Reconstruction of phytoplankton productivity and community structure in the South Yellow Sea

Xi Mei, Ri-hui Li, Xun-hu Zhng,b, Zhong-bo Wng,b, Yong Zhng,b

a Qingdao Institute of Marine Geology, China Geological Survey, Ministry of Natural Resources, Qingdao 266071, China

b Laboratory for Marine Mineral Resources, Qingdao National Laboratory for Marine Science and Technology, Qingdao, 266071, China

c Key Laboratory of Marine Sedimentology and Environmental Geology, First Institute of Oceangraphy, Ministry of Natural Resources, Qingdao 266061, China

Keywords:

Biomarker

Phytoplankton productivity

Phytoplankton community

Ocean warm current

Global climate change

Marine geological survey engineering

Yellow Sea

China

A B S T R A C T

The sedimentary environment and ecological system in the South Yellow Sea (SYS) changed dramatically due to sea level change caused by glacial-interglacial cycles.The authors report the use of marine biomarkers (brassicasterol, dinosterol and C37 alkenones) and terrigenous biomarkers (C28+C30+C32 nalkanols) in core DLC70-3 from the SYS to reconstruct the variation in the phytoplankton productivity and community structure and possible mechanisms during the middle Pleistocene.The results show that the primary productivity and that of single algae presented a consistent trend for the whole core during the middle Pleistocene, which was high during interglacial periods and low during glacial periods, with the highest being in marine isotope stage (MIS) 5-9 and MIS 19-21.The main reason is that the Yellow Sea Warm Current (YSWC) carried much of high temperature, high salinity water into the SYS, causing upwelling and vertical mixing and stirring, which increased the nutrient supply in the photosynthetic layer.The phytoplankton community structure mainly showed an increase in the relative content of haptophytes in MIS 5-9 and MIS 19-21, while the relative content of diatoms and dinoflagellates decreased; there was no evidence for a haptophyte content in other stages.The results reveal a shift from a coccolitho-phoriddominated community during MIS 5-9 and MIS 19-21 to a diatom-dominated community during the other stages, mainly as a result of surface salinity variation, attributed to the invasion of the YSWC during high sea level periods.

1.Introduction

The CO2content of the oceans is nearly 60 times that of the atmosphere (Sigman DM and Boyle EA, 2000); oceans control the marine carbon cycle through the biological pump.Marine phytoplankton, as the major producers in the ocean,controls the efficiency of the biological pump; and changes in the productivity or community structure can affect the pump(Archer D et al., 2000).Therefore, studying the evolution of marine phytoplankton productivity and community structure over geologic time is of vital significance for understanding the evolution of the carbon cycle and mechanisms for climate change.

Marginal seas are connected to the oceans and land; their organic carbon (OC) deposition rate is 8-30 times that of oceans and ＞80% of marine sedimentary organic matter (OM)is buried in them (Hedges JI and Keil RG, 1995).One of the key links is fluvial carbon transportation, which carries terrestrial organic matter (TOM) of 4×1014g C/year to the seas (Schlesinger WH and Melack JM, 1981).The sedimentary OM in marginal seas acts as an important carbon“sink ” and influences the global carbon cycle; it is also responsible for the changes in climate and environment(Muller-Karger FE et al., 2005).

Many methods are used for the reconstruction of marine phytoplankton productivity on a long time-scale; some examples are OC flux (Budziak D et al., 2000), planktonic foraminiferal carbon isotope composition (Thunell R et al.,1992), planktonic foraminiferal Cd/Ca ratio (Lin HL et al.,1999), Cl content (Harris P et al., 1996), biologic Ba flux(Dymond J et al., 1992), biologic opal content (Wang R and Li J, 2003) and calcareous nannofossils (Liu C et al., 2008).However, these indexes have their own problems; for example, total OC (TOC) derives from terrigenous and marine sources, so using it to reconstruct productivity may lead to difficulties (Werne JP et al., 2000).Algal fossils are prone to dissolution in sea-water and not easily preserved in sediments,making it difficult to estimate changes in phytoplankton productivity using fossils (Ragueneau O et al., 2000).In addition, the oxidation -reduction environment of deposition,the dissolution of BaSO4and terrigenous disturbance may offset the biologic Ba signal (Paytan A and Griffith EM,2007).

As geochemical indicators, biomarkers can have a definite source, are relatively stable and are not easily degradable in the geological environment.Many studies have used them, on a glacial-interglacial scale, to reconstruct the changes in marine phytoplankton productivity and community structure in different sea areas, including in the Okinawa Trough (Xing L et al., 2008), the South China Sea (He J et al., 2008, 2013),the sea of Japan (Xing L et al., 2011a), the Arabian Sea(Schubert C et al., 1998), the Indian Ocean (Schulte S and Bard E, 2003), the Tasman Sea (Calvo E et al., 2004), the sea of Okhotsk (Seki O et al., 2004) and the North Atlantic (Zhao M et al., 2006).Studies of the shelf seas in China began relatively late, mainly because of the strong sea-land interaction in the shelf sea area, multiple influencing factors and the complex mechanisms for community structure change.Only a few reports are therefore available on the use of biomarkers to reconstruct the changes in marine phytoplankton productivity and community structure of the shelf seas on a relatively short geologic time scale (Xing L et al., 2009, 2012; Wang F et al., 2012; Feng X et al., 2013;Yuan Z et al., 2013; Zhao X et al., 2013), and no reports of glacial-interglacial scale studies.

In this study, the authors chose a 71.2 m core collected from the abyssal region of the northern SYS and discuss the changes in phytoplankton productivity and community structure and factors influencing them, especially the response of the YS circulation formation and evolution on a glacialinterglacial scale in the deep water area of the SYS, using a series of biomarker indexes.The biomarker indicators used were: brassicasterol for diatoms, dinosterol for dinoflagellates, C37alkenones for haptophytes and long chainn-alkanols (C28+C30+C32) for higher land plants (Schubert C et al., 1998; Zhao M et al., 2006; Xing L et al., 2008, 2011a,2012).

2.Geological background

The YS is a typical semi-closed shelf sea, with the sediments coming mainly from mainland China and to some extent from the Korean Peninsula; sedimentation and the sedimentary structure are restricted by the marine dynamics(Yang SY et al., 2003).The northward YSWC, the southward coastal current and seasonal YS cold water mass (YSCWM)make the ecological environment of the YS diverse and complex.The effects of terrigenous input, confluence of cold and warm masses, and upwelling result in high productivity;for example, the average productivity has been estimated to be 672.4 mg C/m2/d in September and 835.6 mg C/m2/d in May from 1998 to 2003 (Son S et al., 2005).OC burial was enhanced by a high deposition rate and high productivity,making the YS an important carbon sink (Chen CTA and Borges AV, 2009).Therefore, studies of the ecological system of the YS are of significance.

Productivity in the SYS also has a clear spatial distribution and seasonal variation, related primarily to the nutrient supply, which is driven by the seasonal alternation of monsoon winds, river input and water circulation (Fu M et al.,2009).In summer, the weaker south wind, at a 1.5 m/s, is prevalent in the SYS (Mask AC et al., 1998).An obvious frontal in the horizontal direction divides the SYS into a shore mixed area and an open sea stratification area; below the thermocline of the open sea is the low temperature and high salinity YSCWM, which is relatively stable (Qiao F et al.,2006).Under the influence of the YSCWM, high primary productivity sites are distributed in the area, which is greatly influenced by the mainland coastal water; primary productivity is low in the open sea stratification area, which indicates that, in summer, primary productivity is controlled mainly by the change in nutrient sources caused by the thermocline block (Fu M et al., 2009).In contrast, the strong north wind is prevalent in the winter, and in January the average wind speed can be up to 10 m/s (Yuan Y and Su J,1984).As a compensation flow, the YSWC from the East China Sea intrudes into the SYS from SE to NW and the water undergoes vertical mixing (Zhang S et al., 2008); as a result, the bottom water becomes rich in nutrients vs the upper level, and the nutrient deficiency no longer remains the limiting factor for primary productivity.However, the primary productivity of phytoplankton is much less than that predicted by the nutrient concentration; the reason may lie in the change in illumination conditions caused by the change in the mixedlayer depth in winter (Fu M et al., 2009).

3.Sampling and methodology

The DLC70-3 core from the northern SYS (36°38'15' N,123°32'56" E, water depth 73 m) was drilled in September 2009 (Fig.1).It is 71.20 m long and the coring recovery rate was 93.0% on average; the sediments consist mainly of silt,sandy silt and silty sand.

Fig.1.Location of core DLC70-3 and other cores mentioned in the text, with surface currents in the South Yellow Sea (SYS) superimposed(modified from Liu J et al., 2010).Black arrows indicate paths of currents includng Kuroshio Current (KC); Yellow Sea Warm Current(YSWC); Tsushima Warm Current (TWC); The gray area indicates the mud distribution area of south Yellow Sea including the central south Yellow Sea Mud area (CS), southeastern Yellow Sea Mud area (SE), the mud area southwest off Cheju Island (CJ).NYS-the North Yellow Sea.

Samples were taken at 20 cm intervals from the core.Measurements for biomarkers included sample processing and instrumental analyses, which were carried out at the Key Laboratory of Marine Chemistry Theory and Technology,Ocean University of China, Ministry of Education/Qingdao Collaborative Innovation Center of Marine Science and Technology.

3.1.Sample processing

About 2 g freeze-dried sample was extracted ultrasonically 4 times with dichloromethane (DCM)/MeOH(v/v, 3:1), using a C24deuterium-substitutedn-alkane and a C19alcohol as internal standards (ISs).The combined extract was hydrolysed (KOH -MeOH), and then separated into three fractions using silica gel chromatography.The neutral lipid fraction (including alkenones and sterols) was eluted with 12 mL 5% MeOH (v/v, 95:5), dried under a gentle N2stream, and derivatised using N, O-bis (trimethyl-sily)-trifluoroacetamide(BSTFA) at 70℃ for 1 h before gas chromatography (GC)measurements were obtained.

3.2.Instrumental analysis

A thermo gas chromatography-mass spectrometry (GCMS) system was used for compound assignment and an Agilent 6890 N GC instrument for quantification.In both cases, an HP-1 column (50 m × 0.32 mm × 0.17 μm) was used with H2as carrier gas at 1.3 mL/min.The oven was programmed from 80℃ (held 1 min) to 200℃ at 25℃/min,then to 250℃ at 4℃/min, then to 300℃ at 1.7℃/min and finally to 310℃ (held 8 min).The content of each biomarker was calculated from the ratio of its GC peak integration vs that of the IS in its fraction.

4.Chronological framework

Using the magnetostratigraphic results together with the AMS14C dates, a chronostratigraphic framework was established for the sedimentary succession of core DLC70-3.Geomagnetic polarity boundaries or subchrons were recorded in the core (Mei X et al., 2016; Fig.2): the Matuyama-Brunhes chron boundary (M/B, 781 ka) at about 59.08 m.The core was estimated to cover the last about 0.90 Ma.It was divided into 4 depositional units (DU) according to lithology,color, sedimentary structure and grain-size parameters (Mei X et al., 2013, 2016).

Fig.2.Chronological frame of DLC70-3 core.The GPTS (Fig.2a) is from Gradstein FM et al.(2012), (L-Laschamp; Ja-Jaramillo; B/M:Brunhes-Matuyama chron boundary), sea-level estimate (Miller KG et al., 2005; Fig.2b, c).Magnetostratigraphy of the DLC70-3 core in the South Yellow Sea.The ages near the top of the cores were obtained via AMS 14C dating, data of core DLC70-3 are from Mei X et al.(2013,2016).

(i) DU1 (0-4.5 m).This unit is dominated by gray sandy silt, intercalated with silty beds containing millimeter-scale layers of silty clay to clayey silt.Shell fragments of oysters and other mollusks occur in the in the middle part.

(ii) DU2 (4.5-26.8 m).The unit is made up of gray,greenish gray, or dark gray silt to sandy silt with poor sorting,intercalated with beds of dark gray clayey silt.Cross-bedding is common in the sandy silt beds, which exhibit an erosional contact with the underlying muddy sediment.

(iii) DU3 (26.8-37.8 m).This unit is dominated by gray,dark yellowish gray, sandy silt to silty sand with poor sorting.On the basal erosional surface is a shell fragments of oysters bed of gray to dark gray fine sand, referred to as a deepening downward environmental transition from littoral zone to shelf.In this unit, the sand content is usually higher than 30%(average on 43%), and coarse fraction all show the highest values of the whole core, indicating that this unit was deposited in a generally high energy environment.

(iv) DU4 (37.8-71.2 m).This unit is composed mainly of dark gray to dark greenish gray silt with good sorting.In this unit, the sand content is usually lower than 20% (average on 8%), and coarse fraction indicating that this unit was deposited in a generally low energy environment.

5.Results

The content of biomarkers changed by 2-3 orders of magnitude since late Pleistocene as follows: The brassicasterol content (Fig.3a) changed from 0 to 401.2 ng/g(mean 85.0 ng/g), dinosterol (Fig.3b) from 0 to 801.9 ng/g(mean 102.3 ng/g), (C37:2+C37:3) alkenones (Fig.3c) from 0 to 926.8 ng/g (mean 96.6 ng/g) and cholesterol (Fig.3d) from 0 to 314.2 ng/g (mean 85.6 ng/g).

Fig.3.Contents of biomarkers, and other proxy records for DLC70-3.a-Content (ng/g) of brassicasterol; b-content (ng/g) of dinosterol; c-content (ng/g) of C37 alkenones; d-content (ng/g) of cholesterol; e- content (ng/g) of marine organic matter (MOM, sum of brassicasterol, dinosterol and alkenones); f-content (ng/g) of the terrestrial organic matter (TOM, C28+C30+C32 n-alkanols); g-ratio of MOM content to TOM content.Marine isotope stages (MISs) are labeled on the top panels.

The total content of brassicasterol, dinosterol and(C37:2+C37:3) alkenones, which acts as an indicator of the input of marine OM (MOM), changed from 4.3 ng/g to 1870.1 ng/g,with a mean of 284.9 ng/g (Fig.3e); the content of the(C28+C30+C32)n-alkanols, a biomarker of terrigenous OM(TOM) input, changed from 172.7 ng/g to 18723.7 ng/g, with a mean of 2453.5 ng/g (Fig.3f).The ratio of MOM/TOM changed from 0.01 to 1.38 (mean 0.18), a change of two orders of magnitude (Fig.3g).

The contents of brassicasterol and dinosterol followed a relatively consistent trend (Fig.4a;R2=0.53), with the highest values in MIS 5-9 and MIS 19-21, and the lowest in MIS 4 and MIS 5a-5d.The content of alkenones showed high values inMIS 5-9 and MIS 19-21; these components were not detected in the other units because of their extremely low content.

Fig.4.Correlation between concentration of brassicasterol and dinosterol (a); cholesterol and MOM (b); TOM and MOM (c) , in DLC70-3 core.

6.Discussions

6.1.Change in the biomarker content and productivity

Both biomarker content and sedimentation flux are considered as indicators of primary productivity (Zhao M et al., 2006; Xing L et al., 2008, 2011a, 2012).Sedimentation flux represents the influence of accumulation rate change on the content of biomarkers, so it can reflect well the change in primary productivity.However, calculation of the sedimentation flux needs sufficient amount of dating and density data from dry samples; hence, the samll number of age control points here would introduce significant errors into the calculation.Therefore, biomarker content was the only choice as a productivity index; this type of application has been found to be successful in studies of the surface layer and the Holocene drilling cores in the YS (Xing L et al., 2011b,2012).

As an indicator of MOM, the total content of brassicasterol, dinosterol and (C37:2+C37:3) alkenones changed in a manner similar to each these three individual biomarkers,showing a trend of higher values in MIS 5-9 and MIS 19-21 and lower values in other stages (Fig.3).In addition, as cholesterol is produced by various marine zooplankton species, it can be used as an approximate indicator of the total quantity of marine zooplankton, and thus be an indirect indicator of the total primary productivity.Irrespective of whichever community dominates the primary productivity, it will lead to the growth of zooplankton.As a result, making cholesterol a representative of the total primary productivity can avoid an interfering effect due to the difference between the relative content of phytoplankton biomarkers and the influence of the dominant biomarker in charging the gross primary productivity; the variation in MOM content and cholesterol content were basically similar (Fig.4b;R2=0.61).In general, primary productivity showed a trend from the high values in MIS 5-9 and MIS 19-21 to low values in other layers, being lowest in MIS10 and MIS 18.

The primary productivity of the West Pacific region was higher in glacial periods than in the interglacial periods, the main reason being the effects of the East Asian monsoon and terrestrial nutrient change (Zhang J et al., 2007).The YS, East China Sea and South China Sea are all important marginal seas of the West Pacific.Although there are a number of studies of the changes in productivity and community structure of phytoplankton in the South China Sea on a glacial-interglacial scale, only a handful have reported the use of multi-parameter biomarkers for the reconstruction of the productivity and community structure changes in the YS and the East China Sea.Xing L et al.(2008) reconstructed the evolution of the community structure and productivity of phytoplankton in the middle Okinawa trough at the edge of the east China sea shelf, and pointed out that productivity was higher in the deglacial period than in the Holocene due to the strengthening of the winter monsoon and the relatively lower sea level.They also reported that the proportion of coccoliths decreased, whereas that of dinoflagellates and diatoms increased (Xing L et al., 2008).In the SYS, earlier researchers reconstructed the primary productivity and the community structure of phytoplankton in the YE-2 core since 8200 ka,and found low productivity in the early Holocene; the productivity increased gradually with the rise in sea level and the formation of the YS circulation (Xing L et al., 2012).A study of the ZY1 and ZY2 cores relating to phytoplankton primary productivity since 6000 ka showed that it increased since 3000 ka, mainly because of the enhanced winter monsoon and the YS circulation (Zhao X et al., 2013).Biomarker studies of the SYS surface sediments also found low offshore phytoplankton productivity, and a high phytoplankton productivity in the deep water area of the SYS basin (Xing L et al., 2011b).

In summary, change in phytoplankton productivity is affected mainly by change in nutrient supply; for this core,two main factors affect the nutrient change: flux change in terrigenous input, and formation and evolution of the YS circulation.

6.1.1.Variation in terrestrial material input flux

Research on the nearby CC02 core showed that in the last deglacial period, sea level dropped by 56 m and and the sedimentation rate in the central part of the YS was higher than at present (Kim D et al., 1998).Research on the nearby YE-2 core also found that the terrigenous input index of long chainedn-alkane content in the early Holocene was 3×that in the mid-late Holocene; the reason of this is that the sea level in the early Holocene was low, leading to an increase in the fluvial input of terrestrial detritus (Xing L et al., 2012).

The TOM value of the DLC70-3 core can be an index of terrigenous input.During all the glacial periods, such as MIS 2-4, MIS 10 and MIS 18, the TOM values were high (Fig.3f)due to the low sea level, resulting in the core site coming much closer to land, which in turn increased the terrigenous input; this finding is in line with the previous research results.

On the other hand, MOM values, indicators of the change in the content of the phytoplankton productivity, showed no similarity trend (Fig.4c;R2=0.01).Unlike TOM values,MOM values appeared to be low during MIS 2-4, MIS 10 and MIS 18 (Fig.3e).The MOM/TOM ratio also indicated high phytoplankton productivity and low terrigenous input during MIS 5-9 and MIS 19-21 (Fig.3g).Results from the studies, of the surface sediments in the SYS (Xing L et al., 2011b) and core sediments (Xing L et al., 2012), indicated that an increase in terrigenous input material did not lead to a corresponding increase in phytoplankton productivity; on the contrary, in the deep zone or during high sea level periods when the transportation of terrigenous material was less, the phytoplankton productivity values were high.

In short, terrigenous input was not the major factor to result in a significant change in phytoplankton productivity of the DLC70-3 core on glacial-interglacial scales.

6.1.2.Formation and strengthening of YS circulation

The most important controlling factor for primary productivity in the modern SYS is the change in nutrient sources; in addition to the terrestrial river input, upwelling supplements are also an important source of the nutrients.Marine ecologists found that the primary productivity and community structure of phytoplankton are strongly influenced by the YSWC (Huo Y et al., 2012); hence, it was deduced that the formation and evolution of the YSWC on a longer time scales are also important factors in influencing phytoplankton productivity.Studies of the productivity of phytoplankton from piston core 10694 of from the SYS showed that the major driver of the primary productivity during the past 70 ka has been the enhanced cold vortex caused by the strengthening of the YS circulation (Xing L et al., 2009).On a longer time scale, the productivity of the phytoplankton from the ZY1 and ZY2 cores obviously increased since 3000 ka,with in a total record of 6000 ka, which was controlled mainly by the enhanced winter monsoon and the YS circulation(Zhao X et al., 2013).The productivity of the phytoplankton from the YE2 core was much higher in the mid-late Holocene than that in the early Holocene; this is because the modern YS circulation pattern was established after the formation of the YSWC (Xing L et al., 2012).

The distribution characteristics of the benthic foraminifera community in the DLC70-3 core indicated that the dominant species was a cold water species,Buccella frigida,in the layers (Fig.5d) corresponding MIS 5-9 and MIS 19-21 (Mei x et al., 2013, 2016), representing a similar cold water environment on the cold vortex edge to that of the present day.These findings showed that the paleo-cold water mass of the SYS existed during MIS 5-9 and MIS 19-21, indirectly suggesting that the YS circulation in that period was similar to the modern circulation pattern (Mei X et al., 2016).All the phytoplankton productivity indices, including MOM (Fig.3e),MOM/TOM (Fig.3g) and cholesterol content (Fig.3d) in core DLC70-3 showed two peaks at MIS 5-9 and MIS 19-21,which exactly matched the cold water horizon indicated by the benthic foraminifera, i.e.the corresponding periods of YSWC existence.It was speculated that upwelling and vertical mixing agitators driving the underlying supply caused by the strong YSWC increased the nutrient input, in turn enhancing productivity.

Fig.5.Normalized (total content) biomarker contents for DLC70-3 core.a-Brassicasterol; b-dinosterol; c-alkenone; d-abundance of cold water species (Buccela frigida); e-abundance of foraminifera;Marine isotope stages (MISs) are labeled on the top panel.

In summary, the formation and enhancement of YS circulation are throught to be the main factors influencing the high phytoplankton productivity recorded in the DLC70-3 core during MIS 5-9 and MIS 19-21.

6.2.Change in the ratios of biomarkers and phytoplankton community structure

Brassicasterol, dinosterol and alkenones are components of the cell membranes of diatoms, dinoflagellates and coccoliths, respectively, which can be used to indicate particular source organisms, and have similar properties; i.e.from the photosynthetic layer through sedimentation and diagenesis, the ratios between individual biomarkers and the sum basically remains unchanged, and so the ratio can also be used to indicate changes in phytoplankton community structure (Zhao M et al., 2006; Xing L et al., 2011a).

Changes in the proportion of biomarkers with depth are shown in Fig.5; it can be seen that the variation trends for brassicasterol and dinosterol are similar, while that of the alkenones is different.During MIS 5-9 and MIS 19-21, the proportion of alkenones was high, about 0.6 and 0.3 respectively, but 0 in the other layers (Fig.5c).In the high sea level periods of MIS 5-9 and MIS 19-21, the contribution from haptophytes was high, whereas those from diatoms and dinoflagellates were significantly lower; however, at other times, the phytoplankton community was fully occupied by diatoms and dinoflagellates.Significant changes in the phytoplankton community structure took place on a glacialinterglacial cycle; this situation is quite different from that in the Arabian Sea (Schubert C et al., 1998), Indian Ocean(Schulte S and Bard E, 2003) and South China Sea (He J et al., 2008, 2013), where the phytoplankton community structure was stable on the glacial-interglacial cycle, but is consistent with results from the Sea of Okhotsk (Seki O et al.,2004) and Sea of Japan (Xing L et al., 2011a).Studies of the Sea of Japan found that the large change in the phytoplankton community structure on a glacial-interglacial cycle was controlled mainly by a large salinity variation caused by sea level changes (Xing L et al., 2011a).

Phytoplankton community structure is influenced not only by nutrient concentration, but also by factors such as temperature and salinity.For haptophytes in particular,because the formation of the CaCO3shell depends significantly on the concentration of HCO3-in the sea water,low salinity is conducive to the growth of coccoliths (Wang P and Cheng X, 1988); in the oceans, the main species of coccolith,Emiliania huxleyi, has never been reported as living under a salinity of ＜ 11 ‰ (Van der Meer MTJ et al., 2008).Studies have shown that the salinity of the YS was low before the formation of the South YSWC.Research on the YE2 core has shown that, for 8.4-6.9 ka, the salinity of the SYS was ＜15‰, which started to gradually increase only after the formation of the modern YS circulation during 5-6 ka and to the current level (Xiang R et al., 2008).Studies of calcareous nannofossils in surface sediments from the SYS have shown that the distribution characteristics of the calcareous nannofossil content and the proportion of the main haptophyte species,Emiliania huxleyi, relatively consistent; the content is low in the nearshore shallow waters but high in the central zone, indicating that the distribution is affected mainly by water depth and the YSWC (Rui X et al., 2011).A study of the alkenone content of 36 surface sediment samples from the SYS has shown the same distribution characteristic as the calcareous nannofossils, high values in the deepwater area in the SYS basin and low value in the shallow water area along the coast (Tao S et al., 2012).The main species of calcareous nannofossils in the SYS,Emiliania huxleyiandGephyrocapsa oceanica, are the primary sources of linear alkenones, which can only be secreted by calcareous nannofossils (Wakeham SG et al., 2002).As a result, haptophyte productivity increased in the areas affected by the YSWC; although low salinity is conducive to the growth of coccolith, the higher salinity water brought in by the YSWC is advantageous for the growth of haptophytes.

In core DLC70-3, during the high sea level periods of MIS 5-9 and MIS 19-21, when the paleo-warm current existed, the phytoplankton community structure was changed by the influence of the warm current intrusion, and as a result the contribution of haptophyte contribution increased markedly.

Interestingly, with the disappearance of the YSWC, in early MIS 11-17 and early MIS 2-4, diatoms quickly became the main contributor of the phytoplankton community structure (Fig.5a), and the contributions of dinoflagellates and coccolithophorides almost became zero.The rapid fall of the sea level and disappearance of the YSWC reduced in salinity, resulting in extremely low coccolith content.Moreover, dinoflagellate content was also very low,indicating that in the process, diatoms were more competitive than dinoflagellates.In the community structure, light and nutrient are the most important resources; communities with different levels of primary productivity have different nutrient requirements; for example, diatoms and coccoliths are more sensitive to nitrogen and phosphorus, while dinoflagellates are more sensitive to carbon and phosphorus.The relationship between photosynthesis and irradiance of marine algae is also an important factor influencing the community structure and productivity; stronger light intensity is more suitable for dinoflagellates than for diatoms (Chen YLL et al., 2007).As for DLC70-3 core record, during early MIS 7-11 and early MIS 2-4 when the YSWC disappeared, why diatoms are more competitive than dinoflagellates is not clear and still needs a deep research.

7.Conclusions

Individual and total phytoplankton productivity in the DLC70-3 core show a consistent trend on the whole, high in the interglacial periods and low in the glacial periods, the highest being in MIS 5 and MIS 19-21 periods.The main reason for this is that the YSWC carried plenty of hightemperature and high-salinity water into the SYS, and caused upwelling and vertical mixing and stirring, increasing the nutrient supply in the photosynthetic layer greatly.

Biomarker ratios reveal a shift from a coccolitho-phoriddominated community during MIS 5 and MIS 19-21 to a diatom-dominated community during the other stages, mainly as a result of surface salinity variations in the SYS, which attributed to the invasion of the YSWC during the high-sealevel periods.

Acknowledgment

The authors are grateful to the crew of the R/V Kan 407 for their assistance with sample collection.Special thanks are also extended to Prof.Zhao Meixun and Dr.Xing Lei for help with biomarker measurements and advice, to the anonymous reviewers and the Executive Editor-in-Chief Dr.Yang Yan for their comments and suggestions, which significantly improved the quality of the manuscript.The work was jointly supported by the China Geological Survey (DD20160137,DD20190208) and the National Natural Science Foundation of China (No.41502175).