A reduced thermodynamic model on the formation of the Nansha warm water

Acta Oceanologica Sinica 2004, Vol. 23, No. 2, p.239-249http://www.oceanpress.com.cnE-mail:hyxbe@263.netA reduced thermodynamic model on the formationof the Nansha warm waterCAI Shuqun!*, GAN Zijun', LI Chiwai?, LONG Xiaomin', DONG Danpeng'1. LED, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China2. Hong Kong Polytechnic University, Hong Kong, ChinaReceived 11 February 2004; accepted 20 April 2004AbstractA reduced vertically integrated upper mixed layer model is set up to numerically study the thermodynamic processof the formation of the“Nansha warm water"(NWW) in the Nansha Islands sea areas in spring. According to thenumerical experiments, it is shown that, in spring, the formation of the NWW is mainly due to the sea surface net heatflux and the local weak current strength; the contribution from temperature advection transport and warm waterexchange with the outer seas (Sulu Sea or south of Sunda shelf) is very litle. In the sea areas where the curent isstrong, the advection may also play an important role in the temperature field.Key words: numerical modeling, warm water, thermodynamic process, Nansha Islands, South China Seaand sea, thus it is called the Nansha warm wa-1 Introductionter (NWW). Since the sea surface temperatureThe Nansha Islands and its adjacent sea(SST) in the deep basin of the SCS is higher allarea (about within 4~15°N, 109~119°E) arethe year round, and the isotherm is on the north-located in the southerm South China Sea (SCS).east-southwestward trend, thus some research-ers argue that it is certainly related to the warmIt connects with the Sulu Sea through theMindoro and Balabac Straits and with the Javawater in the Nansha Islands sea area and calledSea through the Sunda shelf (see Fig. 1). Dur-the NWW the SCS warm water (Jia et al,ing the investigation period of the 1990s, it is2000).Up to date, there are some research worksfound that (Zhao et al, 1998), within the ellip-about the NWW's characteristics and its rela-tic sea area about 7~14°N, 109~1 16°E, theretions with ENSO (Zhao et al., 1998; Chen etexists an all-year-round high temperature warmal, 1998; Xie et al, 1998; Xu and Sun, 1998).waters with horizontal uniformity and consis-However, lttle work on the thermodynamictent annual variation. It is less affected by theland runoff and the heat difference between landmechanism of the NWW's formation has beenreported (Jia et al., 2000). It is also found that* Corresponding author,E -mail: caisq@scsio.ac.cnthe mean warmest water temperature of the中国煤化工MYHCNMHG240CAI Shuqun et al. Acta Oceanologica Sinica 2004, Vol. 23, No. 2, p.239-24928.100° . 104° 108 112。 116° 120° 124° EU2(Uu), a(Vu)dt dsdy24*-=-g(h+n)an_ g(h+n)2 dρ." dx2ρ。dx20%-4Luzon Srait(T°-ry)1ρ+A.V2U，(1)fuzgh60-\!dV. a(Uv), 0(Vv)” ?Indo-Chin Peninsul+ fU120At'axpNansia Is=-g(h+η)η_ g(h+η)2 ap,Balabac sutai”dy2po dBomeo(T}-zq)/ρ+A,V°V，(2)aη。aUdV=0,(3)AtdxFig. 1. Computational domain (double dashed lineoTaT dl+K,V2T+-H。(4stands for open boundary, and the pointsA, B, C, D~ax' aC,p(h+m)'and E are referred in the text).asaS+KgV2s， (5)atdx' dy'NWW appears in May, but, the annual meannet heat flux in the Nansha Islands sea area isρ= P(T,S,p),(6lower than that in the other sea areas in thesCS (Wang, 1998). How is the NWW formed?where p and P。are the sea water density andIs it related to the local circulation? To answerreference density; (U, V) and (u, v) are thethese questions, the numerical model study iswhole-depth current speed components and ver-certainly needed.tical mean speed components in the directionsIn this paper, to simplifty the questions, aofx and y respectively; τ and τ are thereduced vertically integrated upper mixed layerwind stress and bottom frictional stress, whichmodel is set up, and some numerical experi-are computed using a quadratic law (Cai et al,ments are carried out to study the thermody-2001); h is the upper mixed layer thicknes; pnamic process of the NWW's formation.is the static pressure; η the free surfaceanomaly (FSA in short); f the Coriolis2 Set ψp and choice of parameters of the modelparameter; T and S are the vertical mean tem-perature and salinity of water; Hg is the net heatUnder the conditions of hydrostaticflux; A, K, and K。are the lateral frictionequilibrium, Boussinesq and β -planecoefficient, horizontal diffusion coefficients ofapproximation, the upper vertically integratedtemperature and salinity respectively.upper mixed layer model's controlling equationsIn the computation, Eqs (1)~(3) are solvedareon the“Arakawa C”" grid using a semi implicit中国煤化工MYHCNMHGCAI Shuqun et al. Acta Oceanologica Sinica 2004, Vol. 23, No. 2, p.239~249241scheme. The obtained discrete equations withmean 1°x1° COADS data (da Silva et al, 1994)boundary conditions are solved by the column-of termperature, salinity, wind stress and net heatmain-element Gauss elimination method. Theflux are interpolated into the computationaltime differencing is leapfrog, thus an Euler dif-domain.ference is used for the start and restart to pre-In the paper, we major in the study thatvent development of the time spltting associ-why the SCS SST gets warmer and warmerated with this scheme, i.e., every 2 d (24 timefrom February (late winter) to May (late spring),steps in the model) the solution for 2 adjacentthus in the model, the initial temperature fieldtime steps is averaged. The time differencing(see Fig.2a) and salinity field are set to thein Eqs (4)~(5) is explicit, while the Euler-mean fields in February, whilst the final tem-Lagrangian approximation is used for the ad-perature field (see Fig. 2b) and salinity field arevection term (Casulli, 1990). Centered differ-set to the mean fields in May, and the drivenence scheme is used for the horizontal diffu-net heat flux field is set the mean field in Maysion term in all of the equations.(see Fig. 3). The numerical experimental casesIn the computation, as in spring the ob-are listed in Table 1. The temperature and salin-servational mean upper mixed layer thicknessity at the inflow open boundary are prescribedis less than 50 m (Cai and Gan, 2000), thus theas those in May, whilst those at the outflowmaximum water depth in the model is 50m, open boundary (except in experiment E4) arewhile the minimum water depth is 20 m. Thegiven as:space and time steps are (1/6)° and 2 hrespectively. The inflow of Kuroshio, that is4 x106 m/s here, is given through the openwhere中represents T orS ; and v。is theboundary east of Luzon Island, and it is exactlyvelocity normal to the boundary.compensated by outflow through the openboundary north of Taiwan Island. The width3 Simulation experimen tal results and discussionof inflow is (11/6)°, the flow angle is 170°, andthe inflow core is l°distant from east shore ofEach experiment commences from theLuzon Island. The inflow velocities are uniform,state of rest and becomes quasi equilibrium af-while the outflow velocities are geostrophicallyter running about 45 d. The model is run con-adjusted by the model itself (Cai et al, 2001). .tinuously for 120 d and the data, after 90 d, areMoreover, the zero gradient condition is usedsaved for analysis.for the FSA and velocity in the south openboundary at the Sunda shelf, and the no-slip3.1 Experiment E1: to study whether the SCScondition is prescribed on the tangential flowSST can get warmer from the coldest inin the other solid boundaries (Cai et al, 2001).February into the warmest in May or notThe coefficient of bottom friction is 2x103; theby the actions of the external forceslateral friction coefficient, A, is 100 m2/s, thus(Kuroshio inflow, winds and net heat fluxthe grid Renold number, Re= vOx/A >10, soin May)that the severe damping effect by friction onthe solution could be prevented (Preller, 1986);Distributions of the upper current field andand it is supposed that K=K;=A,. The monthly temperature field after running 90 d are shown中国煤化工MYHCNMHG242CAI Shuqun et al. Acta Oceanologica Sinica 2004, Vol. 23, No. 2, p.239~24927°r -102 105° 108 1”114° 117” 120 123° 126E102” 105° 108 1”114" 117” 120 123° 126'E24*-24"27.521°,25广22225- 21"|-28.5-7广-23< 21518"|1825-26309.5-20.5.foh的Fig. 2. Distribution of initial temperature field (°C)obtained from observational datain February (a), in May ().102° 105° 108° 11' 114 117” 120 123° 126ETable 1. Model' s running cases and theirNexperimental parameters.。 -70-24'如-80EI Driven fores are the winds, net heat flux in May andthe Kuroshio infow21'100E2 Same as E1, except the maximum mixed layer thick-18-ness is 200 m,120-E3 Same as E1, except the Balabac Strait is opened and15set 0.25 x10* m/s inflow and the Mindoro Strait is100--B0opened and set 0.25 x106 m'/s outlow, the tempera-ture near both boundaries is set 30.2 C80一4 Same as E1, except the Sunda shelf is set 1.0 x10 m2/s9h 2inflow and the temperature near this boundary is set30.2 C, whilst the open boundary north of TaiwanIsland is set 5.0 x10* m/s outflowE5 Same as E1, except in the model, the advection termin Eq. (4) is ignored.80、Note: the other parameters are the same except thoseFig. 3. Distribution of driven net heat fux(W/m)given in the table.field in May obtained from observational datain Figs 4a and b. In the Luzon Strait, most ofout of the SCS, along the eastern shore of Tai-the Kuroshio water meanders slightly aroundwan Island. However, a part of the Kuroshiothe southern part ofTaiwan Island; it then flowswater enters the northeastem SCS, spreading中国煤化工MYHCNMHGCAI Shuqun et al. Acta Oceanologica Sinica 2004, Vol. 23, No. 2, p.239 -24924327102 105° 108° 111 114 117” 120” 123° 126'E102° 105° 108” 111 114° 117° 120° 123° 126EN242121°18151299hr26Fig. 4. Distribution of current field ( cm/s) (a) and temperature field (b) after running 90 d in experiment E1.westward. Since in May the southwest mon-in Fig.2b by observation, although there existssoon begins to dominate the SCS, in the south-some difference, e.g., the simulated tempera-ern SCS, a large amount of water in the Sundature in the northeastern sCs cannot get soshelf is driven northward by the southwestwarm as that by observation in Fig.2b. To an-monsoon. In the southeast shore of the Indo-swer this question, experiment E2 is carried out.China Peninsula, it keeps flowing northward,forming a western boundary current, the so-3.2 Experiment E2: to study if the simulationcalled“Summer Southeast Vietnam Offshoreresults are related to the mixed layer thick-Current”(Fang et al., 2002). It then flowsnessnorthward and crosses the mouth of Beibu Gulf,flowing northeastward along the northeasternIn Experiment E2, the maximum mixedcontinental shelf of the SCS into the Taiwanlayer thickness is thickened to 200 m. The simu-Strait. However, in the other sea areas includ-lated temperature field after running 90 d ising the middle and southern SCS, the currentshown in Fig. 5, in which the temperature offield is very weak.the SCS is much lower than that in ExperimentIn Fig.4b, the temperature in most of theE1. This demonstrates that, the maximum mixedcentral sea basin is larger than 29 °C. In thelayer thickness chosen in the model can causenortheastern sCs, the isotherms trend north-the difference between the simulated and ob-east- southwestward and the water temperatureservational temperatures. In fact, from the heatincreases northward. It is noticed that the tem-source term in the Eq.(4), it is shown that, forperature field actually gets warmer by compari-the positive net heat flux (it is the case in May),son of Fig.4b with Fig.2a, and the isotherm dis-the heat source term is in inverse proportion totribution trend in Fig.4b is very similar to thatthe water depth. Thus, when the maximum中国煤化工MYHCNMHG244CAI Shuqun et al. Acta Oceanologica Sinica 2004, Vol. 23, No.2, p.239~24927"°r102* 10S" 108” 1心114 117” 120" 123" 126E the both boundaries is set at 30.2 C. In Ex-N202251periment E4, the Sunda shelf is set 1.0 m2/s24*|inflow, and the temperature near this boundary1~ 2is set at 30.2 C, whilst the open boundary north21"of Taiwan Island is set 5.0 m3/s outflow, soz23that the sum of inflow and outflow in the whole! 8°computational domain is 0. The correspondingresulted temperature fields are shown in Figs6~7.In Experiment E3, the obtained current fieldindicates that near the Balabac Strait, the west-罗ward current speed can get about 0.33 m/s,which is much larger than that from observa-tional data (Cai et al, 2004). However, evenfor such a large current speed, the temperatureFig. 5. Distribution of temperature('C) field afterfield in the Nansha Islands sea areas (Fig.6)has lttle change except near the Balabac Strait,running 90 d in Experiment E2where the 29。C -isotherm extrudes into themixed layer in the model is thickened, the simu-SCS, showing that the effect of the warm wa-lated temperature would get much lower thanter from the Sulu Sea through the Balabac Straitthe observational one. In fact, this can be wellon the temperature field of the Nansha Islandsunderstood, since Fig.2b only shows the tem-sea areas is very limited and local.In Experiment E4, when compared withperature at sea surface. However, in experimentEI, the maximum mixed layer thickness is cho-sen as 50 m.1029 105° 108° 11 1149 117”120° 123" 126EIs it possible that the outer warm water Nflows into the Nansha Islands sea areas and 24causes the sea water temperature there to get21so high? For instance, the Sulu Sea water flowsthrough the Balabac Strait or the Java Sea wa-18ter flows through the Sunda shelf into the SCS?153.3Experiment E3~E4: can the water ex-change with the adjacent outer seas bringabout the formation of the NWW9*39~-20/In Experiment E3, the Balabac Strait isopened and set 0.25 x10° m/s inflow andMindoro Strait is also opened and set 0.25 m2/sLeoutflow (so that the water amount isFig. 6. Distribution of temperature("C) field afterconservative), meanwhile the temperature nearrunning 90 d in Experiment E3.中国煤化工MYHCNMHGCAI Shuqun etal. Acta Oanologica Snica 2004, VolL. 23. No.2, p.239-249245102 105" 108 u 114 11”120" 123°126E . the eastern sea area near the Balabac Strait,NTwhere there is water flowing from the Sulu Sea;whilst the waters from the south of Sunda shelf4°into the sCs is mainly limited to the southern1"|Nansha Islands sea area. Moreover, from Mayto June 1990, it is observed that, in the Nansha8"Islands sea area, the SST in the southwest isthe highest, then it begins to reduce in the north5°of the Balabac Strait (Huang, 1994). These ob-2-/servational facts show that the NWW in the厂-27Nansha Islands sea area cannot be formed by9*h n28-the intrusion of the outer sea water, otherwisethe temperature distribution in the Nansha Is-lands sea areas would show the changing trendfrom high temperature near the Balabac Strait/Sunda shelf to low temperature in the other seaFig. 7. Distribution of temperature(C) field afterareas distinctly.running 90d in Experiment E4.Then, what causes the formation of theNWW? Let's review the Eq. (4) about tempera-that in Experiment EI (see Fig 4b), the obtainedture and evaluate the contribution of each termtemperature field (Fig.7) also shows that thein the right of the Eq. to the temperaturedistinct variation is limited near the Sunda shelf,variation. In the model experiments, 5 pointswhere the local temperature is somewhat higher(see Fig.1), i.e., Point A (19°N, 116°E) in thedue to the effect of the warm water flowingnortheastern SCs, Point B (14N,110°E) tofrom the south of Sunda shelf; meanwhile, alongthe east of middle Indo- China Peninsula, Pointthe east shore of the Indo- China Peninsula, sinceC(12°N, 1189E ) to the west of Mindoro Strait,the northward flowing western boundary cur-Point D (10°N, 115°E) to the west of Palawanrent gets stronger, the corresponding tempera-Island and E( 7°N, 114°E) to the southwest ofture field also changes.Balabac Strait are chosen for the calculation ofBoth of the above experimental resultsthe term contribution. In these points the mixeddemonstrate that the water exchange with thelayer thickness in the model is the same, 50 m.outer seas has lttle contribution to the forma-Table 2 gives the mean values of each term intion of NwW, i.e, either the warm water fromEq.(4) and the change of temperature after run-the Sulu Sea or south of Sunda shelf only makesning from 80 to 90 d at each point in somethe adjacent local water temperature increase.numerical experiments. It is shown that, inThe results agree with the observation on hand.general, the heat source term (the 4th term onFor example, during the late April to middle Maythe right side of the equation) is the largest,in 1987, it is observed that (Synthetic Scien-while that of the dispersion term (the 3rd termtific Research Group of the Nansha Islands ofon the right side of the equation) is the smallest.Chinese Academy of Sciences, 1989), the wa-In the experiments, since the current is stron-ter temperature in the western and northernger in Points A and B, thus the induced tem-Nansha Islands sea area is higher than that in perature advection term (the 1st and 2nd termns中国煤化工MHCNMHG246CAIShuqun etal. Acta Ocenologica Sinica 2004,0 Vol.23. No.2, p.239-249Table 2. Mean values of each term in Eq.(4) and the change of temperature after runningfrom 80 to 90 d at each point in some numerical experimentsExperiment Point No.Advection termDispersion termHeat source termTemperatureNo./X10*Cs1/X10* °Cs'/X10+ °Cs1increment /°CE7.88-0.5155.605.67B-47.580.9947.901.42c-0.82-0.2833.803.01-0.110.0232.902.99-0.0833.303.00E29.63-0.4013.802.406.28-6.9112.001.07.08-0.058.450.76-0.100.048.210.740.23-0.068.320.77E37.85-0.52-48.210.661.141.38-0.043.090.340.15-0.94-0.29E4-56.011.1156.601.55-48.800.0948.810.47-4.22_0.7334.43D-0.3233.453.02-0.120.1233.923.03in the right side of the equation) there is larger,in the northern open boundary increases, thesomewhat less than that of the heat source term;current structure in Point A changes, thus thewhilst in Points C, D and E (within the Nanshacorresponding negative advection term reduces,Islands sea areas), since the current is weak,and the temperature increment also reduces. Inthus the induced temperature advection termPoint B, the heat source term is almost counterthere is far more less than that of the heatto that of the advection term, therefore, the tem-source term.perature increment there is the least. The majorConsidering that the heat source term isdifference between Points A and B in the Ex-positive in all of the points, thus, in PointA, inperiment E1 is mainly due to the advection term,Experiments E1 and E3, the heat source termsince the eastward speed component u in Pointis much larger than the other terms, therefore,B is almost 0 (see Fig.4a), thus -ugis aboutthe temperature increment there is the largest;"ahowever, in Experiment E4, since the outflow0, whilst the northward speed component v is中国煤化工MHCNMHGCAI Shugun et al. Acta Oceanologica Sinica 2004, Vol. 23, No. 2, p.239-249247positive and large,,Tis a large negative102* 105 108 11" 114 117° 120° 123" 126'EN厂magnitude and counter to that of the heat source24°term, thus the temperature increment is lttle.However, in Point A, since the current field is 21|different, -.T_1may be still positive, soF一V8*{that the sum of the advection term and heatsource term is the largest, thus the temperature5+increment is the largest. In the other points (C，12*{/D and E), -u_ vT approaches 0, thus thecorresponding temperature increment is mediumand almost the same. In Experiment E2, sincethe mixed layer thickness in the model increases,thus the heat source term is reduced and thelrtemperature increment is correspondinglyFig. 8. Distribution of temperature(C) field afterdecreased. In Experiments E3~E4, the waterrunning 90 d in Experiment E5.exchange from the Sulu Sea or south of theSunda shelf has lttle effect on Points C, D andfield is shown in Fig.8. As demonstrated in theE, showing that it has lttle contribution to theabove discussion, in the strong-current sea ar-NWW formation in the whole of the Nanshaeas such as in the Gulf of Tonkin, continentalIslands sea area.shelf of the northerm SCS and near the Gulf ofThus, it can be inferred that, the NWW'sThailand, the temperature distribution is veryformation in the Nansha Islands sea areas isdifferent from that in Experiment El, whilst inmainly due to the heating by the local net heatthe other sea area, especially in the Nansha Is-flux; meanwhile, it is important that, the cur-lands sea areas, the temperature field is almostrent there is weak and the temperature advec-the same as that in Experiment E1.tion transport is correspondingly weak, so thatthe water temperature can keep so high and4 Conclusionsconservative. In the sea areas where the cur-rent is strong, the temperature advection trans-In the paper, a reduced vertically integratedport is correspondingly strong, which mayupper mixed layer model is set up, and somemake the temperature increment reduced (e.g., numerical experiments are carried out to studyin Point B, although where the heat source termthe thermodynamic process of the formationis large).of the NWW in the Nansha Islands sea areas inspring. According to the above experimental3.4 Experiment E5: to test if the temperature results, the following conclusions can beadvection transport is importantreached:It is shown that, in spring, the water in theIn Experiment E5, the advection term inNansha Islands sea areas gets heat from theEq. (4) is ignored. The resulted temperature air. Although the obtained positive heat in those中国煤化工MYHCNMHG248CAI Shuqun et al. Acta Oceanologica Sinica 2004, Vol. 23, No. 2, p.239-249sea areas may be not larger than that in the otherActa Oceanologica Sinica(in Chinese), 22(3): 7~14sea area, however, the current here is very weak,Cai s, Long X, Chen R, et al.2004. Study on the circula-and the corresponding temperature advectiontion stnucture in the Nansha Islands sea areas, in spring.term is lttle, thus the heat is conserved and theJ Tropical Oceanogr(in Chinese), 23(2): 37--44water temperature gets high. The formation ofCai Shuqiun, Su Jilan, Gan Zijun, et al.2001. The nu-the NWW is mainly due to the sea surface netmerical study on the dynamic mechanism of the Southheat flux. Supposing that there is warm waterChina Sea upper circulation in winter. Actaflowing into the southern SCS through theOceanologica Sinica(in Chinese ), 23(5): 14~-23Balabac Strait or south of Sunda shelf, this kindCasuli V.1990. Semi implicit finite difference methodsof water exchange can only make the adjacentfor the two-dimensional shallow water equations. Jlocal water temperature increase, but it con-Computational Physics, 86: 56~74tributes lttle to the formation of the NWW. TheChen Y, Zhang Q, Zhao Y.1998. Variation of the Nanshatemperature dispersion is very lttle by and large.warm water and its coupled-oscillation with ENSOIn the sea areas where the current is strong,and the subtropical high in the westerm Pacific. Studythe advection may also play an important roleof Air-Sea Interaction and Climatologic Characteris-on the temperature field. The simulation resultstics in the Nansha Sea Areas. Beijing: Science Press,agree basically with those from observation.80-89The shortcoming of the reduced verticallyda Silva A, Young C, Levitus S.1994. Atlas of Surfaceintegrated upper mixed layer model is that theMarine Data v.1: Algorithms and Procedures. NOAAbaroclinic effect and the vertical convection ofAtlas NESDIS 6. Wash D C: U s Gov Printing Office,heat cannot be reflected. Thus, the thermody-83namic mechanism of the formation of the NWWFang w, Fang G, Shi P, et al.2002. Seasonal structuresin the Nansha Islands sea areas should be fur-of upper layer circulation in the southerm South Chinather studied by a better model.Sea from in situ observations. J Geophys Res, 107:3202, doi: 10.1029/2002JC001343AcknowledgementsHuang Q.1994. Distribution and change of the tempera-ture and salinity in the Nansha sea areas. ProceedingsThe authors are indebted to the anonymous refer-of the Study on the Physical Oceanography in theees for helpful comments. This work was supported byNansha Islands Sea Areas (). Beijing: China Oceanthe Knowledge Innovation Frontier Project of SouthPress, 39~63China Sea Institute of Oceanology, Chinese AcademyJia Y, Liu Q, SunJ.2000. Characteristics of the seasonalof Sciences under contract No. LYQY200310, Nationalchange of the warm water in the South China Sea andScience Foundation of China under contract Nosits numerical simulation. Oceanologia et Limnologia40376003 and 40276004 and National Special KeySinica(in Chinese), 31(4): 354- -362Project of China under contract No.2001DIA50041.Preller R H.1986. A numerical model study of theAlboran Sea gyre. Progr in Oceanogr, 16: 113~146ReferencesSynthetic Scientific Researtch Group of the Nansha IS-lands fromChinese Academy of Sciences. 1989. Syn-Cai Shuqun, Gan Zijun.2000. The application of a threethetic Investigation Reports of the Nansha Islandsdimensional baroclinic shelf sea model: the seasonaland its Adjacent Sea Areas (T) . Beijing: Science Press,variation of the South China Sea upper mixed layer.334-352中国煤化工MYHCNMHGCAI Shuqun et al. Acta Oceanologica Sinica 2004, Vol. 23, No. 2, p.239-249249Xie Q, Zhang Q, Hu J.1998. Characteristics of the longWang D.1998. Characteristics of the annual cycle of theterm change of the sea water temperature in theSST and the sea surface heat flux in the South ChinaNansha Islands and warm pool sea areas and itsSea. Study of Air Sea Interaction and Climatologiccoupled-oscillation. Study of Air Sea Interaction andCharacteristics in the Nansha Sea Areas. Beijing: Sci-Climatologic Characteristics in the Nansha Sea Areas.ence Press, 129~141Beijng: Science Press, 99-107Zhao Y, Zhang Q, Chen Y.1998. Basic characteristicsXu L, Sun L.1998. Relation between the monthly meantemperature, rainfall and EI Nino in the Nansha Is-ENSO. Study of Air Sea Interaction and Climatologiclands sea areas. Study of Air-Sea Interaction andCharacteristics in the Nansha Sea Areas Beijing: Sci-ence Press, 67-79Beijing: Science Press, 158~162中国煤化工MYHCNMHG