Summer monsoon–induced upwelling dominated coastal sea surface temperature variations in the northern South China Sea over the last two millennia

Abstract

The South China Sea (SCS), situated to the north of the Indo-Pacific Warm Pool (IPWP), is under the strong influence of the Asian monsoon system. However, coastal sea surface temperature (SST) records from the SCS, which are of vital importance to exploring ocean-atmosphere-land interactions behind the Asian monsoon system, remain scarce. Here, we use a sediment core collected at the coast of northern SCS to investigate alkenone-SST variations over the past two millennia. On multi-centennial timescale, SST changes in our record exhibit an opposite pattern to that of Northern Hemisphere temperature and solar irradiance, for example, relatively cool SST during the Medieval Warm Period (MWP) and warm conditions during the Little Ice Age (LIA). Together with alkenone content and existing records, we suggest that the regional SST changes result from a strengthening (weakening) of wind-driven coastal upwelling, associated with variability of the Asian summer monsoon intensity during the MWP (LIA).

Keywords

alkenone Asian summer monsoon coastal upwelling late-Holocene South China Sea

Introduction

The Indo-Pacific Warm Pool (IPWP), usually characterized by sea surface temperature (SST) exceeding 28°C and also called the ‘heat engine’ of the Earth, is fundamental to atmospheric circulation and thus the hydrological cycle (e.g. Fasullo and Webster, 1999; Neale and Slingo, 2003). To the north of the IPWP, the South China Sea (SCS) beneath the rising branch of the Pacific Walker circulation (Figure 1) plays an important role in ocean-atmosphere-land processes and is known to supply, for example, a large fraction of moisture carried by the Asian summer monsoon (ASM) (e.g. Guo et al., 2017; Wang et al., 2004; Zhang et al., 2017). In turn, the SCS oceanic conditions, variations of which strongly interact with the ASM and Asian winter monsoon (AWM) (e.g. Chu and Wang, 2003; Lau and Nath, 2009; Liu et al., 2004; Liu and Zhu, 2016; Wang et al., 2003; Xie et al., 2003), have attracted broad attention from multiple scientific disciplines over the past decades (e.g. Tao and Chen, 1987; Wang et al., 2009; Xie et al., 1998). For example, both the ASM and AWM, although governed by the southwesterly and northeasterly winds, respectively, induce upwelling of subsurface waters and enhance vertical mixing over the SCS (e.g. Jing et al., 2009; Ning et al., 2004; Su, 2004; Yuan et al., 2011). As such, winter SSTs along the northern SCS coasts could cool down to ~20°C, while in summer season, upwelling-related cooling is identified in mosaic regions along the coasts (e.g. Jing et al., 2009) (Figure 1). Moreover, the winds prevailing alongside the ASM and AWM also drive clockwise and anti-clockwise surface water circulations in the northern SCS coasts, respectively (e.g. Caruso et al., 2006; Chen et al., 2016; Su, 2004; Wang et al., 1995).

Figure 1.

Regional environment. (a) averaged July sea surface temperature (SST) from the World Ocean Atlas 2013 (Zweng et al., 2013), (b) site of our core HKGS-A (red dot) and other published records from the Mirs Bay and South China Sea (black dots) as mentioned in the text, and (c) comparison between $U_{37}^{K^{'}}$ -SST value of topmost sample from core HKGS-A (blue) and long-term (1991–2016 AD) monthly/annual mean in-situ SST (black dots/black dashed line) and sea surface salinity (SSS) (gray dots/gray dashed line) from four monitoring stations (data are available at www.epd.gov.hk).

Abundant datasets including instrumental and satellite-based records have been compiled to investigate the evolution of the Asian monsoon system (Wang et al., 2017). However, detailed knowledge about the ASM and/or AWM, as triggered by the land-ocean thermal contrast, also requires a thorough understanding of the contribution of oceanic counterpart (e.g. Thompson et al., 2000; Wang et al., 2005; Yancheva et al., 2007). Recent studies, based on analyses of observational data and numerical simulations, indicate the primary control of anthropogenic-induced warming on SST signal across the IPWP and SCS since ~1950s (Rao et al., 2012; Ratna et al., 2016; Weller et al., 2016). Geochemical proxies from marine archives offer an excellent opportunity to differentiate human-forced/natural variability of oceanic conditions and associated impacts over the globe (e.g. Abram et al., 2016; McGregor et al., 2015). For the SCS, despite a large number of studies presenting SST reconstructions from its open ocean (e.g. Kong et al., 2017; Lin et al., 2006; Oppo and Sun, 2005; Shintani et al., 2008; Steinke et al., 2011; Wang et al., 1999; Wei et al., 2007; Zhou et al., 2012), very few records have been achieved in the coastal region. A gravity core HKGS-A collected from the southeastern waters off Hong Kong suitably fills this gap, allowing an in-depth examination of ocean-atmosphere-land interactions along with the ASM/AWM changes.

Long-chain alkenones in marine sediments are exclusively produced by haptophytes (e.g. Marlowe et al., 1984; Volkman et al., 1980) and broadly used to reconstruct SST changes by its unsaturation index ( $U_{37}^{K^{'}}$ ) (e.g. Brassell et al., 1986; Prahl et al., 1988). In this study, we generate a $U_{37}^{K^{'}}$ -SST record from sediment core HKGS-A, which presents a relatively high-resolution SST signal in the northern SCS over the last two millennia. Combined with alkenone concentration in the same core and previously published SST records, our study also provides further insight into the spatial characteristics of interaction between the SCS and Asian monsoon system against well-identified climate anomalies. Our results show that the strength of wind-driven coastal upwelling, probably associated with the ASM intensity due to corresponding climate background, dominantly controlled past SST variations in the southeastward offshore region of Hong Kong.

Materials and methods

Study region

A 60.2-m-long gravity core HKGS-A (22°09′09′′N, 114°26′55′′E, 30 m water depth) was retrieved from the southeastern waters off Hong Kong via a Hong Kong Geological Survey Project (Figure 1). In-situ measurements of four monitoring stations surrounding our core site (data are available at www.epd.gov.hk) show prominent seasonal variations of both SST (~27.5°C in summer season (June-July-August, JJA) and ~18.5°C in winter season (December-January-February, DJF)) and sea surface salinity (SSS) (~29.5 psu in JJA and ~32.5 psu in DJF) (Figure 1c). At this site, seasonal SSS variations demonstrate the influence of the Pearl River discharge in summer, which, notably, exerts little impact on SST changes (e.g. Chen et al., 2017; Jiang and Wang, 2018; Jing et al., 2009; Kong et al., 2015). Furthermore, investigations of three nearby cores in the Mirs Bay (Kong et al., 2015) have indicated that regional SST variations over the past four centuries, as mainly controlled by changes in coastal upwelling, are able to reflect the ASM intensity (Kuang et al., 2011; Liu et al., 2013).

Age model

Lithological and seismic profiles suggest that the upper ~10 m of this core is composed of continuously deposited soft sediments which contains bioturbation and shells (and shell fragments), roughly covering the Holocene as constrained by using three radiocarbon (¹⁴C) dates (Fyfe et al., 1999; Owen et al., 1998). In this study, we present the results of upper ~3.0 m from core HKGS-A down which alkenone analysis was implemented (see Section 2.3). To develop the chronology, we performed additional ¹⁴C measurements (at the Beta Analytic Inc., USA) of other seven samples based on shells (and shell fragments) (Table 1 and Figure 2). These 10 ¹⁴C dates were calibrated into calendar ages using the Marine13 curve (Reimer et al., 2013) in the Calib7.1 software (Stuiver and Reimer, 1993), with consideration of a regional correction in reservoir age (ΔR = 200 ± 50 (error) years) (e.g. Southon et al., 2002). Ages at discrete depths were then calculated via polynomial regression (with degree of 2) of all calibrated ¹⁴C dates by fixing the core-top to 1993 AD when our core was drilled. To corroborate the age model, we also operated the BACON software (Blaauw and Christen, 2011) with default parameters (Figure 2). An ensemble of 10,000 age–depth realizations incorporated into this approach allows us to include all six ¹⁴C ages within our target interval (upper ~3.0 m) to calculate mean ages and associated 2σ uncertainty. The results are in good agreement with those derived from polynominal regression, substantiating robustness of our age model. However, it is worth noting that the ²¹⁰Pb/¹³⁷Cs analyses were not conducted due to insufficient sediment materials from the top section (actually dried up and shrunk already).

Table 1.

Raw ¹⁴C conventional dates and calibrated ages of upper 11 m in sediment core HKGS-A.

Depth (cm)	Material	Lab. No^a	¹⁴C AMS ages ± 1σ error (yr BP)	Calibrated ages ± 2σ error (yr BP)
60	Shell	Beta − 363309	940 ± 30	695 ± 116
175	Shell	Beta − 363310	1320 ± 30	1076 ± 137
226	Shell	Beta − 363311	1350 ± 30	1108 ± 136
254	Shell	Beta − 451084	2030 ± 30	1832 ± 147
287	Shell fragments	Beta − 451083	2390 ± 30	2260 ± 154
300	Shells	OxA-5400	2025 ± 65	1825 ± 208
477	Shells	OxA-5401	3575 ± 60	3715 ± 208
813	Shells	OxA-5402	4870 ± 75	5423 ± 194
921	Shell	Beta − 363312	5520 ± 30	6130 ± 142
1102	Shell	Beta − 363313	8550 ± 40	9420 ± 126

OxA denotes published ¹⁴C dates (Fyfe et al., 1999) with recalibrated ages (Section 2.2), and Beta represents new data from this study.

Figure 2.

Age models of sediment core HKGS-A based on polynominal regression approach (red line) and BACON algorithm (black line) of all calibrated ¹⁴ C ages (black dots). Gray lines represent the 2σ uncertainty derived from the BACON output (see Section 2.2 for details).

Alkenone analysis

We analyzed alkenones following the procedure as described in Kong et al. (2014). In general, the sediment core was sampled continuously with a ~5 cm step for upper 1 m, and a ~2 cm step between 1 m and 3 m. Then, samples of bulk sediment were dried (~5 g), powdered and soaked to extract total lipids using organic solvent of dichloromethane (DCM): methanol (MeOH) 9:1 solution, under ultrasonic waves in 40°C-water bath for three cycles (lasting 30 min each). The extract was hydrolyzed with 6% KOH in MeOH to remove alkenoates, and subsequently separated into three fractions using silica gel column chromatography with eluents of n-hexane, DCM and MeOH. Finally, alkenones were analyzed on the Agilent 7890 Gas Chromatography equipped with a flame ionization detector, at the University of Hong Kong. An internal standard (n-C₃₆ alkane) was used to quantify alkenones, and replicate injections were also performed every 10 samples.

The index $U_{37}^{K^{'}}$ was calculated according to the equation by Prahl et al. (1988): $U_{37}^{K^{'}}$ = C_37:2/(C_37:2 + C_37:3) where C_37:2 and C_37:3 are concentrations of di- and tri-unsaturated C₃₇ alkenones, respectively. Afterwards, mean annual SST was estimated using the regional calibration equation: $U_{37}^{K^{'}}$ = 0.031×SST + 0.092 (Pelejero and Grimalt, 1997), which was developed specifically for the SCS. Analytical uncertainties are ~0.005 unit for ratio-based $U_{37}^{K^{'}}$ index (which is equivalent to about 0.2–0.3°C for calculated SST values) and 5% for C₃₇ concentration, respectively.

Results

The $U_{37}^{K^{'}}$ -based SST values vary from 26.0°C to 27.6°C over the last ~1900 years. If four discrete lower SSTs are omitted (as depicted by the scattered points in Figure 3a), our SST record fluctuates between 26.8°C and 27.6°C, with relatively lower values during ~100–400 AD, ~900–1200 AD and after ~1850 AD, but higher values during ~400–900 AD and ~1300–1800 AD. Mean SST values for these intervals are 27.12 ± 0.12°C, 27.11 ± 0.15°C, 26.95 ± 0.11°C, 27.38 ± 0.15°C and 27.21 ± 0.09°C, respectively. Despite a narrow range of changes (slightly < 1°C), our SST record displays a general pattern opposite to Northern Hemisphere temperature (Moberg et al., 2005) and total solar irradiance (TSI) changes (Steinhilber et al., 2009) (Figure 3a), within the chronological uncertainty (Figure 2). For the last two centuries, around 0.5°C cooling (instead of warming) at our site is also less than the 1–2°C cooling as identified further north in the Mirs Bay (Kong et al., 2015) (Figure 4b). Variations of the C₃₇ contents, between 60 and 160 ng/g over the investigated period, are characterized by relatively high C₃₇ contents with cool SSTs, and vice versa (Figure 3b). For instance, C₃₇ content was lower during ~400–900 AD and ~1300–1800 AD (including the Little Ice Age, hereafter ‘LIA’, ~1400–1800 AD), relative to higher values during ~100–400 AD and ~900–1200 AD (within the Medieval Warm Period, hereafter ‘MWP’, ~800–1400 AD). In addition, both SST values and C₃₇ contents after ~1850 AD are roughly comparable with their values during the MWP, respectively.

Figure 3.

The $U_{37}^{K^{'}}$ -based SST record (a) and C₃₇ concentration (b) for sediment core HKGS-A. Total solar irradiance (TSI) (Steinhilber et al., 2009) is superimposed on the SST for comparison (the red arrow at left side marks increased SST downward). Color bars highlight the intervals discussed in the text, for example, the Medieval Warm Period (MWP, ca. 800–1400 AD) (red) and Little Ice Age (LIA, ca. 1400–1800 AD) (green). Please note that scattered blue dots denote discrete SST values.

Figure 4.

The $U_{37}^{K^{'}}$ -SST records of sediment cores HKGS-A (a) and NS02G (c), as well as (b) published SST records from the Mirs Bay (T3, T6 and M10) since 1750 AD (Kong et al., 2015) and (d) magnetic susceptibility at the Huguang Maar Lake (Yancheva et al., 2007) as an indicator of Asian winter monsoon (AWM). The stalagmite δ¹⁸O record from the Jhumar-Wah Shikar Cave (Sinha et al., 2011), an indicator of Asian summer monsoon (ASM), is also superimposed on SST record for comparison (a). Color bars represent the intervals as depicted in Figure 3.

Discussion

The multi-centennial-scale SST changes in sediment core HKGS-A are apparently anti-phased with Northern Hemisphere temperature (Moberg et al., 2005), for example, marked cool conditions for the MWP and relatively warm SSTs for the LIA (Figure 3a). In particular, a declining SST trend since ~1850 AD, corroborated by previously published SST records from nearby short sediment cores (T3, T6 and M10) in the Mirs Bay (Kong et al., 2015) (Figure 4b), strongly conflicts with the recent warming background. Furthermore, combined with the opposite patterns between our SST record and TSI (Steinhilber et al., 2009) (Figure 3a), it is very likely that regional features overturned global/Northern Hemisphere temperature changes. Identifying this underlying mechanism would help to clarify the role of how the SCS interacted with the Asian monsoon system. To this end, we first need to address the chronological uncertainty and alkenone-SST signal.

Chronological uncertainty

To calibrate conventional ¹⁴C dates, we use a regional correction of radiocarbon reservoir age (ΔR = 200 ± 50 years, Kong et al., 2014; Southon et al., 2002) that coincides well with the estimation of conventional age at core-top (ΔR = 213 years) via polynomial regression of all these conventional ¹⁴C dates together (Figure 2). Both the polynomial regression and BACON methods yield rather consistent age models, except for the lower part, for example, about 200 years difference at 350 AD (254 cm depth). However, there is only slight discrepancy (less than 50 years) for the age at about 900 AD which was outlined as the MWP interval. The dating uncertainty thus does not modify the time window of the MWP. The LIA time window indeed contains uncertainty of about 150 years and hence allows a possible shift of ages (in particular at 60 cm), which, notably, still maintains the general pattern of our SST record. Furthermore, the sedimentation rate of its upper 60 cm, as roughly estimated by ¹⁴C date in Table 1, is also comparable with that of several nearby cores derived from the ²¹⁰Pb results (e.g. Kong et al., 2015; Owen and Lee, 2004).

More importantly, the cooling trend over the last two centuries observed at our study site, which is corroborated by previously published SST records from nearby sediment cores in the Mirs Bay (Kong et al., 2015) and unlikely influenced by chronological uncertainty, should be a robust regional feature. If the cooler SSTs in the MWP were shifted toward the LIA (the typical cooling interval), one could then see warmer SSTs during the MWP (Figure 3a). However, any reasonable modifications of our current chronology, like applying different ΔR values, could not accommodate this shift. Furthermore, such ‘warmer’ conditions during the MWP at our study site, as compared to cool SSTs after ~1850 AD, would also require a completely different mechanism to account for. It hence enables us to consider the above-mentioned shift of cooler SSTs at MWP toward LIA very unlikely. Altogether, our chronology in this study, despite its uncertainty, is able to capture the general pattern of SST changes during the typical MWP and LIA, and should not affect our conclusion.

Alkenone-SST signal

At (or near) surface waters of northern SCS open ocean, the species of coccolithophore Gephyrocapsa oceanica and Emiliania huxleyi, producing alkenones, dominantly bloom in the winter season when AWM prevails (e.g. Chen et al., 2007). In contrast, a set of modern surveys show that, at the Pearl River estuary, phytoplankton and haptophyte are primarily abundant in the summer season (Dai et al., 2008; Le et al., 2008; Song et al., 2010). Moreover, $U_{37}^{K^{'}}$ -SST values from a transect of core-top sediments in the inner shelf off the Pearl River estuary document the euphotic zone temperature in spring and summer (Zhang et al., 2013). Indeed, the near-surface sample in sediment core HKGS-A has an alkenone-SST value of 27°C, which is slightly lower than long-term composited summer (JJA) SST (~27.5°C) but clearly higher than winter (DJF) SST (~18.5°C) as observed from in-situ measurement nearby (Figure 1c). The $U_{37}^{K^{'}}$ -SST values, although converted into annual mean temperatures, are considered to mainly represent summer temperature due to a large fraction of weighted alkenone production within this season. In addition, very good agreement of mean values among our $U_{37}^{K^{'}}$ -SST record (27.2°C for the investigated period) and previously published SST records in the northern SCS (Figure 1), such as sediment cores 17940 (26.98°C) (Wang et al., 1999) and NS02G (27.04°C) (Kong et al., 2017), also reinforce the validity of this interpretation. Yet, seasonal dynamics of alkenones production in the coastal regions still need to be further investigated.

It is necessary to evaluate whether the alkenone signal at our study site can be influenced by lateral transport, for instance, from the Taiwan Strait or Pearl River discharge. We note that the AWM northeasterly winds, which drive anti-clockwise currents in the northern SCS, are able to induce strong vertical mixing and bring subsurface nutrients to the euphotic zone (e.g. Ning et al., 2004; Su, 2004). However, this mechanism gives rise to thermal homogenization of the water column over our study area in winter (from ~15 to 20°C, Kong et al., 2015), thus being difficult to reconcile with the range of our $U_{37}^{K^{'}}$ -based SST values (Figures 1 and 3). Also, weakened winter coastal currents, as expected when the AWM intensity was reduced during the MWP (e.g. Yancheva et al., 2007; Figure 4), are hard to induce the cooler SSTs at our study site correspondingly. On the other hand, the cold winter coastal currents can transport large amounts of detrital material/sediment from the Taiwan Strait into the northern SCS (Chen, 1978; Tamburini et al., 2003; Wang et al., 1999), which, as a result, dilutes alkenone content at our site and causes relatively low alkenone contents during ~400–900 AD and the LIA. However, no considerable change in sedimentation rate is identified at these two intervals (Figure 2). Overall, we thus consider that the lateral transport of coastal currents, driven by the AWM, does not play a dominant role in controlling the alkenone signal.

Another important point is that the Pearl River discharge, although exerting little impact on our SST changes (e.g. Chen et al., 2017; Jiang and Wang, 2018; Kong et al., 2015), delivers a large amount of fluvial nutrients into the northern SCS coasts in summer (e.g. Dai et al., 2008; Le et al., 2008; Song et al., 2010). At the same time, prevailing clockwise surface currents are also able to transport such nutrient-enriched freshwater to our site and hence allow algal blooms and maximal primary productivity (Le et al., 2008). It is likely that both current-carried fluvial nutrients and upwelling-induced subsurface nutrients are favorable for alkenone production at our site. However, two alkenone records from cores HKUV16 (22°17.5′N, 113°52′E) and NS01 C (21°51′N, 113°49′E), located at the Pearl River estuary (Kong, 2014; Kong et al., 2014), bear little similarity with C₃₇ record from our core. For those two cores, warmer SSTs and no increase in C₃₇ contents during the MWP, in contrast to the cooler SSTs and increased alkenone contents at our site, definitely rule out a major control of the Pearl River discharge on the alkenone signal in core HKGS-A.

Summer monsoon-induced upwelling

Our topmost SST value is close to summer temperature, but there is still a possibility that relative SST changes might be largely ascribed to winter temperature changes. Such case has been reported previously across the Yellow Sea and Bohai Bay (He et al., 2014). However, at our site, the opposite SST changes (Figure 3) cannot be explained by the AWM intensity since the AWM strengthened at cool intervals and would only have reinforced the global temperature signal (Figure 4). Regarding substantially cool SSTs in winter (~20°C) in our study region, winter temperature signal could essentially contribute to regional SST change. However, it is other factors that ultimately overprint the winter temperature signal. By comparing with global temperature, the AWM and ASM changes over the past two millennia (Figures 3 and 4), summer monsoon-induced upwelling stands out as a viable driver (perhaps the only one) of the opposite SST changes at our site. In summer, the prevailing southwesterly winds from the ASM circulation blow over the northern SCS. A stronger ASM leads to intensification of wind-induced upwelling, which, subsequently, decreases the coastal SSTs of northern SCS (e.g. Kong et al., 2015; Kuang et al., 2011; Liu et al., 2013). Moreover, an enhanced upwelling also promotes vertical mixing of nutrient-rich subsurface waters that allows favorable conditions for haptophyte production (Villanueva et al., 1998) and would thereby explain higher C₃₇ concentrations (Figure 3b). By contrast, a weaker ASM results in increased coastal SSTs and lower C₃₇ concentrations (Figure 3). Although C₃₇ content changes could be attributed to multiple factors, summer monsoon-induced upwelling remains as the most likely explanation in our case (see Section 4.2). Thus, coupled variations of our SST record and alkenone content effectively reflect the ASM intensity.

In fact, additional analysis of observational data shows negative correlation between in-situ measured summer SSTs nearby our core site and the 850 hPa meridional winds over central India and adjacent ocean (Figure 5), which apparently facilitates usage of $U_{37}^{K^{'}}$ -SST record in this region to infer the ASM intensity. On this basis, our SST and C₃₇ records together support an intensification of the ASM during the MWP, and a reduction during the LIA. Such explanation is substantiated by comparing our SST and existing stalagmite δ¹⁸O record from the Jhumar-Wah Shikar Cave (Figure 4a), a powerful indicator of the ASM from central India (Sinha et al., 2011). Accordingly, those four scattered lower SST values also imply extremely enhanced upwelling at the corresponding episodes (Figure 3a). Under the control of the ASM within the same season, SST signals across the SCS display spatial heterogeneity. For instance, a newly published SST record from sediment core NS02G (19.8°N, 113.9°E), about 200 km to the south of our core site, resembles the pattern of Northern Hemisphere temperature over the past two millennia (Figure 4c; Kong et al., 2017). Coexistence of SST variations between the Mirs Bay and SCS open ocean suggests that the upwelling system driven by the ASM mainly occurs at the coastal regions (e.g. Jing et al., 2009; Kuang et al., 2011; Liu et al., 2013).

Figure 5.

Spatial pattern of correlation coefficients between averaged (June-July) 850 hPa meridional winds (NASA MERRA) and in-situ SST measurements nearby site HKGS-A during 1991–2016 AD (colorful rectangle marks regions where p < 10%). The negative relationship, although existing at our site (red dot) as expected, is more significant over central India and adjacent ocean (black dot marks location of stalagmite δ¹⁸O record in Sinha et al., 2011).

Modern El Niño–Southern Oscillation (ENSO) variability also modulates SST changes across the SCS (e.g. Yang et al., 2015; Zhou and Chan., 2007), for example, ~0.1°C as evidenced by observational data (not shown). However, we are unable to quantify the ENSO contribution to SST changes at our site, because available paleorecords support prevailing conditions of either La Niña (Yan et al., 2011) or El Niño (Rustic et al., 2015) over the same interval, such as the LIA. Additional works by model simulations may shed further insight on this topic, which is beyond the scope of this study. Nevertheless, we stress a specific point that the Pearl River discharge would suppress the wind-driven coastal upwelling to the southwest of Hong Kong (e.g. Chen, 2014; Chen et al., 2017; Jing et al., 2009; Wang et al., 2012; Zu and Gan, 2015). Finally, an enhanced upwelling since ~1850 AD calls for more attention to future ASM change under global warming scenarios, and its possible impact in the coastal region.

Conclusion

We have reconstructed alkenone-SST variations over the last two millennia from sediment core HKGS-A, collected from the southeast waters of Hong Kong. Multi-centennial-scale SST variations display an opposite pattern to that of Northern Hemisphere temperature, for example, relatively cool conditions during the MWP and warm SSTs during the LIA. Together with alkenone content changes, our SST record indicates that the strength of wind-driven coastal upwelling in the northern SCS, associated with summer monsoon intensity, could overturn the original signal of global temperature changes. Comparison with published records also allows further insight on the spatial characteristics of SST changes across the northern SCS. In the context of global warming scenarios, future ASM change and the associated impact upon the coastal upwelling are particularly timely topics.

Footnotes

Acknowledgements

We thank two anonymous referees for constructive comments that greatly improved this manuscript. RJS also appreciates the permission from Head of Geotechnical Engineering Office and Director of Civil Engineering and Development Department.

Funding

This work was supported by the National Key Research and Development Program of China (2016YFA0601204) and HK Research Grants Council (17325516 and 707612 P).

ORCID iD

Yancheng Zhang

References

Abram

McGregor

Tierney

et al . (2016) Early onset of industrial-era warming across the oceans and continents. Nature 536: 411–418.

Blaauw

Christen

(2011) Flexible paleoclimate age-depth models using an autoregressive Gamma Process. Bayesian Analysis 6: 457–474.

Brassell

Eglinton

Marlowe

et al . (1986) Molecular stratigraphy: A new tool for climatic assessment. Nature 320: 129–133.

Caruso

Gawarkiewicz

Beardsley

(2006) Interannual variability of the Kuroshio intrusion in the South China Sea. Journal of Oceanography 62: 559–575.

Chen

et al . (2016) Provenance discrimination of the clay sediment in the western Taiwan Strait and its implication for coastal current variability during the late-Holocene. The Holocene 27: 110–121.

Chen

P-Y

(1978) Minerals in bottom sediments of the South China Sea. GSA Bulletin 89: 211–222.

Chen

Chung

(2007) Seasonal variability of coccolithophore abundance and assemblage in the northern South China Sea. Deep-Sea Research Part II: Topical Studies in Oceanography 54: 1617–1633.

Chen

(2014) Coastal upwelling study: Observations, dynamic analysis and modelling. PhD Thesis, University of Delaware.

Chen

Jiang

Liu

et al . (2017) Development of upwelling on pathway and freshwater transport of Pearl River plume in northeastern South China Sea. Journal of Geophysical Research 122: 6090–6109.

10.

Chu

Wang

(2003) Seasonal variability of thermohaline front in the central South China Sea. Journal of Oceanography 59: 65–78.

11.

Dai

Zhai

Cai

W-J

et al . (2008) Effects of an estuarine plume-associated bloom on the carbonate system in the lower reaches of the Pearl River estuary and the coastal zone of the northern South China Sea. Continental Shelf Research 28: 1416–1423.

12.

Fasullo

Webster

(1999) Warm pool SST variability in relation to the surface energy balance. Journal of Climate 12: 1292–1305.

13.

Fyfe

Selby

Plater

et al . (1999) Erosion and sedimentation associated with the last sea level rise offshore Hong Kong, South China Sea. Quaternary International 55: 93–100.

14.

Guo

Klingaman

Vidale

et al . (2017) Contribution of tropical cyclones to atmospheric moisture transport and rainfall over East Asia. Journal of Climate 30: 3853–3865.

15.

Zhou

Liu

et al . (2014) Weakened Yellow Sea Warm Current over the last 2–3 centuries. Quaternary International 349: 252–256.

16.

Jiang

Wang

(2018) Modeling the ecosystem response to summer coastal upwelling in the northern South China Sea. Oceanologia 60(1): 32–51.

17.

Jing

Hua

et al . (2009) Numerical study on the summer upwelling system in the northern continental shelf of the South China Sea. Continental Shelf Research 29: 467–478.

18.

Kong

(2014) Climatic changes in the northern South China Sea since the last glacial maximum. PhD Thesis, University of Hong Kong. DOI: 10.5353/th_b5312314.

19.

Kong

Tse

Y-Y

Jia

et al . (2015) Cooling trend over the past 4 centuries in northeastern Hong Kong waters as revealed by alkenone-derived SST records. Journal of Asian Earth Sciences 114: 497–503.

20.

Kong

Wei

Chen

et al . (2017) Northern South China Sea SST changes over the last two millennia and possible linkage with solar irradiance. Quaternary International 459: 29–34.

21.

Kong

Zong

Jia

et al . (2014) The development of late-Holocene coastal cooling in the northern South China Sea. Quaternary International 349: 300–307.

22.

Kuang

Lee

JHW

Harrison

et al . (2011) Effect of wind speed and direction on summer tidal circulation and vertical mixing in Hong Kong waters. Journal of Coastal Research 27: 74–86.

23.

Lau

N-C

Nath

(2009) A model investigation of the role of air-sea interaction in the climatological evolution and ENSO-related variability of the summer monsoon over the South China Sea and Western North Pacific. Journal of Climate 22: 4771–4792.

24.

Ning

Liu

et al . (2008) Standing stock and production of phytoplankton in the northern South China Sea during winter of 2006. Acta Ecologica Sinica 28: 5775–5784 (in Chinese).

25.

Lin

Liu

Fang

et al . (2006) Millennial-scale changes in terrestrial sediment input and Holocene surface hydrography in the northern South China Sea (IMAGES MD972146). Palaeogeography, Palaeoclimatology, Palaeoecology 236: 56–73.

26.

Liu

Zhu

(2016) A possible precursor of the South China Sea summer monsoon onset: Effect of the South Asian High. Geophysical Research Letters 43: 11072–11079.

27.

Liu

Jiang

Xie

S-P

et al . (2004) A gap in the Indo-Pacific warm pool over the South China Sea in boreal winter. Journal of Geophysical Research 109: C07012.

28.

Liu

Peng

Shen

C-C

et al . (2013) Recent 121-year variability of western boundary upwelling in the northern South China Sea. Geophysical Research Letters 40: 3180–3183.

29.

McGregor

Evans

Goosse

et al . (2015) Robust global ocean cooling trend for the preindustrial Common Era. Nature Geoscience 8: 671–677.

30.

Marlowe

Green

Neal

et al . (1984) Long-chain (n-C37-C39) alkenones in the Prymnesiophyceae: Distribution of alkenones and other lipids and their taxonomic significance. British Phycological Journal 19: 203–216.

31.

Moberg

Sonechkin

Holmgren

et al . (2005) Highly variable Northern Hemisphere temperatures reconstructed from low- and high-resolution proxy data. Nature 433: 613–617.

32.

Neale

Slingo

(2003) The Maritime Continent and its role in the global climate: A GCM study. Journal of Climate 16: 834–848.

33.

Ning

Chai

Xue

et al . (2004) Physical–biological oceanographic coupling influencing phytoplankton and primary production in the South China Sea. Journal of Geophysical Research 109: C10005.

34.

Oppo

Sun

(2005) Amplitude and timing of sea-surface temperature change in the northern South China Sea: Dynamic link to the East Asian monsoon. Geology 33: 785–788.

35.

Owen

Neller

Shaw

et al . (1998) Late Quaternary environmental changes in Hong Kong. Palaeogeography, Palaeoclimatology, Palaeoecology 138(1–4): 151–173.

36.

Owen

Lee

(2004) Human impacts on organic matter sedimentation in a proximal shelf setting, Hong Kong. Continental Shelf Research 24: 583–602.

37.

Pelejero

Grimalt

(1997) The correlation between the

U_{37}^{K^{'}}

index and sea surface temperatures in the warm boundary: The South China Sea. Geochimica et Cosmochimica Acta 61: 4789–4797.

38.

Prahl

Muehlhausen

Zahnle

(1988) Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions. Geochimica et Cosmochimica Acta 52: 2303–2310.

39.

Rao

Dhakate

Saha

et al . (2012) Why is Indian Ocean warming consistently? Climatic Change 110: 709–719.

40.

Ratna

Cherchi

Joseph

et al . (2016) Moisture variability over the Indo-Pacific region and its influence on the Indian summer monsoon rainfall. Climate Dynamics 46: 949–965.

41.

Reimer

Bard

Bayliss

et al . (2013) IntCal13 and Marine13 radiocarbon age calibration curves 0–50,000 years cal BP. Radiocarbon 55: 1869–1887.

42.

Rustic

Koutavas

Marchitto

et al . (2015) Dynamical excitation of the tropical Pacific Ocean and ENSO variability by Little Ice Age cooling. Science 350: 1537–1541.

43.

Shintani

Yamamoto

Chen

M-T

(2008) Slow warming of the northern South China Sea during the last deglaciation. Terrestrial, Atmospheric and Oceanic Sciences 19: 341–346.

44.

Sinha

Berkelhammer

Stott

et al . (2011) The leading mode of Indian Summer Monsoon precipitation variability during the last millennium. Geophysical Research Letters 38: L15703.

45.

Song

Liu

Huang

et al . (2010) Distribution characteristics of basic biological production and its influencing factors in the northern South China Sea in summer. Acta Ecologica Sinica 30: 6409–6417 (in Chinese).

46.

Southon

Kashgarian

Fontugne

et al . (2002) Marine reservoir corrections for the Indian Ocean and Southeast Asia. Radiocarbon 44: 167–180.

47.

Steinhilber

Beer

Fröhlich

(2009) Total solar irradiance during the Holocene. Geophysical Research Letters 36: L19704.

48.

Steinke

Glatz

Mohtadi

et al . (2011) Past dynamics of the East Asian monsoon: No inverse behaviour between the summer and winter monsoon during the Holocene. Global and Planetary Change 78: 170–177.

49.

Stuiver

Reimer

(1993) Extended ¹⁴C database and revised CALIB radiocarbon calibration program. Radiocarbon 35: 215–230.

50.

(2004) Overview of the South China Sea circulation and its influence on the coastal physical oceanography outside the Pearl River Estuary. Continental Shelf Research 24: 1745–1760.

51.

Tamburini

Adatte

Föllmi

et al . (2003) Investigation the history of East Asian monsoon and climate during the last glacial-interglacial period (0–140,000 years): Mineralogy and geochemistry of ODP Sites 1143 and 1144, South China Sea. Marine Geology 201: 147–168.

52.

Tao

Chen

(1987) A review of recent research on the East Asian summer monsoon. In: Chang

C-P

Krishnamurti

(eds) Monsoon Meteorology. London: Oxford University Press, pp. 60–92.

53.

Thompson

Yao

Mosley-Thompson

et al . (2000) A high-resolution millennial record of the South Asian Monsoon from Himalayan ice cores. Science 289: 1916–1919.

54.

Villanueva

Grimalt

Labeyrie

et al . (1998) Precessional forcing of productivity in the north Atlantic ocean. Paleoceanography 13: 561–571.

55.

Volkman

Eglinton

Corner

EDS

et al . (1980) Long-chain alkenes and alkenones in the marine coccolithophorid Emiliania Huxleyi. Phytochemistry 19: 2619–2622.

56.

Wang

Clemens

Liu

(2003) Contrasting the Indian and East Asian monsoons: Implications on geologic timescales. Marine Geology 201: 5–21.

57.

Wang

Huang

et al . (2009) Multi-scale climate variability of the South China Sea monsoon: A review. Dynamics of Atmospheres and Oceans 47: 15–37.

58.

Wang

Lin

Zhang

et al . (2004) Definition of South China Sea Monsoon Onset and commencement of the East Asia Summer Monsoon. Journal of Climate 17: 699–710.

59.

Wang

Zhuang

Xie

et al . (2012) Coastal upwelling in summer 2000 in the northeastern South China Sea. Journal of Geophysical Research 117: C04009.

60.

Wang

Sarnthein

Erlenkeuser

et al . (1999) Holocene variations in Asian monsoon moisture: A bidecadal sediment record from the South China Sea. Geophysical Research Letters 26: 2889–2892.

61.

Wang

Cheng

et al . (2017) The global monsoon across time scales: Mechanisms and outstanding issues. Earth-Science Reviews 174: 84–121.

62.

Wang

Bian

et al . (1995) Late Quaternary paleoceanography of the South China Sea: Surface circulation and carbonate cycles. Marine Geology 127: 145–165.

63.

Wang

Cheng

Edwards

et al . (2005) The Holocene Asian monsoon: Links to solar changes and North Atlantic climate. Science 308: 854–847.

64.

Wei

Deng

Liu

et al . (2007) High-resolution sea surface temperature records derived from foraminiferal Mg/Ca ratios during the last 260 ka in the northern South China Sea. Palaeogeography, Palaeoclimatology, Palaeoecology 250: 126–138.

65.

Weller

Min

Cai

et al . (2016) Human-caused Indo-Pacific warm pool expansion. Science Advances 2: e1501719.

66.

Xie

Chung

Y-S

Liu

et al . (1998) The interannual variations of summer monsoon onset over the South China Sea. Theoretical and Applied Climatology 59: 201–213.

67.

Xie

Wang

et al . (2003) Summer upwelling in the South China Sea and its role in regional climate variations. Journal of Geophysical Research 108: 3261.

68.

Yan

Sun

Wang

et al . (2011) A record of the Southern Oscillation Index for the past 2,000 years from precipitation proxies. Nature Geoscience 4: 611–614.

69.

Yancheva

Nowaczyk

Mingram

et al . (2007) Influence of the intertropical convergence zone on the East Asian monsoon. Nature 445: 74–77.

70.

Yang

Xie

et al . (2015) Interdecadal difference of interannual variability characteristics of South China Sea SSTs associated with ENSO. Journal of Climate 28: 7145–7160.

71.

Yuan

Yin

Cai

et al . (2011) Influence of seasonal monsoons on net community production and CO2 in subtropical Hong Kong coastal waters. Biogeosciences 8: 289–300.

72.

Zhang

Tang

Chen

et al . (2017) Tracing changes in atmosphere moisture supply to the drying Southwest China. Atmospheric Chemistry and Physics 17: 10383–10393.

73.

Zhang

Bai

et al . (2013) Alkenone and tetraether lipids reflect different seasonal seawater temperatures in the coastal northern South China Sea. Organic Geochemistry 58: 115–120.

74.

Zhou

Zheng

Yang

et al . (2012) Climate and vegetation variations since the LGM recorded by biomarkers from a sediment core in the northern South China Sea. Journal of Quaternary Science 27: 948–955.

75.

Zhou

Chan

JCL

(2007) ENSO and the South China Sea summer monsoon onset. International Journal of Climatology 27: 157–167.

76.

Gan

(2015) A numerical study of coupled estuary–shelf circulation around the Pearl River Estuary during summer: Responses to variable winds, tides and river discharge. Deep-Sea Research Part II: Topical Studies in Oceanography 117: 53–64.

77.

Zweng

Reagan

Antonov

et al . (2013) World Ocean Atlas 2013, vol. 2: Salinity (ed. Levitus

Mishonov

). Silver Spring, MD: NOAA Atlas NESDIS 74.