belt were stronger in the winter hemisphere than in
the summer hemisphere.
Significant transequatorial effects first appeared at high
latitudes in the hemisphere opposite the warm pool after
about 2-3 weeks. After a month, the magnitude of the
response was at least as large in the opposite hemisphere
as in the hemisphere of the SST anomaly. The winter
hemisphere appeared to respond more rapidly to SST
anomalies in the opposite hemisphere than did the summer
hemisphere. This can be seen by comparing the rapid
transequatorial response in figure 8, where the opposite
hemisphere is a winter hemisphere, with the slower trans-
equatorial response in figures 2 and 13, where the opposite
hemisphere is the summer hemisphere. Thus, it appears
that the winter hemisphere appears to respond more sensi-
tively than the summer hemisphere, not only to the effects
of thermal forcing within the hemisphere of the warm
oceanic pool, but also to the transequatorial signals trans-
mitted by SST anomalies in the opposite herni~phere.~
The reactions in the Tropics were generally quite small,
and at no time was there evidence of meridional wave
propagation across the Equator. Instead, the transequa-
torial response appeared either as a permanent corrugation
of the isobaric surface or as a standing oscillation in
middle and high latitudes of the opposite hemisphere. The
5 Had this unexpected result been anticipated, we would have planned at least two
additional experiments, one with an SST anomaly in the Southern Hemisphere innorth-
ern winter and one with anomalous warm pools in both hemispheres. But then, there is
no bound to the number ofexperiments one is tempted to perform with dynamics1 models
I-
-12 -
-
-16 -
-20 -
0 5 IO I5 20 25 30 35 40 45 50 55 60 65'70 75 80 85
D A Y
FIGURE 15.-Same as figure 3 for latitude band 6ON-6"S for experi-
ment SHTA.
vacillation between high and low meridional wave number
patterns in the response of the zonally averaged 600-mb
height profile appears to represent an alternation between
the dominance of direct thermal forcing (i.e. inflation) ,
the migration of influences (largely poleward) within the
hemisphere of the warm pool, and the transequatorial
response.
During the 3-mo period of each experiment, the SST
anomaly imposed a small but significant perturbation on
the 600-mb height profile. The effect on the zonal circula-
tion, represented roughly by the slope of the difference
profile, was not computed as part of the experiment but
was undoubtedly of even relatively larger magnitude.
However, the character of the perturbation was quite
different in each experiment despite the identical character
of the SST anomaly. Clearly, the atmospheric response,
as represented by the model computations, is a complex
function of the initial state of the atmosphere and the
hemispheric topography on which the anomaly is super-
imposed and could not have been anticipated from sim-
plistic qualitative reasoning.
In this descriptive study, we have left unanswered the
question of what mechansim is responsible for the propaga-
tion of influence across the Equator. Further diagnostic
studies of the experiment, which are in progress, may
provide a better understanding of that problem.
ACKNOWLEDGMENTS
This study was carried out in cooperation with the Goddard
Institute for Space Studies (GISS) directed by Robert Jastrow,
whose support is gratefully acknowledged. Among the GISS staff
we especially acknowledge the assistance of Milton Halem, who
arranged for computing services, and John Liu, who was responsible
for all the programming and computations carried out for the
project on the GISS IBM 360/95 computer. We also wish to thank
562 J Vol. 101, No. 7 f Monthly Weather Review
Yale Mintz and Akio Arakawa of the University of California,
Los Angeles, for allowing us to use their model and programs for
these experiments. To the National Academy of Sciences, we
express our appreciation for the appointment as National Research
Council Senior Research Associate at GISS during August 1972,
when much of this paper was written. The 600-mb anomaly-control
height difference profile computations were analyzed a t New York
University by Lawrence J. Lewis. Figures were drawn by Gertrude
Fisher.
REFERENCES
Gates, W. L., Batten, E. S., Kahle, A. B., and Nelson, A. B., “A
Documentation of the Mintz-Arakawa Two-Level Atmospheric
General Circulation Model,” Report R-877-ARPA1 The RAND
Corporation, Santa Monica, Calif. Dec. 1971, 408 pp.
Gates, W. L., “The January Global Climate Simulated by the
Two-Level Mintz-Arakawa Model: A Comparison With Observa-
tion,” Report R-1005-ARPA, The RAND Corporation, Santa
Monica, Calif., Nov. 1972, 107 pp.
Mintz, Yale, Katayama, A., and Arakawa, A., “Numerical Simula-
tion of the Seasonally and Inter-Annually Varying Tropospheric
Circulat,ion,” Proceedangs of the Survey Conference, February 16-16
1976, Climatic Impact Assessment Program, Cambridge, Mass., A. E.
Barrington (Editor), U.S. Department of Transportation, DOT-
TSC-OST-72-13, Washington, D.C., Sepf. 1972, 281 pp. (see
Ramage, C. S., and Murakami, T., “Comment on the Paper by
P. R. Rowntree: The Influence of Tropical East Pacific Ocean
Temperatures on the Atmosphere,” @.uarterly Journal of the Royal
Meteorological Society, Vol. 99, No. 420, London, England, Apr.
Rowntree, P. R., “The Influence of Tropical East Pacific Ocean
Temperatures on the Atmosphere,” Quarterly Journal of the Royal
Meteorological Society, Vol. 98, No. 416, London, England, Apr.
Spar, Jerome, “Effects of Surface Anomalies on Certain Model-
Generated Meteorological Histories,” New York University,
Geophysical Sciences Laboratory Report No. GSLTR-72-2,
NASA Grant N GR-33-016-174, Bronx, N.Y., Apr. 1972, 54 pp.
Spar, Jerome, “Some Effects of Surface Anomalies in a Global
General Circulation Model,” Monthly Weather Review, Vol. 101,
No. 2, Feb. 1973, pp. 91-100.
van Loon, Harry, “Mid-Season Average Zonal Winds a t Sea Level
and a t 500 mb South of 25 Degrees South, and a Brief Comparison
with the Northern Hemisphere,” Journal of Applied Meteorology,
van Loon, Harry, “A Climatological Study of the Atmospheric
Circulation in the Southern Hemisphere During the IGY, Part
I: 1 July 1957-31 March 1958,” Journal of Applied Meteorology,
pp. 194-216.)
1973, pp. 393-394.
1972, pp. 290-321.
I
Vol. 3, NO. 5, Oct. 1964, pp. 554-563.
Vol. 4, NO. 4, AUg. 1965, pp. 479-491.
[Received November 6, 1972; revised May 31, 19731
July 1973 I Spar 1 563