孟加拉灣的葉綠素a，海表溫度和風(fēng)速的趨勢

Danushka FERNANDO 唐丹玲 徐華兵

摘要孟加拉灣（BoB）是一個高能量活躍的地區(qū)，其短期內(nèi)的動態(tài)變化將對浮游環(huán)境產(chǎn)生巨大影響.“風(fēng)泵”能夠在BoB海域?qū)е麓怪钡幕旌蠌亩绊懞１頊囟群腿~綠素濃度.本文對2006——2016年的月平均Aqua-MODIS 葉綠素a（chl-a）濃度數(shù)據(jù)和Sea WiFS月度氣候態(tài)數(shù)據(jù)進行了分析，研究了葉綠素濃度的時間/季節(jié)變化和溫度以及風(fēng)速的關(guān)系.基于季風(fēng)期間的chl-a變異與海表溫度（SST），評估了在BoB海域它們之間的關(guān)系和變化.chl-a濃度值的趨勢分析表明，該區(qū)域的垂直混合非常低，冬季最高，夏季最低.冬季最大chl-a濃度值為0.50 mg/m3，并且從2月開始下降到夏季季風(fēng)期間.與冬季季風(fēng)相比，夏季季風(fēng)期間葉綠素表現(xiàn)出較低的濃度.在夏季季風(fēng)期間，特別是在7月和8月，由于云層密集，衛(wèi)星傳感器無法準(zhǔn)確捕獲chl-a濃度值.chl-a濃度和SST之間相關(guān)系數(shù)R2值為0.218 1.

關(guān)鍵詞葉綠素；風(fēng)泵；孟加拉灣，海表溫度；上升流

中圖分類號P724；P71

文獻標(biāo)志碼A

0 導(dǎo)讀

本文原文為英文，希望感興趣的讀者進一步關(guān)注原文.

孟加拉灣（BoB）是一個高能量活躍的地區(qū)，其短期內(nèi)的動態(tài)變化將對浮游環(huán)境產(chǎn)生巨大影響.“風(fēng)泵”是指風(fēng)力驅(qū)動的海洋上層洋流和水體運動以及隨后的生態(tài)效應(yīng)的影響，它可改變海洋中關(guān)鍵元素的運動和循環(huán)，從而影響海洋生態(tài)系統(tǒng)中的初級生產(chǎn)等.本文研究了葉綠素a（chl-a）濃度的時間/季節(jié)變化及其與海表溫度（SST）和“風(fēng)泵”的關(guān)系.

本文對于10年（2006—2016年）的月平均Aqua-MODIS chl-a數(shù)據(jù)和Sea WiFS月度氣候態(tài)數(shù)據(jù)進行了分析.基于季節(jié)的chl-a濃度變異與SST，評估了在BoB海域它們之間的關(guān)系和變化.

chl-a濃度值的趨勢分析表明，該區(qū)域的垂直混合非常低，冬季最高，夏季最低.冬季最大chl-a濃度值為0.50 mg/m3，并且從2月開始下降到夏季季風(fēng)期間.與冬季季風(fēng)相比，夏季季風(fēng)期間葉綠素表現(xiàn)出較低的濃度. 高chl-a濃度可能是由于印度洋地區(qū)冬季水平平流增強造成的，從而影響B(tài)oB海區(qū)的生物生產(chǎn)力.觀測發(fā)現(xiàn)在2013年chl-a濃度的最大值達到0.50 mg/m3，其原因是2013年12月在BoB西南部形成的熱帶氣旋（Madhi）橫穿了BoB西部地區(qū).Ekman抽吸速率指數(shù)是理解這個時期垂直混合的重要指標(biāo).

在9月到次年1月期間，風(fēng)引起的混合導(dǎo)致溫度的降低和chl-a濃度增加.冬季季風(fēng)期間，深層和表層水混合使表層營養(yǎng)物質(zhì)增加，溫度也降低.SST下降，解釋了BoB地區(qū)冬季季風(fēng)（12月）內(nèi)出現(xiàn)最大chl-a濃度的原因，也解釋了表層和深層水的混合導(dǎo)致營養(yǎng)物質(zhì)供應(yīng)到上層并提高表層的生產(chǎn)力的現(xiàn)象.在夏季季風(fēng)期間，特別是7月和8月，衛(wèi)星傳感器無法準(zhǔn)確捕獲chl-a濃度值，原因是夏季季風(fēng)期間云層密集.由于云量覆蓋像素問題，冬季季風(fēng)和季風(fēng)轉(zhuǎn)換期是最適合研究BoB區(qū)域的葉綠素a濃度的時間，因為這種空間驗證的像素可用性遠高于夏季季風(fēng)期間.

本研究顯示了葉綠素a濃度與海表溫度的相關(guān)性（R2=0.218 1，p<0.05）以及風(fēng)速之間的相關(guān)性（R2=0.193 1，p<0.05）.BoB海區(qū)呈現(xiàn)中度的正相關(guān)，最可能的原因是海區(qū)強烈的分層，特別是夏季季風(fēng)期間.和其他印度洋海區(qū)相比，BoB的風(fēng)場類型會刺激較小的上涌，因此，隨著chl-a濃度的增加而出現(xiàn)正相關(guān).與其他海區(qū)一樣，風(fēng)速也被認(rèn)為是BoB引起垂直混合最有效的能量驅(qū)動.

Abstract The Bay of Bengal （BoB） is a high energy active region，dynamics of BoB varies during short term with huge effect over the planktonic environment.“Wind pump” induces vertical mixing in the BoB region，which affects the sea surface temperature （SST） and chlorophyll-a （chl-a）.This study demonstrates the temporal/seasonal variation of chl-a concentration and its relationship to SST and wind speed.Monthly averaged Aqua-MODIS chl-a data for a period of 10 years （2006-2016） and Sea WiFS monthly climatology data were analyzed.Based on monsoonal chl-a and SST variability，we appraised their relationship and variations over the BoB.Trend analysis of chl-a concentration values shows that vertical mixing is very low in this region with weak annual phase，which reaches its maximum in winter and minimum in summer.About 0.50 mg/m3 is observed during winter as the maximum chl-a concentration value and then decreases since February until summer monsoon period.Summer monsoon period is identified as lack of chl-a concentration，compared to winter monsoon period.In summer monsoon period，especially in July and August，satellite sensors couldnt capture chl-a concentration values accurately due to the dense cloud cover.The R2 value for relationship between chl-a concentration and SST is observed to be 0.218 1.

Key words chlorophyll a；Wind Pump；Bay of Bengal；sea surface temperature；upwelling

1 Introduction

Marine plants and phytoplankton are the major photosynthetic sources in the ocean surface，which significantly influence the fluctuation of atmospheric Carbon Dioxide （CO2） （global carbon cycle） and primary productivity of the ocean systems.The major process is the synthesis of organic carbon using inorganic CO2 that varies in different climatic ambience[1-2].When validating or assessing the primary productivity in the ocean，chlorophyll-a （chl-a） concentration is a vital parameter[3].Many nutrients such as nitrite，phosphate，silicate，etc.are highly usable （within the photic zone，where phytoplankton occur in abundance） in phytoplankton environment especially upper layers of the ocean[4].Distribution and density of phytoplankton depend on the nutrient level in the ocean surface and sunlight availability，but the mixing of water，occasional fronts，ocean circulation，cyclones，and upwelling also affect the intensification of phytoplankton distribution[2，5-6].Especially，“wind pump”[7] is a major factor for phytoplankton distribution in the ocean.Defined as the impacts of wind-driven ocean currents on water transport and subsequent ecological effects，“wind pump” changes the movement and the cycling of key elements in the ocean thus affects the primary production，in marine ecosystems[6] these natural phenomena are varying according to their own time scales （annually or seasonally） and chl-a concentration fluctuates almost equally with them.But there is lack of studies and assessments of chl-a relationship with physical properties in different regions of the ocean[8].

Upwelling and vertical mixing induced by “wind pump” are the most significant factors in surface layers of the ocean which regulate the majority of phytoplankton （chl-a） biomass[7].This biomass depends on several limiting factors that is suggested by correlations such as the relationship between wind speed and chl-a concentration，as well as wind speed and SST which are negatively correlated in the world ocean as typically[9].Fluctuations of wind speed regulate the depth of upper mixed layer and also increase vertical nutrient mixing which cools the upper layer of the ocean[10-12].Spreading of SST in the Indian Ocean is very significant because western part of the Indian Ocean is highly abundant with cooler water，although the western part of Pacific and Atlantic oceans are hot.Physical，chemical and biological properties in the Arabian Sea region and the BoB region are comparatively different as western part of the BoB receives an enormous freshwater supply which results in low salinity during monsoonal periods.Therefore，the thermocline is deeper in the western part of the BoB and upwelling is comparatively less considering other regions of the Indian Ocean[13-14].

The BoB is a very significant and important region of the northern Indian Ocean due to the abundant existence of short term and long term seasonal cyclones and eddies formation.Therefore，upwelling is highly induced during these periods[10，14].The BoB overall chl-a is comparatively higher than that in southern Indian Ocean around the equatorial region[15].This study targets to evaluate the temporal （seasonal） variability of chl-a concentration and SST，as well as the effect of wind speed variation，further to study their interrelationship over the BoB region in the northern Indian Ocean territory using statistical methods[2，16].Satellite-derived data and reanalysis data from specific data providers will be used as major data sources.

2 Data and methods

The study area extends around the BoB in the northern Indian Ocean （7-16°N and 82-91°E） and the variability of chl-a and SST in this area were studied （Fig.1）.The moderate-resolution imaging spectroradiometer （MODIS）_Aqua ocean color based monthly composite level-3 standard mapped image （SMI） at 9 km spatial resolution data were downloaded from National Aeronautics and Space Administration （NASA） Ocean Color （http：∥oceancolor.gsfc.nasa.gov）.Time duration is July 2006-Dec 2016.“Chlorophyll concentration data are calculated with MODIS algorithm （OC3M） and averaged in global ocean region.The OC3M algorithm is：log10（CHL）=0.283-2.753R+1.457R2+0.659R3-1.403R4 where R=log10[max（Rrs（443），Rrs（488））/Rrs（551）]”[17].Chl-a values are derived from the recorded radiance using the OC2 algorithm in MODIS-Aq.Chl-a concentration[18].Over 10 mg/m3chl-a values were not evaluated for this analysis because such values are sporadic.Furthermore，Sea-viewing Wide Field-of-view Sensor （Sea WiFS） monthly climatology 1°×1° data were downloaded from University of Hawaii website （http：∥apdrc.soest.hawaii.edu）.

The 0.25°×0.25°-pixels resolution monthly averaged SST data were downloaded from WindSat monthly averaged data products，Version-7.0.1 from the University of Hawaii website （http：∥apdrc.soest.hawaii.edu） for July 2006-December 2016.Since there werent precise values for June and July of 2007，average SST value was used for these two missing values.

Monthly averaged wind speed data were downloaded from National Centers for Environmental Prediction （NCEP） reanalysis data using University of Hawaii website （http：∥apdrc.soest.hawaii.edu）.Pixels resolution is 2.5°×2.5° and time duration was the same as above.

3 Results

3.1 Chl-a concentration variation in the BoB

The area-averaged chl-a concentration in the BoB area during the period of July 2006-December 2016 is showed in Figure 2.Considering this studied time period，the chl-a concentration is varied from 0.10 mg/m3 to 0.50 mg/m3.The lowest of the areal average of chl-a concentration of 0.10 mg/m3 is noted in May 2010 and the highest of 0.50 mg/m3 noted in December 2013.According to Figure 2，five major peaks are identified （0.50 mg/m3 in December 2013，0.45 mg/m3 in August 2006，0.36 mg/m3 in July 2006，0.35 mg/m3 in September 2012 and 0.35 mg/m3 in September 2016）.

3.2 Error of cloud cover pixel

In literature review，many researchers suggested the noticeable cloud cover period during the summer monsoon period （especially June to August）.Therefore，this cloud cover decreases the satellite observations and accuracy of satellite data[12，19].In this study，a lack of pixel data during July and August is also found （figures are not provided） from 2006 to 2016.Figure 3 also shows high chl-a concentrations during July and August.This observation coincides with above researchers suggestions.This problem may be due to the lack of pixel data to take accurate average value.

3.3 SST and chl-a concentration

In this study，SST data during 2006-2016 are also observed to clarify the variability of chl-a concentration.SST variation is in sinusoidal distribution during this time period and varied in the range of 27.9-31.4 ℃ （Fig.4）.The minimum SST value （27.9 ℃） is noted during January 2007 and January 2014，and the maximum value of SST （31.4 ℃） is noted in April 2010.

3.4 Chl-a concentration and wind speed

Chl-a concentration and wind speed are moderately correlated in this area （R2=0.193 1，p<0.05，significant at 95%） （Fig.6）.Chl-a concentration is increasing with increase of wind speed.

4 Discussion

4.1 Variability of chl-a concertation

High chl-a concentration is found during September to January.This high chl-a concentration may be induced by the nutrient upwelling which is caused by the sea level anomalies and winds[12].“Wind pump” influences on water movements which changes the transport and the cycling of major elements in the ocean[6，20].Furthermore，open ocean upwelling which is also motivated by Ekman pumping increases the chl-a concentration by cyclones which are very copious during winter monsoon periods that triaged increase nutrients concentrations to upper layers in the BoB region[14].Variability of chl-a （Fig.2） for period 2006-2016 illustrates that the year 2013 has the maximum value （0.50 mg/m3） of chl-a concentration.In December 2013，category-2 tropical cyclone （“Madhi”） formed in the southwestern BoB and crossed over the western BoB area （https：∥www.nasa.gov/content/goddard/92b-northern-indian-ocean/）.Ekman pumping velocity index is an important character for understanding the vertical mixing in this period[21].This could be the most appropriate reason to explain the above high chl-a value.

Due to the cloud cover pixel error in monthly climatology，chl-a concentration in the lowest part of the study region during July 2006 and August 2006 may be poorly averaged.The resulted chl-a concentration and wind speed correlation shows positive trend （increase） and these results suggest the upwelling induced by wind energy mechanism （moderately）.Kumar et al.，（2002） evaluated the less productivity in the BoB than in the Arabian sea during the summer monsoon period and mentioned that surface layer of the BoB is strongly stratified by frequent rainfall，river inputs and weak wind patterns over the BoB.Though the summer monsoon is impotent to corrode the stratified layer，which leads to the reduction in vertical mixing[22].Figure 3 shows low chl-a concentration during May to June，but during July to September chl-a concentrations are very high as earlier （during summer monsoon）.It could be due to the aforementioned cloud cover pixel issue during the period from July to August.

4.2 Monsoonal effect and SST on chl-a concentration

Upper layer chl-a concentration is identified as lower during the years 2006-2010 compared to the other years and then increases during 2010-2013；the inconsistency of chl-a graph for 2006-2016 displays that the year 2013 has recorded the highest value （0.50 mg/m3） but it tends to decrease from 2014 to 2015，and appears to increase again after June 2016.Due to cloud cover pixel errors winter monsoon and the inter-monsoon period is most suitable to study chl-a concentration in the BoB area because the spatially validated pixel availability is much higher than that in summer monsoon season.During winter monsoon period productivity is higher in the BoB than during summer monsoon period mainly due to the stratified upper water column （excluding natural cyclones formation）.Most probably the occurrence of cold eddies controls the nutrient mixing to the surface ocean during the winter monsoon.The surface cooling is increased by the net heat loss from the sea surface.Relatively heavy winds induce the wind mixing and effect on the water column which also induce a proficient nutrients supply by cold eddies.Chl-a concentration in upper layers is chiefly dependent on stratification and wind mixing.Then，the chl-a concentration in the subsurface layer of the ocean is dependent on the existence of mesoscale eddies[23-24].Importantly，surface water consists high concentration of dissolved oxygen than bottom water.As well，warm water is incapable of holding more dissolved oxygen than cold water，which may explain above results[25].But there is no strong evidence of this impact on BoB region.Overall，a minor increasing trend of chl-a concentration is identified during 2006-2016 over the study area，which is evident from the statistical values （R2=0.000 2，p<0.05）.

During the study period，SST trend is positive but not significant （Fig.4），which is opposed to interpretations over the western Indian Ocean[26].SST is a most important factor （other than wind stress） for vertical upwelling and winter mixing in the western Indian ocean.As well as the SST decline during winter monsoon （about 3 ℃） describes the occurrence of maximum chl-a within the winter monsoon （December） over the BoB region.Reducing SST levels further explain additional mixing of surface water and deep water which cause the supply of nutrients to the upper layers and enhance the productivity of surface layers[27].Kumar et al. suggested that SST be considered as a proxy for factors which induce high chl-a concentration in the central equatorial Indian Ocean[16] but R2 value is observed as 0.22 between chl-a concentration and SST in the current relationship study.Therefore，it couldnt be clearly suggested as a proxy for variables on chl-a concentration in the BoB region as in central equatorial Indian Ocean.

4.3 “Wind pump” effects on chl-a concentration

Wind stress is possibly the most important force effect on the upper layer of the ocean[28].Generally，many scientists suggested that wind speed and chl-a concentration shows a strong positive relationship in open ocean[12，29-30]，while the BoB region showed moderately positive relationship between chl-a and wind speed.Maybe the reason for the latter is the strong stratification which forms especially during the summer monsoon.Mixing behavior of wind and chl-a does not depend on each other alone.The course of winds could induce the upwelling or downwelling[31].Wind patterns in the BoB stimulate minor upwelling considering other parts of the Indian Ocean，therefore，a positive correlation can be observed with wind and the increase in chl-a concentration.This can be considered as one of the “wind pump” effects[7].

5 Conclusion

During 2006-2016 the chl-a concentration illustrates moderate increasing trend in the BoB region.High chl-a concentration possibly occurs due to the enhanced horizontal advection during winter over the Indian Ocean region.This phenomenon affects biological productivity in the BoB region.The behavior of heavy winds and eddy featured cyclone formation are responsible for the nutrient uplifting to the surface areas and mixed layer enhancement through the cool thermocline water，via “wind pump” effects.

During September to January an increase in chl-a concentration occurs by this wind mixing and moderately cooling result.During winter monsoon high chl-a concentration can be observed due to the rich nutrients which are resulted by the deep and surface water mixing led by the decline in SST.Therefore，inhibiting the comparison between winter monsoon and summer monsoon.

This study shows both relationship between chl-a concentration and sea surface temperature （R2=0.218 1，p<0.05），and between chl-a concentration and wind speed （R2=0.193 1，p<0.05）.Moderate positive relationships are shown in BoB region.Most probable reason is the strong stratification especially in summer monsoon.Wind speed is considered as the most effective energy force in the vertical mixing in the BOB like other oceanic regions.

Acknowledgements：This study was funded by Key Project of the National Natural Sciences Foundation of China （NSFC41430968），Project of Guangdong Key Laboratory of Ocean Remote Sensing （LORS）award to DanLing Tang.Danushka Fernando was supported by UCAS scholarship （2017UCAS057）.The authors thank Liu Yupeng of LORS，South China Sea Institute of Oceanology，CAS.

References

[1] Falkowski P G，Barber R T，Smetacek V.Biogeochemical controls and feedbacks on ocean primary production[J].Science，1998，281（5374）：200-207

[2] Behrenfeld M J，OMalley R T，Siegel D A，et al.Climate-driven trends in contemporary ocean productivity[J].Nature，2006，444（7120）：752-755

[3] Antoine D，AndréJ M，Morel A.Oceanic primary production：2.estimation at global scale from satellite （coastal zone color scanner） chlorophyll[J].Global Biogeochemical Cycles，1996，10（1）：57-69

[4] Goldman J C，McCarthy J J，Peavey D G.Growth rate influence on the chemical composition of phytoplankton in oceanic waters[J].Nature，1979，279（5710）：210-215

[5] He Q Y，Zhan H G，Cai S Q，et al.Eddy effects on surface chlorophyll in the northern South China Sea：mechanism investigation and temporal variability analysis[J].Deep Sea Research Part I：Oceanographic Research Papers，2016，112：25-36

[6] Ye H J，Kalhoro M A，Morozov E，et al.Increased chlorophyll-a concentration in the South China Sea caused by occasional sea surface temperature fronts at peripheries of eddies[J].International Journal of Remote Sensing，2017（2）：1-16

[7] Tang D L，Kawamura H，Hai D N，et al.Remote sensing oceanography of a harmful algal bloom off the coast of coutheastern Vietnam[J].J Geophys Res，2004，109（C3），DOI：10.1029/2003JC002045

[8] Sugimoto T，Tadokor O K.Interannual：interdecadal variations in zooplankton biomass，chlorophyll concentration and physical environment in the subarctic Pacific and Bering Sea[J].Fisheries Oceanography，1997，6（2）：74-93

[9] Kahru M，Gille S T，Murtugudde R，et al.Global correlations between winds and ocean chlorophyll[J].J Geophys Res，2010，115（C12），DOI：10.1029/2010JC006500

[10] Chen X Y，Pan D L，Bai Y，et al.Episodic phytoplankton bloom events in the Bay of Bengal triggered by multiple forcings[J].Deep Sea Research Part I：Oceanographic Research Papers，2013，73（3）：17-30

[11] Tseng C M，Wong G T，Lin I I，et al.A unique seasonal pattern in phytoplankton biomass in low-latitude waters in the South China Sea[J].Geophysical Research Letters，2005，32（8）：487-500

[12] Fitch D T，Moore J K.Wind speed influence on phytoplankton bloom dynamics in the southern ocean marginal ice zone[J].J Geophys Res，2007，112（C8），DOI：10.1029/2006JC004061

[13] Vinayachandran P N，F(xiàn)rancis P A，Rao S A.Indian Ocean dipole：processes and impacts[J].Current trends in Science，2009：569-589

[14] Vinayachandran P N，Mathew S.Phytoplankton bloom in the Bay of Bengal during the northeast monsoon and its intensification by cyclones[J].Geophysical Research Letters，2003，30（11）：26-1-26-4

[15] Narvekar J，Kumar S P.Upper ocean variability of the equatorial Indian Ocean and its relation to chlorophyll pigment concentration[C]∥Proceedings of Ocean Obs，2010

[16] Kumar G S，Prakash S，Ravichandran M，et al.Trends and relationship between chlorophyll-a and sea surface temperature in the central equatorial Indian Ocean[J].Remote Sensing Letters，2016，7（11）：1093-1101

[17] Feng J F，Zhu L.Changing trends and relationship between global ocean chlorophyll and sea surface temperature[J].Procedia Environmental Sciences，2012，13：626-631

[18] Oreilly John E，Maritorena S，Mitchell B G，et al.Ocean color chlorophyll algorithms for Sea WiFS[J].J Geophys Res，1998，103（C11）：24937-24953

[19] Kumar S P，Sardesai S，Ramaiah N.A decade of physical and biogeochemical measurements in the Northern Indian Ocean[J].2010

[20] Tang D，Ni I H，Müller-Karger F，et al.Monthly variation of pigment concentrations and seasonal winds in Chinas marginal seas[J].Hydrobiologia，2004，511（1/2/3）：1-15

[21] Stewart R H.Response of the upper ocean to winds[M]∥Stewart R H.Introduction to physical oceanography.Orange Grove Text Plus，2002

[22] Kumar S P，Muraleedharan P M，Prasad T G，et al.Why is the Bay of Bengal less productive during summer monsoon compared to the Arabian Sea？[J].Geophysical Research Letters，2002，29（24）：88-1-88-4

[23] Kumar S P，Nuncio M，Narvekar J，et al.Seasonal cycle of physical forcing and biological response in the Bay of Bengal[J].Indian Journal of Marine Sciences，2010，39（3）：388-405

[24] Pan G，Chai F，Tang D L，et al.Marine phytoplankton biomass responses to typhoon events in the South China Sea based on physical-biogeochemical model[J].Ecological Modelling，2017，356：38-47

[25] Mitchell P，Prepas E E.Atlas of Alberta lakes[M].Edmonton，Canada：University of Alberta，1990

[26] Prakash P，Prakash S，Rahaman H，et al.Is the trend in chlorophyll-a in the Arabian Sea decreasing？[J].Geophysical Research Letters，2012，39（23）：L23605

[27] Prakash S，Ramesh R.Is the Arabian Sea getting more productive？[J].Current Science，2007，92（5）：667-671

[28] Huang R X.Ocean circulation：wind-driven and thermohaline processes[M].Cambridge，UK：Cambridge University Press，2010

[29] George D G，Edwards R W.The effect of wind on the distribution of chlorophyll a and crustacean plankton in a shallow eutrophic reservoir[J].Journal of Applied Ecology，1976，13（3）：667-690

[30] Tang D L，Kawamura H，Hai D N，et al.Remote sensing oceanography of a harmful algal bloom （HAB） off the coast of southeastern Vietnam[J].J Geophys Res，2004，109（C3），DOI：10.1029/2003JC002045

[31] Chen D K，Busalacchi A J，Rothstein L M.The roles of vertical mixing，solar radiation，and wind stress in a model simulation of the sea surface temperature seasonal cycle in the tropical Pacific Ocean[J].J Geophys Res，1994，99（C10）：20345-20359