Journal of Ocean University of Oingdao (Oceanic and

Coastal Sea

Research)

ISSN 1671-2463, October 31,2002, VolA, No.2, pp.130-134 http: // www. ouqd. edu. cn /xbywb / E-maih xbywb@mail, ouqd. edu. cn

Variabilities of Surface Current in the Tropical Pacific Ocean WANG Qi 1)'*, HU Ruijin 1), Anna Zaklikowski 2) 1) Department of Marine Meteorology, Ocean University of Qingdao, Qingdao 266003, P. R. China 2) Department of Civil and Environmental Engineering, Clarkson University, Potsdam, USA

(Received September 11,2002 ; accepted October 16,2002) A b s t r a c t The Simple Ocean Data Assimilation (SODA) package is used to better understand the variabilities of surface current transport in the Tropical Pacific Ocean from 1950 to 1999. Seasonal variation, interannual and decadal variability analyses are conducted on the three major surface currents of the Tropical Pacific Ocean: the North Equatorial Current (NEC), the North Equatorial Countercurrent (NECC), and the South Equatorial Current (SEC). The transport of SEC is quite larger than those of NEC and NECC. The SEC has two maximums in February and August. The NEC has a small annual variation. The NECC has a maximum in October and is very weak in March and April. All currents have remarkable interannual and decadal variabilities. The variabilities of the NEC and the SEC relate to the winds over them well, but the relationship between the NECC and the wind over it is not close. Analysis related to E1 Nifio-Southern Oscillation (ENSO) suggests that before El Nifio (La Nifia) the SEC is weaker (stronger) and the NECC is stronger (weaker), after E1 Nifio (La Nifia) the SEC is stronger (weaker) and the SEC is weaker (stronger). There is no notable relationship between the NEC and ENSO.

Key words the North Equatorial Current ( N E C ) ; Current (SEC); the Tropical Pacific Ocean Number

the North Equatorial Countercurrent ( N E C C ) ;

the South Equatorial

ISSN 1671-2463(2002)02-130-05

1 Introduction The surface currents of the Tropical Pacific Ocean are characterized by three maior equatorial currents: the North Equatorial Current (NEC), the North Equatorial Countercurrent ( N E C C ) and the South Equatorial Current (SEC). Both the N E C and the SEC travel from east to west in response to the easterly winds, while the NECC travels from west to east. The NECC, serving as the boundary, is located between the NEC and the SEC (Talley et al., 1998). Apparently, these three major surface oceanic currents are very important for mass and energy transports in the Tropical Pacific Ocean, thus their variabilities can affect upper ocean thermal structures, which directly relates to E1 Nifio. As the NEC and the SEC connect to the midlatitude, they are also the oceanic passages of the interaction between tropical and mid-latitude regions. However, due to lack of sufficient data, the surface oceanic currents in the Tropical Pacific Ocean has been rarely investigated. Donguy and Meyers (1996) gave a rough review for such studies: the Hawaii-Tahiti Shuttle Experiment provided section information on the Equatorial Current System (Cantos-Figuerola and Corresponding author. E-mail: wangqi@ouqd, edu. cn Tel : 0086-532-2032592 (office) ; 0086-532-2875108 (home)

Taft, 1983) ; Wyrtki and Kilonsky ( 1 9 8 4 ) studied westward currents; other authors used thermal data to document the currents on individual sections (Meyers and Donguy, 1984; McPhaden et al., 1988; Harrisson et al., 1989). Donguy and Meyers (1996) used Expendable bathythermograph ( X B T ) data and the climatological temperature/salinity relationship to study the mean annual variability of dynamic height and geostrophic transport of the major currents in the Tropical Pacific Ocean. They concluded that the NECC had a large annual cycle with a transport-maxim u m during boreal fall and winter. Seasonal variations of NEC are small. Seasonal variations of SEC are slightly larger, and they have considerably different phases from track to track. Recently the Simple Ocean Data Assimilation ( S O D A ) package, which combines a vast amount of information and data, was developed (Carton et al., 1999). It allows us to use this data package to study some variability characteristics of the surface oceanic currents in the Tropical Pacific Ocean. In order to make a preliminary analysis of the variability of the Pacific Ocean, it is important to integrate the fundamental knowledge of oceanic and atmospheric processes. This p a p e r examines the Pacific O c e a n ' s mean sea surface temperature ( S S T ) , wind stress field, and the three major currents in the Tropical Pacific Ocean: the NEC, the NECC, and the SEC on a seasonal basis. Their anomalies will be analyzed by using wavelet method to determine their main peri-

WANG Q. et al.: Variabilities

of Surface Current

in the Tropical Pacific Ocean

131

ods and amplitudes, and the statistical correlation between anomalies will be calculated.

nent (averaged over upper 2 levels, that is upper 30 m) and SST are employed in the present study.

2 Data

3 Seasonal Variability Analysis

The SODA package was developed by Carton et aL (1999) to conduct a retrospective analysis of several components of the global ocean over a period of 50 years. Data was collected monthly from January 1950 to December 1999. The data set includes information regarding global sea temperature, salinity, zonal velocity components, wind stress components, diagnosed sea level, depth of the 2012 contour ( m ) , and heat content. SODA combines both observation and simulated data to compile a complete data set. The domain of SODA is global with 1~ • 1 ~ latitude-longitude horizontal resolution in midlatitude and 0.45 ~ • 1 ~ resolution in the tropic. The model has 20 vertical levels on a stretched grid with 15 m resolution near the surface. Wind stress, surface current velocity zonal compo-

Fig.1 shows the mean SST, wind stress, and surface current velocity zonal component averaged over the 50 years period. The extent ' w a r m pool' is very clearly shown in the western Pacific region, as distinguished by the dark gray color. The 'cold tongue' in the eastern Pacific can be identified by the light gray color. The black contour lines are used to represent current velocity zonal component (interval is 5 cms 1), with solid lines indicating a positive direction of flow (west to east) and dashed lines denoting a negative direction of flow (east to west). White vectors represent wind stress. This diagram clearly displays the location of the NECC, centered at approximately 5~ The NEC and the SEC can be seen directly in the north and the south of the NECC respectively.

30~ 25 ~ 20 ~ 15 ~ 10 ~ 5o EQ 5o 10 ~ 15 ~ 20 ~ 25 ~ 30~ 120~

150 ~

180 ,

18

150~

120 ~

90 ~

]~[]11111

20

22

24

26

28

30

Fig.1 The long-term mean state. The gray colors indicate SST; the contours (interval 5 cms-1) denote zonal component of current velocity, solid lines representing speed from west to east and dashed lines representing speed from east to west; white vectors show wind stress. The NEC is from about 20~ to 8~ and its westward transport increases regularly toward the west. The NECC is from about 8~ to 3~ and there are three maximum eastward transport centers: near western boundary, near eastern boundary and near 135~ The SEC is from about 3~ to 20~ and its westward transport is uniformly strong near the equator. On the west, about 6~ to 10~ there is a weak South Equatorial Countercurrent ( S E C C ) which divides the SEC into two parts. In Fig.1 it is very clear that main currents are asymmetry about the equator. The SEC crosses the equator and reaches 3~ This may be the result due to the in-

tertropical convergence zone ( I T C Z ) located in the North hemisphere. Fig.2 shows the seasonal departures for four different months representing the four seasons: January, April, July, and October. Darkly shaded colors indicate temperatures higher than the annual average, while lightly shaded colors suggest a SST lower than the average. Direction and magnitude of wind stress vectors represent a change relative to the mean state in direction and strength. The current contours represent the original climatological currents, with dashed contours meaning the east to west flow and solid contours the west to east flow.

Journal of Ocean University of Qingdao

132

2002, Vol. 1, No. 2

Based on Figs.1 and 2, we make use of SODA to calculate the transport of the major currents and corresponding windforce. Along each longitude within 170 ~ W and 130~ We sum up for the NEC from 5~ to 30~ every negative current speed multiplied by a longitude width and the mean depth (30 m ) . Thus we get the total NEC transport along this longitude. T h e n average such transports from 170~ to 130~ and get an overall mean transport in the study area,

By examining the temperature departures, it can be seen that the cold tongue has the lowest temperature during October and the highest temperature during April. During the boreal winters the trade winds in the north equatorial Pacific strengthen (vectors pointing west) while in the south equatorial Pacific, they weaken (vectors pointing east). In the boreal summers, the trade winds in the north equatorial Pacific weaken (vectors pointing east) while in the south equatorial Pacific, they strengthen (vectors pointing west). In April, the NECC is almost completely absent, which is indicated by the lack of solid black eastward contours. This is a result of the strengthened northern trade winds that resist the eastward NECC. The NECC is at the strongest stage during October when the trade winds weaken. Analogously, the N E C strengthens during the boreal winter when the trade winds are the strongest and weakens during the boreal summer months when the trade winds weaken. T h e SEC has the opposite behavior since winter and summer are reversed in the two Hemispheres. Fig.2 is consistent with the conclusions made by Donguy and Meyers (1996).

or, T =

(1/n)~(~u•215

where T i s t h e

current transport, n is the number of longitudes, u is the current speed, Ay is latitude width and d is the surface layer thickness (30 m ) . Meanwhile we get wind force over the NEC in unit longitude. T h a t is:

F = (1/n)~(~r

•

• Ad),

where F is the

windforce, r is windstress zonal component and Ace is unit longitude width. In the same way we calculate the transport and windforce for the SEC from 30~ to 5~ and for the NECC from 2~ to 15~ using only positive current speed for the latter current. We define

T.,,r~h = T / ~ Ay and f strenggh = F~ ~ /~y as the transport strength and windforce strength respectively.

25~ 26 ~ 15 ~ 10 ~ 5o EQ 5o 10 ~ 15 ~ 20 ~ 25~ V 0.4

~!!!!!!!}!!!!}!!!! -3 -2

~} -1

0

1

2

3

Fig.2 The seasonal departures of SST (gray color) and windstress (white vector) from long-term mean; the climatological ocean currents (contour lines as indicated in Fig.l) Fig.3 shows the seasonal variations of current transports and windforce defined above. Note that all signs of transports and winds are negative except transport of NECC, but in Fig.3, these signs are changed to positive. From Fig.3 we can see that the transport and strength of SEC are significantly larger than those of NEC and NECC, and the windforces over the NECC are the smallest. The N E C transport has small annual variation with a m a x i m u m in April and a minimum in August. But the annual variation of N E C ' s strength is

rather large and this implies that the width annual variation of NEC is not so small. The windforce over the NEC has large annual variation with also a maxim u m in April and a minimum in August. So apart from amplitude, the annual variation of the NEC is quite consistent with the wind over it. For the NECC, both annual variation and strength of transport have a m a x i m u m in October and a minimum in March. As for the windforce over this current, it has a maximum in September and a minimum in Febru-

WANG Q. et al.: Variabilities of Surface Current in the Tropical Pacific Ocean ary. This explains why as the easterly wind over the NECC is weak, the NECC gets strong subsequently. The windforce is weak in April and this cannot explain the strong NECC during this period. It is interesting that the SEC transport has 2 maximums in February and August while the wind over it has only one maximum near August. It remains to explain why there is a SEC maximum in February. 12.0

;x ,

SEC

\

10.0

/

\~

8.o NECC

6.0 NEC

0.3

/ \-

\

/

0.2

~-~-'~

'~ 4.0

sEc, \

NECC . . . . . . . --

NEC 0.1

2.0 0.0

FMA

MJ

1.6

J A S OND

0.0

FMAMJ

J A S OND

SEC

SEC

0,04

NEC

/"

1.2

0.03 0,8 0.4 " ' ' - , 0.0

NECC

FMAMJ

JASOND

Time(month)

/"

most notable character is that the periods of strength of the strongest NEC become longer from the end of the 1970' s. The third period band of high energy is about 10 years or so, which shows a steady increase in the strength of oscillation from the 1960' s, reaching its maximum around 1990. The strength of NECC anomaly is larger than that of NEC. The variation strength of about 10 year period increases steadily from the 1 9 5 0 ' s , with a maximum around 1995; The maximum signals of interannual variability appear in the 1 9 7 0 ' s and late 1 9 8 0 ' s (not shown). The spectrum of SEC anomaly is similar to that of NECC, except for the spectrum strength being the largest (not shown). Table 1 gives some correlation results for current transport and windforce, In which 'SSTa in Nino3' means SST anomaly in the area of Nino3 (5~ 5~ 150~176 . Table 1

0.02

T h e correlation b e t w e e n current t r a n s p o r t ,

windforce and SST anomalies

0,0l 0.00

133

Term

i~M'A~J'J;~gd~b Time(month)

Fig.3 T h e seasonal variations of current transport ( u p p e r l e f t ) , transport s t r e n g t h ( u p p e r - r i g h t ) , windforce over currents ( l o w e r - l e f t ) a n d windforce s t r e n g t h ( l o w e r - r i g h t )

It can be observed that the NECC and the NEC are out of phase. During the earlier half year (from January to June), the NECC gets weak after the SEC gets strong. But during later half year (from July to December), the situation is reversed.

X NEC Transport NECC Transport SEC Transport

y

Maximum Correlation Value~

Lag Time (months)

NEC Wind

0.43

0

NECC Wind

0.22

X lags Y, 1

0.46

0

SEC Wind

NECC Transport

NEC Transport

- 0.22

0

NECC Transport

SEC Transport

0.29

0

NEC Transport

SEC Transport

0.13

X lags Y, 1

SEC Transport

SSTa in Nino3

0.29

X leads Y, 2

SEC Transport

SSTa in Nino3

- 0.37

X lags Y, 8

NECC Transport

SSTa in Nino3

0.21

X leads Y, 2

NECC Transport

SSTa in Nino3

- 0.29

X leads Y, 7

t All values are significant at levels exceeding 99 %.

4 Interannual Variability Analysis To further illustrate the interannual variabilities of NEC, NECC and SEC, wavelet analyses were performed. Fig.4 is the wavelet spectrum for NEC anomaly (Only periods of less than 32 years are drawn). It shows that the signals of interannual period are remarkable. One period band of high energy is around 1 year to 2 years, with several obvious maximum values appearing in middle 1950's, early 1970's, and early 1980's. Another period band of high energy is from 2 years to 6 years. For this interannual period band, the

Fig.4 W a v e l e t power spectrum for N E C t r a n s p o r t

From Table 1 one can see that both the NEC transport and the SEC transport have close correlations with the wind over them, the correlation coefficients being 0.43 and 0.46 respectively. This implies that variabilities of the NEC and the SEC are mainly due to the responses of currents to the winds over them. Because the stronger (weaker) NECC transport is characterized by having a positive (negative) anomaly and a positive (negative) wind anomaly represents a weaker (stronger) wind over the NECC, the correlation coefficient 0.22 signifies that after a weaker (stronger) wind over the NECC, there is a stronger (weaker) NECC transport and the strongest response occurs one month later. T h e negative correlation coefficient of - 0 . 2 2 between the NECC and the NEC anomalies suggests that to some degree a stronger (weaker) NEC is accompanied with a stronger (weaker) NECC. It remains not clear that the correlation coefficient between the NECC and the SEC anomalies is positive. This positive correlation suggests that the stronger ( w e a k e r ) SEC is accompanied with the weaker ( s t r o n g e r )

134

Journal of Ocean University of Qingdao

NECC. The NEC and the SEC have less correlation (the correlation coefficient is only 0.13). Finally we discuss the relationship between surface currents and the E1 Nifio-Southern Oscillation ( E N SO) events. We take Nino3 SST anomaly as an ENS O ' index. The correlation coefficient between the SEC and E N S O is 0.29 when the SEC leads E N S O by two months and - 0 . 3 7 when the SEC lags behind E N S O by eight months. This suggests that before E1 Nifio (La Nifia) the SEC is weaker (stronger) and after E1 Nifio ( L a Nifia) occurs the SEC is stronger ( w e a k e r ) . The correlation coefficient between the NECC and E N S O is 0.21 when the NECC leads ENSO two months before and - 0 . 2 9 when the NECC lags behind the E N S O seven months. It suggests that before El Niflo ( L a Nifia) the NECC is stronger (weaker) and after E1 Nifio ( L a Nifia) the SEC is weaker ( s t r o n g e r ) . There is no notable relationship between the NEC and ENSO.

5 Conclusion The seasonal and interannual variabilities of the three major upper layer currents in the Tropical Pacific Ocean, the NEC, the NECC, and the SEC, were studied in this paper. The seasonal variation of the transport of the SEC is significantly larger than those of the NEC and the NECC. The S E C ' s variation has two maximums in February and August but the wind over it has only one maximum around August. The NEC has small annual variation with a maximum in April and a minimum in August. The NECC has a maximum in October and there is no NECC in the central Pacific during March and April. Analyses show that these currents have remarkable interannual and decad~al variabilities. The variabilities of the NEC and the SEC relate to the winds over them well and this suggests that these currents are controlled mainly by wind. But the relationship between the NECC and the wind over it is not so clear, so the cause of the NECC variability needs further investigating.

2002, Vol. 1, No. 2

It is shown that before E1 Nifio ( L a Nifia) the SEC is weaker ( s t r o n g e r ) and the N E C C is stronger (weaker), after E1 Nifio (La Nifia) the SEC is stronger (weaker) and the SEC is weaker ( s t r o n g e r ) . No notable relationship exits between the N E C and ENSO.

Acknowledgements This research was supported by the National Natural Science Foundation of China ( G r a n t Nos. 40176003 and 40136010). Anna Zaklikowski was stlpported by the funding of the U . S . National Science Foundation and she would like to thank Dr. H . T . Shen and Dr. H . H . Shen of Clarkson University, and Mr. Y. Liu of the Ocean University of Qingdao.

References Cantos-Figuerola, A., and B.A. Taft, 1983. The South Equatorial Current during the 1979 80 Hawaii-Tahiti Shuttle..Tropical Ocean-Atmosphere Newsletter, 19: 6-8. Carton, J.A., G. Chepurin, Cao, X., and B. S. Giese, 1999. A simple ocean data assimilation analysis of the global upper ocean 1950-1995, part 1: methodology. J. Phys. Oceanogr., 30: 294-309. Donguy, J.R., and G. Meyers, 1996. Mean annual variation of transport of major currents in the Tropical Pacific Ocean. Deep-Sea Res., 43(7): 1105-1122. Harrisson, D.E., W.S. Kessler, and B.S. Giese, 1989. Ocean circulation model hindcasts of the 1982-1983 E1 Nifio: thermal variability along the ship-of-opportunity tracks. J. Phys. Oceanogr., 19: 397-418. McPhaden, M.J., A.J. Busalacchi, and J. Picaut, 1988. Observations and wind-forced model simulations of the mean seasonal cycle in tropical Pacific sea surface topography. J. Geosphys. Res., 93(c7): 8131-8146. Meyers, G., and J.R. Donguy, 1984. The North Equatorial Countercurrent and heat storage in the western Pacific Ocean during 1982-83. Nature, 5991: 258-260. Talley, L.D., G. Fryer, and R. Lumpkin, 1998. Physical

oceanography of the Tropical Pacific, Geography of the Pacific Islands. Bess Press, pp. 19-32. Wyrtki, K., and B., Kilonsky, 1984. Mean water and current structure during the Hawaii-Tahiti Shuttle Experiment. J. Phys. Oceanogr., 14: 242-254.