402
4 Results.
The experiment time-mean SST and wind stress is presented in Figure la and shows a realistic
mean state. The cold tongue extends further 1700W and in its most western part is centered
at equator. A well developed warm pool can be seen in the west with SST around 29°C. Warm
temperatures (>28°C) penetrate eastward both south and north of the equator. Figure Ib and
lc show the April and October means for SST and wind stress. In April, a larger warm pool,
penetrating further east, can be observed. In April, the warm region in the eastern Pacific north
of the equator is warmer and a weakened equatorial SST gradient can be seen. In October, a
well-developed cold tongue exists with eastern Pacific temperatures as low as 22°C.
Figure 2 shows an equatorial longitude-depth section of time-mean temperature (Fig. 2a), zonal
current (Fig. 2b) and vertical velocity (Fig. 2c). The slope of the isotherms across the Pacific is
well simulated. Comparing these results with the coupled ARPEGE/OPA experiment (Terray
et al., 1994) a shallower thermocline in the east, closer to observations, is obtained perhaps due
to the stronger zonal wind stress. In the western equatorial Pacific the thermocline is deep and
a uniform temperature layer of about 60 meters deep can be seen. An Equatorial Under Current
(EUC) exceeding 1 m/s can be seen and is comparable with other forced (Belam.ari, 1993 and
Dandin, 1993) and coupled (Terray et a.l., 1994) experiments undertaken with the OPA ocean
model. The core of the EUC is located between 1200W and 130 0 W and at a depth of 80 meters.
The vertical velocity (Fig. 2c) shows upwelling along the equator except in the most western
part. The upwelling maximum is located between 60-80 meters at 120o-130 o W and its strength
is about 3.2 m/day. The values are close to the observations of Hansen a.nd Paul (1987) of
2.6m/dayat 50 meters for 1.5°N-1.5°S/130oW-I00oW.
Figure 3a shows an equatorial longitude-time diagram of the simulated equatorial SST and
figure 3b shows the COLA-CAC observed SST, for the peri~d 1979-93. Seasonal and interannual
variability is well simulated both in the eastern and western Pacific. Also the EI Nino (1982/83,
1986/87 and 1992/93) and La Nina (1983/84) events are qualitatively captured. More high
frequency variability can be seen in the model results than in the observations. Cold periods
less than a few months are evident in the east Pacific. In 1982/83, 1986/87, 1991/92 and 1993
warm SSTs cross the Pacific. This can be seen in model results and observations.
Interannual SST anomalies averaged over the Nino 3 region (90 0 W-150oW, 5°N-5°S) for the
period 1979-93 are shown in Figure 4 and are compared with the observed COLA-CAC data.
Model anomalies are less intense in the two warm events of 1982/83 and 1986/87, but for the
EI Ninos 1992/93 the anomalies are in quite good agreement. The cold event of 1988 is not well
simulated and appears later in 1989. More high frequency variability is present in the model
results.
To summarize, the experimental results appear to be in reasonable agreement with observations
for the time-mean, seasonal cycle and interannual variability. However, further analysis is needed
before using the results in the development of the statistical model.
4 Results.
The experiment time-mean SST and wind stress is presented in Figure la and shows a realistic
mean state. The cold tongue extends further 1700W and in its most western part is centered
at equator. A well developed warm pool can be seen in the west with SST around 29°C. Warm
temperatures (>28°C) penetrate eastward both south and north of the equator. Figure Ib and
lc show the April and October means for SST and wind stress. In April, a larger warm pool,
penetrating further east, can be observed. In April, the warm region in the eastern Pacific north
of the equator is warmer and a weakened equatorial SST gradient can be seen. In October, a
well-developed cold tongue exists with eastern Pacific temperatures as low as 22°C.
Figure 2 shows an equatorial longitude-depth section of time-mean temperature (Fig. 2a), zonal
current (Fig. 2b) and vertical velocity (Fig. 2c). The slope of the isotherms across the Pacific is
well simulated. Comparing these results with the coupled ARPEGE/OPA experiment (Terray
et al., 1994) a shallower thermocline in the east, closer to observations, is obtained perhaps due
to the stronger zonal wind stress. In the western equatorial Pacific the thermocline is deep and
a uniform temperature layer of about 60 meters deep can be seen. An Equatorial Under Current
(EUC) exceeding 1 m/s can be seen and is comparable with other forced (Belam.ari, 1993 and
Dandin, 1993) and coupled (Terray et a.l., 1994) experiments undertaken with the OPA ocean
model. The core of the EUC is located between 1200W and 130 0 W and at a depth of 80 meters.
The vertical velocity (Fig. 2c) shows upwelling along the equator except in the most western
part. The upwelling maximum is located between 60-80 meters at 120o-130 o W and its strength
is about 3.2 m/day. The values are close to the observations of Hansen a.nd Paul (1987) of
2.6m/dayat 50 meters for 1.5°N-1.5°S/130oW-I00oW.
Figure 3a shows an equatorial longitude-time diagram of the simulated equatorial SST and
figure 3b shows the COLA-CAC observed SST, for the peri~d 1979-93. Seasonal and interannual
variability is well simulated both in the eastern and western Pacific. Also the EI Nino (1982/83,
1986/87 and 1992/93) and La Nina (1983/84) events are qualitatively captured. More high
frequency variability can be seen in the model results than in the observations. Cold periods
less than a few months are evident in the east Pacific. In 1982/83, 1986/87, 1991/92 and 1993
warm SSTs cross the Pacific. This can be seen in model results and observations.
Interannual SST anomalies averaged over the Nino 3 region (90 0 W-150oW, 5°N-5°S) for the
period 1979-93 are shown in Figure 4 and are compared with the observed COLA-CAC data.
Model anomalies are less intense in the two warm events of 1982/83 and 1986/87, but for the
EI Ninos 1992/93 the anomalies are in quite good agreement. The cold event of 1988 is not well
simulated and appears later in 1989. More high frequency variability is present in the model
results.
To summarize, the experimental results appear to be in reasonable agreement with observations
for the time-mean, seasonal cycle and interannual variability. However, further analysis is needed
before using the results in the development of the statistical model.