a. Current around the Luzon Strait
Focusing on Kuroshio intrusion, Figure 11 shows the current velocity
distributions at 50 m and SSH around the Luzon Strait. The figure contains 20
panels, 5 rows by 4 columns. Four columns, from top to bottom, represent the
current velocity on 15th of March, June, September and December, while 5 rows,
from left to right, represent years 1997, 1998, 1999, 2000, and 2001. The current
velocity distributions in 1997 are used to illustrate the annual evolution. The
greatest Kuroshio intrusion into the South China Sea occurred in March. The South
China Sea cyclonic flow with strong low SSH center was found immediately west of
the southern Luzon Strait. The intruded Kuroshio flowed with the South China Sea
cyclonic flow. A small portion of the Kuroshio spilled out, entering the Taiwan
Strait. In June, the Kuroshio consistently intruded into the South China Sea and
merged with the cyclonic flow. However, when the South China Sea cyclonic flow
was weak and shifted to the west, the strength of the Kuroshio in the Luzon Strait
weakened, and the current on the northern shelf break of the South China Sea
intensified. The westward shift of the South China Sea cyclonic flow opened a
space that allowed a portion of intruded Kuroshio to turn clockwise south of the
Taiwan Strait. This clockwise flow either entered into the Taiwan Strait or poured
back into the Kuroshio. However, the main source of Taiwan Strait current was from
the northern South China Sea shelf break. In September, the South China Sea
cyclonic flow and low SSH moved further west. West of the Luzon Strait, the flow
became complicated. The Kuroshio in the Luzon Strait also weakened, but it still
intruded into the South China Sea. A large portion of intruded Kuroshio retroflexed
in a clockwise motion, south of the Taiwan Strait, and then partially entered the Strait.
Some exited the South China Sea by flowing out through the northern Luzon Strait.
In December, the South China Sea cyclonic flow re-intensified itself, and expanded
throughout the basin. The speed of the Kuroshio in the Luzon Strait increased, and
the amplitude of the current in the northern South China Sea shelf break decreased.
An annual cycle was completed. Obviously, the location and strength of South
China Sea cyclonic flow has large impact on the annual variation of Kuroshio
intrusion and the flow west of Luzon Strait. Although the pattern of annual evolution
was basically repeated in subsequent years, there were some differences. The most
visible difference was that the South China Sea cyclonic flow (or low SSH center)
was much stronger in 1997 than in 1998, except for December. The weak SSH
center regained its strength gradually from 1999-2001. This interannual variation of
South China Sea Cyclonic flow caused the variation of Kuroshio intrusion and flow
west of Luzon Strait. For example, the low SSH center was greatest and closest to
Taiwan in March of 1997. It caused the Kuroshio to intrude more northward,
directly impinging on the southern opening of the Taiwan Strait. The small
anti-cyclonic flow that was consistently seen south of Taiwan Strait disappeared.
b. Volume Transport across Luzon Strait
Using the model output and taking a slice along 120.75°E from 18.5 to 22°N, the
depth-averaged U in the upper 300 m water column and SSH as a function of time and
latitude, are shown in Fig. 12. The 3 straight lines indicate mooring locations. The
distribution of U in the Luzon Strait could be separated into 3 parts. The U at the
northern and southern parts was essentially eastward, flowing out of the South China
Sea. It was generally quite stable and the annual and interannual variations were
small. The U in the central Luzon Strait varied annually as well as interannually.
The westward current velocity was high in winter and low in summer. The model
result suggests that annual variations at L2 should be obvious, but at L1 and L3, they
might be difficult to discern. However, the annual variations could be hidden by the
interannual variations. The annual variations, even in the model output, were
unclear in 1997 because of the large interannual variations. This might explain why
the observations at L2 in 1997 did not reveal annual variations.
Both annual and interannual SSH variation was small in the northern Luzon
Strait, and large in the central southern Luzon Strait, with the lowest and highest SSH
generally appearing in November-December and July-August, respectively. For the
interannual, the SSH appears to have had its lowest value early in 1997, increasing
for the subsequent two winters, and finally decreasing again for the next two winters.
The lowest SSH had no significant differences between 2000 and 2001. The
position of the lowest SSH moved interannually, but only slightly. The
development/ contraction of South China Sea cyclonic flow is related to these
variations.
Integrating the modeled U in the upper 300 m along 120.75°E from 18.5 to 22°N,
Figure 13 shows the time series of zonal transport across the Luzon Strait. The
negative and positive values represent the westward and eastward transport,
respectively. The zonal transport across the Luzon Strait displayed a clear annual
variation. The massive westward Kuroshio intrusion occurred from November
through December, when the SSH at the central southern Luzon Strait also reached its
lowest value. Its maximum amplitude was over -6 Sv. The westward intrusion
generally stopped, and even reversed in summer. Eastward transport in summer was
not only related to the reduction of Kuroshio, but was also related to the increased
flow in the northern South China Sea shelf break. This shelf flow eventually worked
its way out of the South China Sea through the northern Luzon Strait. The annual
variation of zonal transport in the upper 300 m across the Luzon Strait generally
varied from 0.2 to –5.4 Sv. The zonal volume transport also displayed an
interannual variation. In 1997, the major westward intrusion was extraordinarily
large through late March. No reversed zonal transport was found in the summer.
Westward transport developed unusually small amplitude in winter. Large eastward
transport observed in the summer of 1998. Since then, the annual evolutions in
subsequent years were similar, but the transport gradually increased westwardly. In
2001, the eastward transport was barely seen in summer, and the amount of annual
westward transport was close to that of 1997. The amplitude of interannual variation
of zonal transport could reach as large as 3 Sv. The El Niño occurred in 1997
(Boulanger and Menkes 1999; McPhaden 1999; McPhaden and Yu 1999; Kutsuwada
et al. 2002) and was predicted to occur in 2002 (CPC/NCEP 2002). Apparently, the
interannual variation of zonal transport across the Luzon Strait may have a cycle with
El Niño.
To explore the mechanism causing the variation of zonal transport across the
Luzon Strait, the zonal Ekman transport across the Luzon Strait was calculated first
using the ECMWF wind stress. Figure 14 shows the principle component
(northeast-southwest) of ECMWF wind stress at the center of the Luzon Strait and
zonal Ekman transport across the Luzon Strait. The wind stress varied annually, as
well as interannually. The northeast monsoon, prevalent from September through
April, displayed larger amplitude than the southwest monsoon, which was dominant
from May through August. The monsoon was weaker starting in the late El Niño
Year (1997) and remained weak for about one year. These variations have been
described in detail by several oceanographers, including, Chao and Shaw (1996),
Liang et al. (2000), etc. The zonal Ekman transport, calculated directly from the
wind stress, also varied annually as well as interannually. It had good correlation
with the total zonal transport, but its amplitude was much smaller. The contribution
of Ekman transport to the total zonal transport was generally less than 15%, which is
close to the estimate of Qu (2000). Obviously, some other mechanism is important
for causing the zonal transport variation across the Luzon Strait.
Figure 15 shows the SSH west of the southern tip of Taiwan (marked as A), SSH
west of the northern tip of Luzon (marked as B) and SSH differences between A and
B. Both the model results and the TOPEX/Poseidon data from the WOCE have been
demeaned for comparing. In general, their variations have great resemblance in both
except the intraseasonal fluctuations. In model results, the SSH at A had little
variation; while at B it displayed clear annual and interannual variation.
Consequently, the difference of SSH between A and B also showed annual and
interannual variation. The variation of SSH differences correlates (correlation
coefficient is 0.9) well with the modeled zonal transport across the Luzon Strait.
The SSH difference varied from 0 to 26 cm. It could generate the zonal geostrophic
transport having the same order of the modeled zonal transport across the Luzon Strait.
Apparently, the meridional pressure gradient across the Luzon Strait was the primary
factor causing the intrusion of Kuroshio. Qu (2000) and Metzger and Hurlburt (1996)
had similar findings. However, the present finding indicates that the SSH difference
was mainly due to low SSH north of Luzon. A conclusion, similar to the previous
one, could be reached. The location and strength of South China Sea cyclonic flow
(or low SSH center) has great impact on the Kuroshio intrusion across the Luzon
Strait.