Astronomical Time Scale of the Late Pleistocene in the Northern South China Sea Based on Carbonate Deposition Record

Abstract

Variations in solar insolation caused by changes in the Earth’s orbit—specifically its eccentricity, obliquity, and precession—can leave discernible marks on the geologic record. Astrochronology leverages these markers to establish a direct connection between chronological measurements and different facets of climate change as recorded in marine sediments. This approach offers a unique window into the Earth’s climate system and the construction of high-resolution, continuous time scales. Our study involves comprehensive bulk carbonate analyses of 390 discrete samples from core SCS1, which was retrieved from the deep-sea floor of the northern South China Sea. By utilizing carbonate stratigraphic data, we have developed a carbonate stratigraphic age model. This was achieved by aligning the carbonate sequence from core SCS1 with the established carbonate standard stratigraphic time scale of the South China Sea. Subsequently, we construct an astronomically tuned time scale based on this age model. Our findings indicate that sediment records in this core have been predominantly influenced by a 20,000-year cycle (precession cycle) throughout the Late Pleistocene. We have developed an astronomical time scale extending back approximately 110,000 years from the present, with a resolution of 280 years, by tuning the carbonate record to the precession curve. Time-domain spectral analysis of the tuned carbonate time series, alongside the consistent comparability of the early Holocene low-carbonate event (11–8 kyr), underscores the reliability of our astronomical time scale. Our age model exposes intricate variations in carbonate deposition, epitomizing a typical “Pacific-type” carbonate cycle. Previous research has illustrated that precession forcing predominantly influences productivity changes in the South China Sea. The pronounced precession-related cycle observed in our record suggests that changes in productivity significantly impact carbonate content in the area under study. Furthermore, the clear precession period identified in the carbonate record of core SCS1 reflects the response of low-latitude processes to orbital parameters, implying that carbonate deposition and preservation in core SCS1 are chiefly influenced by the interplay between the Intertropical Convergence Zone (ITCZ) and the monsoon system within the precession band. Our astronomical time scale is poised to enhance paleoceanographic, paleoclimatic, and correlation studies further. Additionally, the independent evidence we provide for using proxy records for astronomical age calibration of marine sediments lends additional support to similar methods of astronomical tuning.

1. Introduction

Deep-sea sediment sequences are crucial documents of past global climate changes, serving as primary carriers of high-resolution, long-term records of marine history during the Pliocene–Pleistocene epochs [1,2,3]. The South China Sea (SCS), in particular, stands out due to its high sedimentation rates and favorable conditions for carbonate preservation, making it a rich archive of both regional and global climate changes. This provides an optimal setting for conducting detailed paleoceanographic studies [4,5]. Research has extensively explored the environmental evolution within the SCS since the Quaternary, revealing significant insights. For instance, analyses of cores ODP 1147/48 and ODP 1146 have delineated variations in monsoonal activity, while core MD12-3429 has shed light on the long-term impacts of sea level changes and Kuroshio intrusion on terrestrial sediment transport in the northeastern SCS [6,7,8]. The Kuroshio is an important part of the western boundary current of the northern Pacific Ocean, with high temperature and salinity characteristics [9,10]. Additionally, the ${\mathrm{U}}_{37}^{{\mathrm{K}}^{\prime }}$SST record has provided valuable data on glacial and interglacial temperature fluctuations over the last 150,000 years [11]. Despite these advancements, dating sedimentary records and analyzing paleoenvironmental climates in the SCS remain challenging due to the complex depositional environment and varied material sources [12,13].

An independent, orbitally scaled chronological framework is pivotal for accurately identifying and understanding shifts in paleoclimate and paleoenvironment [4,14]. To date, several methods have been applied to establish the chronological framework of the SCS. Zhong et al. and Lin et al. used optically stimulated luminescence of quartz to establish the chronological framework of the ChaoShan Delta and Pearl River Delta, respectively [15,16]. However, in marine depositional environments, the depositional process is complex. The transportation distance, particle size, depositional environment, and other factors can make the degree of solar bleaching of quartz or feldspar difficult to evaluate [14,17]. AMS ¹⁴C dating of planktonic foraminifera (preferably single species) has also been widely used to construct chronological frameworks for sedimentary sequences in the SCS [18,19,20,21]. However, merely relying on ¹⁴C to construct accurate chronologies for the SCS records is problematic. For instance, radiocarbon dating dates back to ~50 kyr ago [22]. It cannot be applied to sediments beyond this dating range. In addition, different planktonic foraminiferal species from the same sediment depth may have significantly different radiocarbon ages due to the effects of sediment redeposition, recrystallization, and incorporation of secondary radiocarbon [23]. Studies have also shown that bioturbation, carbonate dissolution, or chemical erosion may significantly affect radiocarbon-based age scales [24,25,26,27]. Finally, converting marine radiocarbon dates to calendar ages requires knowledge of the surface reservoir age. However, the surface reservoir age variations remain poorly constrained [28]. Beyond radiocarbon dating, matching oxygen isotopes from foraminifera to the global oxygen isotope stack offers another strategy for dating, extending back millions of years [29]. However, this method can only be used in areas where foraminifera are a common component of sediments. Furthermore, a characteristic species’ first or last occurrence can constrain the sediment age [4]. However, it is imprecise because of the significant provincialism of organisms used for biostratigraphic datums [2]. Similarly, while absolute ages can sometimes be inferred from ash layers or other distinctive sediments, the practical application of these methods is often hindered by the absence of well-preserved carbonates or suitable microfossils in many oceanic regions [30]. Except as mentioned above, the paleomagnetic dating techniques provide an effective method for deep-sea sediments [31] and a chronological framework for paleoenvironmental studies in the SCS [4,32]. However, paleomagnetic stratigraphic age control points are relatively limited, and traditional paleomagnetic stratigraphic methods are not applicable to sediments younger than the Matuyama–Brunhes boundary (~780 ka) [33]. Even though relative paleointensity (RPI) methods are now becoming increasingly popular for younger sediments (<780 ka), the RPI model remains ambiguous due to variations in depositional rates without further constraints from other dating methods [34,35]. Next, changes in magnetic minerals in anoxic environments can also record chemical remanent magnetization, which can confound previous interpretations of paleomagnetic stratigraphy. Given the varying age ranges and limitations of these dating methods [14], employing multi-parameter approaches is often necessary to ensure the accuracy of marine sediment chronologies [8,36].

In recent years, astrochronology has emerged as a pivotal method for constructing age models across diverse climatic records, including eolian, abyssal sediments, and continental sediments [4,37,38,39,40,41]. The foundational theory of astronomical climate, proposed by Milutin Milankovitch in the early 20th century, has paved the way for the development of precise, high-resolution astronomical chronologies [42,43]. The landmark study by Hays et al. [44] highlighted the significant role of Earth’s orbital geometry in driving climate change, a phenomenon well-documented within sedimentary records. Subsequent research has consistently shown that Quaternary climate changes are closely aligned with variations in the Earth’s orbit, forming the basis for paleoclimatic age models through astronomical calibration [3,45]. Astrochronology involves identifying Milankovitch cycles within climate proxies from sediment records and aligning these orbital signals with astronomical parameter models to establish a continuous, high-resolution geological chronological framework [46]. This method offers potentially higher resolution and accuracy than traditional time scales based on linear interpolation between geomagnetic reversals or radiometrically dated points [47,48]. Most new age models have been created by correlating isotopic records with orbital target curves [2,47,49,50]. In cases where isotope data are unavailable, other climate indicators such as magnetization, natural gamma, color reflectance, mineral content, elemental content, or carbonate content have been successfully employed [2,4,47,51,52,53,54,55]. Notably, astronomically calibrated carbonate stratigraphy has proven to be an effective tool for constructing high-resolution age models for oceanic deposits [52,54,56,57].

Astrochronology has been successfully applied in the SCS, reconstructing astronomical time scales for deep-sea sedimentary sequences [2,4,47,51]. Astronomical tuning age models for the region have utilized various signal curves and target curves, depending on the core location. For instance, cores ODP 1148, IODP U1499, and U1505 were tuned using combinations of natural gamma radiation, color spectral reflectance, and/or magnetic susceptibility, with global oxygen isotope curves, eccentricity, obliquity, and/or precession serving as target curves [2,4,51]. Additionally, cores ODP 1145 and 1143 employed K/Si ratios, benthic foraminiferal oxygen isotope records, and hematite to goethite content ratio (Hm/Gt) as the tuning signals, with the Northern Hemisphere summer insolation record as the primary tuning target [1,47,55].

The suitability of carbonate stratigraphy and astrochronology for core SCS1 in the SCS is supported by several factors: (a) the widespread availability of carbonate content as a classical paleoenvironmental proxy [58,59,60,61,62]; (b) the established link between carbonate content variations and glacial–interglacial cycles [60,63]; (c) the influence of monsoon variability on carbonate deposition and preservation [55], and (d) the significant effects of Milankovitch cycle-induced climatic oscillations on glacial–interglacial cycles, monsoon variability, and ocean productivity [64,65,66,67,68]. The change in carbonate content is affected by dilution and productivity [55]. Both dilution and productivity are related to monsoon variability [55]. Precession forcing has long been considered to be the main force of monsoon variation on orbital time scales [66,68]. It redistributes or splits seasonal incoming solar radiation, especially at low latitudes where monsoon systems prevail [64,65,66,67]. The summer and winter monsoons both have an impact on carbonate content through productivity and dilution [55]. Thus, the cycle of low-latitude solar radiation is also transmitted to the oceanic carbonate deposition record through the monsoon as a link. Furthermore, the periodic changes in solar insolation with latitude and season caused by changes in the Earth’s orbital parameters are the fundamental driving force for climate change and glacial cycles on a millennium to ten-thousand-year scale [69]. The periodicity controlled by the precession of monsoon variability is considered to be influenced by the Pliocene/Pleistocene glacial–interglacial cycle [55]. At the same time, the carbonate content during the late Pleistocene is also related to the glacial–interglacial cycle [55,60]. These factors suggest that Earth’s orbital parameter changes directly impact the deposition and preservation of carbonates in sediments, reflected in the carbonate and other biogeochemical records of ocean sediments. At present, the core SCS1 in the SCS has been obtained; however, different dating methods have different age ranges and constraints, such as the traditional paleomagnetic methods, which can only be relied on to decipher stratigraphic chronology [3]. Thus, the effectiveness of the astronomical tuning method in improving the resolution of the chronostratigraphy of the SCS requires further study.

This study aims to establish a continuous Late Pleistocene astronomically tuned carbonate stratigraphy of core SCS1 from the northern SCS. We present a high-resolution carbonate record from core SCS1, establish an initial untuned carbonate stratigraphic age model, and construct an astronomical time scale by tuning the carbonate content series to Earth’s orbital parameters. Through comparison with low-carbonate events in other SCS cores, we aim to verify the accuracy of the constructed astronomical time scale and discuss the influence of precession forcing on carbonate content variation.

2. Regional Setting

The SCS, framed by China, the Indo-China Peninsula, and the Taiwan–Philippine Island Arc, is a significant geographical feature in the western Pacific (Figure 1). This semi-enclosed deep-sea basin, the largest marginal sea in the western Pacific, spans approximately 3.5 million square kilometers [47]. It boasts an average depth of around 1250 m, plunging to over 5000 m at its deepest points. The basin is characterized by expansive shelf areas in the northern and southwestern regions, deep-sea basins in the central and eastern parts, and narrow straits in the southwest and northeast, facilitating its connection to the East China Sea and the Okinawa Trough via the Taiwan Strait to the north and to the open Pacific Ocean through the Bashi Strait to the east [70].

The climate of the region is predominantly controlled by the East Asian monsoon system, leading to distinct seasonal weather patterns across the SCS (Figure 1). Winters are marked by cold, dry conditions with prevailing northeast winds, while summers bring warm, wet weather with southwest winds, attributed to the seasonal reversal of wind patterns due to the Asian winter and summer monsoon circulations [47]. During winter, as northern Asia cools and high-pressure systems develop, northeast monsoon winds sweep across the sea. Conversely, the arrival of summer sees central China warming, low pressure systems forming, and the onset of gentle southwesterly winds [13]. The intense winter monsoon persists for about six months, from November to April, while the relatively milder summer monsoon spans approximately four months, from mid-May to mid-September [71].

The northern part of the SCS is influenced by a series of surface circulation and the joint influence of the Kuroshio SCS branch and the West Pacific Deep Current (Figure 1). Monsoons play a significant role in the general surface circulation of the SCS. The surface circulation is also strongly influenced by the Kuroshio Current intrusion in the northern part [72]. During the winter, the Kuroshio Current takes a direct route from northeast Luzon toward southwest Taiwan, continuing westward along the northern slope of the SCS. In summer, the current shifts southward, its flow markedly weakening [73]. Additionally, the deep-water current enters the SCS through the Luzon Strait, deflecting northwest before turning southwest along the continental margin of southeastern China, thus forming the SCS Contour Current [74].

Figure 1. Map showing the position of core SCS1, marked by a red star, within the context of the monsoon winds and current systems prevalent in the northern SCS. The depiction of monsoon winds is based on Webster (1994) [75], while the mapping of major surface currents follows the work of Fang et al. (1998) [72]. The surface currents in the Gulf of Tonkin are derived from studies by Xia et al. (2001) and Wu et al. (2008) [76,77], with the Kuroshio Current information adapted from Caruso et al. (2006) [78]. Deep-water currents are inferred from the combined research of Qu et al. (2006), Wang et al. (2011), and Zhao et al. (2014) [74,79,80]. The currents are numbered 1 to 9, indicating winter (depicted in black) and summer (depicted in red) surface currents, which include: (1) Loop Current, (2) SCS Branch of the Kuroshio, (3) NW Luzon Cyclonic Gyre, (4) NW Luzon Cyclonic Eddy, (5) NW Luzon Coastal Current, (6) SCS Warm Current, (7) Guangdong Coastal Current, (8) SE Vietnam Offshore Current, and (9) Gulf of Tonkin Surface Current. The numbers 10 to 12 represent deep-water currents: (10) Luzon Deep Current, (11) SCS Contour Current, and (12) Deep Cyclonic Current. The base map is constructed using ETOPO1 data.

Figure 1. Map showing the position of core SCS1, marked by a red star, within the context of the monsoon winds and current systems prevalent in the northern SCS. The depiction of monsoon winds is based on Webster (1994) [75], while the mapping of major surface currents follows the work of Fang et al. (1998) [72]. The surface currents in the Gulf of Tonkin are derived from studies by Xia et al. (2001) and Wu et al. (2008) [76,77], with the Kuroshio Current information adapted from Caruso et al. (2006) [78]. Deep-water currents are inferred from the combined research of Qu et al. (2006), Wang et al. (2011), and Zhao et al. (2014) [74,79,80]. The currents are numbered 1 to 9, indicating winter (depicted in black) and summer (depicted in red) surface currents, which include: (1) Loop Current, (2) SCS Branch of the Kuroshio, (3) NW Luzon Cyclonic Gyre, (4) NW Luzon Cyclonic Eddy, (5) NW Luzon Coastal Current, (6) SCS Warm Current, (7) Guangdong Coastal Current, (8) SE Vietnam Offshore Current, and (9) Gulf of Tonkin Surface Current. The numbers 10 to 12 represent deep-water currents: (10) Luzon Deep Current, (11) SCS Contour Current, and (12) Deep Cyclonic Current. The base map is constructed using ETOPO1 data.

3. Materials and Methods

3.1. Core Description

Core SCS1, measuring 7.82 m in length, was collected from a deep-sea basin in the northern SCS using a gravity column sampler on 19 June 2017, at coordinates 18°30′22.56″ N, 116°16′5.28″ E and a water depth of 3770 m (Figure 1). The core is lithologically consistent, comprising gray-green clay silt (Figure 2). Its sedimentary profile is marked by continuous sedimentation, devoid of any discernible hiatuses, turbidite layers, or signs of biological disturbance.

3.2. Analysis of Carbonate (%) Content

In the sediments of the SCS, carbonate, primarily derived from calcareous microfossils such as foraminifera and nannoplankton, as well as from coral reefs, is the predominant biogenic material [70]. Core SCS1 was sampled at 2 cm intervals, yielding 390 samples for carbonate content analysis. The prevalent method for determining the carbonate content in sediment involves treating samples with hydrochloric acid, followed by the measurement of carbon dioxide release. This volume of carbon dioxide is then quantified as the carbonate concentration (%), employing the Scheibler method, with an accuracy of within ±1% [81].

3.3. Time Series Analysis

The time series analysis focused on the higher-resolution carbonate data series. The initial step involved using linear interpolation to resample the data, ensuring even spacing. To accentuate astronomical signals by mitigating low-frequency, long-term trends, the carbonate series underwent pre-whitening, subtracting a 35% weighted average through the LOWESS method [82]. The method emphasizes short-term changes to resolve high-frequency parameters from the strong red noise background signal common in stratigraphic data. Subsequently, the presence of cycles in the raw data was assessed using the 2π multi-taper method (MTM) for power spectral analysis [83], with spectral estimations compared against robust red noise models [84]. Evolutionary spectral analysis was conducted using a sliding window fast Fourier transform (FFT) spectrogram, enabling dynamic tracking of frequency shifts in the carbonate series over time [85]. Gaussian bandpass filtering isolated the interpreted Milankovitch cycles from the carbonate series. These procedures were executed using Acycle v2.4.1 software [86], with further details and guidance available at www.mingsongli.com/acycle (accessed on 6 March 2023). The final step involved tuning the carbonate series’ ~20 kyr component to align with the precession astronomical solution, utilizing QAnaly Series v1.5.0 software for the correlation between the carbonate record and the astronomical solution [87,88].

4. Results and Analyses

4.1. Carbonate Stratigraphy Comparison

The carbonate content of core SCS1 is generally low, ranging from 2.8% to 38.0%, with an average value of 13.8%. According to the level and change trend of carbonate content, this variability is categorized into four distinct stages (Figure 2): Stage I (7.82–5.2 mbsf): the carbonate content fluctuates between 10.3% and 15.1%, averaging 12.7%. This stage is marked by relatively low but gradually increasing carbonate levels from 6.58 to 6.02 mbsf, exhibiting minimal variability. Stage II (5.20–3.10 mbsf): here, the carbonate content spans 10.4% to 21.1%, with an average of 15.3%. This stage shows moderate variability and an overall increase in carbonate content compared to Stage. Stage III (3.10–0.98 mbsf): carbonate levels oscillate between 9.5% and 38.0%, with an average of 14.5%. This stage features significant variability, with lower content observed from 3.10 to 1.66 mbsf and higher levels noted between 1.66 to 0.98 mbsf. Stage IV (0.98–0.02 mbsf): this stage sees the carbonate content ranging from 2.8% to 20.1%, averaging 11.8%. Characterized by large fluctuations, the carbonate content generally decreases towards shallower depths, with the lowest value recorded in Stage IV. Periodic fluctuations across various scales in the data support the feasibility of cyclostratigraphy analysis.

Carbonate content is crucial for understanding deep-water sediments and offers valuable insights into marine paleoenvironmental and paleoclimate changes. It also reflects the historical interplay between climate change and land–sea interactions through the relative content of terrestrial detritus and marine authigenic carbonate deposits [89]. Dividing sedimentary strata based on carbonate content offers an effective alternative when foraminiferal oxygen isotope values are absent, a technique extensively applied in the Pacific since the Quaternary [58,59,61,90]. However, the standard carbonate stratigraphic time scale in the Pacific differs from that in the SCS. Given the ventilation of the deep SCS by the same Pacific Ocean deep-water mass through the Luzon Strait, any changes in seawater chemistry are recorded in the basin’s carbonate system [58,62,74]. Consequently, a standard carbonate stratigraphic time scale for the SCS was established by comparing extensive carbonate data with the SO49-37KL core, introducing the naming convention of the high-resolution standard carbonate stratigraphic event as SB events for the SCS, with numbers indicating specific events [58]. It is widely comparable in the SCS [58,90,91].

In this study, we compared the carbonate content of core SCS1 with the established SCS carbonate stratigraphic time scale to identify main-cycle events and establish an initial chronological framework. At a depth of 5.14 mbsf in core SCS1 (Figure 2), AMS ¹⁴C dating of the carbonate fraction in the sediment was conducted, yielding an age of 42,950 ± 610 years. After calibration using the Calib 8.20 program and Marine 20 dataset [92], with a regional reservoir correction, the calibrated age is 42,733 years. Due to some limitations of the AMS ¹⁴C dating of marine sediments mentioned earlier, this dating result serves as a reference for the carbonate stratigraphic comparison. The identified carbonate stratigraphic events—SB2.0, SB3.0, and SB5.0—correspond to ages of approximately 12 kyr, 24 kyr, and 74 kyr, respectively, corresponding to core depths of 98 cm, 310 cm, and 614 cm, in that order (Table 1; Figure 3 and Figure 4). These events align with the known stratigraphic divisions of the adjacent SO50-29KL core, demonstrating a high similarity in fluctuation trends (Table 1; Figure 3 and Figure 4). Further comparison with twenty high-resolution cores from the SCS confirms the reliability of the carbonate stratigraphic division results for core SCS1 (Table 1; Figure 3 and Figure 4). This comparison reveals consistent carbonate content changes across the region, with distinct patterns observed near identified events. However, regional sedimentation and sampling resolution may account for slight variations in the curve changes.

Table 1. Sediment cores from the SCS that were used in this study.

Core	Latitude (N)	Longitude (E)	Water Depth (m)	References
SO49-37KL	17°49′2.15″	112°47′5.63″	2004	[58]
SO50-29KL	18°26′4.80″	115°39′13.20″	3766	[58]
83PC	17°39′31.32″	112°32′37.32″	1917	[93]
SO50-37KL	18°54′0″	115°48′0″	2695	[94]
S08-57	10°37′1.48″	113°42′39.99″	3400	[90]
V36-3	19°0′30″	116°5′36″	2809	[58]
V36-6	19°46′30.00″	115°48′30″	1597	[58]
MD01-2393	10°30′	110°2′60″	1230	[95]
MD972142	12°42′	119°28′	1557	[96]
ODP1144	20°03′	117°25′	2036	[70]
ODP1143	09°22′	113°71′	2772	[70]
ODP1148	18°50′	116°34′	3297	[70]
SO17956	13°51′	112°35′	3387	[70]
TP86	9°23′24″	115°40′12″	1722	[97]
HYD235	12°24′36″	118°21′	2695	[97]
ZJ83	15°46′48″	112°32′24″	1511	[97]
111PC	18°10′12″	112°1′12″	2253	[97]
ZSQD6	18°30′	114°10′48″	3020	[97]
ZSQD289	19°52′12″	119°52′12″	3605	[97]
SCS-15A	10°25′	114°14′	1812	[60]
SO49-14KL	18°18′	114°24′	3624	[60]
17957	10°54′	115°18′	2195	[60]
GIK17937-2	19.50°	117.67°	3428	[98]
ODP1145	19.58°	117.63°	3175	[99,100]
191PC	19.05°	116.22°	2510	[101]
MD05-2904	19.46°	116.25°	2066	[102,103]
MD05-2903	19.46°	116.25°	2047	[104]
MD05-2905	20°08.2′	117°21.6′	1647	[105]
ZHS176	20°00.0′	115°33.3′	1383	[106]
17940	20°07.0′	117°23.0′	1727	[107]
MD972142	12°41.1′	119°27.9′	1557	[60]
MD052901	14°22.5′	110°44.60′	1454	[108]
ODP1144	20°03′	117°25′	2036	[70]
MD01-2393	10°30′	110°2′60″	1230	[95]
SCS1	18°30′22.56″	116°16′5.28″	3770	This study

Table 1. Sediment cores from the SCS that were used in this study.

Core	Latitude (N)	Longitude (E)	Water Depth (m)	References
SO49-37KL	17°49′2.15″	112°47′5.63″	2004	[58]
SO50-29KL	18°26′4.80″	115°39′13.20″	3766	[58]
83PC	17°39′31.32″	112°32′37.32″	1917	[93]
SO50-37KL	18°54′0″	115°48′0″	2695	[94]
S08-57	10°37′1.48″	113°42′39.99″	3400	[90]
V36-3	19°0′30″	116°5′36″	2809	[58]
V36-6	19°46′30.00″	115°48′30″	1597	[58]
MD01-2393	10°30′	110°2′60″	1230	[95]
MD972142	12°42′	119°28′	1557	[96]
ODP1144	20°03′	117°25′	2036	[70]
ODP1143	09°22′	113°71′	2772	[70]
ODP1148	18°50′	116°34′	3297	[70]
SO17956	13°51′	112°35′	3387	[70]
TP86	9°23′24″	115°40′12″	1722	[97]
HYD235	12°24′36″	118°21′	2695	[97]
ZJ83	15°46′48″	112°32′24″	1511	[97]
111PC	18°10′12″	112°1′12″	2253	[97]
ZSQD6	18°30′	114°10′48″	3020	[97]
ZSQD289	19°52′12″	119°52′12″	3605	[97]
SCS-15A	10°25′	114°14′	1812	[60]
SO49-14KL	18°18′	114°24′	3624	[60]
17957	10°54′	115°18′	2195	[60]
GIK17937-2	19.50°	117.67°	3428	[98]
ODP1145	19.58°	117.63°	3175	[99,100]
191PC	19.05°	116.22°	2510	[101]
MD05-2904	19.46°	116.25°	2066	[102,103]
MD05-2903	19.46°	116.25°	2047	[104]
MD05-2905	20°08.2′	117°21.6′	1647	[105]
ZHS176	20°00.0′	115°33.3′	1383	[106]
17940	20°07.0′	117°23.0′	1727	[107]
MD972142	12°41.1′	119°27.9′	1557	[60]
MD052901	14°22.5′	110°44.60′	1454	[108]
ODP1144	20°03′	117°25′	2036	[70]
MD01-2393	10°30′	110°2′60″	1230	[95]
SCS1	18°30′22.56″	116°16′5.28″	3770	This study

Figure 3. The geographic locations of sediment cores were analyzed in this study. Red dots indicate that these cores used for stratigraphic comparison. Yellow dots indicate that these cores are notable for exhibiting an early Holocene low-carbonate event, in contrast to core SCS1. For detailed locations and references, see Table 1. The HydroRIVERS dataset [109] can be accessed at www.hydrosheds.org (accessed on 2 January 2024).

Figure 3. The geographic locations of sediment cores were analyzed in this study. Red dots indicate that these cores used for stratigraphic comparison. Yellow dots indicate that these cores are notable for exhibiting an early Holocene low-carbonate event, in contrast to core SCS1. For detailed locations and references, see Table 1. The HydroRIVERS dataset [109] can be accessed at www.hydrosheds.org (accessed on 2 January 2024).

Figure 4. Downcore variations in carbonate percentage across 23 extended cores within the SCS. For specific locations and references, please refer to Table 1. The numerals following “SB” denote carbonate stratigraphic events.

Figure 4. Downcore variations in carbonate percentage across 23 extended cores within the SCS. For specific locations and references, please refer to Table 1. The numerals following “SB” denote carbonate stratigraphic events.

4.2. Astronomical Calibration of the Time Scale

4.2.1. Initial Age Model and Spectral Analysis

An initial age–depth model for core SCS1 sediments was established using carbonate stratigraphy (Figure 5). Ages were interpolated and extrapolated across samples by assuming uniform sedimentation rates from the core’s top (considered to be of present age) to all established age control points. Linear sedimentation rates for various stages were calculated based on the age–depth relationship (Figure 5), suggesting the base of core SCS1 to be approximately 100 kyr old, with an overall average sedimentation rate of around 8 cm/kyr. This rate aligns with those from other studies in the northern SCS near our core location [21,100,110]. According to the variation in sedimentation rate (Figure 5), we found that the average sedimentation rate of Stage I and II in the carbonate content record is 6.1 cm/kyr, which belongs to the MIS 3–5 period; the average sedimentation rate of stage III is 17.6 cm/kyr, which belongs to the MIS 2 period; and the average sedimentation rate of stage IV is 8.1 cm/kyr, which belongs to the MIS 1 period. The sedimentation rate during the glacial period is higher than that during the interglacial period, with the highest sedimentation rate during the MIS 2 period in the last glacial period.

We used the statistical software program Acycle v2.4.1 [86] to conduct spectral analyses on the carbonate datasets on the initial age–depth model. The power spectra demonstrated that the carbonate record is dominated by 20 kyr precession paced cycles (Figure 6C). Based on the initial age model, the spectral analysis of the carbonate data series shows that the sedimentary variations in core SCS1 correlate with Earth’s orbital changes, marking an orbital-scale climate cycle. This is a crucial basis for employing the precession series as target curves for refining the initial age scales. The detection of the precession cycle implies low-latitude process influences on core SCS1 deposition, with confidence levels for these climate cycles exceeding 99%, indicating reliable spectral analysis.

4.2.2. Astronomical Calibration

An initial age model was developed by aligning the carbonate curve of core SCS1 with the established carbonate stratigraphic time scale of the SCS. However, this initial model, rooted in carbonate stratigraphy, does not suffice for high-resolution studies that scrutinize the climate’s response to orbital forces. A more direct approach involves tuning the carbonate record to Earth’s orbital parameters, a method that not only aligns with the establishment of astronomical time scales but also promises enhanced age precision. The carbonate sequence of core SCS1 was adjusted to match the corresponding orbital parameters, laying the groundwork for an orbital-scale chronological framework. Analysis of the carbonate sequence, based on the initial age model, reveals that precession cycles prominently influence the sedimentary record (Figure 6C). Consequently, we selected Earth’s orbital precession as our astronomical target curve, employing the La2004 precession solution [87] as our reference for tuning. The La2004 orbital solution [87] is calculated using modern input values for the dynamical ellipticity of the Earth and the tidal dissipation in the evolution of the Earth–Moon system. This orbital solution has been improved with respect to La1993 and proved to have an accurate solution for the geological records [2,47,51].

The carbonate record consistently captures the 20 kyr precession cycle throughout the study period, as depicted in Figure 6B, and is thus chosen as the tuning signal. Leveraging the initial age model, we embark on a detailed calibration of the carbonate record to the 20 kyr precession cycle. Prior research suggests that carbonate content is likely a direct regional response to insolation forcing [56], leading us to base our tuning on the assumption of a steady and linear in-phase relationship between the tuning signal and the precession target. In their study, Zeeden et al. [111] tuned the magnetic susceptibility (MS) maxima to precession maxima in the astronomical tuning of core ODP 926. The low carbonate contents usually induce high MS values in the SCS [2,51]. Thus, the high values of MS indicate a relatively low carbonate content, since it is inversely correlated to the carbonate record [54,111,112]. Accordingly, we associate the lowest carbonate content with the highest precession values. Furthermore, Li’s [108] analysis further refines this relationship by demonstrating that peaks in carbonate content align with troughs in precession values, directly linking carbonate maxima to precession minima in the SCS cores. Thus, the phase relationship between carbonate cycles and precession parameters is characterized by a correspondence where the lowest (highest) carbonate levels coincide with the precession index’s peak (trough) values. Precession dominates productivity and Asian winter monsoon changes in the SCS [66,67,68,113,114,115]. Meanwhile, the productivity change in the northern SCS is affected by the Asian winter monsoon [114,116,117]. The minimum value of the precession causes the enhancement of the Asian winter monsoon, which leads to the enhancement of carbonate productivity, and the opposite is true when the precession is at the maximum value. A detailed discussion is presented in Section 5.2.

We tuned our 20 kyr cycle of the carbonate record to the 20 kyr cycle component curve extracted from the precession to develop an astronomical time scale for core SCS1 [87]. Our analysis identified approximately five and a half 20 kyr cycles within the filtered carbonate content (Figure 6A). The precession model for the 0 to 120 kyr interval revealed six 20 kyr cycles (Figure 6D), aligning closely with observations from visual assessments and Gaussian filtering. We manually aligned the carbonate maxima to 20 kyr precession minima by visually selecting tie points (Figure 6) in the QAnalySeries v1.5.0 software [88]. By iteratively refining the initial age model with precession data, we achieved maximum coherence between the filtered carbonate and orbital signals, resulting in a high-resolution time scale. The congruence between the carbonate record’s 20 kyr filter and the precession curve’s 20 kyr component (Figure 6A,D) underscores the reliability of our age model. The tuned carbonate time series date was used to perform a power spectral analysis in Acycle v2.4.1 software. The method was the multi-taper method (MTM) and robust AR (1) red noise model. The power spectrum of the tuned carbonate time series indicates spectral peaks at periodicities of 20 kyr (Figure 6I). We also used Acycle v2.4.1 software to carry out evolutionary spectral analysis of tuned carbonate time series date by fast Fourier transform. In the evolutionary FFT spectrogram (Figure 6H), one can observe persistent power in the 20 kyr band. The core SCS1 records, as per the carbonate-tuned age model, span from 0 to approximately 110 kyr (Figure 6G), with evolutionary spectral analyses of the tuned carbonate time series revealing a pronounced expression of the precession cycle within the dataset (Figure 6H).

The validity and accuracy of the established astronomical time scale can be gauged through time-domain spectral analysis. In the time-domain spectral analysis, if the signals of the Earth’s orbital parameter cycles used for astronomical tuning are strong and the peaks of the other parameter cycles are within the confidence interval, the established astronomical age scale is more reliable [118]. The spectral analysis results after tuning are consistent with those of the spectral analysis of the initial age model. Time-domain spectral analysis of the tuned carbonate series (Figure 6I) confirms the dominance of a 20 kyr precession cycle, with confidence levels for all cycles exceeding 99%. These results corroborate the presence of the expected astronomical cycles, demonstrating the accuracy of both the astronomical cycle signal interpretation and the established astronomical time scale.

5. Discussion

5.1. Low-Carbonate Event

The carbonate record from core SCS1 spans marine isotope stage (MIS) 1 to 5. The average carbonate content is notably low at 13.6%, except for a significant increase during MIS 3 (Figure 6G). Peaks in the carbonate curve are evident in MIS 3 and late MIS 2, whereas the lowest values are observed in MIS 4, 5, and the late phases of MIS 1 and 3. Additionally, the carbonate content experiences a peak in late MIS 2 before declining towards MIS 1 and 3 (Figure 6G). The carbonate content curve largely reflects the “Pacific-type” carbonate cycle [60]. That is, both MIS 5 and 4 record a significantly lower carbonate content, which is a common feature of the equatorial Pacific Ocean; MIS 3 records a high carbonate content, which is consistent with the higher carbonate content in MIS 3 than in MIS 2 of the “Pacific-type” cycle [58,119]. Similar carbonate cycles are observed not only in core SO-50-29 from the northern SCS but also in core TT013-PC72 from the equatorial Pacific [58,120]. Anomalies in carbonate content are particularly notable during the Last Glacial Period within core SCS1. MIS 3, an interglacial step within this period, is characterized by a high carbonate content, while the glacial step (MIS 4 and 2) exhibits relatively lower levels. The Last Glacial Maximum, in particular, recorded the lowest carbonate content within this period. The Holocene’s average carbonate content, at 11.3%, falls below that of the Last Glacial Period. It can be seen that core SCS1’s carbonate record not only follows the overall “Pacific-type” cyclic pattern but also exhibits “Atlantic-type” cyclic characteristics during the Last Glacial Period.

The carbonate content increased from the Last Glacial Period to the Holocene [62,121]. However, an abrupt decrease in carbonate content during the early Holocene is noted, with the content dropping from 20.1% to 7.9%, a 61% reduction (Figure 7A). Following this decrease, the sediment’s carbonate content gradually recovers (Figure 7). This phenomenon is termed the early Holocene low-carbonate event (~11–8 kyr) [62,89,98,106,121,122]. It is widespread in the sedimentary cores of the broad continental slope of the northern SCS. It occurs relatively simultaneously in many sedimentary cores [62], for example, the Xisha Trough [89], the Pearl River Canyon [122], and near the Dongsha Islands [105]. To mitigate the influence of regional characteristics, 13 cores of relatively high-resolution records were selected for comparative analysis (Table 1; Figure 3 and Figure 7), revealing that the early Holocene low-carbonate event is common across both deep and shallow water cores, with consistent results regarding its timing and duration within dating uncertainties (Figure 7). These comparisons further validate the reliability of the astronomical time scale based on carbonate stratigraphy at core SCS1.

The cause of the early Holocene low-carbonate event in core SCS1 prompts investigation. Considering that changes in seawater chemistry impact the carbonate system throughout the SCS, yet no similar event is observed in the western and southern parts of the sea, the hypothesis of enhanced carbonate dissolution is discounted [62]. This event is likely related to dilution, coinciding with increased rainfall from the East Asian summer monsoon, as indicated by sedimentary records from Dongge Cave stalagmites and Lake Huguang Maar (Figure 7I,R), showing peak rainfall during ~11–8.0 kyr [123,124,125]. This increase in rainfall could have intensified erosion in northern Hainan Island and southern China [89,126,127], thereby enhancing river unloading capacity and the dilution of deep-sea carbonate with terrestrial materials. Additionally, sudden cooling [128,129] and volcanic events [58,130] could have further increased terrestrial material input, diluting the carbonate content. Huang et al. [62] attributed the low-carbonate event to sediment dilution from Taiwan and Luzon, a finding further supported by Luo et al. [121]. Besides these factors, a decrease in surface ocean productivity is another potential contributor [98,122]. The δ¹³C_org values in the core, indicating organic-matter sources in sediments [131], show a significant decrease during this period (Figure 7P), alongside a notable reduction in total organic carbon (TOC) content (Figure 7Q). This suggests a decline in surface seawater productivity, impacting the carbonate content. Consequently, the early Holocene low-carbonate event in core SCS1 is attributed to terrestrial material dilution, coupled with reduced surface seawater productivity.

Figure 7. The variability in carbonate content and other relevant proxies around the SCS to illustrate the early Holocene low-carbonate event: (A,J) Display the carbonate content curves for core SCS1. (P,Q) Present the total organic carbon (TOC) and δ¹³C_org curves for core SCS1, respectively. (B–H,K–O) Feature carbonate content curves from various cores, including GIK17937-2, ODP1145, 191PC, MD05-2904, MD05-2903, ODP1144, 17940, MD05-2905, MD97-2142, MD052901, ZHS176, and MD01-2393. Locations and references for these cores are detailed in Table 1. Comparison of low-carbonate event around the SCS. (I) Shows the titanium (Ti) content from the sediment sequence of Lake Huguang Maar, indicating changes in environmental conditions [123]. (R) Represents oxygen isotope values from stalagmites in Dongge Cave, offering insights into past climatic fluctuations [125]. A gray shaded bar across the figure highlights the period of the early Holocene low-carbonate event, visually correlating the data points and illustrating the widespread nature of this event across different locations in the SCS. This comprehensive comparison helps to further understand the timing, extent, and potential causes of the low-carbonate event during the early Holocene.

Figure 7. The variability in carbonate content and other relevant proxies around the SCS to illustrate the early Holocene low-carbonate event: (A,J) Display the carbonate content curves for core SCS1. (P,Q) Present the total organic carbon (TOC) and δ¹³C_org curves for core SCS1, respectively. (B–H,K–O) Feature carbonate content curves from various cores, including GIK17937-2, ODP1145, 191PC, MD05-2904, MD05-2903, ODP1144, 17940, MD05-2905, MD97-2142, MD052901, ZHS176, and MD01-2393. Locations and references for these cores are detailed in Table 1. Comparison of low-carbonate event around the SCS. (I) Shows the titanium (Ti) content from the sediment sequence of Lake Huguang Maar, indicating changes in environmental conditions [123]. (R) Represents oxygen isotope values from stalagmites in Dongge Cave, offering insights into past climatic fluctuations [125]. A gray shaded bar across the figure highlights the period of the early Holocene low-carbonate event, visually correlating the data points and illustrating the widespread nature of this event across different locations in the SCS. This comprehensive comparison helps to further understand the timing, extent, and potential causes of the low-carbonate event during the early Holocene.

5.2. Effects of Precessional Forcing on Carbonate Content Variations in the Low-Latitude Sea

Since the Quaternary, the formation and expansion of the Northern Hemisphere ice sheet have played a dominant role in the global climate. However, as an important part of the global climate system, low latitudes also play an essential role. The apparent precession cycle in the carbonate record of core SCS1 is a typical low-latitude feature. It should be the response of low-latitude processes (such as the East Asian monsoon) to orbital parameters. Thus, variations in carbonate content in core SCS1 respond to precession forcing, i.e., driven by low-latitude processes.

The precession curve of the tuned carbonate time series is consistent with that of the northern hemisphere summer solar insolation (Figure 6E,F). It shows that the change in carbonate content is influenced by solar insolation. The greater the solar insolation, the higher the temperature, and thus, the more carbonate is formed. Moreover, spectral analysis underscores the dominant role of precession forcing in modulating productivity changes within the SCS region [113]. The evident strong precession cycle in core SCS1’s carbonate record suggests that productivity changes significantly affect carbonate preservation at this site [114]. The variability in surface ocean productivity in the SCS, particularly in its northern parts, is closely linked to the intensity of the Asian winter monsoon systems [114,116,117]. Sedimentary records from the SCS further reveal that these monsoon systems are largely influenced by the precession cycle [66,67,68,115]. Analysis of insolation changes induced by orbital parameters over the last 300,000 years [65,132] shows that minimum precession coincides with maximal northern hemisphere summer insolation and minimal winter insolation, and vice versa for maximum precession [69]. When the northern hemisphere winter insolation caused by precession is weakened (enhanced), the temperature decrease (increase) in the southern East Asian continent is significantly greater than the temperature change in the neighboring western Pacific region, thus strengthening (weakening) the latitudinal land–sea thermal contrast, and then strengthening (weakening) the winter monsoon in the southern East Asia region [67,133]. Thus, when the precession is at a minimum, the Northern Hemisphere winter insolation is low, the East Asian winter monsoon is stronger [66,67,115].

Intensification of the East Asian winter monsoon boosts surface ocean productivity through several mechanisms [6,108,113,134,135,136]. On the one hand, the enhanced winter monsoon increases the input of nutrients in dust deposition, increasing the surface seawater productivity [117,137]. On the other hand, the enhanced winter monsoon drives the surface seawater to flow southward, the thermocline in the northern SCS becomes shallow, the subsurface eutrophic water enters the euphotic layer, and the surface productivity increases. At the same time, the intense winter monsoon can accelerate the vertical circulation of seawater and promote stronger vertical mixing. This brings the trophic cold water below the mixed layer to the surface of the seawater to form a wind-driven upwelling. The upwelling is strong, and the deep water rich in nutrients rises to the surface water, which increases productivity [6,134,136]. In addition, the winter monsoon promotes the North Pacific current to flow through the Taiwan Strait and the Bashi Strait into the SCS, forming a coastal current and bringing more nutrients, resulting in increased surface productivity [138]. Furthermore, the strong East Asian winter monsoon is conducive to the seasonal Kuroshio Current (KC) or Taiwan Warm Current flowing into the northern SCS [139,140]; when the KC flows into the SCS, its nutrient-rich deep water is carried to the surface by upwelling driven by different processes, including wind-driven upwelling, vertical mixing, or cyclonic vortices [141,142], further enhancing productivity. The increase in biological productivity increases the absolute deposition of carbonate, the deepening of the lysocline of carbonate, and the decrease in solubility. Thus, higher productivity is conducive to the preservation of carbonate and increases the carbonate content. The conspicuous precession cycle in our records indicates that the positive impact of increased biological productivity on carbonate deposition outweighs the negative effects of other factors. This analysis suggests that a strong winter monsoon enhances surface seawater productivity during minimum precession phases, favoring carbonate deposition and preservation; the opposite occurs during maximum precession phases. It is noteworthy that monsoon dynamics can also be viewed as manifestations of the seasonal migration of the Intertropical Convergence Zone (ITCZ), with the precession cycle influencing the global monsoon through north–south ITCZ displacement [70,113,143]. Consequently, in the precession band, carbonate deposition and preservation in core SCS1 are influenced by the interplay of the ITCZ and monsoon systems.

6. Conclusions

In this study, we have detailed a high-resolution carbonate record from core SCS1, collected from the depths of the SCS. By integrating carbonate stratigraphy with time series analyses of the carbonate record, we have successfully established an age-calibrated astronomical time scale for core SCS1. The initial analysis of an untuned carbonate stratigraphic age model revealed that precession is the most prominently expressed astronomical parameter. Consequently, an astronomical time scale covering the past ~110 kyr was developed by aligning the time-domain carbonate data with the La2004 precession curve. This astronomical time scale reveals that the carbonate record of core SCS1 exhibits a “Pacific-type” carbonate cycle. Moreover, the timing and duration of the low-carbonate event identified in core SCS1 align with findings from previous research. This event is primarily attributed to dilution effects coupled with a reduction in productivity. The pronounced precession cycle observed in the carbonate record of core SCS1 underscores that changes in carbonate content are driven by precessional forcing, indicative of low-latitude processes. Within the precession band, the interplay between the Intertropical Convergence Zone (ITCZ) and the monsoon system significantly influences carbonate deposition and preservation by modulating productivity variations in the SCS. The insights gained from this study significantly enhance our understanding of paleoceanographic and paleoclimatic evolution in the northern SCS, highlighting the intricate relationship between astronomical parameters and marine sedimentation processes. However, this study did not reconstruct the paleoclimate evolution of the northern SCS, so in future studies, we can combine a variety of climate proxy indicators to carry out research on the paleoclimate evolution of the northern SCS.