Calcareous Nannofossils in Surface Sediments of the Central and Northern Parts of the South China Sea

Taxonomic composition and quantitative distribution of calcareous nannofossils in the nothern and central parts of the South China Sea were studied in 146 samples of surface sediments taken from estuary, continental shelf, continental slope and deep-water basin. This paper reports distribution patterns of nannofossils in the area, including nannofossil abundance, species, assemblages and specimen size. Abundance increases from shelf to slope, with a decrease from the lower part of the slope towards the abyssal plain. The assemblages are dominated by Emiliania huxleyi, Gephyrocapsa oceanica and Florisphaera profunda. Ecologically the effects of water temperature, and nutrient inputs can be detected in the distribution of nannofossils, while dilution by terrigenous materials and deep sea carbonate dissolution influence the sedimentological aspect of the samples. The composition of the South China Sea nannofossil assemblage enables it to be grouped with the central zone of the North Pacific in biogeographic zonations of nannoplankton.


INTRODUCTION
In the Chinese sea area, systematic research on calcareous nannofossils from surface sediments has been concentrated in theEastChinasea (Wangand Min, 1981;WangandSamtleben, 1983;Wang and Cheng, 1985;Zhang and Siesser, 1986). In the South China Sea, Okada and Honjo (1975) analysed calcareous nannoplankton in 61 water samples collected south of 11"N, also Chen and Shieh (1982), Okada (1983) analysed calcareous nannofossils in surface sediment samples from the southern part of the Sea. Okada (1983) studied samples from the Gulf of Thailand, Chen and Shieh (1982) mainly worked on the Sunda Shelf and Southern Basin of the Sea. Varol (1985) studied calcareous nannofossils from nearshore localities in Jason Bay. There have been no previous studies of the calcareous nannofossils in the surface sediments of the northern part of the South China, which are the subject of this study.

STUDY AREA
The samples were collected from the the northern part of the South China Sea (12"-23"N, 108"-118"E) (Fig.l), including the eastern part of Beibu Gulf, continental shelf, continental slope, and abyssal plain. Two major rivers flow into the area -the Zhujiang (Pearl) River and the Hanjiang River. The bottom topography of the area is high in the northwest and low in the southeast, the maximum water depth in the area is over 4000m. The shelf / slope boundary is at a depth of approximately 150m. The abyssal plain starts at approximately 3600m (Physical Geography of China Compilation Committee, 1979). From continental shelf to abyssal plain, the clastic sediments fine gradually, the 30% sand contour is at about 200m (Fig.2a); and CaCO, content gradually declines, it is less than 10% below a water depth of 3500m (Fig.2b).

MATERIALS A N D METHODS
146 samples were studied; 100 of these were collected during 1983-84 by the Second Institute of Oceanography, State Oceanic Administration, 46 were collected during 1974-78 by South China Sea Headquarters of Geological Survey, MGMR (Fig.1).
Light microscope slides were prepared by diluting 0.1 grams of sediment in 40 millilitres of distilled water and spreading a drop of the slurry on a cover glass (24 X 32 mm'). All samples were studied under polarised light microscope (Leitz ORTHOLUX 2 POL BK). The number of nannofossils per 10 randomly selected fields of view at X630 magnification was used as the nannofossil abundance of the sample. If this abundance was under 300, a further 10 fields were examined, and the results averaged. In addition, these abundances are converted to specimens per gram of sediment (Tab.1) (specimens per gram of sediment = specimens per 10 fields of view X (24x32mm2 / 10 fields of view area) X (40ml / drop volume) X (lg / 0.lg) = specimens per 10 fields of view X 1300 X 880 X 10 = specimens per 10 fields of view X 11.44 million).
If the difference between specimens per 10 fields of view counted in different times is 1, the difference between specimens per gram of sediment estimated in different times will be 11.44 million. The estimate of specimens per gram of sediment is accurate to at most lo'. Seventy-one samples with more abundant nannofossils were examined with a scanning electron microscope (Hitachi H-8010 SEM part of Hitachi H-800 transmission electron microscope) at X5000 magnification. In forty-eight of them, with the most abundant nannofossils, the assemblages were counted (counts of more than 300 specimens) to assess the relative abundance of taxa and specimen size.  , 1981).

Nannofossil Abundance
The abundance distribution of calcareous nannofossils (Fig.2~) is closely related to water depth. Specimens are widely distributed in the samples, from nearshore at a water depth of few metres to the deep basin at a water depth of more than 4000 metres. However, theabundancesrange from0 to 1664. Abundances over 300 are limited mainly to water depths of 200m to 3500m (Tabs.1 and 2, Fig.2~). Calcareous nannofossils form less of the sediments in the shelf area (<200m) and much less of the sediments in the abyssal plain (>3500m), where almost no nannofossils could be found in most samples.
Sediment grain size and carbonate content are closely related to nannofossil abundance. Nannofossil content is low in coarse sediments and high in fine sediments. For example, at stations on the northern shelf, sediments are coarse and the sand content is high, and most nannofossil abundances are less than 300 (Figs. 2a and 2c). Generally speaking, high nannofossil contents correlates with high carbonate content (Figs. 2b and 2c).

Species Distribution
Twenty-nine calcareous nannofossil species or groups of species, and two species of calcareous dinoflagellates  (Thoracosphaera heimii and T . tuberosa), were identified (Tab.3). For each species, the percentage of samples it occurred in "frequency", and the average percentage of the total counted assemblages in "average relative abundance", are given in Table 3. In addition, some reworked nannofossils, such as Discoaster deflandrei, D. brouweri, Pseudoemiliania lacunosa, Gephyrocapsa protohuxleyi, G. aperta and Sphenolithus abies, occurred sporadically in samples from the slope (Tab.5). Distributional aspects are described below for major species.  , 1967), and is often found in nearshore seas, for example, in marginal seas of the western Pacific (Okada and Honjo, 1975;Chenand Shieh, 1982;Wang and Cheng, 1985). McIntyre and Be (1967) described a "warm water form" and a "cold water form". The "warm water form" has T-shaped elements in both shields and a delicate plate of interconnected rods forming a grillcoversthe proximal side of the central opening.
The "cold water form" has T-shaped elements only in the distal shield and the solid proximal shield has a central plate of thin interlocked elements completely closing the pore. These forms were identified and counted separately in the study, but there was no indication that their occurrence varied with temperature. The relative abundance of the species ( Fig.3a) and the relative abundance of broken specimens (Fig.2d) have a close relation to water depth.
Gephyrocapsa oceanica Kamptner. This is an important species in coccolith assemblages in modern oceans and abundant in marginal seas around the western Pacific Ocean, it is a major component of coccolith assemblages in the East and South China Seas (Okada and Honjo, 1975;Chen and Shieh, 1982;Okada, 1983;Wang and Samtleben, 1983;Wang and Cheng, 1985). G. oceanica (Fig.3b) is also abundant and widespread in this area, and found in all samples counted for relative abundance. The relative abundance varies from 0.6% to 70%. The high values (>25%) are in the northern continental shelf at water depths of less than 150m and in the lower continental slope at water depths of more than 2000m.
Florisphaera profunda Okada & Honjo. F. profunda (Fig.3c), in terms of its dominance, is third only to E. huxleyi and G. oceanica. It is found in all assemblages counted. It is abundant (>lo%) in sediments from water depths of 1000 to 2000m. It has been widely neglected due to its unusual shape, so few references to the species can be found. According to Okada and Honjo (1973), living F. profunda is abundant in the deeper surface waters (100-150m) over a wide temperature range (10-28°C). Okada (1983) suggested that its relative abundance in surface sediments shows a positive correlation with water depth from the lower continental shelf to abyssal depths and that F. profunda dominates the associations in deep basins. This study found a different distribution. The relative abundance of F. profunda increases with water depth to 2000m but decreases below this depth.

Calcidiscus leptoporus (Murray
is distributed in water depth from 1000 to 3500m. Its average relative abundance is 0.79%, with values greater than 0.8% mainly found in water depths from 2000 to 3500m. The different varieties of C. leptoporus proposed by McIntyre et al. ( ,1970 were not distinguished because total numbers counted were low. Umbilicosphaera sibogae (Weber-van Bosse) Gaarder. U.
sibogae (Fig3e) is not abundant (average relative abundance is 2.86%), but it is present in most samples (frequency is 89.58%). There is little regularity of relative abundance variation although most of the stations in which the value is over 4% come from water depths between 1000 and 3000m. McIntyre and Be (1967) distinguished cold water and warmwater forms of U. sibogae. They were not distinguished in this study because of their low relative abundance. The species was also found in the East China Sea (Wang and Cheng, 1985), but with lower relative abundances than those found here.
Syracosphaera pulchra Lohmann. This species is rare (average relative abundance is 0.84%) and there is no obvious pattern to the relative abundance variation (Fig3f).
Helicosphaera carteri (Wallich) Kamptner. The relative abundance of H. carteri (Fig.3h)   Specimen Size E. huxleyi and G. oceunicu are smaller than 6 microns and can be designated "small nannofossils". The percentage of the small nannofossils is higher (over 85%) nearshore and lower seawards (Fig.4b).

Comparison
The nannofossil assemblages in the central and northern parts of the South China Sea can be compared with those in the southern part of the Sea (Chen & Shieh, 1982 and Cheng, 1985) and the study area. They are generally similar, however, the dominance of E. huxleyi and G. oceanicu is higher in the East China Sea. The relative abundances of other species, especially warm water species, U. irregularis, U. sibogue and 0.frasilis are higher in the South China Sea. These differences may result from the difference in latitudes. The South China Sea is situated in the tropics, the variation of surface water temperature there is less than that in the East China Sea. Roth and Coulbourn (1982) recognized four coccolith assemblages (equatorial, central, transitional and subarctic) (Tab.4) in surface sediments of the North Pacific. They found that the coccolith assemblages followed the distribution of surface water masses. The central assemblage is dominated by E. huxleyi and G. oceunicu. Althoughnot a dominant species, U. tenuis was considered an important water-mass discriminator because it is restricted to the central water-mass. Comparing the average nannofossil assemblage in the central and northern South China Sea with the Pacific assemblages, that of the South China Sea is most closely equivalent to the central assemblage of the North Pacific, as shown by the dominance of E. huxleyiand G. oceunica, and the presence of U. tenuis(Tab.4).
Controlling Factors E. huxleyi and G. oceanicu are adapted to different environments. The former is the most ubiquitous species in today's seas, and can be found from tropical to subpolar waters; the latter is a warm water species, occurring in tropical, subtropical and temperate zones. The high water temperature in the South China Sea is responsible for the dominance of G. oceanica and E. huxleyi in the area. E. huxleyi is found in low to highly fertile waters throughout the world's oceans, but G. oceanica dominated in highly fertile and productive waters. Nutrients (eg. phosphate) have an effect on G. oceanica. In the Gulf of Aqaba (Elat), Red Sea, Winter (1982) found that the distribution of G. oceanica was closely related to the distribution of phosphate. While G. oceanica in surface sediments was decreasing with phosphate from south to north, E. huxleyi was increasing. So it is likely that the riverine nutrient input cause the higher productivity of G. oceanica in the coastal area. However, the predominance of G. oceanica over E. huxleyi in the deep water area might be due to dissolution of delicate E. huxleyi. Evidence for this is provided by the increase in broken specimens of E. huxleyiin samples from below 2000m (Fig.2d). Also the increase in C. leptoporus in deep water (Zone 3) are presumably dissolution effects.
The distribution of nannofossils in surface sediments is also influenced by sedimentological factors. Gartner (1981) found that on the continental shelf of the northern Gulf of Mexico coccoliths were relatively rare in the predominantly detrital sediments, even though they were produced in abundance in the water column. He thought this was due to the dilution by detrital materials. The continental shelf of the studied area receivesahighinput of terrigenous materialsfrom thezhujiang and Hanjiang Rivers, so it is likely that the low abundance of nannofossils nearshore is a result of dilution by terrigenous materials. As noted above, the distribution of calcareous nannofossils is closely related to the sediment carbonate content. A low value of abundance often corresponds to a low value of carbonate content. Although there is a close relationship between distributions of nannofossil and carbonate, carbonate content depends not only on nannofossils but also on other factors. Distribution of sediment carbonate in the South China Sea is mainly controlled by three factors: dilution by non-carbonate components, supply of calcareous skeletons of organisms, and carbonate dissolution. Carbonate content is lower near shore (e.g. 12.30%, 11.80%, 18.80% and 16.60% in stations G81, G82, G83 and G86) (Tab.5) due to dilution by terrigenous materials. It is also controlled by the supply of biogenic carbonate. Its distribution is closely related to these fossils, for example, foraminifers and nannofossils. It decreases towards abyssal plain as foraminifers (Zheng, 1987) and nannofossils (Figs.2b and 2c) do. Another important controlling factor of the distribution and content of carbonate is preservation. Carbonate dissolution increases with depth as bottom water becomes more undersaturated in calcium carbonate (Kennett, 1982). So low value of carbonate content including foraminifers and nannofossils (40%) could be attributed to the carbonate dissolution in deep sea. The relation between nannofossil abundance and sediment grain size reflects, to a certain extent, the effect of sorting on grain size, which results in more abundant nannofossils in fine sediments than in coarse ones.

CONCLUSIONS
1. Calcareous nannofossil abundance in surface sediments increases from continental shelf to slope, with a decrease from lower continental slope to abyssal basin.
2. E. huxleyi and G. oceanica are dominant species. The high relative abundance zones of the latter are situated in northern continental shelf and lower continental slope. Highest value of the former is in the area between water depths about 150 to 2000 metres.
3. The nannofossil assemblages in the northern and central parts of the South China Sea are similar to those in the East China Sea and in the central zone of the North Pacific Ocean.
4. There are two main groups of controlling factors: a) ecological factors including water temperature and supply of nutrients, b) sedimentological factors including dilution by terrigenous materials, deep water carbonate dissolution and the sorting effect.