Genesis and controlling factors of Cenozoic dolostones in the South China Sea: a case study from core Nanke-1 in the Nansha Islands

Thick dolostones developed widely in the South China Sea (SCS) during the Cenozoic. However, the dolostone types and the formation mechanism of these dolostones remain controversial. The core Nanke-1, drilled on Meiji Reef in the southern SCS, recovers thick dolostone formed during the Neogene, providing a good opportunity to reveal the mechanism of various SCS dolostones. The dolostones in this core exist in two forms: microsucrosic and mimetic dolostones. Microsucrosic dolostone, composed of high-calcium dolomite (HCD), shows larger crystal size and partial iron stains, while mimetic dolostone, composed of low-calcium dolomite (LCD), shows crystals with smaller size. The dolomitization was mediated by the normal to slightly evaporated seawater. The differences in microsucrosic and mimetic dolostones are mostly attributed to rock porosity and Mg/Ca ratio of the dolomitizing fluid. During the Late Miocene, coarse sediments with rich pores accumulated on the platform. These sediments were exposed and heavily dissolved during the sea-level lowstands, producing abundant dissolution pores. The fast circulation during the highstands mediated the dolomitization process, favoring the development of microsucrosic dolostone. In contrast, during the Pliocene, the relatively higher sea-level caused the accumulation of bioclasts with fewer primary pores and formation of limited dissolution pores. These conditions favored the development of mimetic dolostone. The above processes suggest that sea-level fluctuation played an important role in causing the differences in Upper Miocene and Pliocene dolostones. Moreover, the increase in seawater Mg/Ca ratio facilitated the occurrence of HCD during the Late Miocene and LCD during the Pliocene in the SCS. The development of Neogene dolostone in the SCS was roughly synchronous with the dolomitization events elsewhere, which hints that the formation of these dolostones could be controlled by similar triggers (e.g., sea-level fluctuation). The dolomitization mechanisms in core Nanke-1 discussed herein can thus also benefit further explanation for the Cenozoic dolomitization events in other regions worldwide.


Introduction
Since the discovery of dolomite by French scholar Dieudonne Dolomieu in 1791, the mechanism of natural dolomite and their controlling factors have not yet been clarified due to the complexity of dolostone formation and the superposition of multi-stage diagenesis throughout geological history.Many dolomitization models have been proposed to resolve the so-called "dolomite problem", a challenging and hot topic for geologists (Land 1980;Budd 1997;Warren 2000;Machel 2004;Gregg et al. 2015).Dolostones are characterized by varied textures in geological history, which reflects the different genesis and controlling factors (Sibley 1982;Sibley and Gregg 1987;Gregg et al. 1992;Mazzullo 1992;Budd 1997;Choquette and Hiatt 2008;Maliva et al. 2011).Dolomites with these textures were further classified into the 19 Page 2 of 18 scheme of microsucrosic and mimetic dolomites by Dawans and Swart (1988).They suggested that variations in sedimentary and diagenetic processes determined the distribution patterns of these two distinct dolomites.Furthermore, Murray and Swart (2017) identified the presence of microsucrosic dolomite and mimetic dolomite in the dolostone in Bahamas platforms.This study suggested that differences in temperature and salinity influenced the formation of these two types of dolomite, but both types of fluids were derived from seawater.The dolomitization process in the Bahamas dolostones was believed to be associated with brine reflux.Zhao and Jones (2012) identified two distinct types of dolomite in their study of the Cayman Brac Formation dolostones.The dolostones from the Brac Formation are predominantly composed of high-calcium dolomite (HCD) with larger crystal sizes (50-1500 μm), resembling microsucrosic dolomite.In contrast, the overlying dolostones of the Cayman Formation are dominated by finer-grained (10-20 μm) low-calcium dolomite (LCD) resembling mimetic dolomite.The main dolomitization model for both types is "island dolomitization model".The differences between these two types of dolomite may be attributed to variations in the reflux of dolomitizing fluids within the platform and differences in the Mg/Ca ratios of these fluids.Controlling factors of dolomitization include rock features (e.g., permeability, porosity, texture), sea-level, and paleoceanography (Harris and Meyers 1987;Cander 1994;Kyser et al. 2002;Machel 2004;Gaswirth et al. 2007).
The South China Sea (SCS) is one of the world's largest marginal seas, with a large volume of biogenic reefs developed during the Cenozoic, and the Late Miocene biogenic carbonate sequences bear thick layers of dolostones (Han et al. 2022).These dolostones are excellent research objects for investigating the "dolomite problem" in the marginal sea basin.Previous studies on the dolomitization were primarily conducted in the Xisha Islands, northern SCS.The proposed models for the genesis of dolostones include the following: mixing zone dolomitization (He and Zhang 1990;Xu et al. 1994); syngenetic dolomitization in slightly modified seawater (Wei et al. 2006); brine reflux dolomitization in lagoon environments (Wang et al. 2015); and hydrothermal dolomitization involving thermal convection (Wei et al. 2006).
In contrast, studies on reef dolostones in the Nansha Islands, southern SCS, are constrained by limited data.The early studies were mainly based on the data from cores Nanyong-1 and Nanyong-2 drilled on Yongshu Reef in the early 1990s (Luo et al. 2022;Guo et al. 1993;Yan et al. 2002;Yao and Zhan 2006;Zhao et al. 1996;Zhu et al. 1994;Zhao et al. 1996).There was a lack of systematic investigation on the genesis and controlling factors of reef dolostones in the Nansha Islands with such little information.Core Nanke-1 is the first fully cored borehole in the Nansha Islands, with a drilled core length of 997 m.The reliable data from this core have provided broad insights into analyzing dolomitization in the southern SCS.
In this study, the Upper Miocene-Pliocene dolostones from core Nanke-1 are taken as the research object.We integrated our mineral composition, crystal morphology, carbon, oxygen values, and rare earth element of dolostone to identify the formation mechanism and clarify the differences in the genesis and development of microsucrosic and mimetic dolostones.Our research findings will contribute to understanding the mechanism of dolostone formation in carbonate platforms elsewhere and interpreting the global dolomitization events.

Regional geological background
The SCS is located at the intersection of the Eurasian plate, Indian plate, Philippine plate, and Pacific plate and is one of the largest marginal seas in the world.Rifting developed widely along the South China margin between the Late Cretaceous-Early Paleocene, which made the Dangerous Grounds (i.e., Nansha Islands) migrate away from the China mainland (Steuer et al. 2013(Steuer et al. , 2014)).Subsequent sea-floor spreading during the Oligocene-Miocene in the SCS forced the southward migration of the Dangerous Grounds, providing warm environment for the active carbonate growths (Steuer et al. 2013(Steuer et al. , 2014;;Fig. 1).The development of carbonate platforms in the SCS began during the Paleocene, with a vast range in the northern, southern, and western parts of the SCS (Steuer et al. 2013(Steuer et al. , 2014;;Wu et al. 2016;Ding et al. 2015;Wu et al. 2019aWu et al. , 2019bWu et al. , 2021a)).The oldest carbonate rocks may be deposited in the Palawan area, but available reports on this set of sediments are rare due to limited information.From the Early Oligocene to the Early Miocene, the "Nido limestone" was deposited in the Nansha Islands and northwest Palawan, with shallow-water platform facies in its lower part and open marine facies in its upper part.The carbonate deposits on the Reed Bank began during the Early Oligocene, whereas most parts of the carbonate platform were inundated in the Early Miocene.This set of shallow-water carbonate rocks is of great thickness, with a lagoon to shallow-water marine facies.The carbonate rocks in Sarawak in the southern Nansha Islands were deposited from the Eocene to the end of the Early Miocene, mainly composed of slope-platform facies (Li et al. 2022).
Core Nanke-1 was drilled on Meiji Reef, an isolated atoll in the central Nansha Islands (Dangerous Grounds), adjacent to the western Liyue Bank (Reed Bank) in the southern SCS (Steuer et al. 2014;Fig. 1).A basaltic basement underlies the ring-shaped Meiji Reef.Its area is about 36 km 2 , surrounding a semi-enclosed lagoon with an average water depth of roughly 27 m (Shen and Wang 2008;Zhao et al. 1992;Guo et al. 2021).The magnetic and Sr-isotope stratigraphy indicated that the carbonate rocks in Meiji Reef developed since the Late Oligocene, slightly earlier than the beginning of carbonate rock deposition in the Xisha Islands (Lin et al. 2016;Yi et al. 2021;Fan et al. 2019;Shao et al. 2017;Wu et al. 2021b).Core Nanke-1 consists of both reef and bioclastic carbonates with coral, coralline algal, and foraminiferal fragments.The Cenozoic strata are divided into six formations.The Late Oligocene Liyue Formation is rich in larger benthic foraminifera and red algal fragments.The Lower Miocene Meiji Formation mainly consists of bioclastic limestone.The Upper Miocene Nanwan Formation is composed of abundant corals, red algae, bivalves and bryozoans fragments.The Pliocene Yongshu Formation contains corals fragments.The Quaternary Nansha Formation and Nanhai Formation are composed of bioclastic rudstones, floatstones, and sands (Li et al. 2022).The Middle Miocene strata are absent.The hiatus spanning 8.4 m.y (20-11.6Ma; Fig. 2) in Meiji Atoll is thought to be mainly caused by tectonic uplift in the Dangerous Grounds (Nansha) (Li et al. 2022).After the platform was flooded again during the Pliocene, the Upper Miocene strata of core Nanke-1 restarted.

Materials and methods
The samples used in this study were sampled at a space of 2-3 m, with 336 carbonate rock samples collected throughout the entire depth range.X-ray diffraction measurements, carbon and oxygen isotope analysis, and trace element testing were conducted on these 336 samples.Thin sections were produced and examined under a polarization microscope (ZEISS® AxioScope.A1) to determine the lithologic facies and the corresponding diagenetic environments.The analysis mentioned above is described in detail as given below.

XRD analysis
The mineral compositions of the samples were examined at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (Wuhan).The mineralogical characteristics of carbonate rocks and the positions of characteristic peak d(104) of dolomite and calcite were determined using an X-ray diffraction instrument (X'pert MPD Pro), and the mineral compositions and contents were identified based on the features of the characteristic peaks.The diffraction spectra of the carbonate rock samples were obtained at an ambient temperature of 25 °C and a humidity of 65%.Using Jade 6 software, we performed map stitching and Rietveld refinement for the XRD data based on the standard X-ray diffraction powder patterns published by the Joint Committee on Powder Diffraction Standards (JCPDS cards) to measure the carbonate rock samples quantitatively.Moreover, we calculated the mole percentage of CaCO 3 (N CaCO3 ) of the dolomite samples to evaluate the stoichiometry of dolomites among the carbonate rock samples using the following equation from Lumsden and Chimahusky (1980).
where the N CaCO3 is the mole content of CaCO 3 (%), d is the observed d ( 104) spacing of the dolomite samples in Angstrom units, M is 333.33, and B is -911.999(Lumsden and Chimahusky 1980;Lu et al. 2022).The method and detailed implementation has been described by Holland and Redfern (1997).

Carbon and oxygen isotope analysis
The carbon and oxygen isotope analysis was conducted at the laboratory of the Wuhan Sample Solution Analytical Technology Co., Ltd.Oriented carbonate rock specimens (5-10 g) were cut from the selected samples, washed with de-ionized water, put into sample bags, and then dried in a drying oven at 45 °C for 12 h.Next, the dried samples were ground to a fine powder using an agate mortar.After these preparation steps, the δ 13 C and δ 18 O values were measured using a GV Isoprime II isotope ratio mass spectrometer equipped with a MultiPrep online carbonate digestion system.The dolomite samples need to be digested in anhydrous phosphoric acid (with a density of about 1.93 g/ cm 3 ) at 90 °C, while the carbonate rock samples with mixed calcite and dolomite need to be digested in anhydrous phosphoric acid (with a density of about 1.93 g/cm 3 ) at 75 °C.
The carbon and oxygen isotope values obtained were calibrated relative to the international primary reference material Vienna Pee Dee Belemnite (VPDB), following the international standard NBS-19 and the Chinese national standard GBW04405, and the equation of Kim et al. (2015) was employed to normalize the oxygen isotope values from the VPDB scale to the Vienna standard mean ocean water (VSMOW) values.Kim et al. (2007) adopted the fractionation factor to correct the δ 18 O fractionation during the acid digestion of calcite or aragonite, and the isotopic values of dolomites were calibrated with the fractionation factor of Rosenbaum and Sheppard (1986).

Trace element analysis
The trace element analysis was performed at the laboratory of the Wuhan Sample Solution Analytical Technology Co., Ltd.An inductively coupled plasma mass spectrometer (ICP-MS, Agilent 7700e) was used to determine the whole-rock trace element content for the samples based on the wavelength-dispersive X-ray fluorescence spectrometry (GB/T 14506.28-2010).The samples used for ICP-MS analysis were prepared in the following seven steps.(1) The samples were dried in an oven at 105 °C for 12 h; (2) 50 mg of the powder sample was accurately weighed and placed in a Teflon capsule; (3) l mL of pure nitric acid and 1 mL of highpurity hydrofluoric acid were slowly added into the capsule in turn; (4) the Teflon capsule was put into a steel sleeve, which was then tightened and heated in an oven at 190 °C for 24 h; (5) the capsule was opened after cooling down and dried on an electric hot plate at 140 °C; after that, the sample was added with l mL of pure nitric acid and dried on the hot plate (until there was no liquid on the capsule wall); (6) l mL of high-purity nitric acid, l mL of Milli-Q water, and l mL of internal standard In solution (with a concentration of 1 ppm) were added into the sample; the Teflon capsule was put into the steel sleeve again, which was then tightened and heated at 190 °C for 12 h; (7) the solution in the capsule was transferred into a polyethylene bottle and diluted to 100 g with 2% nitric acid (HNO 3 ) for ICP-MS analysis.

Lithostratigraphy
Based on the Sr isotope and magnetic stratigraphy data, as well as exposure surfaces, the sedimentary sequences since the Miocene in core Nanke-1 are divided into five formations (Fig. 2).
The Lower Miocene Meiji Formation (538.6-616.5 m) is mainly composed of bioclastic deposits.Bioclastic wackestone with foraminifera and gastropods characterizes this formation, and small intergranular dissolved pores are common (Fig. 3a).The Upper Miocene Nanwan Formation (287-538.6m) mainly consists of dolomitic framestone and wackestone.The framestone is characterized by rich corals and rare benthic and planktonic foraminifera.Skeletal and dissolved pores are abundant in the coral framestone (Fig. 3c).The wackestone with benthic foraminifera has abundant large intergranular pores and intergranular dissolved pores (Fig. 3b).The Pliocene Yongshu Formation (210.1-287m) is characterized by alternation of wackestone and framestone as well as algal bindstone.Wackestone contains red algal and foraminiferal fragments with small size (below 0.1 mm).Small moldic pores and intergranular dissolved pores are identified (Fig. 3d).The Nansha Formation (20.5-210.1 m) is composed of framestone, packstone and wackestone with rich corals and some red algal and foraminifera fragments.Dissolved pores are frequent in this formation (Fig. 3e).The Nanhai Formation (above 20.5 m) mainly consists of unconsolidated bioclastic sands.

Distribution and characteristics of exposure surfaces
Reddish iron-oxide strips and nodular texture occurs frequently in core Nanke-1.Dissolution pores and leaching of the bioclasts are also obvious in some short intervals.These features represent the exposed surfaces recorded in core Nanke-1.A total of 27 exposure surfaces are identified in the Miocene-Holocene interval, which are located at depths of 560.9 m, 587.9 m, and 616.

Mineral composition
Three types of minerals (i.e., aragonite, calcite, and dolomite) are identified in the interval of 616.5-5 m in core Nanke-1 (Fig. 5).Aragonite only occurs above 20.5 m.It decreases with depth, while the calcite increases with depth in the upper 20.5 m interval.Calcite is generally dominant in the intervals of 125-10 m and 616.5-538.6 m.Dolomite is dominant in the interval of 538.6-125 m.In addition, calcite also occurs in short intervals around 503 m, 485 m, 450 m, 383 m, and dolomite also occurs in short intervals around 60 m.

Dolomite optical features
Based on the observed microscopic differences, the dolomite from core Nanke-1 was identified into two types: microsucrosic dolomite and mimetic dolomite (Fig. 6).

Carbon and oxygen isotopes
The carbon and oxygen isotope data (δ 13 C and δ 18 O values) of the samples from core Nanke-1 exhibit a variation trend in line with mineral transformations (Fig. 7).Both of the δ 13 C and δ 18 O values are variable and negative in the Lower Miocene (538.6-616.5 m).The low δ 13 C (− 5.7 to 1.7‰) and δ 18 O (− 6.1 to − 1.4‰) of the section exhibit a correspondence with occurrence of calcite (Fig. 7).In the Late Miocene to Early Pleistocene interval (538.6-120.5 m) dominated by dolomite, the values of δ 13 C (1.10-3.82‰)and δ 18 O (0.86-4.17‰) isotopes are generally high and stable.
In general, carbon and oxygen isotopes have a strong correlation with mineral composition in the core Nanke-1.The higher δ 13 C and δ 18 O values match the dolomite intervals and the calcite intervals have the lower values of the isotopes.

Rare earth elements
A PASS normalization for the rare earth element data of core Nanke-1 was conducted in this study.For comparison, the rare earth element data of seafloor hydrothermal fluids, terrigenous clastic sediments, and surface seawater of the SCS are also presented (Alibo and Nozaki 2000;Wan et al. 2017;Douville et al. 1999).The results, as demonstrated in Fig. 8, indicate the normalized rare earth element distribution patterns for the samples from core Nanke-1 and are characterized by prominent negative Ce anomalies, but no positive Eu anomaly.In addition, the carbonate samples from core Nanke-1 retain an REE distribution pattern similar to that of modern seawater of the SCS, which significantly differs from the REE distribution patterns of terrigenous clastic sediments and seafloor hydrothermal fluids (Fig. 8).

Dolomitization fluids
As shown in Fig. 8, the REE pattern of dolostones in core Nanke-1 is similar to that of SCS seawater, instead of hydrothermal fluids.This indicates that seawater, rather than hydrothermal fluids, could be the main fluid mediating the dolomitization process in in Meiji reef.Moreover, in the dolostones of core Nanke-1, positive δ 13 C values and no obvious correlation between δ 13 C and δ 18 O values also imply that the dolomitization fluids are seawater instead of mixed freshwater-seawater.Absence of evaporite (e.g., gypsum) in core Nanke-1 implies that the dolomitization process was not linked to the hypersaline brines.The δ 13 C and δ 18 O values of the dolostones in core Nanke-1 range from 1.86 to 3.17%, and from 0.86 to 4.15%, respectively, which overlaps with the values in other island dolostones in the South China Sea and also elsewhere in the world (Fig. 9).This suggests that the dolomitization mechanisms here could be similar to those of other dolostones (Wang et al. 2018).Studies on the dolostones elsewhere commonly indicated that normal or slightly evaporated seawater caused the dolomitization process (Wang et al. 2018).Thus, we suggest that the dolostones in core Nanke-1 was also influenced by the normal to slightly evaporated seawater.This point is in line with the indication by the δ 18 O values of dolomitization fluid in core Nanke-1 (Fig. 9; Guo et al. 2021).

Factors influencing the differential dolomitization in the SCS
In core Nanke-1, two types of dolostones, namely microsucrosic and mimetic dolostones are identified in the Upper Miocene-Pleistocene interval.Microsucrosic and mimetic dolostones are different in texture, fabric and crystal feature (Fig. 6; Table .1).The Upper Miocene dolostone is mainly microsucrosic, characterized by large and limpid rhombic crystals with higher porosity, while Plio-Pleistocene dolostone is mainly mimetic, characterized by indistinct small crystals with low porosity (Figs. 3,6;Table. 1).Microsucrosic dolostones moderately destroyed the depositional fabrics, while the mimetic dolostones caused limited damages to the depositional fabrics.In addition, the Upper Miocene dolostones are composed predominantly of HCD, while the Pliocene-Pleistocene dolostones are composed of both HCD and LCD (Table .1).

Sea level
Sea-level is a potential factor influencing the dolomitization process (Ren and Jones 2017).In the Meiji Reef, microsucrosic dolostones occur in company with higher porosity, and mimetic dolostones occur in company with lower porosity.The higher porosity help accelerate the migration of seawater through the carbonate system, facilitating the formation of microsucrosic dolostone and also the destruction of depositional fabrics.The difference in porosity could arise from depositional and diagenetic processes.Compared to the Pliocene interval, the upper Miocene interval contains more framestone, which could cause relatively higher depositional porosity.Meanwhile, more frequent exposure horizons occur in the upper Miocene interval, which would produce more dissolution pores (Fig. 10).Generally, coral framestone tends to thrive in shallower water, and exposure horizons could occur when the sea level fell and exposed the carbonate platform.Regional unconformities also occurred widely during the Late Miocene in the southern SCS (Li et al. 2022;Lai et al. 2021;Tsikouras et al. 2021;Lai et al. 2021).Thus, it is suggested that the Upper Miocene interval was formed in shallower water than the Pliocene interval.According to the Sr-dating in Li et al. (2022), the dolomitization time was about 1 Ma later than the corresponding depositional time.Thus, the dolomitization time was from about 10 to 1 Ma.According to Miller et al. (2020), eustatic sea level during the Upper Miocene was lower than that during the Pliocene (Fig. 10).Thus, the differences in Upper Miocene dolostone (microsucrosic dolostone) and Pliocene dolostone (mimetic dolostone) could be related with the eustatic sea-level changes.

Cenozoic Mg/Ca variations trend
The Upper Miocene dolostones are HCD (Fig. 10; Table .1), while the Pliocene to Lower Pleistocene dolostones are mixed HCD and LCD (Fig. 10;Table. 1).According to Kaczmarek and Sibley (2011), the Ca% of dolomite depends largely on the Mg/Ca ratio of dolomitizing fluids.It is possible that the transition from HCD to mixed HCD and LCD could be related to changes in seawater Mg/Ca ratio from the Miocene to the Pleistocene.According to Coggon et al. (2010), the reconstructed seawater Mg/Ca ratio gradually rose from the Miocene to the Pleistocene.Thus, the increase in seawater Mg/Ca ratio could account for the change from HCD to mixed HCD and LCD.

Formation mechanisms of the Neogene dolostones in the SCS
As discussed above, the dolomitizing fluid in Meiji reef could be normal to slightly evaporated seawater.Sea level and seawater Mg/Ca ratio also influenced the dolomitization process.Taking all these clues into account, we propose the following mechanism for the formation of the Neogene dolostones in the SCS (Fig. 11).
During the Late Miocene, coral framestone occurred in the Meiji carbonate platform in the highstands.The platform was exposed in the lowstands, facilitating the occurrence of extensive dissolution pores on the platform tops (Fig. 11a).When the sea level rose, evaporated seawater from lagoon and normal seawater from the open sea circulated fast through the limestone, favored the dolomitization in Meiji platform (Fig. 11b).Dolostones are microsucrosic under conditions of fast circulation.As the seawater Mg/ Ca ratio was low during the Late Miocene, the dolomite was mainly HCD.During the Pliocene, bioclastic limestone began to dominate in the Meiji platform.The porosity of the Pliocene interval was lower due to fewer occurrence of framestone and exposed horizons (Fig. 11c).Seawater could have circulated at lower speed through such carbonate deposits.Thus, more mimetic dolostone occurred in the Pliocene interval (Fig. 11d).As the seawater Mg/Ca ratio increased during the Pliocene, both HCD and LCD occurred in the platform.

Cenozoic dolomitization events in the SCS and their comparison with other sites
Apart from the dolostone in Meiji reef in the Nansha Islands, dolostones also occurred elsewhere in the SCS, such as on Chenhang and Yongxing Island in the Xisha Islands, and also in many regions around the globe (Budd 1997;Xiu et al. 2017;Wang et al. 2015Wang et al. , 2018;;Wei et al. 2008;Zhang et al. 2013;Fig. 13).Dolostone formation was not a local event in the SCS but a series of worldwide events between the Late Miocene to the Early Pleistocene (Budd 1997;Suzuki et al. 2006;Ren and Jones 2018), as shown in Table .2 One to five dolomitization events occurred in different regions (Table 2; Fig. 13).
The Sr isotope dating of dolostone sections in Nanke-1 and other sites (including Xichen-1, Chenke-2, Xiyong-1, GB-2 on Great Bahama Bank, Kato-daito-jima in the Philippines, the ones in Grand Cayman and the western Pacific's Funafuti) exhibits the similar ages (Fig. 13), from about 10 to 1 Ma.Specifically, the two-episode Neogene dolomitization event in the SCS (episodes I and II) revealed by the drill core records of cores Xike-1 (at a reef in the northern   SCS) and Nanke-1 (at a reef in the southern SCS) correspond to the age of the Late Miocene and Pliocene, respectively (Figs. 10,11).Budd (1997) defined seven globally synchronous dolomitization events (A-G) from the Miocene to the Quaternary based on the 87 Sr/ 86 Sr data (Fig. 13).The events B and C roughly correspond to episode I in the SCS and the events D and E correspond to episode II in the SCS.This indicated that Cenozoic dolomitization events could have been commonly linked to global factors (Budd 1997;Jones and Luth 2003;Ren and Jones 2018;Fig. 12).During the Cenozoic, the large fluctuations in eustatic sea-level occurred coevally in the world (Fig. 12; Lear et al. 2000;Zachos et al. 2001;Wu et al 2019a).Sea level generally fell during the Late Miocene due to the occurrence of Antarctic ice sheet.Such a fall in sea level would expose shallowwater carbonate platforms, producing common dissolution pores in the carbonate platform tops.When the sea level rose during the following highstands, the circulation of seawater through the carbonate platform would cause pervasive dolomitization in the world.During the Pliocene, the sea level gently rose, causing fewer exposure surfaces.The circulation of seawater through the carbonate platform was possibly slower, and the dolomitization process was not that active as that during the Miocene.This is also supported by the fact that the Pliocene interval was only partly dolomized at some locations (such as core XC-1, CK-2, XY-1 and XK-1 in the SCS).What's more, the thicknesses of the Cenozoic dolostone sections vary from site to site (Fig. 13), which suggests that eustatic changes in sea-level may not have been the only factor involved (Wei et al. 2008;Zhang et al. 2013;Wang et al. 2015Wang et al. , 2019;;Xiu et al. 2017;Ohde and Elderfield 1992;Jones and Luth 2003;Ohde et al. 2002).For example, it is worth noting from Fig. 13 that the sequence in core Nanke-1 in the southern SCS has thicker dolostone sections than those in cores Xike-1, Xiyong-1, Xichen-1, and Chenke-2 in the northern SCS (Xiu et al. 2017;Wang et al. 2015Wang et al. , 2018;;Wei et al. 2008;Zhang et al. 2013).Such significant lithological differences between the northern and southern SCS may be caused by the higher temperature and higher salinity due to more intense evaporation on the carbonate platforms in the southern SCS near the tropics.The joint effects of high temperature promoted the formation of dolomitizing fluids and the dolomitization process (Budd 1997).As a result, Meiji Reef experienced more significant dolomitization compared to the carbonate platforms in the central-northern SCS.In addition, according to Wang et al. (2019), higher tectonic subsidence rate of the platform could also be beneficial to the increase in thickness of dolostone strata.The Meiji platform experienced obvious subsidence after 10.5 Ma (Li et al. 2022).The thickness of dolostone succession in the Meiji platform is up to 390 m, which is much larger than that in the Grand Cayman and Kita-daijojima with no obvious subsidence (Wang et al. 2019).

Conclusions
In this study, we conducted mineralogical analysis, microscopic thin-section investigation, and geochemical tests on the samples from core Nanke-1 to explore the mechanism of dolostone genesis in the southern SCS.The dolostones in the southern SCS were also compared with those in the northern SCS and other sea areas worldwide.The conclusions obtained are as follows.
(1) The dolostones in the Nansha Islands mainly occurred in the Upper Miocene-Pliocene interval, dominated by thick-layered dolomite with thin interlayers of calcite.
The Upper Miocene interval is mainly composed of thick-layered microsucrosic dolostone with HCD.In contrast, the Pliocene interval is dominated by thicklayered mimetic dolostone with LCD.
(2) The dolomitization fluid is normal to slightly evaporated seawater.High porosity in limestone succession occurred when the sea level was generally low during the Late Miocene, facilitating the seawater circulation during the highstands and the occurrence of microsucrosic dolostone.In contrast, relatively lower porosity in limestone succession occurred when the sea level was higher during the Pliocene, favoring the occurrence of mimetic dolostone.The increase in seawater Mg/Ca ratio facilitated the development of HCD during the Late Miocene and LCD during the Pliocene.
(3) The Neogene dolomitization event in the SCS can be mainly divided into two episodes: the Late Miocene and the Plio-Pleistocene, which can also be correlated with other dolomitizing events around the world.The good temporal correspondence between the dolomitization events in the SCS and those elsewhere reflects the extensive control of sea-level fluctuation and Mg/Ca ratio on the overall dolostone development worldwide.

Fig. 2
Fig. 2 Magnetic Stratigraphy and Lithostratigraphic Units division of core Nanke-1 (Sr Isotope Data from South China Sea Institute of Oceanology, Chinese Academy of Sciences; Geomagnetic Data from Yi et al. 2021) 5 m in the Early Miocene section, 342.7 m, 352 m, 360 m, 380 m, 384 m, 408 m, 441.2 m, 465 m, 480 m, 490 m, 506 m, and 538.6 m in the Late Miocene section, 185 m, 210.1 m, 236 m, and 287 m in the Pliocene section, 20.5 m, 30 m, 34.5, 50 m, 65.5 m, 84 m, 128 m, and 165 m in the Quaternary section, respectively (Figs. 2, 3).Overall, exposure surfaces frequently developed at Meiji Reef, with one exposure surface occurring about every 20 m in the Early Miocene-Pliocene interval.White carbonate deposits are separated by these reddish exposed surfaces (Figs. 2, 4).

Fig. 3
Fig. 3 Thin section photomicrographs under cross-polarized light illustrate lithological features from core Nanke-1.a Bioclastic wackestone in Meiji Formation.The molds of foraminifera and gastropods and small intragranular dissolved pores are abundant; b Wackestone with foraminifera in Nanwan Formation.Abundant large moldic pores and small intergranular dissolved pores are identified; c Framestone in Nanwan Formation.The medium intergranular pores and intercrystalline dissolved pores are common; d Wackestone in

Fig. 6 Fig
Fig. 6 Thin section photomicrographs under cross-polarized light illustrate optical petrography features from core Nanke-1.a Microsucrosic dolomite in Nanwan Formation.Rhombic crystals have clear rims with foraminifera and gastropods fragments; b Microsucrosic dolomite in Nanwan Formation.Foraminifera mold are broken and filled with dolomite cements; c Mimetic dolomite within Yongshu

Fig. 8
Fig.8PASS-normalized rare earth element distribution patterns of the samples from core Nanke-1 (the rare earth element data of seafloor hydrothermal fluids, terrigenous clastic sediments, and surface seawater of the South China Sea are cited fromAlibo and Nozaki 2000;Wan et al. 2017;Douville et al. 1999)

Fig. 11
Fig. 11 Dolomitization models for the Neogene dolostones in the South China Sea

Fig. 12
Fig. 12 Relationship between the dolomitization events in the South China Sea and global dolomitization events and paleoclimate and paleoenvironment.(Modified from Wilson 2008)

Table 1
Comparison of the upper miocene dolomite and pliocenelower pleistocene dolomite

Table 2
Distribution of the neogene dolostones