Pleistocene Calcareous Nannofossil Biostratigraphy and Gephyrocapsid Occurrence in Site U1431D, IODP 349, South China Sea

: We reinvestigated the Pleistocene calcareous nannofossil biostratigraphy of Site U1431D (International Ocean Discovery Program (IODP) Expedition 349) in the South China Sea (SCS). Twelve calcareous nannofossil Pleistocene datums are identiﬁed in the site. The analysis conﬁrms that the last occurrence (LO) of Calcidiscus macintyrei is below the ﬁrst occurrence (FO) of large Gephyrocapsa spp. ( > 5.5 µ m). The FO of medium Gephyrocapsa spp. (4–5.5 µ m) is also identiﬁed in the samples through morphometric measurements, which was unreported in shipboard results. Magnetobiochronologic calibrations of the numerical ages of LO of Pseudoemiliania lacunosa and FO of Emiliania huxleyi are underestimated and need reassessment. Other potential markers such as a morphological turnover of circular to elliptical variants of Pseudoemiliania lacunosa and a small Gephyrocapsa acme almost synchronous with the FO of Emiliania huxleyi may o ﬀ er biostratigraphic signiﬁcance in the SCS. The morphologic changes in Gephyrocapsa coccoliths are also examined for the ﬁrst time in Site U1431D. Placolith length and bridge angle changes are comparable with other ocean basins, suggesting that morphologic changes are most likely evolutionary novelties rather than being caused by local climate anomalies. These span different local changes in the morphology of in SCS may factors of regional perturbations. These results show the importance of both angle size for the determination of certain morphotypes for biostratigraphy.


Introduction
The Pleistocene is characterized by worldwide sea level lows and highs due to increased amplitudes of glacial and interglacial episodes, respectively [1][2][3]. Ice volume fluctuations are well represented in deep sea sedimentary sequences by variations in proxies such as sediment properties, fossil assemblages, and/or isotope concentrations. The basis for proxy correlations is a precise and robust chronology constructed from natural cycles of sediment physical properties, radiometric ages, magnetic reversals, and/or biostratigraphic data. An accurate age model is important because major climatic shifts in the Pleistocene occurred within a few hundred thousand years, as exemplified by glacial-interglacial amplitude variations in δ 18 O records [2]. Furthermore, several periods in the Pleistocene are thought to be analogs to the modern-day climate, such as Marine Isotope Stage (MIS) 11 [4,5] and may contribute to a further understanding of past and future climate change dynamics. These climatic events can be put into geologic context through relative dating by using calcareous nannofossils.
Calcareous nannofossils are mostly made up of remains of marine, unicellular phytoplankton known as coccolithophores. Coccolithophores produce a robust, calcium carbonate exoskeleton called the coccosphere. The coccosphere disintegrates into its coccoliths (individual calcite plates) upon death and is incorporated in fecal pellets and/or marine snow, thus dominating the carbonate fraction of deep-sea pelagic sediments. Calcareous nannofossils are utilized in sediment core studies because of their widespread geographic occurrence and stratigraphic significance. They underwent extinction and speciation events (datums) since the Triassic period, defining brief spans of geologic time [6]. The calibration of nannofossil datums through astronomical tuning and cyclostratigraphy has been substantial in creating a robust time scale for the Pleistocene [7][8][9]. Star-shaped discoasters are mainly used in subdividing the Early Pleistocene [7,[10][11][12][13][14]. Gephyrocapsids are also major components of Pleistocene calcareous nannofossil assemblages [15][16][17][18][19][20][21][22]. Gephyrocapsids have been extensively studied because of complex morphological changes, which have been utilized to subdivide the Pleistocene [7,[21][22][23][24]. However, semi-enclosed marginal seas may have different oceanographic parameters, resulting to a different calcareous nannofossil assemblage. It is therefore important to study the calcareous nannofossil biostratigraphy of marginal seas to verify if standard nannofossil zonations apply to these areas.
The South China Sea (SCS) is the largest marginal sea in the western Pacific area and has been subject to multiple paleoceanographic [25][26][27][28][29][30] and sedimentological studies [31][32][33][34]. The SCS is characterized by better carbonate preservation compared to other western Pacific marginal basins, providing a valuable sediment archive for East Asian Monsoon System (EAMS) variations since the Miocene [35]. Calcareous nannoplankton in surface sediments [36][37][38][39] have also been studied in the northern, western, and eastern SCS. Despite increasing scientific attention on the SCS, calcareous nannofossil biostratigraphic studies of Pleistocene deep sea sediments are limited. Most works are concentrated on the northern flanks of the SCS to acquire a record of the spreading history of the SCS [40][41][42].
International Ocean Discovery Program (IODP) Expedition 349 aimed to establish the ages of cessation of different spreading centers in the SCS Basin [43,44]. Site U1431D was the most successful in providing a continuous archive from the Miocene to Holocene, recovering around 400 m of sediment. Since Ocean Drilling Program (ODP) Leg 184 [42], IODP 349 is one of the very few research expeditions to obtain records since the start of ocean floor spreading of the SCS basin. The acquired cores, therefore, represent a valuable archive in which age constraints for different spreading events in the sub-basins can be studied. Furthermore, the thick sediment recovery provides a continuous sedimentary record of paleoceanographic changes in the SCS. Shipboard data have been able to estimate age dates of different sections of the whole core through biostratigraphic analysis using different microfossils [43]. Calcareous nannofossils have been useful in shipboard analysis to determine sedimentation rates and estimate the cessation of the sub-basin spreading. However, further refining of datum placements is needed to achieve a better time constraint for paleoceanographic changes.
This study reinvestigates the calcareous nannofossil biostratigraphy of the Pleistocene sediments acquired from IODP 349. Temporal size and bridge angle changes in the genus Gephyrocapsa are discussed in detail to assess the reliability of these bioevents in the SCS. The results of the research contribute to the assessment of nannofossil datums and their importance in Pleistocene correlation and chronology, especially in the SCS.

Present Day Circulation of the South China Sea
Surface water circulation in the SCS is currently influenced by the EAMS [30,45,46]. The EAMS is characterized by having a stronger winter and a weaker summer season. The winter season (December-February) is dominated by the northeast monsoon, causing a basin-wide cyclonic gyre ( Figure 1). The winter monsoon is mainly driven by cold winds from Siberia, which runs southward along the coast of East Asia. Mountains near the East Asian coast force the winter monsoon to produce a southwest traveling jet over the SCS. The summer season (June-August) is characterized by a weaker southwest monsoon forming an anticyclonic gyre concentrated at the south and a cyclonic cell at the north ( Figure 1). The ocean jet separates near Vietnam: one branch travels northeast crossing the central basin while the other continues on the northern path. Both branches of the current exit the SCS through the Bashi Strait. The ocean dynamics of the SCS produce localized annual upwelling in Vietnam and in northwest Luzon, although upwelling during winter is generally stronger based on chlorophyll concentrations [30,45,46].
Geosciences 2020, 10, x FOR PEER REVIEW 3 of 26 the current exit the SCS through the Bashi Strait. The ocean dynamics of the SCS produce localized annual upwelling in Vietnam and in northwest Luzon, although upwelling during winter is generally stronger based on chlorophyll concentrations [30,45,46].

Site U1431D
The sediment cores were obtained from the central SCS by IODP Expedition 349 from 1 to 6 February 2014, at Site U1431 ( Figure 1). Site U1431 is located near a relict spreading ridge where the youngest magnetic anomalies in the SCS are located [43]. The primary objective of the expedition was to determine the cessation of seafloor spreading at the East Sub-basin of the SCS. The site also provides an almost continuous sediment record, providing the opportunity to study a continuous, highresolution paleoceanographic evolution of the SCS. Core U1431D was recovered from a mean seafloor depth of 4240.5 m below sea level (mbsl). About 402.11 m of core was recovered in Site U1431D in a 617-m interval (65.2% core recovery). The uppermost 135 m of core material (from section 1H to 15H) was sampled every 20 cm for the first two sections (1H and 2H), and every 150 cm for the rest of the sections. A total of 168 samples were analyzed for this study ( Figure 2).
Eleven (11) lithostratigraphic units were defined by the shipboard team based on visual core descriptions, thin section analysis, biogenic materials, and physical properties. The first two units were investigated in this study, which spans the late Pliocene (section 15H) to the latest Pleistocene (section 1H; Figure 2). Unit I (0-101.16 m below sea floor (mbsf)) is characterized by a sequence of clay, silty clay, and clayey silt that has substantially more silt than Unit II. The silt layers are composed of quartz, feldspars (plagioclases and K-feldspars), and minor amounts of amphiboles, micas, and heavy minerals. The layers with more silt are greenish gray interbeds, usually with graded, fining upward sequences interpreted as turbidites [43]. Nannofossil oozes occur locally and often show a fining upward character, also identified as turbidite deposits. Unit I contains multiple dark gray to brown volcanic ash layers, 0.5-to 5.0-cm thick.

Site U1431D
The sediment cores were obtained from the central SCS by IODP Expedition 349 from 1 to 6 February 2014, at Site U1431 ( Figure 1). Site U1431 is located near a relict spreading ridge where the youngest magnetic anomalies in the SCS are located [43]. The primary objective of the expedition was to determine the cessation of seafloor spreading at the East Sub-basin of the SCS. The site also provides an almost continuous sediment record, providing the opportunity to study a continuous, high-resolution paleoceanographic evolution of the SCS. Core U1431D was recovered from a mean seafloor depth of 4240.5 m below sea level (mbsl). About 402.11 m of core was recovered in Site U1431D in a 617-m interval (65.2% core recovery). The uppermost 135 m of core material (from section 1H to 15H) was sampled every 20 cm for the first two sections (1H and 2H), and every 150 cm for the rest of the sections. A total of 168 samples were analyzed for this study ( Figure 2).
Eleven (11) lithostratigraphic units were defined by the shipboard team based on visual core descriptions, thin section analysis, biogenic materials, and physical properties. The first two units were investigated in this study, which spans the late Pliocene (section 15H) to the latest Pleistocene (section 1H; Figure 2). Unit I (0-101.16 m below sea floor (mbsf)) is characterized by a sequence of clay, silty clay, and clayey silt that has substantially more silt than Unit II. The silt layers are composed of quartz, feldspars (plagioclases and K-feldspars), and minor amounts of amphiboles, micas, and heavy minerals. The layers with more silt are greenish gray interbeds, usually with graded, fining upward Geosciences 2020, 10, 388 4 of 26 sequences interpreted as turbidites [43]. Nannofossil oozes occur locally and often show a fining upward character, also identified as turbidite deposits. Unit I contains multiple dark gray to brown volcanic ash layers, 0.5-to 5.0-cm thick.
Geosciences 2020, 10, x FOR PEER REVIEW 4 of 26 Figure 2. Schematic diagram of core sections (1H to 15H) examined in this study. It is generally composed of clays/claystone with interbeds of minor sand/sandstone, silt/siltstone, clayey silt, silty clay, and foraminiferal oozes. Age dates are determined from shipboard biostratigraphic data. White areas in the "Core Recovery" column approximates coring gaps. Horizontal lines in the "Samples" column approximate positions of the samples utilized for this study Abbreviations listed: CSF = core depth below seafloor, H = advanced piston corer, Plio = Pliocene. Figure modified from [43].
Unit II (101.16-267.82 mbsf) is divided into two sub-units. Only samples from Sub-unit IIA (101. .95 mbsf) are analyzed for calcareous nannofossils. This sub-unit is composed of clay, nannofossil-rich clay, and silty clay. The greenish gray clays are thickly bedded and are usually interbedded with greenish gray medium-to thickly-bedded nannofossil-rich clays and dark greenish gray thin-to medium-bedded silty clay layers. The clayey silt layers often show sharp, planar bases, which grade into clay and are interpreted as turbidite deposits [43]. Moreover, fine sand foraminiferal oozes with sharp, erosive bases, fining upward towards nannofossil-rich layers, are also interpreted as turbidites. Bioturbation from the Nereites ichnofacies are darker colored clay within lighter colored carbonate-rich sediments. Thin volcanic ash layers 0.5-to 2.0-cm thick are also identified but are much less compared to Unit I and are restricted at the upper portions of the sub-unit [43].

Calcareous Nannofossil Analysis
Smear slide replicates from unprocessed sediments were prepared following the standard procedure of [47]. At least 300 random fields of view (FOVs) were scanned using an Olympus BX51 phase contrast polarizing microscope under 1000× magnification. Additional fields of view were analyzed for samples with important calcareous nannofossil taxa. A sample was considered barren of calcareous nannofossils if there were none present after 300 FOVs. Relative abundances of individual calcareous nannofossils identified per slide were estimated using IODP standards [43]: D = dominant (>50%, or 100 specimens per FOV) A = abundant (10-50%, or 10-100 specimens per FOV) Figure 2. Schematic diagram of core sections (1H to 15H) examined in this study. It is generally composed of clays/claystone with interbeds of minor sand/sandstone, silt/siltstone, clayey silt, silty clay, and foraminiferal oozes. Age dates are determined from shipboard biostratigraphic data. White areas in the "Core Recovery" column approximates coring gaps. Horizontal lines in the "Samples" column approximate positions of the samples utilized for this study Abbreviations listed: CSF = core depth below seafloor, H = advanced piston corer, Plio = Pliocene. Figure modified from [43].
Unit II (101.16-267.82 mbsf) is divided into two sub-units. Only samples from Sub-unit IIA (101. .95 mbsf) are analyzed for calcareous nannofossils. This sub-unit is composed of clay, nannofossil-rich clay, and silty clay. The greenish gray clays are thickly bedded and are usually interbedded with greenish gray medium-to thickly-bedded nannofossil-rich clays and dark greenish gray thin-to medium-bedded silty clay layers. The clayey silt layers often show sharp, planar bases, which grade into clay and are interpreted as turbidite deposits [43]. Moreover, fine sand foraminiferal oozes with sharp, erosive bases, fining upward towards nannofossil-rich layers, are also interpreted as turbidites. Bioturbation from the Nereites ichnofacies are darker colored clay within lighter colored carbonate-rich sediments. Thin volcanic ash layers 0.5-to 2.0-cm thick are also identified but are much less compared to Unit I and are restricted at the upper portions of the sub-unit [43].
Nannoliths from Discoaster species are usually reworked specimens in younger sediments and occur both as fragmented and complete fossils within their age ranges. Thus, the last occurrences (LO) of discoasters are difficult to recognize because of reworking. Complete, incomplete, and arm fragments of 100 Discoaster specimens were counted from 135 mbsf to 91 mbsf to semi-quantitatively determine their relative abundances. In this study, specimens that have ≤60% of their original form intact are considered incomplete specimens. One or two arms of discoasters without a central area are considered arm fragments. The LO of the Discoaster is assigned to the midpoint depth between the sample containing >1% of the total count that are complete specimens and the level where it has disappeared. The informal grouping "Discoaster spp." is used for ambiguous stellate nannofossils. The Base Acme (Ba) of D. triradiatus was identified using the definition of [51]. Based on the counts, barren intervals were noted at 133.2-130.2 mbsf, 127.5 mbsf, 124.9 mbsf, 121.9-118.9 mbsf, and 99.9-98.39 mbsf.
Morphological analysis of gephyrocapsids was done using smear slides prepared from unprocessed sediments. The Gephyrocapsa coccoliths were photographed and measured using Image-Pro Plus Version 7.0 mounted on an Olympus BX51 polarizing microscope under 1000× magnification. The Gephyrocapsa coccoliths were measured based on the parameters set by [52] (Figure 3), mainly the coccolith length and width, coccolith central area length and width, and bridge angle. At least 40 specimens per sample were analyzed. In samples with few nannofossils, 200 FOVs were scanned and all Gephyrocapsa placoliths seen were measured. The study adopted the biometric subdivision of [21] which is mainly based on the maximum length of the coccolith. Forms <4 µm are small Gephyrocapsa spp.; medium Gephyrocapsa spp. are those with placolith lengths from ≥4 to <5.5 µm, and large Gephyrocapsa spp. are ≥5.5 µm in length. All measurements are provided in the Supplementary Materials.
Nannoliths from Discoaster species are usually reworked specimens in younger sediments and occur both as fragmented and complete fossils within their age ranges. Thus, the last occurrences (LO) of discoasters are difficult to recognize because of reworking. Complete, incomplete, and arm fragments of 100 Discoaster specimens were counted from 135 mbsf to 91 mbsf to semi-quantitatively determine their relative abundances. In this study, specimens that have ≤60% of their original form intact are considered incomplete specimens. One or two arms of discoasters without a central area are considered arm fragments. The LO of the Discoaster is assigned to the midpoint depth between the sample containing >1% of the total count that are complete specimens and the level where it has disappeared. The informal grouping "Discoaster spp." is used for ambiguous stellate nannofossils. The Base Acme (Ba) of D. triradiatus was identified using the definition of [51]. Based on the counts, barren intervals were noted at 133.2-130.2 mbsf, 127.5 mbsf, 124.9 mbsf, 121.9-118.9 mbsf, and 99.9-98.39 mbsf.
Morphological analysis of gephyrocapsids was done using smear slides prepared from unprocessed sediments. The Gephyrocapsa coccoliths were photographed and measured using Image-Pro Plus Version 7.0 mounted on an Olympus BX51 polarizing microscope under 1000× magnification. The Gephyrocapsa coccoliths were measured based on the parameters set by [52] ( Figure 3), mainly the coccolith length and width, coccolith central area length and width, and bridge angle. At least 40 specimens per sample were analyzed. In samples with few nannofossils, 200 FOVs were scanned and all Gephyrocapsa placoliths seen were measured. The study adopted the biometric subdivision of [21] which is mainly based on the maximum length of the coccolith. Forms <4 μm are small Gephyrocapsa spp.; medium Gephyrocapsa spp. are those with placolith lengths from ≥4 to <5.5 μm, and large Gephyrocapsa spp. are ≥5.5 μm in length. All measurements are provided in the Supplementary Materials.

Results
Twelve (12) calcareous nannofossil datums were detected for the Pleistocene section of Site U1431D. The preservation of calcareous nannofossils ranged from very poor to pristine specimens.

Results
Twelve (12) calcareous nannofossil datums were detected for the Pleistocene section of Site U1431D. The preservation of calcareous nannofossils ranged from very poor to pristine specimens. Overgrowth and fragmentation of calcareous nannofossils (especially among discoasters) are evident. Foraminiferal tests, sponge spicules, radiolarian tests, and diatom frustules were observed in the first 3 m of the core. Samples barren of nannofossils were dominated by detrital materials such as clay minerals, biotite, plagioclase, quartz, and muscovite. Some barren intervals coincide with volcanic ash layers [43].
Calcareous nannofossil abundances were generally low throughout the core, but were high in portions of sections 1H and 2H (0 to 12.2 mbsf). Recrystallization of nannofossil specimens was uncommon throughout the core. Samples with partially to heavily dissolved placoliths coincide with decreases in calcareous nannofossil relative abundances. Figure 4 illustrates important calcareous nannofossil species used as biostratigraphic markers for Site U1431D. in the first 3 m of the core. Samples barren of nannofossils were dominated by detrital materials such as clay minerals, biotite, plagioclase, quartz, and muscovite. Some barren intervals coincide with volcanic ash layers [43]. Calcareous nannofossil abundances were generally low throughout the core, but were high in portions of sections 1H and 2H (0 to 12.2 mbsf). Recrystallization of nannofossil specimens was uncommon throughout the core. Samples with partially to heavily dissolved placoliths coincide with decreases in calcareous nannofossil relative abundances. Figure 4 illustrates important calcareous nannofossil species used as biostratigraphic markers for Site U1431D. Table 1 [43] and this study for International Ocean Discovery Program (IODP) Expedition 349, Site U1431D. Age discrepancies from [43] are due to the incorporation of Equatorial Pacific ages (compiled by [9]) when available. Abbreviations listed: FO-first occurrence, LO-last occurrence, Ba-base acme, reemG-reentrance event of medium Gephyrocapsa, tlG-top (last occurrence) of large Gephyrocapsa, blG-base (first occurrence) of large Gephyrocapsa, bmG-base of medium Gephyrocapsa. The LO of D. surculus was previously assigned to Sample 15H-CC (135.20 mbsf) in the IODP 349 biostratigraphic results [43]. Rare to few specimens are observed from Sample 15H-CC to Sample 15H-3W until they occur sporadically as reworked specimens throughout the younger portions of the sediment record ( Figure 5A). There is a barren interval of nannofossils from 133.2 to 130.2 mbsf, but D. surculus reappeared as well-preserved specimens above this interval (Sample 15H-3W). It again reappears at 112.4 mbsf, but is regarded as reworked in this sample because not only is it higher than that of the bottom acme of D. triradiatus, a younger datum ( Figure 5A), but its occurrence is also anomalously located above samples with consistent absences of D. surculus. Thus, the LO of D. surculus is placed between Samples 15H-3W, 20 to 21 cm and 15H-2W, 20 to 21 cm (127.85 mbsf).  The LO of D. surculus was previously assigned to Sample 15H-CC (135.20 mbsf) in the IODP 349 biostratigraphic results [43]. Rare to few specimens are observed from Sample 15H-CC to Sample 15H-3W until they occur sporadically as reworked specimens throughout the younger portions of the sediment record ( Figure 5A). There is a barren interval of nannofossils from 133.2 to 130.2 mbsf, but D. surculus reappeared as well-preserved specimens above this interval (Sample 15H-3W). It again reappears at 112.4 mbsf, but is regarded as reworked in this sample because not only is it higher than that of the bottom acme of D. triradiatus, a younger datum ( Figure 5A), but its occurrence is also anomalously located above samples with consistent absences of D. surculus. Thus, the LO of D. surculus is placed between Samples 15H-3W, 20 to 21 cm and 15H-2W, 20 to 21 cm (127.85 mbsf).

LO of Discoaster pentaradiatus
The occurrence of D. pentaradiatus is very sporadic as they occur as complete and fragmented specimens throughout the deeper sections of the core (Sample 11H-CC to Sample 15H-CC; from 96 to 135 mbsf; Figure 5B). The LO of D. pentaradiatus may lie between 120 and 125 mbsf based on the sharp decrease in abundance between these depth intervals and the relative placement of the LO of D. surculus and Ba D. triradiatus. However, there is a pervasive occurrence of fragmented specimens above and below this depth. Similar problems regarding the LO of D. pentaradiatus is also reported in ODP 149 in the Iberia Margin (Holes 897C and 898A, [53]) and Deep Sea Drilling Project (DSDP) Site 662 of the eastern North Atlantic [10]. Based on discoaster counts for Site U1431D, the LO of this species is unclear and therefore not utilized in this study.

First Occurrence (FO) of Medium Gephyrocapsa spp. (bmG)
The base of medium Gephyrocapsa (bmG) lies between Samples 10H-7W, 22 to 23 cm and 10H-8W, 20 to 21 cm (87.635 mbsf), which was previously unreported in shipboard results [43]. Placolith length measurements have been useful in determining the nannofossil datum. The gephyrocapsid specimens observed generally have a bridge oriented nearer the long axis of the coccolith (15 • -40 • ) and a relatively open central area. These gephyrocapsids are most likely equivalent to Gephyrocapsa oceanica, G. margereli, and G. muellerae.

LO of Calcidiscus macintyrei
The species C. macintyrei generally occurs sporadically (only 1 or 2 specimens) in Sample U1431D-15H-CC, 10-16 cm to 7H-6W, 20 to 21 cm (135.33 to 57.9 mbsf), except in Samples 14H-CC (125.48 mbsf) and 10H-6W, 20 to 21 cm (85.8 mbsf) where they occur as few specimens. Recognition of C. macintyrei above Sample 10H-CC is difficult because siliciclastic materials dominate the samples and nannofossils are reworked to younger portions of the core. Coccolith preservation is moderate to good in Samples 10H-CC to 10H-6W, while very few, poorly preserved and broken specimens of C. macintyrei are found above the interval and occur sporadically in younger sediment samples. Furthermore, coccoliths~9 µm but similar to the morphology of C. macintyrei are noted in some samples. In this study, the last occurrence is placed between Samples 10H-6W, 20 to 21 cm and 10H-5W, 20 to 21 cm (85.705 mbsf).
The base (first occurrence) of large Gephyrocapsa (blG) had been documented at Sample U1431D 11H-CC (98.15 mbsf) based on IODP shipboard biostratigraphy . However, this study did not observe large Gephyrocapsa at any sample within sections 11H and 10H (from 96.10 to 79.40 mbsf). It is detected much shallower (at Samples 9H-CC (78.91 mbsf) and above) and occurs above the LO of C. macintyrei rather than below it. Few to common large Gephyrocapsa specimens occur at Sample 9H-CC (78.91 mbsf) and they appeared abruptly in the sediment record. The first occurrence of large Gephyrocapsa morphotypes has also been documented above the LO of C. macintyrei in the Shatsky Rise in the northwest Pacific Ocean [54] and in ODP Leg 111 [55] and ODP Leg 138 [56] in the Equatorial Pacific. This pattern is also seen in the Caribbean Sea from ODP Leg 165 [57] and in the South Atlantic at ODP Sites 925 and 926 [58]. Thus, the blG is placed between Samples 9H-CC and 9H-7W, 36 cm (78.985 mbsf), above the LO of C. macintyrei. The distinct first appearance of large gephyrocapsids suggests that they are reliable markers in the SCS.

LO of Pseudoemiliania lacunosa
Shipboard results from IODP Site U1431D placed the LO of P. lacunosa at Sample 2H-CC, 0 to 4 cm (12.15 mbsf; [43]). In this study, the datum is found at younger portions of the core-between Samples 2H-2W, 80 to 82 cm and 2H-2W, 60 to 62 cm (5.42 mbsf). Circular morphotypes are found more abundantly in the deeper sections of the core, while elliptical types are more common in younger sediments. Only the elliptical forms of Pseudoemiliania are documented in the last two samples (Samples 2H-2W, 100 to 102 cm and 2H-2W, 80 to 82 cm; 5.7 and 5.5 mbsf, respectively).

FO of Emiliania huxleyi
The FO of E. huxleyi was previously reported at Sample 1H-CC, 7 to 14 cm (3.22 mbsf; [43]). The datum has been reassigned between Samples 2H-1W, 60 to 61 cm and 2H-1W, 80 to 81 cm (3.905 mbsf). Nannofossils above and below the event show a cyclical trend of barren and well-preserved assemblages, which made the recognition of this bioevent difficult. Dissolution in some of the nannofossils, especially among the placoliths, is also observed.

Morphologic Changes in Gephyrocapsids in Site U1431D
Small Gephyrocapsa is the only morphotype present from 135 to 87.94 mbsf (~2. 58

FO of Emiliania huxleyi
The FO of E. huxleyi was previously reported at Sample 1H-CC, 7 to 14 cm (3.22 mbsf; [43]). The datum has been reassigned between Samples 2H-1W, 60 to 61 cm and 2H-1W, 80 to 81 cm (3.905 mbsf). Nannofossils above and below the event show a cyclical trend of barren and well-preserved assemblages, which made the recognition of this bioevent difficult. Dissolution in some of the nannofossils, especially among the placoliths, is also observed.

Morphologic Changes in Gephyrocapsids in Site U1431D
Small Gephyrocapsa is the only morphotype present from 135 to 87.94 mbsf (~2. 58   Bridge angle changes across the Gephyrocapsa morphotypes range from 30° to 75° (Figure 6). A general decrease in bridge angle is documented from 2.58 to 1.79 Ma (small Gephyrocapsa) while an increase in bridge angle from 1.79 to 0.9 Ma is evident in the record (small and medium Gephyrocapsa).
The highest values for bridge angles for small and medium forms are noted from 1.04 to 0.9 Ma and correlate with the first occurrence of Gephyrocapsa sp. 3. The bridge angle rapidly decreases in value Bridge angle changes across the Gephyrocapsa morphotypes range from 30 • to 75 • (Figure 6). A general decrease in bridge angle is documented from 2.58 to 1.79 Ma (small Gephyrocapsa) while an increase in bridge angle from 1.79 to 0.9 Ma is evident in the record (small and medium Gephyrocapsa).
The highest values for bridge angles for small and medium forms are noted from 1.04 to 0.9 Ma and correlate with the first occurrence of Gephyrocapsa sp. 3. The bridge angle rapidly decreases in value from 0.9 to 0.8 Ma. An overall increase in bridge angle is seen from 0.8 Ma to present, with distinct lows between 0.5 and 0.45 Ma and at 0.2 Ma. In general, the bridge angle changes seem to respond similarly, regardless of the size fraction, and there is no apparent relation between the bridge angle with respect to average placolith length ( Figure 6).
Relative abundances of small, medium, and large Gephyrocapsa morphotypes show oscillations in relative abundances of small and medium forms at around 1.79 and 0.65-0.6 Ma ( Figure 6). Medium Gephyrocapsa comprise 20-30% of the total gephyrocapsid assemblage in this interval. From 0.65 to 0.6 Ma, oscillations of the shifting dominance of small and medium forms became more evident. When medium Gephyrocapsa is dominant, large morphotypes also appear in the record in few numbers, except for a spike in large Gephyrocapsa at around 0.2 Ma. The increasing amplitude fluctuation of medium Gephyrocapsa roughly coincides with the increasing glacial-interglacial variations trending towards younger sections ( Figure 6). However, it is not clear if the driving mechanism of the changes in relative abundance fluctuations is correlated to glacial or interglacial stages, as both increasing and decreasing trends are evident in both conditions.

Magnetobiochronologic Calibration of Nannofossil Datums in Site U1431D
Nannofossil events identified in Site U1431D are tied with magnetostratigraphic data from shipboard results to estimate a numerical age for the calcareous nannofossil datums. Magnetostratigraphic reversals are calibrated against [61], compiled in [9], while the calculation of the numerical ages using magnetostratigraphy is explained in the Supplementary Materials. Low-to mid-latitude Deep Sea Drilling Project (DSDP) and ODP sites are utilized to compare the nannofossil ages of datums observed in Site U1431D. The locations of these DSDP and ODP Sites are listed in Table 2 and are shown in Figure 7. The magnetobiochronologic calibrations are mostly in agreement with the compilation of [9], except for some datums (Table 1). These inconsistencies are focused on the discussion below. The LO of Calcidiscus macintyrei in Pacific, Atlantic, Indian, and Mediterranean sections are correlated to Chron C1r.3r; however, the placement of the datum within the chron varies by location ( Figure 8). Excluding Site 1146, the Indian, Atlantic, and Mediterranean records have younger age estimates than the Pacific and the SCS. In Site U1431D, correlation with shipboard magnetostratigraphy suggests that the event falls just above the top of the Olduvai Chron (C2n; 1.778-1.945 Ma, [9]), but still within the lowermost portions of Chron C1r.3r, while previous works document the LO of C. macintyrei at a slightly higher magnetostratigraphic position (lower C1r.3r but not directly on top of Olduvai- [7,9,22,58,64,[70][71][72]). Discrepancies in the ages of the extinction event (0.2 Ma) may suggest diachroneity.  Although the taxonomic definition of C. macintyrei is simple, several inconsistencies and differing species concepts exist in the literature. In Sites 768 and 769 in the Sulu Sea, [67] cited [73] for their adopted age for C. tropicus to subdivide NN19 of [74]. However, there is no written account of C. tropicus in [73]. The C. tropicus of [67] may be equivalent to the C. macintyrei of [73]. This equivalence may be unlikely because large age discrepancies exist in the literature between the two species (1.48 Ma- [67]; 1.57 Ma- [73]) and thus the differences may be an effect of varying species concepts. The stratigraphic inconsistency for the datum is also reported in the Japan Sea (1.33 Ma; [75]). The LO in the Japan Sea is younger by 0.22 Ma compared to the study of [51] and by 0.27 Ma compared to [7]. Size variations in C. leptoporus (5-8 μm, sometimes reaching up to 10 μm) may have caused confusion whether the larger end of the species belonged to C. macintyrei or a larger variant of C. leptoporus (e.g., [75]).  Although the taxonomic definition of C. macintyrei is simple, several inconsistencies and differing species concepts exist in the literature. In Sites 768 and 769 in the Sulu Sea, [67] cited [73] for their adopted age for C. tropicus to subdivide NN19 of [74]. However, there is no written account of C. tropicus in [73]. The C. tropicus of [67] may be equivalent to the C. macintyrei of [73]. This equivalence may be unlikely because large age discrepancies exist in the literature between the two species (1.48 Ma- [67]; 1.57 Ma- [73]) and thus the differences may be an effect of varying species concepts. The stratigraphic inconsistency for the datum is also reported in the Japan Sea (1.33 Ma; [75]). The LO in the Japan Sea is younger by 0.22 Ma compared to the study of [51] and by 0.27 Ma compared to [7]. Size variations in C. leptoporus (5-8 µm, sometimes reaching up to 10 µm) may have caused confusion whether the larger end of the species belonged to C. macintyrei or a larger variant of C. leptoporus (e.g., [75]).
Studies concerning the LO of Gephyrocapsa sp. 3 are limited only to the Mediterranean. The limited occurrence and distribution of studies of the LO of Gephyrocapsa sp. 3 is the reason [7] did not include the event for global biostratigraphic correlation. The datum age (0.58 Ma) was first identified by [76] and calibrated through magnetostratigraphy. More recent calibrations in the Mediterranean correlated the event to MIS 15 (0.570-0.577 Ma; [64,77]). In the SCS, there is very little documentation on the distribution of this nannofossil downcore [43]. The LO of Gephyrocapsa sp. 3 is placed within the middle portions of Chron C1n (Brunhes Chron; 0-0.780 Ma) in Site U1431D, suggesting an age of 0.44 Ma for the datum (Figure 9). Although events using Gephyrocapsa are used in some studies in the SCS [41,42], this study presents one of the very few evidence for the occurrence and distribution of Gephyrocapsa sp. 3 in SCS Pleistocene sediments. Its clear appearance downcore implies its significance as a calcareous nannofossil marker in the SCS. It must be noted with caution that Gephyrocapsa sp. 3 is still present in the open ocean today [77]; thus, careful analysis of the bridge angles of Gephyrocapsa downcore may be necessary to determine the LO of this species.
C. tropicus in [73]. The C. tropicus of [67] may be equivalent to the C. macintyrei of [73]. This equivalence may be unlikely because large age discrepancies exist in the literature between the two species (1.48 Ma- [67]; 1.57 Ma- [73]) and thus the differences may be an effect of varying species concepts. The stratigraphic inconsistency for the datum is also reported in the Japan Sea (1.33 Ma; [75]). The LO in the Japan Sea is younger by 0.22 Ma compared to the study of [51] and by 0.27 Ma compared to [7]. Size variations in C. leptoporus (5-8 μm, sometimes reaching up to 10 μm) may have caused confusion whether the larger end of the species belonged to C. macintyrei or a larger variant of C. leptoporus (e.g., [75]).   [7] did not include the event for global biostratigraphic correlation. The datum age (0.58 Ma) was first identified by [76] and calibrated through magnetostratigraphy. More recent calibrations in the Mediterranean correlated the event to MIS 15 (0.570-0.577 Ma; [64,77]). In the SCS, there is very little documentation on the distribution of this nannofossil downcore [43]. The LO of Gephyrocapsa sp. 3 is placed within the middle portions of Chron C1n (Brunhes Chron; 0-0.780 Ma) in Site U1431D, suggesting an age of 0.44 Ma for the datum (Figure 9). Although events using Gephyrocapsa are used in some studies in the SCS [41,42], this study presents one of the very few evidence for the occurrence and distribution of Gephyrocapsa sp. 3 in SCS Pleistocene sediments. Its clear appearance downcore implies its significance as a calcareous nannofossil marker in the SCS. It must be noted with caution that Gephyrocapsa sp. 3 is still present in the open ocean today [77]; thus, careful analysis of the bridge angles of Gephyrocapsa downcore may be necessary to determine the LO of this species. Correlation with magnetostratigraphy places the LO of P. lacunosa within the upper portions of Chron C1n (Figure 10), which corresponds to an age of 0.092 Ma in Site U1431D. There is a considerable age discrepancy in the LO of P. lacunosa in Site U1431D compared to the literature. The difference ranges from 0.37 Ma between Sites U1431D and Sites 1146 (SCS; [42]), 925 (Atlantic; [58]), and 964 (Mediterranean; [77]) to as much as 0.55 Ma compared to the southwest Pacific [74,75]. The results of this study differ significantly compared from previous works which suggest the LO of P. lacunosa to be correlated to MIS 12 [17,21,22]. The upper portions of the core are defined mainly by Correlation with magnetostratigraphy places the LO of P. lacunosa within the upper portions of Chron C1n (Figure 10), which corresponds to an age of 0.092 Ma in Site U1431D. There is a considerable age discrepancy in the LO of P. lacunosa in Site U1431D compared to the literature. The difference ranges from 0.37 Ma between Sites U1431D and Sites 1146 (SCS; [42]), 925 (Atlantic; [58]), and 964 (Mediterranean; [77]) to as much as 0.55 Ma compared to the southwest Pacific [74,75]. The results of this study differ significantly compared from previous works which suggest the LO of P. lacunosa to be correlated to MIS 12 [17,21,22]. The upper portions of the core are defined mainly by interbedded to interlaminated clayey silt to silty clay which are interpreted as turbidite sequences [43]. Turbidity currents may have caused the resuspension of coccoliths of P. lacunosa to younger sediments of the core. Uncertainties in magnetostratigraphic calibration may also be a factor in the age discrepancy. Because the age calculated from magnetostratigraphy is too young compared to other sites, the standard age of 0.44 Ma [9] is adopted for the LO of P. lacunosa (Figure 10, gray datum).
Geosciences 2020, 10, x FOR PEER REVIEW 14 of 26 interbedded to interlaminated clayey silt to silty clay which are interpreted as turbidite sequences [43]. Turbidity currents may have caused the resuspension of coccoliths of P. lacunosa to younger sediments of the core. Uncertainties in magnetostratigraphic calibration may also be a factor in the age discrepancy. Because the age calculated from magnetostratigraphy is too young compared to other sites, the standard age of 0.44 Ma [9] is adopted for the LO of P. lacunosa (Figure 10, gray datum).  [61] as compiled by [9]. Oxygen isotope data is used to determine the ages in ODP Sites 1146, 925 and 964. These datums are calibrated to the benthic foraminiferal δ 18 O time scale of [2]. Horizontal bars represent the recalculated datum ages for each site. Black datum in IODP Site U1431D represents age acquired from magnetostratigraphy. Gray datum in IODP Site U1431D represents placement of the datum from this study using the datum age reported in [9]. Med.-Mediterranean.
The age of the FO of E. huxleyi in Site U1431D also has a large discrepancy. Correlation to Site U1431D magnetostratigraphy showed that the FO of E. huxleyi corresponds to the upper portions of Chron C1n at 0.066 Ma (Figure 11, black datum). The oldest age estimates are found in ODP Sites 1146 and 1148 followed by ODP Sites 768 and 769 in the Sulu Sea ( Figure 11). Younger ages are noted from the open ocean (Pacific and Atlantic) and in the Mediterranean Sea; however, the age acquired from Site U1431D is much younger than all other sites. The uncertainty ranges from 0.305 Ma compared to Site 1148 [42] to 0.188 Ma compared to Site 925 [58]. Bioturbation and terrigenous input to the basin may have affected the signal of the FO. Bioturbation causes vertical sediment mixing which may reposition a datum to a slightly deeper depth. However, shipboard results reveal only slight bioturbation (10-30%) in the interval where the FO of E. huxleyi is documented [43] and may have had minor effect in the position of the datum. Terrigenous input may be a problem especially for less robust species like E. huxleyi. Previous reports suggest the sporadic occurrence of E. huxleyi near its first appearance [15,17,78]. Carbonate dilution through terrigenous input might further obscure the datum. Shipboard data suggest that carbonate percentages in Pleistocene sediments of Site U1431D are very low (1-10%; [43]). This interval is also defined by an abundance of plagioclase, mica, and amphibole seen in smear slides, suggesting increased detrital input. Magnetostratigraphic calibration may have also contributed to the age discrepancy recorded in IODP Site U1431D. Thus, the age of 0.29 Ma [9] is adopted for the datum (Figure 11, gray datum).  [61] as compiled by [9]. Oxygen isotope data is used to determine the ages in ODP Sites 1146, 925 and 964. These datums are calibrated to the benthic foraminiferal δ 18 O time scale of [2]. Horizontal bars represent the recalculated datum ages for each site. Black datum in IODP Site U1431D represents age acquired from magnetostratigraphy. Gray datum in IODP Site U1431D represents placement of the datum from this study using the datum age reported in [9]. Med.-Mediterranean.
The age of the FO of E. huxleyi in Site U1431D also has a large discrepancy. Correlation to Site U1431D magnetostratigraphy showed that the FO of E. huxleyi corresponds to the upper portions of Chron C1n at 0.066 Ma (Figure 11, black datum). The oldest age estimates are found in ODP Sites 1146 and 1148 followed by ODP Sites 768 and 769 in the Sulu Sea ( Figure 11). Younger ages are noted from the open ocean (Pacific and Atlantic) and in the Mediterranean Sea; however, the age acquired from Site U1431D is much younger than all other sites. The uncertainty ranges from 0.305 Ma compared to Site 1148 [42] to 0.188 Ma compared to Site 925 [58]. Bioturbation and terrigenous input to the basin may have affected the signal of the FO. Bioturbation causes vertical sediment mixing which may reposition a datum to a slightly deeper depth. However, shipboard results reveal only slight bioturbation (10-30%) in the interval where the FO of E. huxleyi is documented [43] and may have had minor effect in the position of the datum. Terrigenous input may be a problem especially for less robust species like E. huxleyi. Previous reports suggest the sporadic occurrence of E. huxleyi near its first appearance [15,17,78]. Carbonate dilution through terrigenous input might further obscure the datum. Shipboard data suggest that carbonate percentages in Pleistocene sediments of Site U1431D are very low (1-10%; [43]). This interval is also defined by an abundance of plagioclase, mica, and amphibole seen in smear slides, suggesting increased detrital input. Magnetostratigraphic calibration may have also contributed to the age discrepancy recorded in IODP Site U1431D. Thus, the age of 0.29 Ma [9] is adopted for the datum (Figure 11, gray datum).  [61] as compiled by [9]. Oxygen isotope data is used to determine the ages in ODP Sites 1146, 925 and 964. These datums are calibrated to the benthic foraminiferal δ 18 O time scale of [2]. Horizontal bars represent the recalculated datum ages for each site. Black datum in IODP Site U1431D represents age acquired from magnetostratigraphy. Gray datum in IODP Site U1431D represents placement of the datum from this study using the datum age reported in [9]. Med.-Mediterranean.

The Importance and Complexity of Gephyrocapsa Morphometry in Site U1431D
The overall time-transgressive size increase correlates well with those of previous works [19,[21][22][23][24]59,79]. The work of [19] complements the data gathered; however, no large Gephyrocapsa were observed in the lower Pleistocene sections of the core studied. Only small Gephyrocapsa spp. placoliths are documented from Sample 15H-CC (~135.3 m; latest Pliocene) to 10H-8W, 20-21 cm (~87.94 m; early Pleistocene). The Pliocene-Pleistocene was defined at the base of the Calabrian (~1.8 Ma) by [19] and may account for the disparity. Nevertheless, the size increase is documented across ocean basins; thus, the morphologic change in Gephyrocapsa most probably points to an evolutionary trend. The timetransgressive size increase is applicable for biostratigraphic use and may explain the comparable record of Site U1431D sediments to other ocean basins. Global climate changes longer than glacial-interglacial cycles, such as ocean cooling brought on by a steady increase in ice cover, may have affected the longterm changes in the size of Gephyrocapsa coccoliths [79]. It is also noted that the changes in morphological parameters of the gephyrocapsids are not in one-to-one correspondence with glacialinterglacial variations. Higher resolution sampling, correlatable to suborbital variations in δ 18 O, is needed to further explain the response of Gephyrocapsa during these time scales.
Specimens of Gephyrocapsa coccoliths analyzed by [80] from tropical Atlantic waters generally have high bridge angles (parallel to the short axis of the coccolith). This observation is supported by [52], noting further that the bridge angle may be correlated with geographic distribution-the position of the bridge approaches the long axis towards high latitudes, in contrast to a mean bridge angle of 60° within 15 degrees north and south of the Equator. If temperature has a direct effect on bridge angle orientation, then there should be a glacial-interglacial trend with respect to bridge angle, with relatively colder periods shifting to lower bridge angle values. However, in Site U1431D there is very little correlation between the average bridge angle values of all size fractions measured and the δ 18 O record (LR04 Stack of [2]; Figure 6). Most probably, environmental factors other than temperature may have had an influence on the development of the bridge.
The presence of medium-sized forms with high bridge angle values at around 1.0-0.6 Ma is  [61] as compiled by [9]. Oxygen isotope data is used to determine the ages in ODP Sites 1146, 925 and 964. These datums are calibrated to the benthic foraminiferal δ 18 O time scale of [2]. Horizontal bars represent the recalculated datum ages for each site. Black datum in IODP Site U1431D represents age acquired from magnetostratigraphy. Gray datum in IODP Site U1431D represents placement of the datum from this study using the datum age reported in [9]. Med.-Mediterranean.

The Importance and Complexity of Gephyrocapsa Morphometry in Site U1431D
The overall time-transgressive size increase correlates well with those of previous works [19,[21][22][23][24]59,79]. The work of [19] complements the data gathered; however, no large Gephyrocapsa were observed in the lower Pleistocene sections of the core studied. Only small Gephyrocapsa spp. placoliths are documented from Sample 15H-CC (~135.3 m; latest Pliocene) to 10H-8W, 20-21 cm (~87.94 m; early Pleistocene). The Pliocene-Pleistocene was defined at the base of the Calabrian (~1.8 Ma) by [19] and may account for the disparity. Nevertheless, the size increase is documented across ocean basins; thus, the morphologic change in Gephyrocapsa most probably points to an evolutionary trend. The time-transgressive size increase is applicable for biostratigraphic use and may explain the comparable record of Site U1431D sediments to other ocean basins. Global climate changes longer than glacial-interglacial cycles, such as ocean cooling brought on by a steady increase in ice cover, may have affected the long-term changes in the size of Gephyrocapsa coccoliths [79]. It is also noted that the changes in morphological parameters of the gephyrocapsids are not in one-to-one correspondence with glacial-interglacial variations. Higher resolution sampling, correlatable to suborbital variations in δ 18 O, is needed to further explain the response of Gephyrocapsa during these time scales.
Specimens of Gephyrocapsa coccoliths analyzed by [80] from tropical Atlantic waters generally have high bridge angles (parallel to the short axis of the coccolith). This observation is supported by [52], noting further that the bridge angle may be correlated with geographic distribution-the position of the bridge approaches the long axis towards high latitudes, in contrast to a mean bridge angle of 60 • within 15 degrees north and south of the Equator. If temperature has a direct effect on bridge angle orientation, then there should be a glacial-interglacial trend with respect to bridge angle, with relatively colder periods shifting to lower bridge angle values. However, in Site U1431D there is very little correlation between the average bridge angle values of all size fractions measured and the δ 18 O record (LR04 Stack of [2]; Figure 6). Most probably, environmental factors other than temperature may have had an influence on the development of the bridge.
The presence of medium-sized forms with high bridge angle values at around 1.0-0.6 Ma is documented in this study and is comparable to previous works [7,19,21,23,24,59,[81][82][83]. This Gephyrocapsa morphotype is most probably equivalent to Gephyrocapsa sp. 3 of [19], Gephyrocapsa oceanica vertical of [23], and Gephyrocapsa sp. C of [24]. Small forms of Gephyrocapsa with a bridge angle parallel to the short axis are also documented mostly from 1.0 to 0.9 Ma near the "small Gephyrocapsa acme" [23,24]. These studies are correlatable and span different ocean basins, each with their own local climate. Thus, the changes in the morphology of Gephyrocapsa in the SCS may not be factors of regional climate perturbations. These results show the importance of both bridge angle and size for the determination of certain Gephyrocapsa morphotypes for Pleistocene biostratigraphy.

Age-Depth Model for Site U1431D
An age-depth model is constructed based on magnetostratigraphic and nannofossil datums identified in Site U1431D. The datum ages are calculated based on linear interpolation between magnetostratigraphic chrons (Supplementary Materials). The ages for the FO of E. huxleyi and LO of P. lacunosa are adopted from [9] because the calculated site-specific ages are too young compared to published figures. The LO of Gephyrocapsa sp. 3 is also taken from [9] because the only calibration done for this datum is from [76] and [64] in the Mediterranean. Thus, the age-depth model for Site U1431D is a combination of ages compiled by [9] and site-specific numerical ages.
The age-depth curve generated from the nannofossil datums using calcareous nannofossils shows consistency with other microfossil groups ( Figure 12). From 135 mbsf to~80 mbsf, planktonic foraminifera and calcareous nannofossil data generally agree with each other (Figure 12), with slight divergence at the 100-135 mbsf interval. Above 90 mbsf, planktonic foraminiferal datums imply younger ages than the calcareous nannofossils. Further analysis of planktonic foraminiferal biostratigraphy is needed to verify the age discrepancies between the two microfossil groups. The age model from radiolarians for the uppermost layers (0-10 mbsf) agrees with the nannofossils, reflecting a slow sedimentation rate within the interval. Gephyrocapsa morphotype is most probably equivalent to Gephyrocapsa sp. 3 of [19], Gephyrocapsa oceanica vertical of [23], and Gephyrocapsa sp. C of [24]. Small forms of Gephyrocapsa with a bridge angle parallel to the short axis are also documented mostly from 1.0 to 0.9 Ma near the "small Gephyrocapsa acme" [23,24]. These studies are correlatable and span different ocean basins, each with their own local climate. Thus, the changes in the morphology of Gephyrocapsa in the SCS may not be factors of regional climate perturbations. These results show the importance of both bridge angle and size for the determination of certain Gephyrocapsa morphotypes for Pleistocene biostratigraphy.

Age-Depth Model for Site U1431D
An age-depth model is constructed based on magnetostratigraphic and nannofossil datums identified in Site U1431D. The datum ages are calculated based on linear interpolation between magnetostratigraphic chrons (Supplementary Materials). The ages for the FO of E. huxleyi and LO of P. lacunosa are adopted from [9] because the calculated site-specific ages are too young compared to published figures. The LO of Gephyrocapsa sp. 3 is also taken from [9] because the only calibration done for this datum is from [76] and [64] in the Mediterranean. Thus, the age-depth model for Site U1431D is a combination of ages compiled by [9] and site-specific numerical ages.
The age-depth curve generated from the nannofossil datums using calcareous nannofossils shows consistency with other microfossil groups ( Figure 12). From 135 mbsf to ~80 mbsf, planktonic foraminifera and calcareous nannofossil data generally agree with each other (Figure 12), with slight divergence at the 100-135 mbsf interval. Above 90 mbsf, planktonic foraminiferal datums imply younger ages than the calcareous nannofossils. Further analysis of planktonic foraminiferal biostratigraphy is needed to verify the age discrepancies between the two microfossil groups. The age model from radiolarians for the uppermost layers (0-10 mbsf) agrees with the nannofossils, reflecting a slow sedimentation rate within the interval. Figure 12. Comparison of age-depth models using different microfossil groups for Site U1431D. Symbols represent specific microfossil groups found in the site. The nannofossil age model incorporates published ages for FO of E. huxleyi, LO of P. lacunosa, and LO of Gephyrocapsa sp. 3 found in [9] and magnetostratigraphic ages acquired from this study for the older datums. The planktonic foraminifera and radiolarian curves use the ages from [9].
The sedimentation rate in Site U1431D may be divided into three distinct segments based on nannofossil datums (Figure 12). From 135 to 60  Comparison of age-depth models using different microfossil groups for Site U1431D. Symbols represent specific microfossil groups found in the site. The nannofossil age model incorporates published ages for FO of E. huxleyi, LO of P. lacunosa, and LO of Gephyrocapsa sp. 3 found in [9] and magnetostratigraphic ages acquired from this study for the older datums. The planktonic foraminifera and radiolarian curves use the ages from [9].
The sedimentation rate in Site U1431D may be divided into three distinct segments based on nannofossil datums (Figure 12). From 135 to 60

Other Possible Nannofossil Events
There are other possible nannofossil events found in the core which may need further reexamination. These events are recorded as shifts in the abundance values of nannofossil species, or a cross-over of the abundance of variants within a species. These shifts may offer biostratigraphic significance if assessed further and documented across other basins.
The genus Pseudoemiliania had been differentiated based on the number of distal slits [23] and the degree of ellipticity [84]; these morphotypes are not routinely used in biostratigraphy. However, large, circular forms only appeared in the late Pleistocene and the changes in their morphology may be related to temperature fluctuations [48,85]. Absolute abundance counts from Site U1431D show a shift in abundance between Pseudoemiliania circular and elliptical variants between 120 to 130 mbsf ( Figure 13). Circular variants of P. lacunosa are dominant from 135 to 130 mbsf and become replaced by elliptical types. This observation is different from the work of [84] wherein the circular forms of Pseudoemiliania dominate the younger portions of DSDP Leg 4 cores. It is possible that ecological preferences and the marginal basin setting of the SCS may have affected the distribution of Pseudoemiliania compared to the open-ocean setting of DSDP Leg 4 [84].

Other Possible Nannofossil Events
There are other possible nannofossil events found in the core which may need further reexamination. These events are recorded as shifts in the abundance values of nannofossil species, or a cross-over of the abundance of variants within a species. These shifts may offer biostratigraphic significance if assessed further and documented across other basins.
The genus Pseudoemiliania had been differentiated based on the number of distal slits [23] and the degree of ellipticity [84]; these morphotypes are not routinely used in biostratigraphy. However, large, circular forms only appeared in the late Pleistocene and the changes in their morphology may be related to temperature fluctuations [48,85]. Absolute abundance counts from Site U1431D show a shift in abundance between Pseudoemiliania circular and elliptical variants between 120 to 130 mbsf ( Figure 13). Circular variants of P. lacunosa are dominant from 135 to 130 mbsf and become replaced by elliptical types. This observation is different from the work of [84] wherein the circular forms of  The first and last occurrences of the small Gephyrocapsa acme (1.24 to 1.02 Ma, respectively [9]), was not detected in Site U1431D. Possible carbonate dilution through terrigenous input may have been the cause of the absence of the event. However, Zone NN21 [74] consists of several other Gephyrocapsa acme intervals, one of which may be recorded in the SCS. Nannofossil biostratigraphy from the Kings Trough at the northeastern Atlantic Ocean show several acme intervals corresponding to different species of small Gephyrocapsa [85]. They have documented the dominance of G. aperta near the FO of E. huxleyi. In Sample 2H-1W, 60-61 cm (3.80 mbsf), an increase in small Gephyrocapsa placoliths is observed (Figure 14). These small placoliths have a large central area and a bridge near the long axis, in agreement with the taxonomic description of G. aperta. Interestingly, the first occurrence of E. huxleyi is assigned to this sample and correlates well with observations in Kings Trough. A minor small Gephyrocapsa acme event occurs between Zones NN20 and NN21 [86] and was also documented by [24], most likely corresponding to the earlier portions of their Assemblage "B".
The reliability of these secondary events can be assessed through correlation with oxygen isotope stratigraphy or cyclostratigraphy using the physical properties of sediments in the SCS. These The first and last occurrences of the small Gephyrocapsa acme (1.24 to 1.02 Ma, respectively [9]), was not detected in Site U1431D. Possible carbonate dilution through terrigenous input may have been the cause of the absence of the event. However, Zone NN21 [74] consists of several other Gephyrocapsa acme intervals, one of which may be recorded in the SCS. Nannofossil biostratigraphy from the Kings Trough at the northeastern Atlantic Ocean show several acme intervals corresponding to different species of small Gephyrocapsa [85]. They have documented the dominance of G. aperta near the FO of E. huxleyi. In Sample 2H-1W, 60-61 cm (3.80 mbsf), an increase in small Gephyrocapsa placoliths is observed (Figure 14). These small placoliths have a large central area and a bridge near the long axis, in agreement with the taxonomic description of G. aperta. Interestingly, the first occurrence of E. huxleyi is assigned to this sample and correlates well with observations in Kings Trough. A minor small Gephyrocapsa acme event occurs between Zones NN20 and NN21 [86] and was also documented by [24], most likely corresponding to the earlier portions of their Assemblage "B".
The reliability of these secondary events can be assessed through correlation with oxygen isotope stratigraphy or cyclostratigraphy using the physical properties of sediments in the SCS. These minor events have been noted in some studies [24,[84][85][86]. Details of these events in the SCS can be studied further through higher resolution sampling in the intervals where they are noted. Additional data from other ocean basins are needed to verify whether these nannofossil events are recorded on a regional or a global scale.
Geosciences 2020, 10, x FOR PEER REVIEW 18 of 26 data from other ocean basins are needed to verify whether these nannofossil events are recorded on a regional or a global scale.

Conclusions
Twelve (12) calcareous nannofossil datums are documented in Pleistocene sediments of Site U1431D. This study is one of the first attempts to describe the distribution of Gephyrocapsa sp. 3 in SCS sediments and is suggested to be a well-defined marker for the subdivision of the Pleistocene in the marginal sea. The FO of Gephyrocapsa sp. 3 is near the reentrance event of medium Gephyrocapsa (reemG), which is also observed in several studies [7,[21][22][23]60,81,83]. Nannofossil analysis suggests that the FO of large Gephyrocapsa (blG) is found above the LO of Calcidiscus macintyrei, which differs from shipboard results [43]. The LO of Discoaster species are determined using a semi-quantitative method to reduce the effect of reworking; however, the LO of D. pentaradiatus is very sporadic and therefore is not used as a marker for the Pleistocene in Site U1431D. The numerical ages of nannofossil datums are generated through correlation with magnetostratigraphy. The ages of the FO of E. huxleyi and LO of P. lacunosa are underestimated by 0.22 and 0.35 Ma, respectively, while bmG and LO of C. macintyrei are slightly overestimated by 0.12 and 0.15 Ma, respectively. The results of the magnetobiochronological calibration suggests further analysis using other stratigraphies (e.g., oxygen isotope stratigraphy, cyclostratigraphy).
Morphometric analysis of Gephyrocapsa coccoliths from the late Pliocene to Pleistocene sediments of Site U1431D allowed for the recognition and refinement of bioevents of the gephyrocapsids. These events are useful in updating the nannofossil biostratigraphy of the site. The events documented are the FO of medium Gephyrocapsa morphotypes (87.635 mbsf), the FO and LO of large Gephyrocapsa (78.985 and 60.65 mbsf, respectively), the reemG event (58.24 mbsf), and the FO and LO of Gephyrocapsa sp. 3 (57.605 and 25.925 mbsf, respectively). The trends of placolith length and bridge angles of Gephyrocapsa in Site U1431D are correlatable with previous studies on other ocean basins. The data suggest that the change in the size of the gephyrocapsids points to an evolutionary trend rather than being caused by local or regional climate anomalies.

Conclusions
Twelve (12) calcareous nannofossil datums are documented in Pleistocene sediments of Site U1431D. This study is one of the first attempts to describe the distribution of Gephyrocapsa sp. 3 in SCS sediments and is suggested to be a well-defined marker for the subdivision of the Pleistocene in the marginal sea. The FO of Gephyrocapsa sp. 3 is near the reentrance event of medium Gephyrocapsa (reemG), which is also observed in several studies [7,[21][22][23]60,81,83]. Nannofossil analysis suggests that the FO of large Gephyrocapsa (blG) is found above the LO of Calcidiscus macintyrei, which differs from shipboard results [43]. The LO of Discoaster species are determined using a semi-quantitative method to reduce the effect of reworking; however, the LO of D. pentaradiatus is very sporadic and therefore is not used as a marker for the Pleistocene in Site U1431D. The numerical ages of nannofossil datums are generated through correlation with magnetostratigraphy. The ages of the FO of E. huxleyi and LO of P. lacunosa are underestimated by 0.22 and 0.35 Ma, respectively, while bmG and LO of C. macintyrei are slightly overestimated by 0.12 and 0.15 Ma, respectively. The results of the magnetobiochronological calibration suggests further analysis using other stratigraphies (e.g., oxygen isotope stratigraphy, cyclostratigraphy).
Morphometric analysis of Gephyrocapsa coccoliths from the late Pliocene to Pleistocene sediments of Site U1431D allowed for the recognition and refinement of bioevents of the gephyrocapsids. These events are useful in updating the nannofossil biostratigraphy of the site. The events documented are the FO of medium Gephyrocapsa morphotypes (87.635 mbsf), the FO and LO of large Gephyrocapsa (78.985 and 60.65 mbsf, respectively), the reemG event (58.24 mbsf), and the FO and LO of Gephyrocapsa sp. 3 (57.605 and 25.925 mbsf, respectively). The trends of placolith length and bridge angles of Gephyrocapsa in Site U1431D are correlatable with previous studies on other ocean basins. The data suggest that the change in the size of the gephyrocapsids points to an evolutionary trend rather than being caused by local or regional climate anomalies.

Discoaster bellus Bukry and Percival, 1971
Five-rayed discoasters without any central knob or central area development. Straight rays that taper slightly and terminate in points. All rays lie on the same plane. REMARKS: Occurs as reworked specimens.
Discoaster blackstockae Bukry, 1973 four-rayed form with inter-ray angles of 60 • and 120 • . Considered a variant of D. brouweri. REMARKS: Distribution within the core is very limited and does not parallel that of D. brouweri.
Discoaster calcaris Gartner, 1967 Six-rayed asterolith, bifurcating asymmetrically at tip. Longer branch of bifurcation curved sharply proximally but extending only slightly laterally beyond tip of ray. REMARKS: Occurs as reworked specimens.
Discoaster kugleri Martini and Bramette, 1963 Central-area wide and flat; free rays short with notched ends rather than true bifurcations. REMARKS: Occurs as reworked specimens.

Discoaster pentaradiatus Tan, 1927b
Five-rayed acute bifurcations; strongly concavo-convex and birefringent (each ray is a separate unit, with c-axes inclined slightly from vertical toward radial). REMARKS: The distribution of D. pentaradiatus is sporadic in Site U1431D sediments, thus the LO of the species was not used for biostratigraphy. Gartner, 1969c Five-rayed with concavo-convex rays. Central area with prominent distal sutural ridges and large proximal boss. REMARKS: Occurs as reworked specimens.

Discoaster quinqueramus
Discoaster surculus Martini and Bramlette, 1963 Six-rayed and rarely with fewer rays with a stellate knob in the central area. From the knob, small ridges extend to the margin between the rays on one side of the asterolith and along the arms on the side. The slender rays are slightly enlarged at the end, and show a trifurcation, with the central spine extending beyond the arms.
Discoaster tamalis Kamptner, 1967 Concavo-convex rays. Calcite body consists of four arms arranged to form an orthogonal cross. There is no boss in the central area. REMARKS: Occurs as reworked specimens.