搜档网
当前位置:搜档网 › A climatological description of circulation in and around the East China Sea

A climatological description of circulation in and around the East China Sea

A climatological description of circulation in and around the East China Sea
A climatological description of circulation in and around the East China Sea

Deep-Sea Research II50(2003)1065–1084

A climatological description of circulation in and around the

East China Sea

Hung-Jen Lee a,*,Shenn-Yu Chao b

a National Museum of Marine Biology and Aquarium,2Houwan Road,Checheng,Pingtung944,Taiwan

b Horn Point Laboratory,Center for Environmental Science,University of Maryland,Cambridge,MD21613-0775,USA

Accepted12September2002

Abstract

To provide baseline information for the biogeochemical studies,the circulation in the three-sea(Bohai,Yellow Sea and East China Sea)system is derived from climatological forcing,using a three-dimensional general circulation model. In the light of extensive literature dealing with similar subjects,the discussion is focused mostly on previously under-represented parts of the circulation,namely,seasonal variations of circulation in the surface layer,Changjiang plume dispersal,sea-level variations,and upwelling–downwelling locations and intensities.In terms of circulation,notable features include the development of China Coastal Current in fall and winter,appearance of a southward coastal jet off the west coast of Korean Peninsula in winter,and northward expansion of Taiwan Warm Current in summer.In response to the seasonally changing circulation,the Changjiang plume dispersal is primarily southward in winter and northeastward in summer.Two upwelling centers near the shelfbreak are worth noting.One is to the northeast of Taiwan and the other is to the southwest of Kyushu.Both are associated with anticyclonic meanders of the Kuroshio Current.Along the Kuroshio axis,a series of upwelling and downwelling centers develops in response to topographic forcing.East of Ryukyu Islands,the model also predicts a persistent upwelling location east of Okinawa Island.

r2003Elsevier Science Ltd.All rights reserved.

1.Introduction

Off the east coast of China,the East China, Yellow and Bohai seas constitute a series of interconnected marginal seas extending from 25 N to41 N(Fig.1).Among them,the East China Sea is the largest and deepest member, receiving more than90%of the river runoff entering the three-sea system(Wong et al.,2000). Being adjacent to the western Paci?c to its east and the Japan Sea to its northwest,the East China Sea provides an important conduit channeling terrige-nous and anthropogenic materials from northern Asia to the North Paci?c.

The East China Sea receives most of its river input from the Yangtze River(Changjiang);the meandischarge is about928km3yearà1(Zhang et al.,1990).By comparison,drainage from Minjiang(58.4km3yearà1)and Qiantangjiang (35.3km3yearà1)is at least one order of magni-tude smaller.Circulationinthe East Chin a Sea is modulated by the East Asia monsoon.Fig.2, derived from Hellerman and Rosenstein(1983), shows the climatological wind-stress?elds in October,December,April and August.The north-

*Corresponding author.

E-mail address:lec@https://www.sodocs.net/doc/ce9661428.html,.tw(H.-J.Lee).

0967-0645/03/$-see front matter r2003Elsevier Science Ltd.All rights reserved. doi:10.1016/S0967-0645(03)00010-9

erly winds in winter last from September to April,peaking in December.The summer monsoon from the south is much weaker and considerably shorter in duration,lasting from July to August.Winter winds intensify toward the south and follow the contour of the three-sea system,being northwest in the northern reaches but northeast in southern reaches.

The circulationinthe East Chin a Sea,drivenby local winds and remote forcing from the Kuroshio and Taiwan Strait,is of suf?cient strength.In consequence,the dispersal of Changjiang plumes does not exactly follow a path dictated by the Coriolis de?ectioneffect.na relatively quiescen t coastal ocean,river runoff turns anticyclonically until it impinges on the coast and thereafter intrudes forward as a density current along the coast (see,for example,Chao and Boicourt,1986).For Changjiang plumes,the shelfwide circulation in summer offsets this rotational tendency.In-stead,observations (Beardsley et al.,1985)an d models (Chao,1991)indicated a northeastward dispersal of Changjiang plumes in summer.

Over and beyond the shelfbreak (about 200m deep),the northward ?owing Kuroshio mostly follows curved isobaths,leaving the East China Sea from the Tokara Strait.Mesoscale variations are large wherever there is a sharp bend in the curvature of isobaths.These mesoscale features will be discussed later in conjunction with model results.

To provide baseline information for the biogeo-chemical studies of the East China Sea in this special issue,circulationinthe Bohai–Yellow Sea–East China Sea system is derived herein from the climatological forcing,using a three-dimensional numerical ocean model.Emphases are on the surface circulation,dispersal of Changjiang plumes,sea-level variations,and seasonal varia-tions of upwelling/downwelling locations.

Several

Fig.1.(a)Model domain and bottom topography (in meters)of the East China Sea and vicinity.In?ow and out?ow locations are marked in (b).Shading in (b)indicates a sponge-layer area in which horizontal mixing for temperature only is enhanced to improve computation stability.Open-ocean boundaries are stippled in (b).

H.-J.Lee,S.-Y.Chao /Deep-Sea Research II 50(2003)1065–1084

1066

longitude

l a t i t u d e

longitude

l a t i t u d e

longitude

longitude

(a)(b)

(d)

(c)Fig.2.Monthly climatological wind-stress ?elds in:(a)October,(b)December,(c)April and (d)August.

H.-J.Lee,S.-Y.Chao /Deep-Sea Research II 50(2003)1065–10841067

modeling studies of the East China Sea were undertaken prior to this effort;some were process oriented and some were deliberately simpli?ed to meet various levels of inquiries.For brevity,early numerical investigations using barotropic models are not discussed here.Models dealing with the Kuroshio path but not the shelf circulation are also excluded.Of some relevance to this paper is the work of Chao(1991),who derived summer and winter circulation patterns using idealized bottom topography,wind?elds and Changjiang forcing; seasonal transitions were excluded.A more realis-tic circulationmodel for the three-sea system was later developed by Hirose et al.(1999)to validate parameterizationschemes for the air–sea heat ?uxes.Their model forcing includes the monthly meanwin d stress derived from Na et al.(1992),the Kuroshio,the Changjiang out?ow,and?ows through the Taiwanan d Tsushima Straits.Circu-lation patterns and their relation to the Chang-jiang plume dispersal are not discussed.Further, Ryukyu Islands were treated as a vertical barrier with no gap in between,inhibiting zonal exchanges betweenthe Kuroshio water an d westernPaci?c Ocean through narrow passages.The present model skill is compatible with Hirose et al. (1999),except for a few improvements.However, our objective is to provide background informa-tionfor the follow-up biogeochemical studies in this issue,focusing on circulation patterns,upwel-ling/downwelling and Changjiang plume dispersal inthe vicin ity of the East Chin a Sea.

2.Circulation model

The numerical ocean model solves the three-dimensional temperature,salinity,momentum and mass equations under Boussinesq and hydrostatic approximations.It is similar to the general circulationmodel of Semtner(1974),except for the additionof a free surface.The model had been used previously to investigate circulation in the South China Sea(Chao et al.,1996)and tide-induced upwelling in the southern tip of Taiwan (Lee et al.,1999).The computationdomain (Fig.1)is from23 N to41 N and from116 E to134 E with a horizontal grid spacing of1/6 .There are33levels inthe vertical directionwith improved resolutiontoward the oceansurface. The top layer is initially5m thick but subject to sea-level?uctuations subsequently.Beneath the surface layer,the thickness of each layer is85%of that immediately below,yielding a maximum basin depth of6018m.

A few islands in the Ryukyu chain have dimensions comparable to or smaller than the grid scale.Our strategy is to treat a small island as a tall seamount if it is much below the grid resolution,or as an island occupying one mesh area if it is comparable to the grid resolution.Note that a tall seamount and an island of comparable horizontal dimensions exert similar disturbances to the mean?ow.Each small islan d is actually the tip of a‘‘cone’’,supported below by a much larger landmass,which is adequately resolved by the computationcells.f a small islan d is treated as a tall seamount,for example,the consequent con-straint as a result of potential vorticity conserva-tionaffects the mean?ow inways similar to that by the small island.

Horizontal mixing is achieved by using Lapla-cian mixing coef?cients of8000m2sà1inmomen-tum equations and800m2sà1intemperature an d salinity equations.Vertical viscosity and diffusivity are calculated from the Richardsonn umber according to the formulae of Pacanowski and Philander(1981).All solid boundaries are im-permeable,impenetrable and no-slip except at the bottom,where quadratic stress with a dimension-less drag coef?cient of0.0025is enforced.At the ocean surface,the monthly climatological wind stress of Hellermanan d Rosen stein(1983)is prescribed.Sea-surface temperature and salinity are slowly nudged by a Newtonian relaxation scheme(Anthes,1974)toward monthly climatolo-gical values.In consequence,heat and salinity ?uxes are set to zero at the oceansurface to avoid overspeci?cation.

The modeled ocean is initially motionless,with prescribed distributions of January climatological temperature and salinity?elds(Levitus and Boyer, 1994).Time integration is done using the typical mode-splitting technique(Simons,1974).The external and internal time steps are8and400s, respectively.Spinning up from a motionless state,

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–1084 1068

the modeled temperature and salinity?elds are subsequently nudged toward monthly climatolo-gical values givenby Levitus and Boyer(1994). There is no nudging between50and250m depths, allowing the thermocline to move more freely. Above and below the thermocline,the restoration rate is25dayà1for temperature and50dayà1for salinity.Sarmiento and Bryan(1982)?rst used the removal of nudging in the thermocline as a way to improve poleward transport of heat in a North Atlantic circulation model.It is used here mainly to stabilize the Kuroshio in?ow from the southern openboun dary.Details about this stability issue will be givenlater inthis section.Cyclical equilibrium on an annual time scale is reached in about3years;thereafter,the year-to-year changes are barely visible.The third year’s results are discussed below.

Except for a few landmasses,the southern, eastern and northern boundaries are open.The Sea of Japanis trun cated,but the Tsushima Straits are opento allow for the out?ow of Tsushima Current.Open-ocean boundaries are stippled in Fig.1b.The open-ocean boundary to the east is chosento be suf?cien tly away from Ryukyu Islands,allowing for water exchanges through narrow passages.The Kuroshio path is known to meander on seasonal and interannual time scales to the south of Japan.In this light,the eastern open-ocean boundary is chosen to coincide with the longitude of a nodal point for the Kuroshio meander(Shoji,1972),so that the out?ow location for the Kuroshio canbe?xed inthe model.Onthe western ocean boundary,the Changjiang mouth is open to release monthly varying freshwater discharge.

On all open-ocean boundaries,the depth-aver-aged velocity normal to a boundary(U n)is prescribed,while the depth-averaged tangential component(U t)is obtained through advection by the normal?ow.For the in?ow,U t is set to zero upstream of an open boundary.Deviations of currents from their depth averages(U n and U t)are extrapolated from the adjacent model interior under the assumption of zero normal gradients. The differential treatment of external and internal velocity components is meant to?x the transport but not the vertical pro?le of the in?ow and out?ow.Sea-level?uctuation s onopenboun daries are similarly extrapolated by assuming zero normal gradients.Temperature and salinity on openboun daries are derived from advectionby the normal?ow.In the case of in?ow,boundary temperature and salinity values are prescribed according to Levitus’s monthly climatology on all open-ocean boundaries.For the Changjiang in?ow,the monthly temperature and salinity values were inferred from a variety of sources. The set of open-boundary conditions generally works well.Near the southernboun dary,however, spurious temperature responses often appear because of the mismatch betweenclimatological temperature section at the southern boundary and the in?ow width.Climatological temperature data are spatially averaged over coarse meshes with one-degree resolution and therefore under-repre-sent the sharp Kuroshio front.Ideally,the in?ow width east of Taiwan(chosento be108km inthis study)must be geostrophically supported by a sharper Kuroshio front.In this model,tempera-ture?elds vary intime inrespon se to curren ts everywhere except at in?ow locations where temperature?elds are speci?ed.The mismatch betweentemperature an d in?ow onthe southern boundary presents a destabilizing factor,which can be suppressed by enhanced horizontal mixing. As illustrated by shading in Fig.1(b),a sponge-layer area is imposed near the southern boundary to suppress spurious temperature responses.The horizontal mixing coef?cient for temperature increases linearly southward inside the sponge-layer area,becoming15times as large on the southernboun dary.

I

nadditionto the use of a sponge layer,the removal of nudging in the thermocline layer also stabilizes because it allows for the formationof a sharper temperature fron t near the southern boundary to support the Kuroshio in?ow.

Table1lists the model-imposed monthly volu-metric?uxes through all openboun daries.These estimates are not completely objective;interpola-tions from a few seasonal estimates are often necessary.Further,transports for the Kuroshio out?ow and Tsushima Current are adjusted slightly to balance the sum of all incoming currents.Transports for the Kuroshio in?ow east

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–10841069

of Taiwan were synthesized using estimates from Schultz (1994),Kagimoto and Yamagata (1997),and Yang (1999).As a reference,Table 2lists transports of the Kuroshio in?ow from the aforementioned sources.These transport estimates vary with season,latitude and zonal extent of the section.Nevertheless,variations among these estimates are relatively small if the zonal section is restricted to regions with active Kuroshio https://www.sodocs.net/doc/ce9661428.html,rge transport estimates result for a much wider zonal section such as the one chosen by Kagimoto and Yamagata (1997)at 25 N.Leaving the wide section at 25 N aside,the monthly transport used in the present model (Table 1)is a smooth blend of previous estimates.I nadditionto the last three estimates,estimates of the Kuroshio out?ow were further based on Huang (1979),Bingham and Talley (1991),an d Sekine and Kutsuwada (1994).

Monthly transports through the Taiwan Strait in Table 1were estimated from Jan(1995).For comparison,the Taiwan Strait transport estimated by Jan(1995)is listed in Table 3.His estimates were slightly smoothed in Table 1for implementa-tioninthis model.The tran sports of Tsushima Current in Table 1were estimated from a host of sources (Yi,1966;Mizuno et al.,1990;Isobe et al.,1994;Katoh et al.,1996;Isobe,1997).Monthly discharge from Changjiang (Table 1)is obtained from the Global Runoff Data Center (GRDC)of the Federal Institute of Hydrology in Germany (http://www.bafg.de/grdc.htm ).

Table 1

Monthly transports (in Sv)through open ocean boundaries Month

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

TaiwanStrait 0.50.5 1.5 1.5 2.2 2.3 2.3 2.4 2.4 1.50.50.5Kuroshio in?ow 22.022.022.523.024.023.523.528.028.022.021.022.0Kuroshio out?ow à20.24à20.62à22.27à22.63à23.96à23.04à22.53à26.74à26.59à19.83à18.2à19.72Tsushima Straits à2.27à1.89à1.75à1.89à2.28à2.80à3.32à3.71à3.85à3.71à3.33à2.80Changjiang 0.010.0130.0160.0230.0370.0420.050.0450.0430.0370.0250.016Positive and negative values are for in?ow and out?ow,respectively.

Table 2

Estimated transport (in Sv)of Kuroshio in?ow from the existing literature Source

Section

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Yang (1999)(122E, 24.5N)_(123E, 24N)Kagimoto and Yamagata (1997)(124.5E, 30N)_(128.5E 27.5N)Kagimoto and Yamagata (1997)(120E, 25N)_(137E, 25N)Schultz (1994)

(121E, 23N)_(124E, 23N)

1732

20

40

Table 3

Estimates of transport (in Sv)through Taiwan Strait by Jan(1995),and monthly temperature (in o C)and salinity (psu)values for the Changjiang in?ow Month

Jan Feb Mar Apr May Jun Jul

Aug Sep Oct Nov Dec Taiwan Strait transport Changjiang temperature 8.09.012.014.018.022.025.025.023.018.014.010.0Changjiang salinity

18.0

18.017.0

16.0

15.0

13.0

12.0

10.0

12.0

13.0

15.0

17.0

0.5

1.50.5H.-J.Lee,S.-Y.Chao /Deep-Sea Research II 50(2003)1065–1084

1070

Table3also lists model-imposed monthly temperature and salinity values for the Changjiang in?ow.These estimates were made from Zhang et al.(1987),Lazure and Girardot(1988),Wang (1988)and Le(1988).Seasonal characteristic is captured to the lowest order.For example,the in?ow is the freshest during the peak discharge from June to September.

3.Surface circulation

Fig.3shows the model-derived surface circula-tioninthe midst of October,December,April an d August.The currents are depth-averaged over the top830m of the ocean.If the local water depth is shallower than830m,the depth average is over the entire water column.The model-produced circula-tioncompares favorably with that derived from surface drifters released during the World Ocean CirculationExperimen t(Lie et al.,1998).Several features over the broad shelf of the three-sea system are worth noting.In October,after1 month of winter monsoon,a southward coastal jet develops mostly to the south of Changjiang, extending toward the Taiwan Strait.The so-called China Coastal Current is wind-driven and en-hanced by the strati?cation provided by the Changjiang runoff.A similar coastal jet was produced earlier in Chao(1991).By December, the northern extent of China Coastal Current is visibly expanded to35 N,about450km north of the Changjiang mouth.The northern extension is mostly driven by the local northerly wind,which intensi?es considerably from October to Decem-ber.The China Coastal Current peaks in Decem-ber and diminishes with the winter monsoon in March.Farther north,the southward wind-driven coastal jet is pronounced along the west coast of Korea Peninsula from December to February.A southward coastal jet also exists off the west coast of Kyushu as part of a Kuroshio meander;its seasonal variation is not clear.The shelf circula-tion is quite weak during summer and between monsoons in northern reaches of the three-sea system.I nsouthernreaches of the East Chin a Sea, however,the northward in?ow from the Taiwan Strait begins to increase in April,soon after the relaxation of the winter monsoon.The so-called TaiwanWarm Curren t peaks from Jun e to August,entering the East China Sea from Taiwan Strait while turning anticyclonically.

The model-derived shelf circulationis also consistent with historical data.The winter emer-gence of China Coastal Current was documented as early as in1935by the Japan ese Naval Hydrographic Of?ce(e.g.,Nitani,1972).naddi-tionto the Chin a Coastal Curren t,win ter south-ward coastal jets off the west coasts of Korea and Kyushu were also documented in a Chinese Marine Atlas(Chen,1992).The summer expan-sion of the Taiwan Warm Current into the East China Sea is also a pronounced feature in the Chinese Marine Atlas.

Mesoscale features of the Kuroshio Current are noted below.Northeast of Taiwan,the shelfbreak becomes almost zonal and perpendicular to the northward?owing Kuroshio.As the Kuroshio impinges on the shelfbreak,an anticyclonic mean-der is generated.Historical data suggest that the anticyclonic meander coming from east of Taiwan appears to be modulated by the monsoon wind forcing,expanding in fall and winter months but shrinking in summer.The largest meander appears from October to December(He and White,1987; Sun,1987).However,the seasonal variation is not obvious in Fig.3,because the out?ow from the Taiwan Strait also contributes to the meander and cannot be separated from the signal.Recent shipboard ADCP surveys(Tang et al.,2000) revealed additional features.While the anticyclo-nic meander and associated Kuroshio intrusion onto the shelf are prevalent in fall and winter,a cyclonic eddy on the shoreward side of Kuroshio oftenappears off n ortheast Taiwaninsummer. The summer cyclonic eddy is about50km in diameter,too small to be resolved by the present model.It seems necessary to use models with?ner resolutionto resolve this eddy.This effort is, however,beyond our present scope.

Around the southern tip of Kyushu,the shelfbreak protrudes southward and the Kuroshio forms a large anticyclonic meander on the west side of Kyushu before following the southern coast of Japann ortheastward.This feature is quite distinct from tracks of surface drifters released

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–10841071

(a)(b)

Fig.3.Model-derived depth-averaged surface currents in:(a)October,(b)December,(c)April and(d)August.The depth average extends from sea surface to830m or bottom,whichever is shallower.

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–1084

1072

during the World Ocean Circulation Experiment (Lie et al.,1998).Let f be the Coriolis parameter and h be the local water depth.The large meander is steered by the f =h contours in an effort to conserve potential vorticity.In a preliminary experiment,a barotropic version of the present model had beencon ducted to verify this.The northward excursion of Kuroshio meander to the west of Kyushu appears to be maximal from

August to October.Isobe (1997)suggested that

this meander crest is the origin of the Tsushima Warm Current during fall and winter months.The minor branch of the Kuroshio entering the Tsushima Straits seems to separate from the main stream around 29 N,also consistent with surface drifter tracks obtained from the World Ocean CirculationExperimen t (Lie et al.,1998).

Whenaveraged over the top 830m of the ocean ,the ?ow to the east of Ryukyu Islands is mostly northward throughout the year.Surface-current charts compiled by the Japanese Naval Hydro-graphic Of?ce in1935(see Nitani,1972)indicated a generally northward current in summer and a southward current in winter off the east coasts of Ryukyu Islands.The near-surface currents in the present model contain a similar seasonal variation;results are not shown herein for brevity.The winter appearance of southward surface ?ow on the easternside of Ryukyu I slan ds is con sisten t with a wind-driven coastal jet.The chain of Ryukyu Islands is not continuous.With gaps in between,the island chain is only partially effective to support a wind-driven coastal jet.Under intense northeast monsoon in winter,a downwind coastal jet is expected off the east coasts of the island chain.

4.Changjiang plume dispersal

Fig.4shows model-produced surface-salinity ?elds inthe midst of October,December,April and August.These salinity ?elds over the shelf re?ect mostly the dispersal of Changjiang plumes.Note that only Changjiang discharge is imposed in this model.Effects of other smaller rivers may be felt through weak nudging toward monthly climatological salinity ?elds,but are expected to

be quite small because climatological data sets are averaged over coarse meshes with one-degree resolution.From October to April,the Changjiang discharge forms a narrow band of fresher water following the China coastline southward.The maximum expansion is around December and January,extending well into the Taiwan Strait.North of the Changjiang mouth,fresher waters also accumulate near the China coast in winter.Since winter currents are southward over the shelf,most of freshwater anomalies to the north of Changjiang mouth in winter are remnants of Changjiang plumes from the preceding summer.By April,the freshwater boundary layer has retreated out of the TaiwanStrait an d the primary plume dispersal is toward the northwest of Changjiang mouth.By August,the plume dis-perses to the north and east of the Changjiang mouth.Some of the freshwater anomalies enter the Sea of Japanthrough Tsushima Straits mostly from June to October.The remnants stay near the China Coast north of the Changjiang mouth in the following winter,completing an annual cycle.The model reproduced essential features of sea-surface salinity distributions documented by the Chinese climatological atlas (Chen,1992).The agreement not only lends credibility to the model but also indicates the dominance of the Chang-jiang plume in the East China Sea,noting that other river inputs are excluded in this model.The northward and eastward dispersal of Changjiang plumes insummer comes as a result of n orthward current over the shelf and upwelling-favorable winds.Dynamically,it is known that upwelling-favorable winds induce seaward dispersal of estuarine plumes (Chao,1988).The northward Taiwan Warm Current further hinders the would-be southward excursion of Changjiang plumes in summer,resulting in a counterintuitive north-eastward dispersal.

5.Sea-surface temperature

Fig.5shows the model-derived sea-surface temperature ?elds inthe midst of October,December,April and August.The temperature ?elds are loosely constrained by Levitus’s

H.-J.Lee,S.-Y.Chao /Deep-Sea Research II 50(2003)1065–10841073

(a)(b)

Fig.4.Model-derived sea-surface salinity ?elds (in psu)in:(a)October,(b)December,(c)April and (d)August.

H.-J.Lee,S.-Y.Chao /Deep-Sea Research II 50(2003)1065–1084

1074

(a)(b)

Fig.5.Model-produced sea-surface temperature(in C)in:(a)October,(b)December,(c)April and(d)August.

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–10841075

climatology with one-degree resolution and there-fore miss a few mesoscale features revealed inthe Chinese Marine Atlas.Nevertheless,the overall patterns of the seasonally changing temperature ?eld inthe three-sea system are captured.nthe northern reaches(Bohai and Yellow Sea),cooling starts in fall and continues to the end of March; temperature reaches minimum around March.By comparison,sea-surface temperature in southern reaches of the three-sea system varies little with seasondue to the heat supply from the Kuroshio. In consequence,the meridional temperature gra-dient is mostly controlled by the amount of cooling in Bohai and Yellow Sea.The sharpest north–south temperature gradient appears in March and April,centered about31 N.In summer,the meridional temperature gradient is markedly reduced by warming in Bohai and Yellow Sea.

A few mesoscale features are missing in the model and the discrepancy is caused by the coarse resolutionof Levitus’s climatology.nthe n orth-ernreaches of Bohai,for example,the model-produced sea-surface temperature inwin ter is about2 C warmer thanthat revealed inChin ese Marine Atlas or the Marine Environment Atlas compiled by the JapanOcean ographic Data Center(1978).Also,inwin ter,the observed sea-surface temperature inthe Yellow Sea appears to be colder thaninthe model inthe shallow reaches along the coasts of Korea and China.These discrepancies point out the need for high-resolu-tionclimatological data sets inthe future inorder to better predict temperature?elds inthe three-sea system.

Leaving a few missing localized features aside, the model-derived sea-surface temperature?elds closely resemble the observed?elds.Fig.6shows AVHRR sea-surface temperature?elds inOcto-ber,December,April and August of1999(cour-tesy of National Center for Ocean Research, Taiwan).Each image is an average of all available images ina mon th.The overall features,such as the nearly isothermal condition in summer and winter emergence of cold temperature in the Bohai and Yellow Sea,are well produced by the model (Fig.5).Major meander patterns in Fig.6are also reproduced by the model.For example,the northward excursion of the Kuroshio meander west of Kyushu is visible inDecember an d April. The model to some extent also reproduces a meander front stretching from the Yangtze River mouth to the southern tip of the Korean Peninsu-la.AVHRR images inOctober,December an d April indicate a double-crest structure along the meander front,separated by a southward-intrud-ing trough between China and Korea.The model reproduced the double crests inApril(Fig.5c).In October and December,however,the two meander crests become one in the model.Note that the northward meander crest off the west coast of Korea Peninsula in winter(Fig.6)is associated with warm water intrusion into the Yellow Sea. This process is under-represented by the model in October and December,pointing out the need for better temporal resolutionof the atmospheric forcing and better spatial resolution to resolve mesoscale topographic features.AVHRR images also reveal southward intrusion of cold water from the Yangtze River southward along the coast of China from October to April.The so-called China Coastal Current was captured by the model in Fig.5.

6.Sea-level distribution

Fig.7shows model-derived sea-level?elds inthe midst of October,December,April and August. These?elds have beendemean ed so that the annual mean of basin-averaged sea-level height is zero.

I

ngen eral,the sea level is lower over the broad shelf of the three-sea system,except for the narrow band of high anomaly induced by the China Coastal Current in winter.Over the Kuroshio region,sea level rises seaward to support the northward current.Farther east of the Ryukyu Islands,sea level may drop occasionally in some regions to support southward countercurrents. Speci?cally,Okinawa Island appears to be the center of a high anomaly of sea level especially in winter,supporting the southward?ow to its east. Leaving zonal gradients aside,the meridional sea-level gradient over the broad shelf of three-sea system generally supports a higher sea level to the south in all months of the year.The meridional

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–1084 1076

(a)(b)

(d)

(c)Fig.6.Monthly averaged sea-surface temperature derived from AVHRR for:(a)October,(b)December,(c)April and (d)August of 1999.Color bars in each panel indicate temperature in C.

H.-J.Lee,S.-Y.Chao /Deep-Sea Research II 50(2003)1065–10841077

(a)(b)

Fig.7.Model-derived sea-level contours(in cm)in:(a)October,(b)December,(c)April and(d)August.Positive and negative values indicate sea-level rise and drop,respectively.The sea-level?elds have been demeaned,so that the annual mean of the basin-averaged sea level is zero.Contour intervals are normally10cm and negative contours are dashed.An additional contour of35cm around Okinawa Island in winter is included to highlight a high anomaly.

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–1084

1078

slope indicates the predominance of winter mon-soon,which moves waters to the south.Season-ally,sea level inthe n orthernreaches of the three-sea system drops considerably during the winter monsoon,and rebounds in summer months. Analyses of satellite altimetry data generally support the existence of a positive sea-level anomaly around Ryukyu Islands.Fig.8,derived from a hybrid product of satellite altimetry data and the Naval Research Laboratory(NRL) Layered OceanModel(http://www7300.nrlssc. https://www.sodocs.net/doc/ce9661428.html,/altimetry),shows the sea-surface height inthe regionof in terest on(a)October21,(b) December22,(c)April20and(d)August21of 1999.Note the persistence of a positive sea-level anomaly around Ryukyu Islands.The anomalous high is onthe seaward side of the Kuroshio. Conceivably,the positive anomaly is associated with the anticyclonic shear of the Kuroshio current on its seaward side.East of Okinawa Island,the positive anomaly persists throughout the year,but weakens markedly in April.The seasonal trend is qualitatively similar to the model-produced sea-level?elds in Fig.7.However,the present model is heavily constrained by monthly climatology forcing and therefore cannot be used to make quantitative comparisons with daily snapshots derived from a vastly different product. Leaving the anomalous high around Okinawa aside,Fig.8also shows the persistent Kuroshio meander around the southern tip of Kyushu. Similar features are also evident from the present model(Fig.7).

7.Upwelling and downwelling

Fig.9shows model-derived upwelling–down-welling locations and intensities in the midst of October,December,April and August.The depth surface is chosenat the mid-depth of the water columnfor water depths shallower than200m, and at100m depth in waters deeper than200m. When presented as such,vertical motions are much more pronounced over and beyond the shelfbreak.By comparison,weak vertical motions over shallow shelves often become insigni?cant. Further,vertical motions near open boundaries to the east and south may be in?uenced by the imperfect boundary conditions and less reliable. The discussionis therefore limited to vertical motions over the shelfbreak,in the Kuroshio Current and slightly seaward of Ryukyu Islands. Seven quasi-stationary upwelling and downwelling centers are commented on below.Other centers are not discussed because they are either too remote from the East China Sea or too close to open-oceanboun daries.

Near the shelfbreak or200m isobath,two prominent upwelling centers stand out.One is to the northeast of Taiwan and marked by A in Fig.9.From our model results,the upwelling occurs year-round with an average upwelling rate of about13m dayà1at the center.This upwelling center is extensively documented in the literature (e.g.,Liu et al.,1992;Wong et al.,2000),and the average upwelling rate was estimated to be about 5m dayà1based onpast observation s.While the observed and modeled upwelling intensities appear to have the same order of magnitude,the model seems to have displaced this upwelling center somewhat closer to the northeastern tip of Taiwan. The other upwelling center,marked by B in Fig.9, is associated with the anticyclonic meander to the southwest of the Kyushu Island.This center occurs in waters shallower than200m.Its intensity reaches maximum from August to October, consistent with the observations of Isobe et al. (1994).

I

nregion s occupied by the mainstream of Kuroshio Current,there is a series of upwelling and downwelling centers if one follows the warm current northeastward.Most of these vertical motions appear to be correlated with topography forcing.For example,the downwelling center marked by C is to the east of Taiwan,where the Kuroshio Current begins to enter the deep basin of the Okinawa Trough(see Fig.1).As the Kuroshio turns eastward at about24 N(see Fig.3),the rising topography produces the upwelling center D in Fig.7.Farther northeastward,the downwelling center(E)occurs as the Kuroshio enters the central deep basin of the Okinawa Trough. Farther northward,upwelling center(F)is pro-duced as the Kuroshio impinges on a rise in the Okinawa Trough.All these vertical motions are

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–10841079

(a)(b)

(d)

(c)Fig.8.Sea-surface height (incm)for:(a)October 21,(b)December 22,(c)April 20,an d (d)August 21of 1999,derived from a hybrid product of satellite altimetry data and the Naval Research Laboratory (NRL)Layered Ocean Model.Shallow shelf areas are masked by the gray color.

H.-J.Lee,S.-Y.Chao /Deep-Sea Research II 50(2003)1065–1084

1080

essentially consistent with rising and deepening bottom topographies along the Kuroshio path.Among four centers (C –F ),the upwelling center F is closest to the shelfbreak and the Kuroshio Current also manifests a small anticyclonic mean-der at about the same location(Fig.3).

East of Ryukyu Islands,a persistent upwelling center (G )east of Okinawa Island is also worth noting.According to the Chinese Marine Atlas edited by Chen(1992),this is the area of low sea-surface temperature.Ship drift data compiled by the Japanese Naval Hydrographic Of?ce (Nitani,1972;Fig.2)suggested that this area is often dominated by southward countercurrent or antic-yclonic vorticity.Further,sea-level height ?elds (Figs.7and 8)indicate that this is an area of positive anomaly.From a theoretical perspective,upwelling from depths and anticyclonic vorticity aloft should be inter-related,so the model-produced upwelling makes sense physically.

(b)

(a)Fig.9.Model-derived upwelling/downwelling locations and intensities in (a)October,(b)December,(c)April and (d)August.The depth surface is inthe middle of the water columnfor water depths shallower than200m,an d at 100m inwaters deeper than200m.The thick line indicates 200m isobath.Thin solid and dashed contours indicate upwelling and downwelling,respectively.Contours start from 0.005cm s à1with increments of 0.005cm s à1.Upwelling and downwelling centers marked by A –G are discussed inthe text.

H.-J.Lee,S.-Y.Chao /Deep-Sea Research II 50(2003)1065–10841081

Nevertheless,future veri?cationof this upwellin g center is very much in need.It also should be emphasized that all upwelling and downwelling centers commented on herein are not related to Ekman pumping associated with the wind-stress curl.This had beenveri?ed through prelimin ary analyses of monthly?elds of wind-stress curl(not shown).

8.Conclusions

Using climatological forcing to drive a general circulationmodel,the three-dimen sion al circula-tion in and around the East China Sea and Changjiang plume dispersal were derived and described herein.In terms of circulation over the shelf,notable features include the winter develop-ment of the China Coastal Current,intensi?cation of a southward coastal jet along the west coast of the Korean Peninsula in winter,and the north-ward expansion of the Taiwan Warm Current from the TaiwanStrait insummer.Over the shelfbreak,the model reproduces two anticyclonic Kuroshio meanders.One is to the northeast of Taiwanan d the other is to the southwest of Kyushu.Both manifest sizable upwelling over the shelfbreak.Along the main axis of the Kuroshio, the model produces a series of upwelling and downwelling centers.The upwelling and down-welling motions vary little with seasons and appear to be induced by topographic forcing.Seaward of the chainof Ryukyu I slan ds,the model predicts a persistent upwelling center east of the Okinawa Island.This upwelling center appears to be related to the anticyclonic vorticity or southward counter-currents revealed from limited observations(Nita-ni,1972;Chen,1992).While the crude model-data comparison appears to be encouraging,future observationefforts are very much n eeded to verify this upwelling center.

To the lowest order,the model is able to reproduce the seasonal Changjiang plume disper-sal,which is essentially southward in winter but northeastward in summer.While this is encoura-ging,model improvements in the future will depend on better observation efforts with re?ned temporal and spatial resolutions.It is our inten-tionto improve the model’s capability to simulate Changjiang plume dispersal using results from intensive observations to be launched by Taiwan oceanographers in the next few years.The improved circulationmodel will thenbe used as a basis to build a biogeochemical model of the East China Sea.At issue here is the future impact of the Three-Gorge Dam in Changjiang on primary productionof the East Chin a Sea.A biogeochemical model,if developed successfully, can be used to synthesize and evaluate changes of the East China Sea after the Dam is built.As it stands now,this remains as a distant goal. Acknowledgements

Author H.J.L.was supported by the National Science Council(Taiwan)through a postdoctoral fellowship under a Grant NSC88-2611-M002-010-AP7.Author H.J.L.also thanks Dr.C.-S.Chern for encouragement.Author S.Y.C.was supported by a subcontract from the National Taiwan Ocean University(NSC89-2611-M019-004-K2).We thank Dr.K.-K.Liu for providing essential background materials that led to this investigation and Dr.G.-C.Gong for his administrative support.

References

Anthes,D.L.T.,1974.Data assimilation and initialization of hurricane-predicting models.Journal of Atmospheric Sciences31,702–719.

Bingham, F.M.,Talley,L.D.,1991.Estimates of Kuroshio transport using an inverse technique.Deep-Sea Research I 38,S21–S43.

Beardsley,R.C.R.,Limeburner,H.Y.,Cannon,G.A.,1985.

Discharge of the Changjiang(Yangtze River)into the East China Sea.Continental Shelf Research4,57–76.

Chao,S.Y.,1988.Wind-driven motion of estuarine plumes.

Journal of Physical Oceanography18,1144–1166. Chao,S.Y.,1991.Circulationof the East Chin a Sea,a numerical study.Journal of the Oceanographical Society of Japan46,273–295.

Chao,S.Y.,Boicourt,W.C.,1986.Onset of estuarine plumes.

Journal of Physical Oceanography16,2137–2149. Chao,S.-Y.,Shaw,P.-T.,Wu,S.Y.,1996.El Nin o modulation of the South China Sea circulation.Progress in Oceano-graphy38,51–93.

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–1084 1082

Chen,G.,1992.Marine atlas of Bohai Sea,Yellow Sea and the East China Sea.China Ocean Press,Beijing,524pp.

He,Y.,White,W.,1987.Interannual variability of the Kuroshio frontal structure along its western boundary in the North Paci?c Oceanassociated with the1982ENSO event.Journal of Physical Oceanography,1494–1506. Hellerman,S.,Rosenstein,M.,1983.Normal monthly wind stress over the world oceanwith error estimates.Journ al of Physical Oceanography13,1093–1104.

Hirose,N.,Lee,H.-C.,Yoon,J.-H.,1999.Surface heat?ux in the East China Sea and the Yellow Sea.Journal of Physical Oceanography29,401–417.

Huang,J.C.K.,1979.Numerical simulation studies for oceanic anomalies in the North Paci?c basin:II.Seasonally varying motions and structures.Journal of Physical Oceanography 9,37–56.

Isobe,A.,1997.The determination of the volume transport distribution of the Tsushima Warm Current around the Tsushima/Korea Straits.Continental Shelf Research17, 319–336.

Isobe, A.,Tawara,S.,Kaneko, A.,Kawano,M.,1994.

Seasonal variability in the Tsushima Warm Current, Tsushima–Korea Strait.Continental Shelf Research14, 23–35.

Jan,S.,1995.Seasonal variations of currents in the Taiwan Strait.Ph.D.Thesis,Institute of Oceanography,National Taiwan University,Taipei,Taiwan,unpublished.

Japan Oceanographic Data Center,1978.Marine Environment Atlas,Northwestern Paci?c Ocean II(seasonal and monthly).Japan Hydrographical Association,Japan, 157pp.

Kagimoto,T.,Yamagata,T.,1997.Seasonal transport varia-tions of the Kuroshio:an OGCM simulation.Journal of Physical Oceanography27,403–418.

Katoh,O.,Teshima,K.,Kubota,K.,Tsukiyama,K.,1996.

Downstream transition of the Tsushima Current west of Kyushu in summer.Journal of Oceanography52,93–108. Lazure,P.,Girardot,J.P.,1988.Hydrodynamics of the Changjiang estuary area.In:Yu,G.et al.(Eds.)Proceed-ings of the International Symposium on Biogeochemical Study of the Changjiang Estuary and its Adjacent Coastal Waters of the East China Sea.China Ocean Press,Beijing, pp.38–61.

Le,K.,1988.A preliminary study of basic hydrographic features and the current structures off the Changjiang river mouth in the dry season.Progress in Oceanography21,387–400. Lee,H.-J.,Chao,S.-Y.,Fan,K.-L.,Kuo,T.-Y.,1999.Tide-induced eddies and upwelling in a semi-enclosed basin:Nan Wan.Estuarine,Coastal and Shelf Science49,775–787. Levitus,S.,Boyer,T.,1994.World oceanatlas1994,Vol.4.

NOAA Atlas NESDIS4,US Department of Commerce, Washington,DC,https://www.sodocs.net/doc/ce9661428.html,/ SOURCES/.LEVITUS94/

Lie,H.J.,Cho,C.H.,Lee,J.H.,Niiler,P.,Hu,J.H.,1998.

Separation of the Kuroshio water and its penetration onto the continental shelf west of Kyushu.Journal of Geophy-sical Research103,2963–2976.Liu,K.-K.,Gong,G.-C.,Shyu,C.-Z.,Pai,S.-C.,Wei,C.-L., Chao,S.-Y.,1992.Response of Kuroshio upwelling to the onset of the northeast monsoon in the sea north of Taiwan: observations and a numerical simulation.Journal of Geophysical Research97,12511–12526.

Mizuno,S.,Kawatate,K.,Nagahama,T.,Miita,T.,1990.

Measurements of East Tsushima Current in winter and estimation of its seasonal variability.Journal of the Oceanographical Society of Japan45,375–384.

Na,J.Y.,Seo,J.W.,Han,S.K.,1992.Monthly-mean sea surface winds over the adjacent seas of the Korean Peninsula.Journal of the Oceanological Society of Korea 27,1–10inKoreanwith En glish abstract.

Nitani,H.,1972.Beginning of the Kuroshio.In:Stommel,H., Yoshida,K.(Eds.),Kuroshio,Physical Aspects of the Japan Current.University of Washington Press,Washington DC, pp.129–163.

Pacanowski,R.C.,Philander,S.G.H.,1981.Parameterization of vertical mixing in numerical models of tropical ocean.

Journal of Physical Oceanography11,1443–1451. Sarmiento,J.L.,Bryan,K.,1982.An ocean transport model for the North Atlantic.Journal of Geophysical Research87, 394–408.

Schultz,J.R.,1994.A numerical model of the Kuroshio in the East China Sea.Ph.D.Thesis,Department of Oceanogra-phy,The Florida State University,Florida,USA,143pp. Sekine,Y.,Kutsuwada,K.,1994.Seasonal variation in volume transport of the Kuroshio south of Japan.Journal of Physical Oceanography24,261–272.

Semtner, A.J.,1974.An oceanic general circulation model with bottom topography.Technical Report9,Department of Meteorology,University of California,Los Angeles, 99pp.

Shoji,D.,1972.Time variationof the Kuroshio south of Japan.

In:Stommel,H.,Yoshida,K.(Eds.),Kuroshio,Physical Aspects of the Japan Current.University of Washington Press,Washington,DC,pp.217–234.

Simons,T.J.,1974.Veri?cation of numerical models of Lake Ontario:Part I:circulation in spring and early summer.

Journal of Physical Oceanography4,507–523.

Sun,X.,1987.Analysis of the surface path of the Kuroshio in the East China Sea.In:X.Sun(Ed.),Essays on the Investigation of Kuroshio,China Ocean Press,Beijing, pp.1–14(Chinese with English abstract).

Tang,T.Y.,Tai,J.H.,Yang,Y.J.,2000.The?ow pattern north of Taiwan and the migration of the Kuroshio.Continental Shelf Research20,349–371.

Wang,W.,1988.Yangtze brackish water plume—circulation and diffusion.Progress in Oceanography21,373–385. Wong,G.T.F.,Li,Y.-H.,Chao,S.-Y.,Chung,Y.-C.,2000.The Kuroshio Edge Exchange Processes(KEEP)study—an introduction to hypotheses and highlights.Continental Shelf Research20,335–347.

Yang,Y.,1999.Mesoscale eddies’in?uence on the Taiwan Current(Kuroshio)volume transport.Ph.D.Thesis,In-stitute of Oceanography,National Taiwan University, Taipei,Taiwan,unpublished.

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–10841083

Yi,S.-U.,1966.Seasonal and secular variations of the water volume transport across the Korea Strait.Journal of Oceanography Society Korea1,7–13.

Zhang,Q.H.,Janowitz,G.S.,Pietrafesa,L.J.,1987.The interaction of estuarine and shelf waters:a model and application.Journal of Physical Oceanography17,455–469.Zhang,J.,Letolle,R.,Martin,J.M.,Jusserand,C.,Mouchel, J.M.,1990.Stable oxygenisotope distributioninthe Huanghe(Yellow River)and the Changjiang(Yangtze River)estuarine systems.Continental Shelf Research10, 369–384.

H.-J.Lee,S.-Y.Chao/Deep-Sea Research II50(2003)1065–1084 1084

相关主题