Miocene to Pliocene Paleoceanography of the Western Equatorial Pacific Ocean Based on Calcareous Nannofossils , ODP Hole 805 B

We describe in detail the Miocene to Quaternary paleoceanography of the western equatorial Pacific Ocean based on calcareous nannofossils of Hole 805B. The relative abundance of Discoaster which lived in the lower photic zone under the stable sea with nutricline and thermocline, decreased step by step at NN5/NN6 and NN10/NN11 boundaries. Although the size of Reticulofenestra which is strongly influenced by nutrient, increased five times throughout the section, it drastically decreased in NN4-5 zone, NN10, NN12, and in NN15/NN16 boundary. On the basis of the relationship between Discoaster abundance and Reticulofenestra size change, collapse of the stability of the sea surface in the western equatorial Pacific Ocean progressed step by step throughout the Miocene to Quaternary.


Introduction
Calcareous nannofossils, which have been widely distributed in the world's oceans since the Mesozoic, are well known as a useful tool for paleoenvironmental analysis as well as for age determination.A living nannofossil species, Florisphaera profunda, which is presently found in lower photic zone [1], is a proxy to reconstruct the stability of the Quaternary sea surface condition with nutricline and thermocline.However, as F. profunda is not found in the Pliocene or older sections, it was difficult to reconstruct the sea surface conditions during the Neogene Period.
Recently, it became gradually clear that the size changes of coccolith and productivity of nannoplankton are strongly influenced by sea surface conditions such as the presence of nutricline, thermocline or upwellings [2,3].Furthermore, as Discoaster which last appeared in 2.0 Ma [4], is indicated as a lower photic zone species same as F. profunda, abundant occurrence of Discoaster indicates the sea surface stratification and the oligotrophic condition.On the basis of these phenomena, firstly we describe in detail the calcareous nannofossil assemblages of Ocean Drilling Program (ODP), Hole 805B located in the western equatorial Pacific Ocean, and then we discuss the sea surface stability of the ocean from the Miocene to the Pleistocene, focusing on both the cocco-lith size and Discoaster abundance changes.

Oceanographic Setting
Hole 805B is situated in 3188 m of water on the northwestern margin of the Ontong Java Plateau, close to the equator of the western equatorial Pacific Ocean, (latitude 1˚13.68'N,longitude 160˚31.76'E).
This site is influenced under variable world ocean circulation systems (Figure 1), such as, North Equatorial Current (NEC), North Equatorial Countercurrent (NECC), and Indonesian Throughflow (IT), in association with the New Guinea Coastal Undercurrent (NGCUC), the South Equatorial Current (SEC), and Equatorial Undercurrent (EUC), respectively.The western equatorial Pacific Ocean is dominated by the Western Pacific Water Pool (WPWP), which is characterized by a thick mixed layer, a deep, well-developed thermocline, and a deep maximum chlorophyll [5].

Samples and Methods
The sedimentary section at Site 805 is stratigraphically complete, covering the Pleistocene to the upper Oligocene.The sedimentary sequence was devided into two subunits (IA and IB) based on the degree of consolidation.The subunit IA from 805B-1H to 805B-30X is composed of nannofossil ooze to foraminifer nannofossil ooze, and subunit IB 805B-31X to 50X is characterized by nannofossil chalk to foraminifer nannofossil chalk [6].We examined Hole 805B of this site.
A total of 317 samples from ODP Hole 805B were made samples preparation.The sample of 0.0400 g (after pounded to be powder) was placed in a breaker, and then 50 ml of water was added.After suspension, 0.5 ml of suspension was pippetting to be evenly distributed on a cover glass (18 mm × 18 mm).The cover glass was carefully dried on an electric hot plate set at 40˚C, and placed on a slide glass with embedded medium to make slides.The slide was examined under a binocular polarizing microscope with an oil-immersion objective at lens at 1.500× magnification.
Several thousands coccolith and Discoaster were observed to recognized both their last and first occurrence horizons.After that, to obtain their floral change through the examined geological time, 200 specimens were counted at random.In addition, number of coccolith and Discoaster were counted in the area of one survey line (18 mm × 10 μm) to be estimated the quantitative amount of accumulation in a unit of sample weight (the number/gram).The size distribution of Reticulofenestra spp. was also measured to the 50 to 100 Reticulofenestra specimens in each sample.

Biostratigraphy
The zonal scheme by Martini [7], is applied for in this study.Calcareous nannofossil flora are characterized by abundant occurrence of Reticulofenestra spp.throughout the section (Figure 2).Coccolithus pelagicus, which is commonly distributed in the high latitude regions near Greenland [8,9], occurs abundantly below 220 m.Calcidiscus leptoporus are found commonly in the middle to upper section of Hole 805B, whereas the samples in the lower part are characterized by the Miocene marker species such as Cyclicargolithus floridanus, Sphenolithus moriformis and Sphenolithus heteromorphus.
On the basis of distributions of marker species me ned above, the sequence in this study is correlated to NN2 to NN20 Zone from the early Miocene to Pleistocene.
Coccolith and Discoaster are phytoplank amount of these microfossils are roughly reflected to the primary productivity.The size variation of Reticulofenestra through time is also well known worlwide, but there are a few quantitative data.
We examined the coccolith and Discoaster produc as shown in Figure 3.
From NN2 Zone to the number of coccolith productivity decreased around 3.0 × 0 7 N/g through less then 1.0 × 10 7 N/g. of NN4-5 Zone, the average decreased step-wise from 1 Otherwise the highest pe 10% in the upper part of NN4-5 to lower part of NN11 olith size of Reticulofenestra speify the size distributions of NN4-5 Zone, N Discoaster productivity was recognized at a maximum in this period, reaching approximately 6.0 × 10 6 N/g (~50%).From the upper part of NN4-5 Zone through the upper part of NN8-10 Zone, the coccolith is likely to increase in number.On the other hand, Discoaster in the sequence of the upper part of NN4-5 to NN7 Zones decreased abruptly.The section above the NN7/NN8 boundary is marked by increasing Discoaster again until approximately 4.0 × 10 6 N/g, and then its number gradually decreased towards the upper Pliocene.
From the upper part of NN8-10 to NN16, the amount of coccolith increased in NN11, indicating ak around 3.5 × 10 7 N/g.The relative abundance of Discoasters is also shown in Figure 3 as an indicator the surface water stability.The relative abundance of Discoaster specimens steps down throughout the sequence.Although the average of the relative abundance is scored at about 15% in the interval from NN2 to the upper part Zone, to 3% in NN11 to NN16 Zone.

Size Variation
We measured the cocc cimens in each samples to clar them, throughout the sequence (Figure 4).
The increase trend of the maximum size is recognized in the following four intervals, NN2 to N4-5 to NN8-10 Zone, NN11 to NN12 Zone, and NN13 to 15 Zone.As shown in Figure 4, the statistical modal size of them shows the repetition of apparent gradual increasing period and abrupt size decresing.The left column of Figure 4 shows that the majority of Reticulofenestra showed change in size in associated with change in modal size, suggesting the flora kept the statistical Gauss distribution in size throughout time.It is noteworthy that the change in the maximum size of as like that the mode of the Reticulofenestra size is characterized by positive correlation with maximum size.Not only does Figure 4 show the size change in the mode, but also the bimodal distributions are marked every interval before and after the termination of maximization (Figure 4).The bimodal patterns are composed of both unchanged small groups and maximized large groups.
Although the smaller size mode is not distinct in the early to middle Miocene sequence, 2 -3 micron Reticulofenestra specimens mark a clear mode in the interval of NN11 to NN20 Zone.However, these two small and large groups are not always concurrently present throughout the time.The small size mode is not detected in the intervals in NN2 to NN10 Zone of the early to middle Miocene, while the maximized phenomena have been unextinguished since the late NN16.

Backgroun Nannoplan
The relation between the structure of the sea surface and ecology of coccolith previous studies [1,2,10].Molfino and McIntyre [1] described that Florisphaera profunda lives in lower photic zone in recent ocean, so that its relative abundance decreases in upwelling environment, because of the upward moving the nutricline.Hagino et al. [2] studied the ecology of the coccolithophorids in the western equatorial Pacific Ocean and clarified that the lower photic zone species disappear in unstable surface conditions such as upwelling environment.Takahashi and Okada [3] reconstructed the paleoenvironment of the Quaternary southeastern Indian Ocean based on the abundance of F. profunda.Although the abundance of F. profunda is a good proxy for reconstruction of sea surface stability, the species does not occur in the Miocene sequence.Instead, Dis-coaster species is useful to similar purpose.
Discoaster species occur abundantly in the Paleogene to Neogene intervals and had been interpreted as l warm water species [11].Recently, Aubry [12], Flores et al. [13], Stoll et al. [14] and Sato and Chiyonobu [15] summarized that the Discoaster species are a typical lower photic zone species same as F. profunda.Sato and Chiyonobu [15] studied the relationship between the size changes of coccolith and Discoaster abundance in the Paleogene to Neogene sequences, and described the changes of stability of the sea surface conditions briefly.
Coccolith size changes in the Neogene sequence were discussed by Young [16] and Kameo and Bralower [17] d indicated as a proxy of stratigraphic utility.Recently, the size changes and productivity of coccolith are dis-tions are related to the development of both thermocline and nutricline, since they strongly influence to the productivity and size of coccolith.

Paleoceanography of the Western Equatorial Pacific Ocean
Our results indicate that lower coaster decreased step by step (Figure 5).This means that thermo-and nutri-cline developed in NN2 to the r part of NN4-5 nutri-cline collapsed step by step from the upper part of NN4-5 zone to the NN10/NN11 boundary.The change in relative abundance is also closely related with the stepwise change in the modal size of Reticulofenestra (Figure 5).The termination 2 shown in Figure 5 is characterized by discontinuity in size changes, which is situated roughly in the same horizons of significant changes in Discoaster abundance.This spontaneous event can be explained by the stepwise collapse of the thermo-and nutri-cline collapsed step by step at NN4-5 Zone and NN10/NN11 boundary.As a result, the sea surface of the western equatorial Pacific Ocean changed from oligotrophic mode to eutrophic condition.
The maximum size of coccolith (Reticulifora) changed four times at the termination 1, 3, 4, and 5 (Figure 5).Differing from termination 2, it is unlikely to be positively compared with the changes in nce.This will be interpreted that minor paleoceanographic changes occured across the termination 1, 4, and 5, respectively.

Conclusions
We describe in detail the nannofossil biostratigraphy of the Hole 805B is Ocean.The sequence is c Zones of early Mio nofossil datums.
Collapse of the changes of Discoaster relative abundance and the coccolith size indicates that the thermoand nutri-cline progressed step by step at NN4-5 Zone and NN10/NN11 e western equatorial Pacific Ocean has been changed from oligotrophic to eutrophic conditions during the Miocene to Quaternary.

Figure 3 .
Figure 3.Comparison between Coccolith and Discoaster number (N/g), and relative abundance of Discoaster (%).Grey solid line indicates the mean of Discoaster percentage.

Figure 4 .Figure 4 .
Figure 4. Comparison between distribution of size variation of Reticulofenestra specimens (%) and the maximum and mode size of Reticulofenestra.

Figure 5 .
Figure 5. Simplified illustrations of percentage and mean of Discoaster, and also shown in left figure both maximum size and mode of Reticulofenestra.