Seasonal variation in he Yellow Sea derived from satellite a

Pergamon

ContinentalShelf Research,Vol. 17, No. 6, pp. 655~564,1997© 1997ElsevierScienceLtd Printed in Great Britain. All rightsreserved PII: S0278-4343(96)00054-4 0278-4343/97$17.00+0.00

Se~tsonal variation in surface circulation of the East China Sea and the Yellow Sea derived from satellite altimetric dataT E T S U O Y A N A G I, * A K I H I K O M O R I M O T O * and KAORU ICHIKAWA*(Received 8 March 1996;accepted 30 August 1996)

Abstract--Seasonal variation in the surface circulation of the East China Sea and the YellowSea is investigated using altimetric data of TOPEX/POSEIDON and numerical model output. In the Yellow Sea an anticlockwise circulation develops during summer and a clockwise one during winter In the East China Sea an anticlockwise circulation occurs during winter. Such results coincide well with those obtained by the numerical experiment and drifter buoys experiment.© 199"7Elsevier Science Ltd. All rights reserved

1. I N T R O D U C T I O N Satellite altimetry has been very useful for studies of open ocean dynamics (e.g. Ichikawa and Imawaki, 1994) but has not been used for the study of coastal ocean dynamics. This is due to the fact that the tidal signals cannot be correctly eliminated from altimetric data in the coastal sea where the tidal signal and its spatial gradient are very large and where the existing global tidal models do not have sufficient accuracy. We have developed a new procedure:for correctly estimating tidal signals in the coastal sea from the altimetric data of T O P E X/ P O S E I D O N (Yanagi et al., 1997). In this paper, we try to determine the seasonal variation in the sea surface circulation of the East China Sea and the Yellow Sea by use of the altimetric data of T O P E X/ P O S E I D O N after correctly eliminating tidal signals. 2. A L T I M E T R I C D A T A PROCESSING

The satellite T O P E X/ P O S E I D O N was launched in August 1992 and has continued to obtain the altimetric data about every 10 days along seven observation lines in the East China Sea and the Yellow Sea as shown in Fig. l(b). The data are provided as Merged Geophysical Data Record ( M G D R ) by the Physical Oceanography Distributed Active Archive Center at Jet Propulsion Laboratory, U.S.A. We used the data taken by only T O P E X altimeter from Cycle 1 (September 1992) until Cycle 108 (August 1995) in this analysis. Since the P O S E I D O N altimeter had an unknown bias of the order of 20 cm to T O P E X one, we did not use the data taken by P O S E I D O N altimeter in this study (this *Department of Civil and Ocean Engineering, Ehime University, Matsuyama790, Japan. 655

656

T. Yanagi et al.

©©

©

oO

v

Z

Z

o

o

©

z

~

z

~~

,~

,~

.~~

i:~~

~ -~~ j~

Satellite altimetric data

657

problem was solved by NASA in July 1996 and we can also use POSEIDON data now). Standard data correction including electromagnetic bias correction, ionospheric correction, dry and wet trospheric correction

and solid earth tide correction were made using the values provided in the MGDR. After these corrections, tidal signals of M2,$2, K2, N1, O1 and P~ constituents, which have more than 10 cm amplitude in the East China Sea and the Yellow Sea, are eliminated on the basis of the results of Yanagi et al. (1997). As an example, the sea surface height anomaly from the mean sea surface during the whole observed period derived from the altimetry data (SSHA), estimated tidal signal (TIDE) and the sea surface height anomaly after the elimination of tidal signals (Corrected SSHA) at Sta A in the central part of the Yellow Sea[see Fig. l(b)] are shown in Fig. 2. Temporal variation with the period of about 60 days is dominant in SSHA and this is due to the aliasing of Me constituent. The tidal signal is well eliminated from SSHA and the seasonal variation is dominant in Corrected SSHA as seen in Fig. 2, that is, Corrected SSHA is hi[gh in summer and low in winter. The altimetric data are originally obtained about every 1 s (corresponding to about every 6.2 km along the subsatellite track, which is the projection of satellite track on the sea surface) but the data points differ every cycle. Therefore, we linearly interpolated every data point in a cycle at the fixed points within a 6.2 km interval along the subsatellite track and nine data points are further averaged for the following procedure in order to reduce small-scale phenomena observed along tracks. The distribution of basic data points are shown in Fig. l(b). The estimated sea surface height after the tidal correction at the point r at time t, S(r,t), is expressed by the following formula:

S(r,t)=~ (r,t)+ (N(r)+ en(r)}+{es(r)+ 6r(O) -~-~rn(t)

(1)

where~(r,t) denotes the sea surface dynamic topography, N(r) the geoid, e,,(r) the error of geoid, e~(r)+ e~(t) the spatial and temporal components of orbit error and em(t) the measurement error. The sea surface dynamic topography~(r,t) is divided into the temporal averaged height~m(r) and the anomaly~'(r,t):

~(r,t)= Cm(r)+~'(r,t)

(2)

Altimetric data S(r,t) has to be divided into the temporal averaged value S,~(r) and the anomaly S'(r,t) due to that the accuracy of geoid data is not sufficient:

S(r,t)= Sm(r)+ S'(r,t)From equations (1) to (3), we get

(3)

S'(r,t)=~'(r,t)+ l?r(0 q- Sm( 0

(4)

We ignore the temporal average of er(t ) and em(t) because they are small.~'(r,t) is divided into the temporal mean during the period of q,~'mq(r), and the anomaly~"(r,t):

¢'(r,t)=¢~nq(r)+~"(r,t)Substituting equation (5) into equation (4), we get

(5)

S'(r,t)

~'mq(r)+~"(r,t)= H'q(r)+ E'(r,tq)=

+

e~(t)

+

gm(t)(6)

658

T. Yanagi

et al.123.296

(a)LAT. 34.850 LON.

(m)SSHA1

0

t'vv~~V~il,i "l ......... I ......... I ......... ! ......... ! ......... ! ......... I ''''|''' '1 ........

A,, II,,30 40 50 60

,~~/~ I~170 80 90 100

-10 10 20

cyc I e

(b)

(~)TIDE

]/~.,A~f~/,/r,, RA./~]'/ vI,V'vv~'

vv~l I~I-10 10 20 30 40 50 60 70 80 90 100

c yc 1 e

(c)

(m)

0.1°'~

o.,O-'1:.:::.. . o°°e 0 ................ lO 0 20 30 40

CORRECTED

".'::'::60 70 80

SSHA

e~e

...'o

.o.-o.-o-, i, 1.ii)i:::?-0.4

,. . . . . . . . 50

,.......,...

i i.i: .i:::::,....., ... 90 I00

cvcle

'O'N'D''F'~'A'~'j'A'S'O'N'D' j j' j'F'~'A'~'$'A'S'O~N'D' j' j'F'~'A'~'j'A j' ''1993 1994 1995

Fig. 2. (a) Observed sea surface height anomaly (SSHA) by TOPEX altimeter. (b) Estimated tide (TIDE). (c) Sea surface anomaly after the tidal correction (Corrected SSHA) at Sta A shown in Fig. 1. where

I-I'q(r)=~'mq(r) E'(r,tq)=~"(r,t)+ er(t )+ em(t )

(7) (8)

W e e s t i m a t e H'q(r) w i t h t h e e r r o r o f E' (r,tq) b y t h e o p t i m a l i n t e r p o l a t i o n m e t h o d ( I m a w a k i et al., 1 9 9 1 ) w i t h u s e o f o b s e r v e d S'(r,tq):

Satellite altimetric data

659

(a)I II I

40x

x

x

~~

20

~x x

x

O

xx x x x

~~

0~

~0~ v× . .~

~x~

xxnx~

~x x x

~~

~ -~0~~

~r=0.74

t i i t=0.9

10I

-40I~ I~

b ia, I

s=

0.6

-40

-20

0

20

40

TIDE

GAUGE

(b)40

(cm)

tr~ -20~ -400Ii i i

.10I i i

20i i i I

30I I I

40I i I

50I i i i

60i i i

70i i i

80i I[

CYCLEOND J F M A M J J A S O N D J F M A M J J A S O N D 1993 1994 Fig. 3. (a) Correlation of altimetric data and tide gauge data. (b) Temporal variation of altimetric data ( 0 ) and 10-day averaged tide gauge data (O) at Lusi (see Fig. 1). Bar in circle means the standard deviation of daily mean sea level data during 10 days.

S'(r,tq)= H~(r)+ E'(r,tq)N

(9) (101

H'q(r)=~ a~xS'(x,tq)x=l

Here, arx denotes the matrix which we want to get:N

arx=~s=l

CrxA~x I

(11)(121

Crs= W ( r -- Is)

where r denotes the point where we want to estimate H~(r), r~ the points with the number

660

T. Yanagi et al.

MEAN

40N

35N<

30N

120E Fig. 4.

125E

130E

LONG I TUDESeasonally averaged sea surface dynamic topography estimated by the diagnostic numerical calculation from Yanagi and Takahashi (1993).

of s[shown in Fig. l(b)] where the observed data exist. W(r - rs) denotes the covariance function of signal and it is expressed as W(IRI)= w~ exp{ - (]RI/L) 2} (13)

where IRI denotes the distance between two points, L the decorrelation scale (= 300 km in this case because the distance between neighboring subsatellite tracks is about 260 km) and Wo the magnitude of signal (= 7 cm in this case). Asx= W(r s - rx)+ q~(At) (14)

where Asx denotes the covariance function of obtained data at rs and r x, q~(At) the covariance function of noise and it is expressed by q~(at)=~g~(At) (15)

Here, o0 denotes the magnitude of noise (= 9 cm in this case), 6(At) is 1.0 at At= 0 and 0 in other cases. The magnitude of W0 and o o are decided on the auto-correlation function. The reason why a0 is large is that oo includes the temporal variation of the sea surface dynamic topography during about 10 days as show

n in Fig. 3(b), where the bar with an open circle denotes the standard deviation of daily mean sea surface variation during 10 days observed by tide gauge. The obtained H'q(r) is compared with the observed sea surface height at tide gauge station Lusi along the Chinese coast[see Fig. l(b)]. Sea surface height at tide gauge station

Satellite altimetric data

661

is obtained by 10-day averaging of the daily mean sea surface height anomaly from the mean sea surface during the same period of satellite observation. The result is shown in Fig. 3. The,correlation coefficient 0.74 between the altimetric data and tide gauge data is much better than that of 0.47 obtained along the southern coast of Japan with Geosat by Ichikawa arid Imawaki (1996) and better than that of 0.68 obtained at the tropical Pacific with Geosat by Cheney et al. (1989) but worse than that of 0.85 at the tropical Pacific with TOPEX/POSEIDON by Cheney et al. (1994). In order to estimate the absolute seasonal variation of sea surface dynamic topography, we have to know the seasonal averaged sea surface dynamic topography which corresponds to~,,(r) in equation (2). In this case, we adopt the seasonally averaged sea surface dynamic topography estimated by a diagnostic numerical model of the East China Sea and the Yellow Sea (Yanagi and Takahashi, 1993) which is shown in Fig. 4. The seasonally averaged sea surface dynamic topography is high along the Kuroshio and low along the western coast of Korea.

SPRING.~,,,,~,,,,,

SUMMER~~,~,,

~.)40N

( c m )

4<I~~

,

,/~

40N

\'

8-" 35NI

<\~

-<-

........ g::.-..:~"_

--*'/-'[I *.C'"~'~"

30N/

~,~~

120E

125E

130E

120E

F. . . .

)-'.

)I8

g~N

.... 12'5~ '

130E

LONG I TUDEAUTUMN ; . . . .~ . . . .~ ~

LONG ITUDE

WINTER~

....

(c

m)

f 4-0N

40N

',~ 2 0

_

.,

3 5 N

~-~ .< ..~

~i;::::L,,~./~::, .. !":=2 i~. 8 5 N

.

',,"v1;"

~

30N

30N

.

. . . . . . . . .

120E

125E

i30E

120E

125E

130E

LONG I TUDE Fig. 5.

LONG I TUDE

Seasonal variation in the composite sea surface dynamic topography estimated from altimetric data.

662SPRING

T. Yanagi et al.

,~,,,,I~,,,~o

~[~,,,~,,,,40N 40N

SUMMER

~~-,~.

..~

::l~::v35N

~

;::~U=" .........:~_~,(~:::2;;;;~30N

~~-,~-~T~,.~L~\\Nx~~ 1:~~-~.,~-~t,~t~~~_-=~Y-~r~:% z:~:~r-~--:

..,~,.~,

35N

30N

~

,,,z130E

~120E

,125ELONGITUDE WINTER

./130E

I~0E I~SE LONGITUDE

~ ....~ ....

AUTUMN

~I..,I,,,,~F-~. 40N

'~~11,

Ns\Jo..,~, . . . . . . .,

40N

.:::I~?:..35N

~ ' ' ' ' '- .~ * ''~D 35N

':iil;iii:,} .ll~.~mo~--~~~*III~30N"

30N

~o~Fig. 6.

i~'~"

i~ 130E

LONGITUDE

120E 125E LONGITUDE

130E

Seasonal variation in the surface geostrophic current estimated from altimetric data.

3. R E S U L T S The seasonal variation in composite dynamic sea surface topography, which is obtained by combining estimated H'q(r) and Fig. 4, is shown in Fig. 5. Here, spring means from March

to May, summer from June to August, autumn from September to November and winter from December to February. The estimated error E'(r,tq) is about 3 cm in the sea and about 5-6 cm near the coast throughout the year because the satellite altimeter cannot measure the sea surface height near the coast. The spatial difference of composite dynamic sea surface topography during four seasons of 10 cm to 18 cm shown in Fig. 5 is much larger than the estimated error and we can expect the good reproduction of geostrophic surface current from Fig. 5. The obtained surface current patterns from the altimetric data using the geostrophic relation during four seasons are shown in Fig. 6. An anticlockwise circulation with the speed of about 5 cm s- 1 exists in the Yellow Sea during summer. On the other hand, a small clockwise circulation exists in the Yellow Sea and an anticlockwise one in the East China Sea during winter. Such results well coincide qualitatively with the calculated current

Satellite altimetric data

663

j~-~

Upper Level

~i. . .:~~~._~.:;r:~:~~_

)~priinig.}.:"":;:';;;;~;'.;:

t=,~c. ----

:.~=

:.~,= . . . .

~:i!!:)i?i~:~:~q-~.:;:::-_:;;:;:.~1 .~.:;:~>~._~.~.~'~'~'~-~.~ I I/~..~,/.~,~'.~.~=._._,

~f(~~:.,~,,..,,~~4~q~3~~t~::;

~//~~ I l//~~ .~

~//~~/~; X q 'h~ )~ (",,~

i. i,/,;,~ .~;,~ .,,~ - ., (,,~ i~ ' -~ .~~~x~,

,: . .;; .~z~: Z,// . .~/l/~ Z

rf~.....

Upper Levelutumn

~~~;%"i~::~2".222"~......... i

~=. . . .~

: .....

~!i -2_io¢1~u¢.~

~,;;~ . .~:,:: .~i~,,.'"~ .i".~"~ N;:,, ............~~~ . . . . . . l x~'~~ '~,~ .....]~~, . .~~.~

~,~ x, .,~ t~ '~(~:~.~..x~x~,~xt~~ ......~ .~ X~ x~~ 1~.~ ......,~XXN~~~ . . . . . .

~~

~ . x~ N X~ x.~~ . X~XXXX~

~

....

~X~~~

~ --~

~

~ x X~~,%% X~~~~~x~~ x~

~

~

~

,,

,,,x,,

x~

. ,,

~x~~~?~ I I I~

~~~,.~Fig. 7.

~,~-..., .-.-.'.w.ttt-~::~,...............

t,

Seasonal variation in the diagnostically calculated sea surface circulation by Yanagi and Takahashi (1993).

664

T. Yanagi et al.

patterns by the diagnostic numerical model (Yanagi and Takahashi, 1993) which is shown in Fig. 7. The speed of summer circulation in the Yellow Sea by the altimetric data (about 5 cm s -1) is higher than that by numerical model (about 1 cm s -a) but our estimated current speed quantitatively coincide with the observed one by the satellite tracking buoys during July to October 1986 at the central part of the Yellow Sea (about 5 cm s -1 from Choi and Lie, 1992). The reason why the current speed is weak in the diagnostic numerical model may be that the averaged water temperature and salinity data in 40 years are used in the calculation (Yanagi and Takahashi, 1993). There is no remarkable sea surface circulation during spring and autumn from Fig. 6. The current patterns in spring are similar to those in winter except the northward flow in

the northeastern part of the Yellow Sea. On the other hand, the current patterns in autumn are similar to those in winter except the northeastward flow in the southwestern part of the East China Sea. 4. C O N C L U S I O N We have successfully estimated the seasonal variation of the sea surface circulation in the Yellow Sea and the East China Sea with use of altimetry data by T O P E X/ P O S E I D O N . We may estimate temporal variation in the sea surface circulations every 10 days with use of the altimetric data by T O P E X/ P O S E I D O N . To do so, we have to use another covariance function of equation (13) including temporal term. However, we do not have enough information on the phase speed of sea surface variation in the East China Sea and the Yellow Sea now. In the near future, we will try to reveal the variation of sea surface circulation with a period of 10 days with use of altimetric data after we get sufficient information on the phase speed of short-term sea level variation from the field observations. Acknowledgements--The TOPEX/POSEIDON altimetric data were provided by the Physical Oceanography Distributed Active Archive Center at Jet Propulsion Laboratory, U.S.A. This study is a part of MASFLEX project sponsored by the Agencyof Science and Technology,Japan. REFERENCES Cheney R. E., B. C. Douglas and L. Miller (1989) Evaluation of Geosat altimeter data with application of tropical Pacificsea level variability. Journal of Geophysical Research, 94(C4), 4737--4747. Cheney R. E., L. Miller, R. Agreen, N. Doyleand J. Lillibridge (1994)TOPEX/POSEIDON:the 2-cmsolution. Journal of Geophysical Research, 99(C12), 24,555-24,563. Choi B. H. and H. J. Lie (1992) Physical oceanographyprogram of the East China Sea and the East Sea (Japan Sea) dynamicsin Korea. Proceedingsof PORSEC-92, 1-28. lchikawa K. and S. Imawaki(1994) Life historyof a cyclonicring detached from the KuroshioExtension as seen by the Geosat altimeter. Journal of Geophysical Research, 99, 15,953-15,956. IchikawaK. and S. Imawaki (1996) Dynamictopographyfrom Geosat altimetry data. Journal of Oceanography, 52, 43-68. lmawaki S., K. Ichikawa and H. Nishigaki (1991) Mapping the mean sea surface elevation field from satellite altimetry data using optimal interpolation. Marine Geodesy, 15, 31-46. Yanagi T. and S. Takahashi (1993) Seasonal variation of circulations in the East China Sea and the Yellow Sea. Journal of Oceanography, 49,503-520. Yanagi T., A. Morimoto and K. Ichikawa (1997) Co-tidal and co-range charts in the East China Sea and the Yellow Sea derived from satellite altimetry data. Journal of Oceanography (in press).