ArticlePDF Available

Assessing the intercolony δ18O proxy calibration in a coral microatoll and its implication for ENSO reconstruction in the northern South China Sea

Authors:
  • South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou

Abstract and Figures

Massive Porites corals are widely used for paleoclimate reconstruction, however the potential of Porites coral microatoll δ¹⁸O as high-resolution archive of paleoclimate has only been recognized recently, and the systematic chemistry-climate relationship has yet to be well characterized in the western Pacific. In this study, we examined the reproducibility of microatoll δ¹⁸O records against two adjacent Porites corals from the Xisha Islands in the northern South China Sea (SCS) and evaluated the reliability of microatoll δ¹⁸O as a proxy for reconstructing regional climate and El Niño-Southern Oscillation (ENSO) variability. The seasonal to interannual variability in microatoll δ¹⁸O was primarily controlled by local sea surface temperature (SST), while the δ¹⁸O signal might be suppressed by sea surface salinity (SSS) variations on interannual timescale. Despite the overlapping coral δ¹⁸O records exhibited similar patterns of variability, the mean values were consistently offset by ~0.2‰ and the sensitivity of δ¹⁸O proxy to climate also varied across different coral colonies. The microatoll δ¹⁸O exhibited relatively high proxy-SST sensitivities and amplitude of the seasonal variabilities. These results suggested that intercolony δ¹⁸O variability was a significant source of uncertainty in coral-based paleoclimate reconstructions. Microatoll δ¹⁸O anomaly appeared to serve as a sensitive and relatively reliable proxy for ENSO variability, although the imprints of weak-to-moderate ENSO events could not be fully captured due to the complex relationship between the East Asian Monsoon and ENSO, as well as the local seawater salinity changes. This study further strengthened the evidence for microatoll as an alternate climate archive in the SCS and highlighted its potential in helping resolve poorly understood paleoclimate before instrumental observations.
Content may be subject to copyright.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
Available online 6 May 2022
0031-0182/© 2022 Published by Elsevier B.V.
Assessing the intercolony δ
18
O proxy calibration in a coral microatoll and
its implication for ENSO reconstruction in the northern South China Sea
Fei Tan
a
,
b
,
d
, Hongqiang Yang
a
,
b
,
c
,
*
, Xiyang Zhang
a
,
b
, Huilong Xu
a
,
b
, Qi Shi
a
,
b
, Shichen Tao
a
,
b
,
Hongqiang Yan
a
,
b
, Guan Wang
a
,
b
a
Key Laboratory of Ocean and Marginal Sea Geology, South China Sea Institute of Oceanology of Sciences, Chinese Academy of Sciences, Guangzhou 510301, China
b
Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou), 511458, China
c
Nansha Marine Ecological and Environmental Research Station, Chinese Academy of Sciences, Sansha 573199, China
d
University of Chinese Academy of Sciences, Beijing 100049, China
ARTICLE INFO
Edirtor: Dr. Howard Falcon-Lang
Keywords:
Porites coral microatoll
δ
18
O
ENSO
Intercolony offsets
South China Sea
ABSTRACT
Massive Porites corals are widely used for paleoclimate reconstruction, however the potential of Porites coral
microatoll δ
18
O as high-resolution archive of paleoclimate has only been recognized recently, and the systematic
chemistry-climate relationship has yet to be well characterized in the western Pacic. In this study, we examined
the reproducibility of microatoll δ
18
O records against two adjacent Porites corals from the Xisha Islands in the
northern South China Sea (SCS) and evaluated the reliability of microatoll δ
18
O as a proxy for reconstructing
regional climate and El Ni˜
no-Southern Oscillation (ENSO) variability. The seasonal to interannual variability in
microatoll δ
18
O was primarily controlled by local sea surface temperature (SST), while the δ
18
O signal might be
suppressed by sea surface salinity (SSS) variations on interannual timescale. Despite the overlapping coral δ
18
O
records exhibited similar patterns of variability, the mean values were consistently offset by ~0.2and the
sensitivity of δ
18
O proxy to climate also varied across different coral colonies. The microatoll δ
18
O exhibited
relatively high proxy-SST sensitivities and amplitude of the seasonal variabilities. These results suggested that
intercolony δ
18
O variability was a signicant source of uncertainty in coral-based paleoclimate reconstructions.
Microatoll δ
18
O anomaly appeared to serve as a sensitive and relatively reliable proxy for ENSO variability,
although the imprints of weak-to-moderate ENSO events could not be fully captured due to the complex rela-
tionship between the East Asian Monsoon and ENSO, as well as the local seawater salinity changes. This study
further strengthened the evidence for microatoll as an alternate climate archive in the SCS and highlighted its
potential in helping resolve poorly understood paleoclimate before instrumental observations.
1. Introduction
Paleoclimate archive reconstructions provide a method to extend
climate data before observation records and anthropogenic inuence
(Gagan et al., 2000). The geochemical signals preserved in the massive
coral skeletons are reliable proxies for reconstructing former ocean hy-
drology condition and climate change (Cobb et al., 2003). One of the
most extensively used geochemical tracers for coral-based climate re-
constructions is the ratios of oxygen isotopic (δ
18
O), a proxy that reects
combined variations in sea-surface temperature (SST) and seawater
oxygen isotopic ratio (δ
18
O
sw
) (Corr`
ege, 2006; Gagan et al., 1994).
However, the absolute intercolony δ
18
O shifts observed among Porites
spp. colonies in the same site may be as large as ~0.4(Dassi´
e et al.,
2014; Linsley et al., 2008; Felis et al., 2003), which therefore introduce
large uncertainties in coral-based climate reconstructions. It is still un-
clear whether these intercolony shifts differ by sites, or more likely,
represent the poor understanding of the intrinsic biomineralization
mechanisms in Porites spp. within the eld (Sayani et al., 2019). Mutual
intercolony calibrations and the replication of coral geochemical records
are crucial for improving the quality of coral climate proxies and
discriminating localized non-climatic inuences from regional climatic
signals (Sayani et al., 2019; Ramos et al., 2017; Dassi´
e et al., 2014).
Porites lutea is the most routinely used corals for paleoclimate
reconstruction, as it can provide monthly resolved records that can span
* Corresponding author at: Key Laboratory of Ocean and Marginal Sea Geology, South China Sea Institute of Oceanology of Sciences, Chinese Academy of Sciences,
Guangzhou 510301, China.
E-mail address: hqyang@scsio.ac.cn (H. Yang).
Contents lists available at ScienceDirect
Palaeogeography, Palaeoclimatology, Palaeoecology
journal homepage: www.elsevier.com/locate/palaeo
https://doi.org/10.1016/j.palaeo.2022.111031
Received 31 October 2021; Received in revised form 29 April 2022; Accepted 29 April 2022
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
2
several centuries (Corr`
ege, 2006). Porites coral is generally character-
ized by a domed-shaped growth form and can grow up to several meters
in height (Knutson et al., 1972). However, the domed structure is not the
only growth form for Porites corals. The at-topped discoid Porites
microatolls, common in the Indo-Pacic tropical region (Meltzner and
Woodroffe, 2015), provide an alternative proxy for high-resolution sea
level change (Majewski et al., 2018; Stoddart and Scofn, 1979) and
paleoclimate information (Mcgregor et al., 2013; Mcgregor et al., 2011;
Mcgregor and Gagan, 2004; Woodroffe et al., 2003; Woodroffe and
Gagan, 2000). Coral microatolls grow on shallow reef ats (12 m water
depths) and the upper at surface of the colony is constrained by the
airsea interface, which is usually related to the air exposure at Mean
Low Water Spring tide (Meltzner et al., 2010). Microatolls are sensitive
to changes in the relative sea level change, indicating that they may have
signicant potential for reconstructing climate variations, such as SST
and SSS (sea surface salinity), related to El Ni˜
noSouthern Oscillation
(ENSO) and ocean dynamic height (McGregor et al., 2011; Woodroffe
and McLean, 1990). The Sr/Ca-SST calibrations of two modern living
microatolls collected from the Xisha Islands in the northern South China
Sea (SCS) showed a strong correlation with the instrumental SST and
domed-shaped Porites records for the same period (Chen et al., 2018).
The modern microatoll skeletal δ
18
O in the Ni˜
no 3.4 region of the central
Pacic has shown comparable reproducibility and delity in climate
proxy ability to that of the conventional doom-shaped Porites corals.
Moreover, the microatoll δ
18
O record indicated predominantly ENSO-
induced SST and rainfall variability (McGregor et al., 2011). δ
18
O in
fossil Porites microatolls from Christmas Island in the central Pacic
provided high-resolution proxy records of Mid-late Holocene ENSO
variability as well (Mcgregor et al., 2013; Mcgregor and Gagan, 2004;
Woodroffe et al., 2003; Woodroffe and Gagan, 2000). Both living and
fossil microatolls appear suitable for extending paleoclimate recon-
struction, however, the systematic characteristics of the chemistry-
climate relationship between microatoll δ
18
O and Porites corals have
not been thoroughly investigated in the tropical western Pacic. Hence,
further examination and assessment of the reproducibility of SCS
microatoll δ
18
O composition and its reliability as a proxy for climate
reconstruction, such as ENSO, are required.
ENSO is a tropical oceanatmospheric phenomenon and leading
model of global interannual climate variability (Ropelewski and
Halpert, 1987). It oscillates irregularly every 28 years and signicantly
disrupts global rainfall and SST anomalies patterns (McPhaden et al.,
2006; Ropelewski and Halpert, 1987). The teleconnection regulations of
ENSO in the global ocean climate system show evident spatiotemporal
heterogeneity in terms of intensity, spatial scope, onset, duration, and
cessation (Freund et al., 2019). The SCS is located on the northern edge
of the Western Pacic Warm Pool (WPWP), and its interannual climate
change is primarily regulated by ENSO and associated climate systems
(Xie et al., 2016; Zhou and Chan, 2010; Wang et al., 2000). Skeletal δ
18
O
records from corals and Tridacna spp. from the northern SCS have
captured the imprints of ENSO variability on the frequency domain since
the mid-Holocene, conrming the teleconnection relationship between
SCS interannual climate and the ENSO (Jiang et al., 2021; Hu et al.,
2020; Han et al., 2020; Wang et al., 2018; Yan et al., 2017). Neverthe-
less, recent studies suggested that modern Porites coral δ
18
O in the Xisha
Islands may not be able to record entire the intensity of ENSO events
(Han et al., 2020; Wang et al., 2018). There is uncertainty concerning
whether the northern SCS corals can reliably reconstruct historical
ENSO variability.
In this study, we presented a 34-year, monthly resolved δ
18
O record
of a living Porites coral microatoll from the Xisha Islands in the northern
SCS. We explored the systematic relationship between the microatoll
δ
18
O signal and the instrumental SST and SSS information of the local
seawater on seasonal to interannual scales, and investigated the inter-
colony δ
18
O reproducibility and offsets relative to two adjacent con-
ventional P. lutea corals. Finally, the applicability and reliability of the
SCS microatoll δ
18
O signal as a proxy for reconstructing ENSO activity
were assessed.
2. Regional setting
The SCS is the largest semi-enclosed marginal sea of the Western
Pacic and is located at the center of the Asian-Australian monsoon
(Fig. 1). Its unique geographical location means it is synchronously
affected by both the Asian monsoon and ENSO climate systems (Wang
et al., 2009). The East Asian Monsoon (EAM) is the most important
feature of regional climatology and signicantly controls the formation
and variation of seasonal climatic circulation across the SCS and sur-
rounding areas (Wang et al., 2000). In boreal winters (DJF: December to
Fig. 1. (a) Location of the study site in the
Xisha Islands, northern SCS. (b) The sam-
pling site of microatoll GJ-1 (blue star) in
Guangjin Island. (c) P. lutea XS1 (yellow
star) (Wang et al., 2018) and YXN1 (red star)
(Han et al., 2020) were collected in Qilianyu
reef and Yongxin Island, respectively. (d)
The X-radiograph positive image of of Porites
coral microatoll GJ-1. Clear light and dark
growth line indicate the annual density
bands. The white lines indicate the main
coral growth axes and transects for monthly
coral δ
18
O sub-sampling. (For interpretation
of the references to colour in this gure
legend, the reader is referred to the web
version of this article.)
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
3
February), the dry and cold winter monsoon blows towards the southern
SCS and Southeast Asia, resulting in an increase in the SSS and colder
SST. In boreal summer (JJA, June to August), the prevailing summer
monsoon brings warm surface water and vapor from the low-latitude
tropical ocean, causing a reduction in SSS and warmer SST in the
northern SCS. Typically, the highest monthly mean SST in the Xisha
Islands occurs in June and the lowest in January. The monthly mean
maximum rainfall and minimum SSS synchronously occur in October,
with a four-month lag behind the maximum SST (Fig. 2a).
The tropical climate signal feedback of SST anomaly (SSTA) and
precipitation anomaly in the SCS are signicantly regulated by ENSO
teleconnection (Xie et al., 2016; Wang et al., 2006; Wang et al., 2000).
The El Ni˜
no events result in both marked positive SST (Fig. 2c) and
negative precipitation anomalies (Fig. 2d) in the boreal winter.
Conversely, La Ni˜
na events produce negative SST and increased pre-
cipitations due to the strengthened EAM and convection associated with
anticyclonic circulation. Such asynchronous changes in temperature and
precipitation in the SCS exert opposite effects on coral δ
18
O variations
and may weaken the reliability of coral δ
18
O-based ENSO reconstruction
in the northern SCS (Han et al., 2020). Interannually, the climate
anomalies of SCS are closely related to the ENSO variability (Han et al.,
2020; Yan et al., 2017; Wang et al., 2006). The 28 years band-pass
ltered SST and the precipitation in the SCS are signicantly corre-
lated with the Ni˜
no 3.4 SSTA with coherences exceeding the 95% sta-
tistical condence level, indicating consistent changes in these climate
variables on the interannual scale (Fig. 3).
3. Materials and methods
3.1. Coral sampling and δ
18
O analysis
The living Porites microatoll (GJ-1) was collected at a water depth of
~1.5 m from the southwest shallow reef at of Guangjin Island
(16.45N, 111.70E) in the Xisha Islands during the low tides in
November 2013 (Fig. 1b). The microatoll GJ-1 was then transported to
the lab and cut into 10 cm thick slab along the largest diameter axis.
Slices approximately 0.8 cm thick were then cut from the slabs along the
maximum growth axes of the coral. Radiographs of the coral slices were
produced by X-ray analysis to visualize the skeleton density bands and to
determine the major horizontal growth axis along which the subsamples
were collected using a micro-drilling system. The X-radiograph of GJ-1
showed clear annual density bands, with an average linear extension
rate of 1.7 cm/year for the period of 19792013 CE (Fig. 1d). The
growth bands of the coral core were not well dened due to bioerosion,
and were therefore not sampled for geochemical analysis. When
growing, the top surface of the microatoll may die off or develop dis-
continuities due to extremely exposures, however, no obvious evidence
of growth discontinuity or growth hiatus was found on the GJ-1 record.
The sampling transects and orientation of the microatolls were opti-
mized to avoid the possibility of intra-colony variations in δ
18
O isotopic
fractionation from the parts that might be exposed during lower water
levels. Finally, along an optimal sampling path on the the right-hand
side of the GJ-1, a 33-year transect with clear skeletal bands was
selected for powder subsamples micro-drilling and δ
18
O isotope analysis
Fig. 2. (a) The instrumental monthly mean SST from 1981 to 2013 CE (green), monthly SSS (blue) and precipitation (red) from 1981 to 2007 CE at Xisha Islands. (b)
Cross-spectral analysis between the instrumental SST and SSS at Yongxing Island, northern SCS from 1979 to 2013 CE. The cross-spectral analysis was performed
byARANDsoftware (Howell et al., 2006). The blue dashed line indicates the 95% condence level. The SST and SSS show signicant coherence at 3.85.0 years
interannual cycles. (c) Spatial correlations between monthly NCEP OI SST v2 (Reynolds et al., 2002) and Ni˜
no 3.4 SST from 1981 to 2020 CE (with SST lags by 6
months). (d) Spatial correlations between monthly GPCP v2.3 precipitation (Adler et al., 2003) and Nino 3.4 SST anomaly from 1981 to 2020 CE (Rayner et al.,
2003). The blue stars indicate the location of this study. Only correlations with p <0.05 are colored. The maps were created using the KNMI Climate Explorer (Trouet
and Van Oldenborgh, 2013). (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
4
(Fig. 1d).
Before subsampling and isotope analyses, coral slices were individ-
ually immersed in 10% H
2
O
2
for 48 h to remove organic materials,
followed by thorough ultrasonic cleaning using MilliQ water to
completely remove the surface contaminants and residual salt. The slices
were then dried in an oven at 40 C for 48 h. The slices were then
continuously micro-sampled along the major growth axis using a micro-
milling system at intervals of 0.81.0 mm, yielding a sampling resolu-
tion of ~12 samples per year, equivalent to monthly resolution
sampling.
Microatoll δ
18
O was measured in a Thermo Fisher MAT-253 isotope
ratio mass spectrometer coupled with an automated Finnigan Kiel IV
carbonate device at the Key Laboratory of Ocean and Marginal Sea
Geology, South China Sea Institute of Oceanology, Chinese Academy of
Sciences. For each sample aliquot, the coral powder was acidied with
105% H
3
PO
4
at 70 C. All the δ
18
O results were calibrated relative to the
Vienna Pee Dee Belemnite (VPDB) using the National Bureau of Stan-
dards (NBS) 18 standard (δ
18
O = − 23.2 ±0.1) (Stichler, 1995).
Fig. 3. Comparison of the Nino 3.4 index (Rayner et al., 2003) with the SST and precipitation records in the SCS. (a) Correlation between the 28 years band-pass
ltered SST in the SCS and Nino 3.4 SST, with the SCS SST lags for 6 months. (b) Correlation between the 28 years band-pass ltered precipitation and Nino 3.4 SST.
(c) Cross-spectral analyses between the SCS SST and Ni˜
no 3.4 SST records. (d) Cross-spectral analyses between the SCS SST and precipitation records. (e) Cross-
spectral analyses between the SCS precipitation and Ni˜
no 3.4 SST records. The precipitation data was obtained from the monthly GPCP v2.3 precipitation data-
set (Adler et al., 2003) and the SST data was retrieved from the OI SST v2 (Reynolds et al., 2002). The precipitation and SST data of the SCS are based on the average
values over the grid 100
120E, 2.522.5N. The cross-spectral analyses were performed by ARAND software (Howell et al., 2006). Before analyses, the data were
interpolated at a step of 0.083 year (monthly). The 95% condence level is shown with the horizontal red dashed line. The gray shaded bars represent the signicant
correlation (>95%) period within the ENSO cycles of 28 year. (For interpretation of the references to colour in this gure legend, the reader is referred to the web
version of this article.)
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
5
Analytical precision based on repeated analyses (n =80) of the NBS 18
standard yielded a standard deviation within ±0.08(1
σ
) for δ
18
O. A
total of 520 δ
18
O isotopes were analyzed with a monthly resolution for
microatoll GJ-1.
For comparative analysis between the different type of corals, we
obtained the coral δ
18
O data for the two previously published P. lutea
which were collected from Yongxing Island at a 6 m water depth
(16.84N, 112.33E) (coral YXN1, Han et al., 2020) and Qilianyu Reef at
a 2.5 m water depth (16.96N, 112.31E) (coral XS1, Wang et al., 2018)
in Xisha Islands, respectively (Fig. 1c).
3.2. Chronology development
The microatoll chronology was developed by the cross-validation of
visually counting the annual growth bands and the pronounced annual
cycle of coral δ
18
O records. The maxima and minima of the annual δ
18
O
cycles were assigned to the instrumental SST minima (January) and SST
maxima (June or July) each year respectively, to anchor the age-depth
model. The δ
18
O series were then linearly interpolated into monthly
resolution data using MATLAB software. Since there is approximately
12 months variation in the exact timing of the coldest and warmest
months in an individual year, this may cause a ~ 2 months age error in
the monthly-resolved coral δ
18
O sequence, but is assumed to be
noncumulative between years.
3.3. Data sources and analysis
The monthly instrumental SST and SSS, as well as precipitation
observation data, were obtained from the Meteorological Observatory
located on Yongxing Island, approximately 80 km away from our sam-
pling site. The monthly 1×1grid square resolution SST data from the
Optimum Interpolation SST v2 centered within 111E112E and
16N17N (Reynolds et al., 2002) were used for SST calibration and
comparison over 34 years period (19792013 CE). The gridded 0.5×
0.5resolution Simple Ocean Data Assimilation (SODA) 3.3.1 salinity
data (Carton et al., 2018) (centered at 16.25N and 111.75E) at a depth
of 5 m were used for SSS evaluation. Precipitation data were retrieved
from the monthly Global Precipitation Climatology Project v2.3 com-
bined with satellite-gauge precipitation data for a gridded 2.5square
resolution (Adler et al., 2003). The Ni˜
no 3.4 index anomaly was calcu-
lated from the area-averaged SST in the Ni˜
no 3.4 region (5S5N, 170
W120W) and obtained from the NOAA Climate Prediction Center
(Rayner et al., 2003).
A 28 years band-pass ltering of coral δ
18
O time series in this study
was performed using a simple Fast Fourier Transform (FFT) lter on
ORIGIN software to isolate the interannual variability. Spectral analysis
of SST, precipitation, and ENSO index time series were performed by
using Redt 3.8 software (Schulz and Mudelsee, 2002). The cross-
spectral analysis was applied to the instrumental SST and SSS using
the Crospecsoftware in the ARAND software package (Howell et al.,
2006).
4. Results and discussion
4.1. Seasonal to interannual climate change recorded by microatoll δ
18
O
Coral δ
18
O are predominately affected by changes in the SST and
ambient SSS (McCulloch et al., 1994; Gagan et al., 1994). Therefore, the
seasonal-interannual quantitative relationship between the microatoll
δ
18
O and SST and SSS was investigated to assess whether δ
18
O can
reliably record hydroclimatic information of ambient seawater. The
mean δ
18
O value and standard deviation for GJ-1 were 5.71 ±0.42
(1
σ
) and ranged from 6.54to 4.73, with an amplitude of 1.82,
which was similar to previously reported variation (1.57to 1.95) of
coral records in Xisha Islands (Han et al., 2020; Wang et al., 2018; Song
et al., 2012). However, the average δ
18
O showed a maximum systematic
offset of ~0.3compared with conventional Porites coral δ
18
O of the
same period (Han et al., 2020; Wang et al., 2018). The intercolony mean
coral δ
18
O offsets observed in the northern SCS were comparable with
offsets reported from Christmas Island in the central Pacic (e.g.,
Fig. 4. Microatoll δ
18
O-SST calibrations and
SST regressions. (a) Regression of coral
microatoll δ
18
O records against instrumental
SST in the Xisha Islands for the period of
19792013 CE. The dashed lines and green
shading represent the 95% prediction
bounds for calibration. (b) Comparison of
the monthly microatoll δ
18
O (orange) re-
cords and instrumental SST (blue). (For
interpretation of the references to colour in
this gure legend, the reader is referred to
the web version of this article.)
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
6
0.250.39%) (McGregor et al., 2011). To conrm the relationship
between the coral δ
18
O and the SST and SSS, we calibrated the coral
δ
18
O against the local SST and SSS on seasonal to interannual timescales
(Fig. 4 and Fig. 5). The least squares linear regression of monthly GJ-1
δ
18
O from instrumental SST yielded the following equations:
δ18O() = 0.1532 ( ± 0.0058) × SST (C) − 1.4789 ( ± 0.1594),r
= − 0.80,p<0.01,n=414 (1)
The calibration slope of the microatoll δ
18
O-SST was 0.15/C
(Fig. 4), which was close to the SST dependence of 0.15/C for the
Christmas Island microatolls (McGregor et al., 2011), and within the
range of other published coral Porites δ
18
O-SST slopes in the SCS
(0.22to 0.09) (Han et al., 2020; Ramos et al., 2017; Shen et al.,
2005; Sun et al., 2005; Yu et al., 2005). Moreover, this range lies on the
lower end of the range of calibration slopes (0.08 to 0.22) for
typical Porites in other tropical regions (Gagan et al., 2012; Corr`
ege,
2006; Lough, 2004).
Seasonally, the monthly microatoll δ
18
O was negatively well-
correlated with instrumental SST from 1979 to 2013 CE (r = − 0.80, p
<0.01) and gridded OI SST (r = − 0.80, p <0.01) from 1981 to 2013 CE
(Fig. 5a). Conversely, both the instrumental (r =0.11, p <0.01) and the
Fig. 5. Comparison of the microatoll δ
18
O
record with the local SST and SSS records on
seasonal to interannual timescales. (a)
Relationship between the monthly coral
δ
18
O (gray), instrumental SST (red) and
gridded OI SST V2 data (blue) (Reynolds
et al., 2002). (b) Relationship between the
monthly coral δ
18
O (gray), instrumental SSS
(green) and gridded SODA 3.3.1 SSS (or-
ange) (Carton et al., 2018). (c) Comparison
of the three-year moving averaged micro-
atoll δ
18
O anomaly record (gray), local
instrumental SST anomaly (red), and grid-
ded OI SST v2 anomaly data (blue). (d)
Comparison between three-year moving
average microatoll δ
18
O anomaly record
(gray) and instrumental SSS anomaly
(green) and SODA 3.3.1 SSS anomaly (or-
ange). (For interpretation of the references
to colour in this gure legend, the reader is
referred to the web version of this article.)
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
7
SODA monthly SSS (r =0.09, p =0.07) showed a weak positive corre-
lation with GJ-1 δ
18
O (Fig. 5b). In the study region, the seasonal SST and
SSS variations from 1980 to 2013 CE was 8.8 C and 2.56 psu respec-
tively. The seasonal SST variations could introduce 1.59variations of
δ
18
O and theoretically account for approximately 87.2 of the coral
δ
18
O variations based on the empirical δ
18
O-SST slope of 0.18/C
previously reported in the northern SCS (Shen et al., 2005; Sun et al.,
2005). Calibration among the microatoll δ
18
O, SST, and SSS indicated
that seasonal SST variation is the dominant controlling factor for coral
δ
18
O in the Xisha Islands, rather than the SSS.
The δ
18
O records of microatoll GJ-1 were further investigated to
explore their relationship with interannual SST and SSS variations.
Interannually, the three-year moving average of δ
18
O anomaly showed a
signicant negative correlation with both instrumental SSTA (r = − 0.42,
p <0.01) and grid OI-SSTA (r = − 0.60, p <0.01) (Fig. 5c). In com-
parison, the three-year moving average δ
18
O anomaly was signicantly
correlated to the instrumental SSS anomaly (r =0.37, p <0.01) and
SODA SSS anomaly (r =0.59, p <0.01) (Fig. 5d), which might weaken
the inuence of SST on interannual coral δ
18
O variations. The amplitude
of the three-year moving average δ
18
O anomaly and SSS anomaly was
~0.56 (ranging from 0.26 to 0.30) and ~ 0.88 psu (ranging
from 0.52 to 0.36 psu), respectively. This result indicated that ~36%
of the interannual δ
18
O variations could be attributed to SSS based on a
slope of 0.23/psu (Yan et al., 2013; Hong et al., 1997). Moreover, the
cross-spectral analysis (Howell et al., 2006) of the instrumental annual
SST and SSS (19802013 CE) also indicated that the amplitude of SST
variability was concentrated on interannual timescales, whereas the SSS
variation was strong on interdecadal timescales (Fig. 2b). The signicant
(at 95% condence level) coherence of the cross-spectral analysis at a
scale of approximately 3.85.0 years indicated that a common variance
may have existed between the instrumental records during this
Fig. 6. Comparison of the microatoll GJ-1
δ
18
O record (black line) with previously re-
ported coral δ
18
O records of P. lutea YXN1
(Han et al., 2020) (red line) and XS1 (green
line) (Wang et al., 2018) from the Xisha
Islands. (a) Monthly variability in coral δ
18
O
from cores GJ-1, YXN1 and XS1. (b) Re-
lationships between the monthly δ
18
O
anomaly of coral GJ-1, YXN1 and XS1. (c)
Comparison of interannual variability in
coral δ
18
O anomaly records. The interannual
coral δ
18
O variability was isolated by
removing the seasonal cycle and applying a
three-year moving average lter to each re-
cord. (For interpretation of the references to
colour in this gure legend, the reader is
referred to the web version of this article.)
Table 1
Intercolony correlations of the monthly mean δ
18
O and interannual δ
18
O
anomalies between different corals in Xisha Islands. The interannual coral δ
18
O
anomalies were applied by a three-year moving average lter to each record. All
records showed a statistically signicant correlation (p <0.05) with each other.
Monthly δ
18
O correlations
Coral Monthly GJ-1
δ
18
O ()
Monthly YXN1
δ
18
O ()
Monthly XS1
δ
18
O
Monthly GJ-1
δ
18
O ()
/ r =0.66 r =0.56
Monthly YXN1
δ
18
O ()
r =0.66 / r =0.62
Monthly XS1 δ
18
O r =0.56 r =0.62 /
Interannual δ
18
O anomaly correlations
Coral GJ-1 δ
18
O
anomaly ()
YXN1 δ
18
O
anomaly ()
XS1 δ
18
O
anomaly ()
GJ-1 δ
18
O
anomaly () / r =0.45 r =0.60
YXN1 δ
18
O
anomaly () r =0.45 / r =0.78
XS1 δ
18
O anomaly
() r =0.60 r =0.78 /
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
8
interannual periodicity (Fig. 2b). Therefore, these results support our
argument that the interannual variation in coral δ
18
O was primarily
controlled by SST, yet SSS contribution was non-negligible. Regardless,
the signicant correlation of the coral microatoll δ
18
O composition with
the interannual SST and SSS variations demonstrates that coral δ
18
O can
record local seawater conditions and has the potential to capture
interannual climatic variations regulated by interannual ENSO vari-
ability. However, the SST signal strength caused by ENSO activity may
be partially affected by SSS changes.
4.2. Reproducibility among coral δ
18
O records
To assess the variation and reproducibility of δ
18
O signals among
different corals, the microatoll δ
18
O record was compared to two pre-
viously published Porites colonies (YXN1 and XS1) in the Xisha Islands
(Han et al., 2020; Wang et al., 2018). Monthly coral δ
18
O records from
GJ-1, YXN1, and XS1 exhibited similar patterns of variability and were
well correlated to each other (r =0.56 to 0.66, p <0.01), as well as with
the local instrumental SST (r = − 0.58 to 0.80, p <0.01) (Fig. 6a and
Table 1). Despite exhibiting similar variability, signicant systematic
intercolony absolute (mean) δ
18
O offsets were observed for the period
where the microatoll GJ-1 and both two adjacent conventional dome-
shaped Porites records from core XS1 and YXN1 overlap (19792013
CE) (Fig. 6b and Table 2). For the common period (19792008 CE), the
average δ
18
O of Porites YXN1 and XS1 showed positive offsets of 0.23
and 0.20, respectively, relative to microatoll GJ-1. The δ
18
O differ-
ence between Porites YXN1 and XS1 was relatively small at 0.03. Since
cores XS1 and YXN1 were situated geographically close to one another,
the slight offset between them may indicate similar seawater properties
for the two cores. Between 1979 and 2013 CE, the mean δ
18
O value for
core XS1 was 0.20higher than that of microatoll GJ-1, which is
equivalent to an offset of 1.3 C if interpreted purely in terms of tem-
perature (Table 2). Despite the the differences in absolute mean δ
18
O
Table 2
Comparison of the growth rates, the mean δ
18
O, the average amplitude of the seasonal cycle (δ
18
O standard deviation), and the interannual variability in δ
18
O records
from the overlapping intercolony corals. The intercolony offsets are shown as values relative to the microatoll δ
18
O record.
Coral δ
18
O ()
Coral Record Span (AD) Mean () Offset () Seasonal amplitude
(std. dev.) ()
Interannual amplitude() Growth rate(mm/year)
Microatoll GJ-1 19792008 5.674 / 0.42 0.12 1.73
P. lutea YXN1 19792008 5.443 0.23 0.26 0.07 1.31
P. lutea XS1 19792008 5.477 0.20 0.39 0.17 ~2.0
Microatoll GJ-1 19792013 5.708 / 0.42 0.13 1.71
P. lutea XS1 19792013 5.505 0.20 0.38 0.16 ~1.9
Table 3
Relationships between growth rates, the δ
18
O record of microatoll GJ-1 and the
local SST records.
δ
18
O SST
Summer Winter Annual Summer Winter Annual
Coral
growth
rate
r =
0.34
r =
0.29
r =
0.30 r =0.10 r =
0.28
r =
0.14
p =
0.051
p =
0.10
p =
0.09 p =0.59 p =
0.12
p =
0.46
Fig. 7. (a) Seasonal amplitudes of coral
δ
18
O in microatoll GJ-1 (black), P. lutea
coral YXN-1 (green) (Han et al., 2020), and
XS1 (red) (Wang et al., 2018) for the period
of 19792013 CE they overlapped. (b) Coral
δ
18
O-SST and (c) δ
18
O-SSS calibrations for
microatoll GJ-1 (blue), the P. lutea YXN1
(green) and XS1 (red) cores. R1, R2, R3 in-
dicates the correlation of coral GJ-1, YXN-1,
and XS1 with the instrumental SST and SSS
records, respectively. (For interpretation of
the references to colour in this gure legend,
the reader is referred to the web version of
this article.)
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
9
values during overlapping intervals, these differences are within the
previously reported range of δ
18
O offsets (0.150.40) between in-
dividual corals on a given reef in the tropical Pacic region (Stephans,
2004; Guilderson and Schrag, 1999; Linsley et al., 1999), and are less
than the 0.39intercolony offsets observed among the microatolls and
Porites in Christmas Island (McGregor et al., 2011).
Across the same interval, coral δ
18
O from cores GJ-1 (0.42) and
XS1 (0.38) have similar seasonal amplitudes, which were larger than
the amplitude of seasonal variations in core YXN1 (0.26) (Table 2).
The interannual coral δ
18
O variability was isolated by removing the
seasonal cycle and applying a 3-year moving average lter to each re-
cord. The interannual amplitude of δ
18
O variability was estimated by
computing the standard deviation of each ltered record (Sayani et al.,
2019). The microatolls GJ-1 did not show signicant increased inter-
annual cycle amplitude resulting from their growth on the shallow reef
at, as has been suggested elsewhere (Fig. 6c) (Swart and Coleman,
1980). Interannual variability was consistent between the cores GJ-1
(0.16, 1
σ
) and XS1 (0.13, 1
σ
) for the period of 19792013 CE,
whereas core YXN1 exhibited somewhat muted interannual variations
(0.07, 1
σ
) across the periods where they overlap (Fig. 6c and Table 2).
The standard deviation of interannual variations offsets among three
corals (±0.05, 1
σ
, n =3) are within the range of analytical uncer-
tainty (±0.08, 1
σ
) and thus not signicant. Longer δ
18
O records would
be required from each coral to evaluate whether there is a stable sys-
tematic offset in the recording of interannual coral δ
18
O variability for a
given coral core. Unfortunately, the limited timescales of the microatolls
GJ-1 and core XS1 do not allow us for assessing the longer-scale changes.
McGregor et al. (2011) hypothesized that these differences could
reect non-climatic factors or a real spatial difference in local SST and
SSS variations between different locations. However, in situ SST and SSS
data containing the exact location and time span of the corals did not
exist in the study area. Thus, further analyses of water δ
18
O are required
to test this hypothesis. The skeletal extension rates have been invoked to
explain, and in some cases correct for the absolute intercolony δ
18
O
offsets in a variety of coral species (Sayani et al., 2019; McGregor et al.,
2011; Maier et al., 2004). Nevertheless, the skeletal extension rate
appeared to have minimal inuence on GJ-1 δ
18
O. The extension rates
did not show any statistically signicant relationships with the winter (r
= − 0.29, p =0.10), summer (r = − 0.34, p =0.051) and annual mean
coral δ
18
O (r = − 0.30, p =0.09) (Table 3). Given the northern SCS is
located close to the boundary of the Porites distribution area, the
extremely cold winter SST tending to the limits of the optimum tem-
perature may affect coral growth. Whereas, no signicant correlations
were observed among the microatoll extension rates, summer (r =0.10,
p =0.59), winter (r =0.28, p =0.12) and annual mean SST (r =0.14, p
=0.46) (Table 3). Further, the coral microatoll X-ray radiographs
showed no obvious evidence of growth discontinuity or hiatus in the
tract we sampled for isotope analyses between 1979 and 2013 CE
(Fig. 1d). Moreover, since there was no obvious difference in extension
rates among different corals presented here, and their extension rates
(1.32.0 mm/year) (Table 2) are above the threshold for growth related
effects for Porites spp. (<0.6 mm/year) (Felis et al., 2003; Suzuki et al.,
2003; McConnaughey, 1989), the isotope fractionation effect caused by
the growth rate was negligible and cannot be used to explain these
differences.
The discrepancy in the sensitivity of individual corals to SST was
another important factor likely affecting the absolute mean value and
seasonal variation of coral δ
18
O (Sayani et al., 2019; Gagan et al., 2012).
The δ
18
O-SST slopes were used to assess whether δ
18
O was equally
sensitive to climate variability across different coral colonies (Sayani
et al., 2019; Gagan et al., 2012). Microatoll GJ-1 had a steeper δ
18
O-SST
calibration slope at 0.15 ±0.01 /C, while cores YXN1 and XS1
exhibited very similar but shallower slopes at 0.09 ±0.00/C and
0.10 ±0.01/C, respectively (Fig. 7b). In terms of SST alone, δ
18
O-SST
slopes for cores GJ-1, YXN1 and XS1 would produce about 0.30.4 C
temperature difference for a 0.050.06change in coral δ
18
O. Both the
Fig. 8. Comparison of the monthly microatoll δ
18
O anomaly with the Ni˜
no 3.4
Index, precipitation anomaly and SSTA records. (a) The monthly Ni˜
no 3.4 Index
anomaly calculated from the area-averaged SST in the Ni˜
no 3.4 region (5S5
N, 170W120W) (Rayner et al., 2003). (b) Monthly microatoll anomaly δ
18
O
series. (c) Monthly instrumental SST anomaly series obtained from observation
records of the Meteorological Observatory located on Yongxing Island (purple
line) and OI SST v2 (green line) (Reynolds et al., 2002). (d) Monthly precipi-
tation data was retrieved from the monthly Global Precipitation Climatology
Project v2.3 dataset (Adler et al., 2003). The microatoll δ
18
O anomaly and SSTA
records were processed with a lag of 6 months. The gray bars indicate the strong
and super El Ni˜
no events (ONI >1.5). (For interpretation of the references to
colour in this gure legend, the reader is referred to the web version of
this article.)
Fig. 9. Spatial correlation maps of monthly microatoll δ
18
O anomaly record in
Xisha Islands with OI SST v2 anomaly extended from 1981 to 2013 CE. Only
correlations with p <0.05 are colored. Map was computed using KNMI Climate
Explorer (http://climexp.knmi.nl/) (Trouet and Van Oldenborgh, 2013). The
blue star indicates the location of the microatoll analyzed in this study. (For
interpretation of the references to colour in this gure legend, the reader is
referred to the web version of this article.)
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
10
absolute slopes and intercolony slope offsets were well within the range
of published δ
18
O-SST slopes (Sayani et al., 2019; McGregor et al., 2011;
Grottoli and Eakin, 2007). For the absolute seasonal variations in the
individual coral record from 1979 to 2013 CE, the mean summer (JJA)
δ
18
O value of GJ-1 (average of 6.08) exhibited remarkable negative
offsets comparable to core YXN1 (0.39) and XS1 (0.33), whereas
only small isotope fractionations (0.06) existed between core YXN1
and XS1. In boreal winter, the seasonal δ
18
O of GJ-1 exhibited similar
offsets for core YXN1 (0.01) and XS1 (0.09), suggesting that the
microatoll δ
18
O was more sensitive to summer SST changes (Fig. 7a). In
comparison, the coral δ
18
O-SSS slopes were also largely consistent
among different corals over their respective overlaps, indicating that SSS
may not be the dominant cause of seasonal δ
18
O shifts (Fig. 7c). In
summary, our results demonstrated that overlapping coral δ
18
O records
from Xisha Islands exhibited similar patterns of variability on monthly
to interannual time scales, but exhibited different proxy-SST sensitivities
and amplitude of the seasonal variabilities. Although their mean values
were consistently offset, the coral δ
18
O were signicantly correlated
with each other and the local SST variations.
4.3. Microatoll δ
18
O-ENSO signal and reconstruction
A solid understanding of the isotopic signals recorded in corals
underpinned climate reconstruction based on any species or growth
forms of coral (Ramos et al., 2017; McGregor et al., 2011). As the
sampling location was situated far away from the mainland estuary and
was not disturbed by any ocean circulations, we determined that the
coral δ
18
O was predominately affected by SST and the δ
18
O
sw
signals
related to precipitation. The microatoll anomaly δ
18
O series were
compared to SSTA, precipitation anomaly, and the Ni˜
no 3.4 index to
evaluate the reliability of this proxy for recording ENSO signals (Fig. 8).
The SSTA (r =0.45, p <0.01) and precipitation anomaly (r = − 0.42, p <
0.01) (19792013 CE) were signicantly correlated with Ni˜
no 3.4 index
(three-months moving average lter), indicating that the seasonal
climate in the Xisha Islands is sensitive to ENSO variations.
In terms of coral δ
18
O, the monthly microatoll δ
18
O showed a sig-
nicant negative relationship with local SST (r = − 0.80, p <0.01) and
precipitation (r = − 0.37, p <0.01). When seasonal cycles were
removed, the δ
18
O anomaly was weakly but still signicantly correlated
with OI SSTA (r = − 0.37, p <0.01) and similar to that of coral YXN1 (r
= − 0.33, p <0.01) (Han et al., 2020), whereas no correlation with
precipitation anomalies was found (r = − 0.02, p =0.74). At larger
spatial scales, the microatoll δ
18
O anomalies exhibit signicantly
negative correlations (ranging from 0.2 to 0.5) with SSTA across the
entire SCS region (Fig. 9). On a seasonal scale, however, the microatoll
δ
18
O anomaly and Ni˜
no 3.4 SST anomaly were not signicantly corre-
lated, which is consistent with other coral δ
18
O records reported in the
Xisha Islands (Han et al., 2020; Wang et al., 2018). In comparison, the
δ
18
O of microatolls located at Christmas Island showed a stronger cor-
relation with the Ni˜
no 3.4 SST anomaly (McGregor et al., 2011), sug-
gesting that the coral-based ENSO reconstruction records in the non-
ENSO core regions are restricted, and the seasonal signals of ENSO
cannot be fully captured. Despite that, the microatoll δ
18
O anomaly
recorded mostly strong-super El Ni˜
no events (with ONI >1.5) over the
past 30 years. The negative bias of δ
18
O in microatolls corresponds well
to positive SST anomalies and precipitation reduction during strong-
super El Ni˜
no events (Fig. 8).
The spectral analyses applied to the δ
18
O records and Ni˜
no 3.4 index
also revealed signicant periodicities of 2.33.5 year within the ENSO
domain frequency, demonstrating that the ENSO signals were captured
in the microatoll δ
18
O record (Fig. 10a). Moreover, signicant (at the
90% condence level) cross-spectral coherence within the 2.62.9 years
cycle was observed between the microatoll δ
18
O and Ni˜
no 3.4 SST
(Fig. 10b), supporting our argument that the interannual microatoll
δ
18
O variations were dominated by interannual ENSO activity. The
ENSO-related interannual SST variations were also captured by the
conventional Porites corals in this region (Han et al., 2020; Wang et al.,
2018). The δ
18
O record for core YXN1 presented signicant ENSO pe-
riodicities of 2.3 to 3.9 years since 1851 CE (Han et al., 2020), which
highlights the long-term stability of the ENSO cycle in the northern SCS.
Interannually, the microatoll δ
18
O was well-correlated to the OI-V2
SSTA (r = − 0.60, p <0.01), compared with similar correlation for
Porites YXN1 (r = − 0.66, p <0.01), but lower for Porites XS1 (r =0.39, p
<0.01). To reconstruct the occurrence of ENSO events, microatoll δ
18
O
anomaly records were 28 year band-pass ltered to isolate the ENSO-
dominated frequencies. The band-pass ltered δ
18
O anomalies for
microatoll GJ-1 were also negatively correlated with the Ni˜
no 3.4 SSTA
variability (r = − 0.40, p <0.01, lagging 6 months) (Fig. 11a). The
interannual variability of the coral YXN1 δ
18
O exhibited similar corre-
lation with the Ni˜
no 3.4 SSTA (r = − 0.39, p <0.01) (Fig. 11b, Han et al.,
2020). The signicant correlation for the microatoll δ
18
O anomaly
signal against SST and Ni˜
no 3.4 SSTA suggests that the interannual
ENSO variance dominated the microatoll δ
18
O and local SST in the
northern SCS.
We future analyzed the ENSO detection skill with respect to the coral
δ
18
O anomaly from 1979 to 2013 CE using the empirically calibrated
threshold method proposed by Hereid et al. (2013). The threshold level
was then independently optimized for each individual record to mini-
mize record errors and maximize detection skill. A threshold value of
±0.060was ultimately calculated (Fig. 11a). The peaks that exceeded
Fig. 10. (a) The spectral analysis of the microatoll GJ-1 δ
18
O records. The numbers denote the cycles of signicant spectral peaks (at 90% condence level). (b) The
cross-spectral analysis between the microatoll GJ-1 δ
18
O and Ni˜
no 3.4 index record (Rayner et al., 2003). Dotted red line indicates signicance at the 90% condence
level and gray shaded band represents signicant coherence period within the interannual ENSO band. (For interpretation of the references to colour in this gure
legend, the reader is referred to the web version of this article.)
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
11
the set threshold value were dened as ENSO events. Given that the
maximum SST anomaly in the SCS usually lags ~36 months behind the
peaks in the ENSO core region (Klein et al., 1999), the ENSO events
identied in the coral δ
18
O record may have occurred in the same year
or early in the following year. In this study, a lag of 6 months was used as
the time error to evaluate the accuracy of the microatoll δ
18
O recording
of ENSO events.
A total of 11 El Nin˜
o and 10 La Ni˜
na events were captured by
instrumental Ni˜
no 3.4 SSTA records for 19792013 CE (Fig. 11d). Based
on the threshold level of ±0.060, the microatoll δ
18
O from the Xisha
Islands correctly identied 8 El Ni˜
no events and 5 La Ni˜
na events,
showing an overall detection skill of 73% (8/11) for El Ni˜
no events and
50% (5/10) for La Ni˜
na events, respectively (Fig. 11a). Compared to the
La Ni˜
na events, the coral δ
18
O in northern SCS appeared to favor El Ni˜
no
expression. The mean ENSO detection skill was 62%, obtained by
averaging the El Ni˜
no and La Ni˜
na detection skills, which was broadly
similar to the coral records from the edge region of WPWP and the
Central Pacic region (Hereid et al., 2013). At the local scale, this
detection skill was generally consistent with that for core YXN-1 δ
18
O
(~65%) over the past 157 years (Han et al., 2020), as well as the ENSO
detection skill (60%70%) for instrumental SSTA records from Xisha
Islands during the period of 19002000 CE (Jiang et al., 2021).
Fig. 11. ENSO reconstruction based on
coral δ
18
O anomaly. (a) 28 years band-pass
ltered microatoll δ
18
O anomalies (orange
line) for the period of 19792013 CE. The
blue line indicates the ltered δ
18
O anomaly
series with a lead of 6 months. The dash
lines highlight the threshold of ±0.060.
The solid triangles represent correctly iden-
tied El Ni˜
no (red) and La Ni˜
na (green)
events. The hollow triangles indicate the El
Ni˜
no (red) and La Ni˜
na (green) events
captured by Ni˜
no 3.4 SST, but not identied
in the microatoll reconstruction. (b) 28
years band-pass ltered δ
18
O anomaly for
coral YXN1 during 19792008 CE (Han
et al., 2020). (c) 28 years band-pass ltered
instrumental SSTA (green) and OI SST v2
anomaly (blue) from the Xisha Islands. The
instrumental SSTA was obtained from the
meteorological observatory located on the
Yongxing Island. The gridded OI SST was
centered within 111E112E and 16
N17N (Reynolds et al., 2002). (d) The
Ni˜
no 3.4 SSTA record (red) and its 28 years
band-pass ltered series is also presented
(black). The numbers denote the El Ni˜
no
(black) and La Ni˜
na (blue) events, respec-
tively. The Ni˜
no 3.4 index anomaly was
calculated from the area-averaged SST in the
Ni˜
no 3.4 region (5S5N, 170W120W)
and obtained from the NOAA Climate Pre-
diction Center (Rayner et al., 2003). (For
interpretation of the references to colour in
this gure legend, the reader is referred to
the web version of this article.)
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
12
Moreover, we found that the strong (1.5 <|ONI| <2.0) and super (|
ONI| 2.0) ENSO events revealed signicantly better reconstruction
skills than the weak to moderate ENSO events (0.5 |ONI| 1.5). The
detection skills for strong or super El Ni˜
no and La Ni˜
na events were
100% and 75% respectively, whereas that for weak or moderate events
they were reduced to 50% and 33%. These results indicated that the
δ
18
O signal of microatolls was more sensitive to strong ENSO events and
can accurately capture these events. This observation was also
conrmed by two other Porites records, YXN1 and XS1, in the Xisha
Islands (Han et al., 2020; Wang et al., 2018). At least six weak-moderate
ENSO events were missed in coral YXN1 from 1979 to 2007 CE period
(Han et al., 2020).
Apart from the misjudgment and omission of ENSO events, the in-
tensity of the El Ni˜
no event recorded in the microatoll δ
18
O records also
showed partial disagreement with the instrumental records. The super El
Ni˜
no events that occurred during 19821983 CE and 19971998 CE on
instrumental records did not correspond well to the microatoll δ
18
O
record, which were incorrectly identied as moderate El Ni˜
no events
(Fig. 11a). The mismatch in ENSO intensity during the 19821983 CE El
Ni˜
no event was also observed in the coral δ
18
O record for core YXN1
(Fig. 11b, Han et al., 2020) and the instrumental SSTA (Fig. 11c).
Several factors may have contributed to the observed discrepancies. We
speculate that changes in local salinity near the sampling site may have
suppressed ENSO-related SST signals in coral δ
18
O records. The climate
in the SCS tended to be drier and warmer during the El Ni˜
no events,
while relative colder and wetter conditions occurred during the La Ni˜
na
events. Such reverse changes in SST and precipitation exerted opposite
effects on changes in coral δ
18
O, potentially impairing the accuracy of
coral δ
18
O records in the northern SCS for reconstructing ENSO vari-
ability (Han et al., 2020). An alternative explanation for the disagree-
ment in ENSO activity is likely related to the local environment where
the microatolls grow. Large freshwater inputs from extreme precipita-
tion were also likely to result in local seawater salinity changes, espe-
cially for the microatolls growing in ponding environments, which may
partially offset the effects of SST variation on microatoll δ
18
O. However,
continuous long-term in-situ observation records are required to further
verify these explanations. The suppression of the δ
18
O signal by salinity
changes was also conrmed by the interannual correlation between the
microatoll δ
18
O and local SSS (Fig. 5). The reconstruction error of ENSO
events from the coral δ
18
O records of the Xisha Islands may also be
related to the complex monsoon-ENSO system relationship, which is
strongly linked to the strength of ENSO variability (Zhou and Chan,
2010). The weak ENSO events may not be recorded in corals from the
northern SCS due to possible decoupling of the connection between the
EAM and ENSO (Han et al., 2020; Webster and Yang, 1992). Further-
more, given that the northern SCS is located in the non-core region of
ENSO, the signal of weak-to-moderate ENSO events may not be captured
by the corals there (Wang et al., 2018). The ENSO signal recorded by
microatoll δ
18
O was suppressed to a certain extent, and the ability of
microatoll δ
18
O to record the weak-to-moderate ENSO events was con-
strained due to the above challenges. However, our records demon-
strated that the microatoll δ
18
O from the northern SCS appeared to be a
sensitive and relatively reliable proxy for regional interannual climate
change and ENSO reconstruction. Composite δ
18
O records from multiple
microatolls growing in well-ushed reef at may facilitate more accu-
rate reconstruction of ENSO activity in the northern SCS.
5. Conclusions
This study investigated the microatoll δ
18
O records relative to two
adjacent Porites corals from the Xisha Islands in the northern SCS, to
examine the intercolony reproducibility of δ
18
O signals and evaluate the
reliability of microatoll δ
18
O as paleoclimate proxies for reconstructing
regional climate and ENSO variations. The seasonal and interannual
variations of coral microatoll δ
18
O were primarily controlled by local
SST, while the contribution of SSS in δ
18
O (accounting for ~36%)
becomes non-negligible on the interannual timescale. Despite the
overlapping coral δ
18
O records exhibited similar patterns of variability
on monthly to interannual time scales, the mean values among different
coral colonies were consistently offset by ~0.2. The microatoll δ
18
O
exhibited relatively high proxy-SST sensitivities and amplitude of the
seasonal variabilities. The comparisons of the microatoll δ
18
O anomaly
against SST and Ni˜
no 3.4 index records indicate that the interannual
ENSO variance dominated the microatoll δ
18
O and local SST in the
northern SCS. Microatoll δ
18
O can serve as a sensitive and relatively
reliable proxy for ENSO variability, with an overall a detection skill of
73% for El Ni˜
no events and 50% for La Ni˜
na events. This is generally
consistent with that for the conventional Porites corals (~67%) and the
ENSO detection skill (~71%) for instrumental SSTA records from Xisha
Islands. However, we also found that the reliability for recording the
weak-to-moderate ENSO events was constrained due to the complex
relationship between the EAM and ENSO, as well as the local seawater
salinity changes. Composite δ
18
O records from multiple microatolls
growing in well-ushed reef ats may facilitate a more accurate
reconstruction of ENSO activity in the northern SCS. The intra-colony
reproducibility of the microatoll δ
18
O signal and other geochemical
proxies may also require further investigation.
Declaration of Competing Interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have inuenced the work
reported in this paper.
Acknowledgements
This work was supported by the National Natural Science Foundation
of China (41506061, 41702122); National Key R&D Program of China
(2021YFC3100603); Key Special Project for Introduced Talents Team of
Southern Marine Science and Engineering Guangdong Laboratory
(Guangzhou) (GML2019ZD0206); Strategic Priority Research Program
of the Chinese Academy of Sciences (XDA13010103). We would like to
thank Dr. Guohui Liu for his help with eld sampling and the sectioning
of the core. We also thank Dr. Tao Han for his assistance with the data
analysis and manuscript revision. The authors sincerely appreciate two
reviewers for their careful works and constructive suggestions, which
greatly improved this manuscript.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.palaeo.2022.111031.
References
Adler, R.F., Huffman, G.J., Chang, A., Ferraro, R., Xie, P.P., Janowiak, J., Rudolf, B.,
Schneider, U., Curtis, S., Bolvin, D., Gruber, A., Susskind, J., Arkin, P., Nelkin, E.,
2003. The version-2 global precipitation climatology project (GPCP) monthly
precipitation analysis (1979-present). J. Hydrometeorol. 4, 11471167.
Carton, J.A., Chepurin, G.A., Chen, L.G., 2018. SODA3: a new ocean climate reanalysis.
J. Clim. 31, 69676983.
Chen, T., Cobb, K.M., Roff, G., Zhao, J., Yang, H., Hu, M., Zhao, K., 2018. Coral-derived
western Pacic tropical sea surface temperatures during the last millennium.
Geophys. Res. Lett. 45, 35423549.
Cobb, K.M., Charles, C.D., Cheng, H., Edwards, R.L., 2003. El Ni˜
no Southern Oscillation
and tropical Pacic climate during the last millennium. Nature. 424, 271276.
Corr`
ege, T., 2006. Sea surface temperature and salinity reconstruction from coral
geochemical tracers. Palaeogeogr. Palaeoclimatol. Palaeoecol. 232, 408428.
Dassi´
e, E.P., Linsley, B.K., Corr`
ege, T., Wu, H.C., Lemley, G.M., Howe, S., Cabioch, G.,
2014. A Fiji multi-coral δ18O composite approach to obtaining a more accurate
reconstruction of the last two-centuries of the ocean-climate variability in the South
Pacic Convergence Zone region. Paleoceanography 29, 11961213.
Felis, T., P¨
atzold, J., Loya, Y., 2003. Mean oxygen-isotope signatures in Porites spp.
corals: inter-colony variability and correction for extension-rate effects. Coral Reefs
22, 328336.
F. Tan et al.
Palaeogeography, Palaeoclimatology, Palaeoecology 598 (2022) 111031
13
Freund, M.B., Henley, B.J., Karoly, D.J., McGregor, H.V., Abram, N.J., Dommenget, D.,
2019. Higher frequency of Central Pacic El Ni˜
no events in recent decades relative to
past centuries. Nat. Geosci. 12, 450455.
Gagan, M.K., Chivas, A.R., Isdale, P.J., 1994. High-resolution isotopic records from corals
using ocean temperature and mass-spawning chronometers. Earth Planet. Sci. Lett.
121, 549558.
Gagan, M.K., Ayliffe, L.K., Beck, J.W., Cole, J.E., Druffel, E.R.M., Dunbar, R.B., Schrag, D.
P., 2000. New views of tropical paleoclimates from corals. Quat. Sci. Rev. 19, 4564.
Gagan, M.K., Dunbar, G.B., Suzuki, A., 2012. The effect of skeletal mass accumulation in
Porites on coral Sr/Ca and δ
18
O paleothermometry. Paleoceanography. 27, PA1203.
Grottoli, A.G., Eakin, C.M., 2007. A review of modern coral δ
18
O and Δ
14
C proxy records.
Earth Sci. Rev. 81, 6791.
Guilderson, T.P., Schrag, D.P., 1999. Reliability of coral isotope records from the western
Pacic warm Pool: a comparison using age-optimized records. Paleoceanography 14,
457464.
Han, T., Yu, K., Yan, H., Jiang, W., Yan, H., Tao, S., 2020. Coral δ18O-based
reconstruction of el Ni˜
no-Southern Oscillation from the northern south China sea
since 1851 AD. Quat. Int. 550, 159168.
Hereid, K.A., Quinn, T.M., Okumura, Y.M., 2013. Assessing spatial variability in el Ni˜
no-
Southern Oscillation event detection skill using coral geochemistry.
Paleoceanography 28, 1423.
Hong, A., Hong, Y., Wang, Q., Ke, J., 1997. Distributive characteristics of δ
18
O isotope of
the northeastern South China Sea in the summer of 1994. Trop. Oceanol. 16, 8290
(in Chinese with English abstract).
Howell, P., Pisias, N., Balance, J., Baughman, J., Ochs, L., 2006. ARAND Time-Series
Analysis Software. Brown Univ, Providence RI.
Hu, Y., Sun, X., Cheng, H., Yan, H., 2020. Evidence from giant-clam δ
18
O of intense El
Nin˜
oSouthern Oscillation-related variability but reduced frequency 3700 years ago.
Clim. Past 16, 597610.
Jiang, L., Yu, K., Tao, S., Wang, S., Han, T., Jiang, W., 2021. ENSO variability during the
medieval climate Anomaly as recorded by Porites corals from the northern South
China Sea. Paleoceanogr. Paleoclimatol. 36 e2020PA004173.
Klein, S.A., Soden, B.J., Lau, N.C., 1999. Remote sea surface temperature variations
during ENSO: evidence for a tropical Atmospheric bridge. J. Clim. 12, 917932.
Knutson, D.W., Buddemeier, R.W., Smith, S.V., 1972. Coral chronometers: seasonal
growth bands in reef corals. Science 177, 270272.
Linsley, B.K., Messier, R.G., Dunbar, R.B., 1999. Assessing between-colony oxygen
isotope variability in the coral Porites lobata at Clipperton Atoll. Coral Reefs 18,
1327.
Linsley, B.K., Zhang, P., Kaplan, A., Howe, S.S., Wellington, G.M., 2008. Interdecadal-
decadal climate variability from multicoral oxygen isotope records in the South
Pacic Convergence Zone region since 1650 A.D. Paleoceanography 23. PA2219, n/
an/a.
Lough, J.M., 2004. A strategy to improve the contribution of coral data to high-resolution
paleoclimatology. Palaeogeogr. Palaeoclimatol. Palaeoecol. 204, 115143.
Maier, C., Felis, T., P¨
atzold, J., Bak, R.P.M., 2004. Effect of skeletal growth and lack of
species effects in the skeletal oxygen isotope climate signal within the coral genus
Porites. Mar. Geol. 207, 193208.
Majewski, J.M., Switzer, A.D., Meltzner, A.J., Parham, P.R., Horton, B.P., Bradley, S.L.,
Pile, J., Chiang, H.-W., Wang, X., Ng, C.T., Tanzil, J., Müller, M., Mujahid, A., 2018.
Holocene relative sea-level records from coral microatolls in western Borneo, South
China Sea. Holocene 28, 14311442.
McConnaughey, T., 1989.
13
C and
18
O isotopic disequilibrium in biological carbonates: I.
Patterns. Geochim. Cosmochim. Acta 53, 151162.
McCulloch, M.T., Gagan, M.K., Mortimer, G.E., Chivas, A.R., Isdale, P.J., 1994. A high-
resolution Sr/Ca and δ
18
O coral record from the great barrier Reef, Australia, and the
19821983 El nino. Geochem. Cosmochim. Acta 58 (12), 27472754.
McGregor, H.V., Gagan, M.K., 2004. Western Pacic coral δ
18
O records of anomalous
Holocene variability in the El Ni˜
no-Southern Oscillation. Geophys. Res. Lett. 31,
L11204.
McGregor, H.V., Fischer, M.J., Gagan, M.K., Fink, D., Woodroffe, C.D., 2011.
Environmental control of the oxygen isotope composition of Porites coral microatolls.
Geochim. Cosmochim. Acta 75, 39303944.
McGregor, H.V., Fischer, M.J., Gagan, M.K., Fink, D., Phipps, S.J., Wong, H.,
Woodroffe, C.D., 2013. A weak El Ni˜
no/Southern Oscillation with delayed seasonal
growth around 4,300 years ago. Nat. Geosci. 6, 949953.
Mcphaden, M.J., Zebiak, S.E., Glantz, M.H., 2006. ENSO as an integrating concept in
earth science. Science 314, 17401745.
Meltzner, A.J., Woodroffe, C.D., 2015. Coral microatolls. In: Shennan, I., Long, A.J.,
Horton, B.P. (Eds.), Handbook of Sea-Level Research. John Wiley & Sons, Ltd,
Chichester, UK, pp. 125145. https://doi.org/10.1002/9781118452547.ch8.
Meltzner, A.J., Sieh, K., Chiang, H.-W., Shen, C.-C., Suwargadi, B.W., Natawidjaja, D.H.,
Philibosian, B.E., Briggs, R.W., Galetzka, J., 2010. Coral evidence for earthquake
recurrence and an AD 13901455 cluster at the south end of the 2004
AcehAndaman rupture. J. Geophys. Res. 115, B10402.
Ramos, R.D., Goodkin, N.F., Siringan, F.P., Hughen, K.A., 2017. Diploastrea heliopora
Sr/Ca and δ
18
O records from Northeast Luzon, Philippines: an assessment of
interspecies coral proxy calibrations and climate controls of sea surface temperature
and salinity. Paleoceanography. 32, 424438.
Rayner, N.A., et al., 2003. Global analyses of sea surface temperature, sea ice, and night
marine air temperature since the late nineteenth century. J. Geophys. Res. 108 (14),
4407.
Reynolds, R.W., Rayner, N.A., Smith, T.M., Stokes, D.C., Wang, W.Q., 2002. An improved
in situ and satellite SST analysis for climate. J. Clim. 15, 16091625.
Ropelewski, C.F., Halpert, M.S., 1987. Global and regional scale precipitation patterns
associated with the El Ni˜
no/southern oscillation. Mon. Weather Rev. 115,
16061626.
Sayani, H.R., Cobb, K.M., DeLong, K., Hitt, N.T., Druffel, E.R.M., 2019. Intercolony δ
18
O
and Sr/Ca variability among Porites spp. corals at Palmyra Atoll: toward more robust
coral-based estimates of climate. Geochem. Geophys. Geosyst. 20, 52705284.
Schulz, M., Mudelsee, M., 2002. REDFIT: estimating red-noise spectra directly from
unevenly spaced paleoclimatic time series. Comput. Geosci. 28, 421426.
Shen, C.-C., Lee, T., Liu, K.-K., Hsu, H.-H., Edwards, R.L., Wang, C.-H., Lee, M.-Y.,
Chen, Y.-G., Lee, H.-J., Sun, H.-T., 2005. An evaluation of quantitative
reconstruction of past precipitation records using coral skeletal Sr/Ca and δ
18
O data.
Earth Planet. Sci. Lett. 237, 370386.
Song, S., Peng, Z., Zhou, W., Liu, W., Liu, Y., Chen, T., 2012. Variation of the winter
monsoon in South China Sea over the past 183 years: evidence from oxygen isotopes
in coral. Glob. Planet. Chang. 9899, 131138.
Stephans, C.L., 2004. Assessing the reproducibility of coral-based climate records.
Geophys. Res. Lett. 31, L18210.
Stichler, W., 1995. Interlaboratory Comparison of New Materials for Carbon and Oxygen
Isotope Ratio Measurements, Paper Presented at Reference and Intercomparison
Materials for Stable Isotopes of Light Elements, 13 Dec. IAEA Austria, IAEA, Vienna.
Stoddart, D.R., Scofn, T.P., 1979. Microatolls: review of form, origin and terminology.
Atoll Res. Bull. 224, 117.
Sun, D., Gagan, M.K., Cheng, H., Scott-Gagan, H., Dykoski, C.A., Edwards, R.L., Su, R.,
2005. Seasonal and interannual variability of the Mid-Holocene East Asian monsoon
in coral δ
18
O records from the South China Sea. Earth Planet. Sci. Lett. 237, 6984.
Suzuki, A., Gagan, M.K., Fabricius, K., Isdale, P.J., Yukino, I., Kawahata, H., 2003.
Skeletal isotope microproles of growth perturbations in Porites corals during the
19971998 mass bleaching event. Coral Reefs 22, 357369.
Swart, P.K., Coleman, M.L., 1980. Isotopic data for scleractinian corals explain their
palaeotemperature uncertainties. Nature. 283, 557559.
Trouet, V., Van Oldenborgh, G.J., 2013. KNMI climate Explorer: a web-based research
tool for high-resolution paleoclimatology. Tree Ring Res. 69, 313.
Wang, B., Wu, R.G., Fu, X.H., 2000. Pacic-East Asian teleconnection: how does ENSO
affect East Asian climate? J. Clim. 13, 15171536.
Wang, C., Wang, W., Wang, D., Wang, Q., 2006. Interannual variability of the South
China Sea associated with el Ni˜
no. J. Geophys. Res. 111, C03023.
Wang, B., Huang, F., Wu, Z., Yang, J., Fu, X., Kikuchi, K., 2009. Multi-scale climate
variability of the South China Sea monsoon: a review. Dyn. Atmos. Oceans 47,
1537.
Wang, X., Deng, W., Liu, X., Wei, G., Chen, X., Zhao, J.-X., Cai, G., Zeng, T., 2018. Super
instrumental El Ni˜
no events recorded by a Porites coral from the South China Sea.
Coral Reefs 37, 295308.
Webster, P.J., Yang, S., 1992. Monsoon and ENSO: selectively interactive systems. Q. J.
R. Meteorol. Soc. 118, 877926.
Woodroffe, C.D., Gagan, M.K., 2000. Coral microatolls from the Central Pacic record
late Holocene El Ni˜
no. Geophys. Res. Lett. 27, 15111514.
Woodroffe, C., McLean, R., 1990. Microatolls and recent sea level change on coral atolls.
Nature. 344, 531534.
Woodroffe, C.D., Beech, M.R., Gagan, M.K., 2003. Midlate Holocene El Ni˜
no variability
in the equatorial Pacic from coral microatolls. Geophys. Res. Lett. 30, 1358.
Xie, S.-P., Kosaka, Y., Du, Y., Hu, K., Chowdary, J.S., Huang, G., 2016. Indo-western
Pacic Ocean capacitor and coherent climate anomalies in post-ENSO summer: a
review. Adv. Atmos. Sci. 33, 411432.
Yan, H., Shao, D., Wang, Y., Sun, L., 2013. Sr/Ca prole of long-lived Tridacna gigas
bivalves from South China Sea: a new high-resolution SST proxy. Geochim.
Cosmochim. Acta 112, 5265.
Yan, H., Liu, C., Zhang, W., Li, M., Zheng, X., Wei, G., Xie, L., Deng, W., Sun, L., 2017.
ENSO variability around 2000 years ago recorded by Tridacna gigas δ
18
O from the
South China Sea. Quat. Int. 452, 148154.
Yu, K.-F., Zhao, J.-X., Wei, G.-J., Cheng, X.-R., Chen, T.-G., Felis, T., Wang, P.-X., Liu, T.-
S., 2005. δ
18
O, Sr/Ca and Mg/Ca records of Porites lutea corals from Leizhou
Peninsula, northern South China Sea, and their applicability as paleoclimatic
indicators. Palaeogeogr. Palaeoclimatol. Palaeoecol. 218, 5773.
Zhou, W., Chan, J.C.L., 2010. ENSO and the South China Sea summer monsoon onset.
Int. J. Climatol. 27 (2), 157167.
F. Tan et al.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
The El Niño–Southern Oscillation (ENSO) dominates interannual climate variability worldwide and has important environmental and socio‐economic consequences. However, determining the evolution of ENSO variability and its long‐term response to climate forcing remains an ongoing challenge owing to the limited instrumental records. In this study, we quantified ENSO variability via an empirically calibrated threshold and sliding variance windows using monthly sea‐surface temperature (SST) anomalies based on Porites coral Sr/Ca records from the Xisha Islands in the northern South China Sea. Instrumental SST anomalies from the Xisha Islands correctly captured increasing ENSO variability in the twentieth century, with ENSO detection skills similar to those for Niño3.4 regions. Coral Sr/Ca‐SST anomalies can also serve as sensitive and robust proxies for ENSO variability. Sub‐fossil coral Sr/Ca‐SST anomalies indicated intensified ENSO variability at the end of the Medieval Climate Anomaly (MCA) from 1149 to 1205 ± 4.9 (2σ) Common Era (CE). Combining our records with other ENSO‐sensitive proxy reconstructions from the tropical Pacific, we observed fluctuating ENSO variability during the MCA and intensified ENSO variability for the late MCA. Considering the fewer and low intensity fluctuations associated with external climate forcing and the absence of a coherent temporal correspondence of ENSO activity with solar irradiance and volcanic eruption during the MCA, we hypothesized that the internal dynamics of the climate system play a prominent role in modulating ENSO variability and its evolution, which is supported by unforced climate model simulations and coral reconstructions across the tropical Pacific.
Article
Full-text available
El Niño-Southern Oscillation (ENSO) is the largest source of interannual climate variability over the globe. Knowledge of ENSO variability back to the pre-instrumental period (and earlier) can help to enhance our understanding of its mechanisms and impacts. Here we use the coral record from the northern South China Sea, where the interannual climate variability is associated with ENSO activity, to reconstruct ENSO variability over the past 150 years. Our record, together with other ENSO chronologies, indicated not only the weakened ENSO activity between ∼1930s and ∼1960s but also the larger variations of El Niño activity relative to that of La Niña based on the sliding window method. The comparison between the SCS and Pacific ENSO band coral records revealed the relatively stronger climatic coupling between the SCS and the equatorial central-eastern Pacific compared with that between SCS and western Pacific, indicating the reliability of teleconnection relationship between the SCS climate and ENSO. Our result indicated that the coral δ¹⁸O record from the SCS could contribute to the pan-Pacific ENSO reconstructions.
Article
Full-text available
Giant clams (Tridacna) are the largest marine bivalves, and their carbonate shells can be used for high-resolution paleoclimate reconstructions. In this contribution, δ18Oshell was used to estimate climatic variation in the Xisha Islands of the South China Sea. We first evaluate sea surface temperature (SST) and sea surface salinity (SSS) influence on the modern resampled monthly (r-monthly) resolution of Tridacna gigas δ18Oshell. The results obtained reveal that δ18Oshell seasonal variation is mainly controlled by SST and appears to be insensitive to local SSS change. Thus, the δ18O of Tridacna shells can be roughly used as a proxy of local SST: a 1 ‰ δ18Oshell change is roughly equal to 4.41 ∘C of SST. The r-monthly δ18O of a 40-year-old Tridacna squamosa (3673±28 BP) from the North Reef of the Xisha Islands was analyzed and compared with the modern specimen. The difference between the average δ18O of the fossil Tridacna shell (δ18O =-1.34 ‰) and the modern Tridacna specimen (δ18O =-1.15 ‰) probably implies a warm climate, roughly 0.84 ∘C, 3700 years ago. The seasonal variation 3700 years ago was slightly lower than that suggested by modern instrumental data, and the transition between warm and cold seasons was rapid. Higher amplitudes of reconstructed r-monthly and r-annual SST anomalies imply an enhanced climate variability during this warm period. Investigation of the El Ninõ–Southern Oscillation (ENSO) variation (based on the reconstructed SST series) indicates reduced ENSO frequency but increased ENSO-related variability and extreme El Ninõ winter events 3700 years ago.
Article
Full-text available
El Niño events differ substantially in their spatial pattern and intensity. Canonical Eastern Pacific El Niño events have sea surface temperature anomalies that are strongest in the far eastern equatorial Pacific, whereas peak ocean warming occurs further west during Central Pacific El Niño events. The event types differ in their impacts on the location and intensity of temperature and precipitation anomalies globally. Evidence is emerging that Central Pacific El Niño events have become more common, a trend that is projected by some studies to continue with ongoing climate change. Here we identify spatial and temporal patterns in observed sea surface temperatures that distinguish the evolution of Eastern and Central Pacific El Niño events in the tropical Pacific. We show that these patterns are recorded by a network of 27 seasonally resolved coral records, which we then use to reconstruct Central and Eastern Pacific El Niño activity for the past four centuries. We find a simultaneous increase in Central Pacific events and a decrease in Eastern Pacific events since the late twentieth century that leads to a ratio of Central to Eastern Pacific events that is unusual in a multicentury context. Compared to the past four centuries, the most recent 30 year period includes fewer, but more intense, Eastern Pacific El Niño events.
Article
Full-text available
This paper describes version 3 of the Simple Ocean Data Assimilation (SODA3) ocean reanalysis with enhancements to model resolution, observation, and forcing datasets, and the addition of active sea ice. SODA3 relies on the ocean component of the NOAA/Geophysical Fluid Dynamics Laboratory CM2.5 coupled model with nominal 1/ 4° resolution.Ascheme has also been implemented to reduce bias in the surface fluxes. A 37-yr-long ocean reanalysis, SODA3.4.2, created using this new SODA3 system is compared to the previous generation of SODA (SODA2.2.4) as well as to the Hadley Centre EN4.1.1 no-model statistical objective analysis. The comparison is carried out in the tropics, the midlatitudes, and the Arctic and includes examinations of the meridional overturning circulation in the Atlantic. The comparison shows that SODA3.4.2 has reduced systematic errors to a level comparable to those of the no-model statistical objective analysis in the upper ocean. The accuracy of variability has been improved particularly poleward of the tropics, with the greatest improvements seen in the Arctic, accompanying a substantial reduction in surface net heat and freshwater flux bias. These improvements justify increasing use of ocean reanalysis for climate studies including the higher latitudes.
Article
Full-text available
Reconstructions of ocean temperatures prior to the industrial era serve to constrain natural climate variability on decadal to centennial timescales, yet relatively few such observations are available from the west Pacific Warm Pool. Here we present multiple coral-based sea surface temperature reconstructions from Yongle Atoll, in the South China Sea over the last ~1,250 years (762-2013 Common Era [CE]). Reconstructed coral Sr/Ca-sea surface temperatures indicate that the "Little Ice Age (1711-1817 CE)" period was ~0.7°C cooler than the "Medieval Climate Anomaly (913-1132 CE)" and that late 20th century warming of the western Pacific is likely unprecedented over the past millennium. Our findings suggest that the Western Pacific Warm Pool may have expanded (contracted) during the Medieval Climate Anomaly (Little Ice Age), leading to a strengthening (weakening) of the Asian summer monsoon, as recorded in Chinese stalagmites.
Article
Full-text available
The 2–7-year periodicities recorded in fossil coral records have been widely used to identify paleo-El Niño events. However, the reliability of this approach in the South China Sea (SCS) has not been assessed in detail. Therefore, this paper presents monthly resolution geochemical records covering the period 1978–2015 obtained from a Porites coral recovered from the SCS to test the reliability of this method. The results suggest that the SCS coral reliably recorded local seawater conditions and the super El Niño events that occurred over the past 3 decades, but does not appear to have been sensitive enough to record all the other El Niños. In detail, the Sr/Ca series distinctly documents only the two super El Niños of 1997–1998 and 2014–2016 as obvious low values, but does not match the Oceanic Niño Index well. The super El Niño of 1982–1983 was identified by the growth hiatus caused by the coral bleaching and subsequent death of the coral. Three distinct stepwise variations occur in the δ¹³C series that are coincident with the three super El Niños, which may be related to a substantial decline in endosymbiotic zooxanthellae density caused by the increase in temperature during an El Niño or the selective utilization of different zooxanthellaes that was required to survive in the extreme environment. The increase in rainfall and temperatures over the SCS during El Niños counteracts the effects on seawater δ¹⁸O (δ¹⁸Osw) and salinity; consequently, coral Δδ¹⁸O series can be used as a proxy for δ¹⁸Osw and salinity, but are not appropriate for identifying El Niño activity. The findings presented here suggest that the method to identify paleo-El Niño activity based on the 2–7-year periodicities preserved in the SCS coral records might not be reliable, because the SCS is on the edge of El Niño anomalies due to its great distance from the central equatorial Pacific and the imprints of weak and medium strength El Niño events may not be recorded by the corals there.
Article
Full-text available
The Indo-Pacific coral Diploastrea heliopora reveals regional multi-decadal to centennial scale climate variability using coral carbonate δ18O (δ18Oc) as a combined proxy for sea surface temperature (SST) and sea surface salinity (SSS). However, to assess the coral's full potential in resolving climatic events, an independent SST proxy would be more advantageous. We examined both Sr/Ca and δ18O of Diploastrea against an adjacent Porites lobata core collected from northeast Luzon, Philippines. Winter Sr/Ca data from Diploastrea show a significant correlation to SST (r = -0.41, p < 0 .05, RMSR = 0.81 °C) and provide a proxy with similar sensitivity as Porites (r = -0.57, p < 0.05, RMSR = 0.62 °C). An inter-species SST record is shown to be robust and used for a reconstruction of the Pacific Decadal Oscillation (PDO) during boreal winter (r = -0.70, p = 0.02). While we were unable to generate a robust Diploastrea δ18O-SSS calibration at inter-annual timescale, the freshening trend towards the present, commonly observed in the region, is qualitatively captured in Diploastrea δ18O. Comparison with Porites δ18O and instrumental SSS records shows that the magnitude of freshening is consistent between coral species. Wet and dry season Porites δ18O provide support for the relative influence of El Niño Southern Oscillation (ENSO) events and local precipitation to SSS variability at our site. The multi-proxy, multi-species approach of this study further strengthens the evidence for Diploastrea as an alternate climate archive in the Indo-Pacific region, and seals its potential in helping resolve less understood global-scale climate phenomena.
Article
The spatial variability of Holocene relative sea level (RSL) in the South China Sea is unknown, with data restricted to Thailand, the Malay Peninsula, and a few other isolated sites. In this study, we present new continuous RSL records for Borneo using surveyed and U–Th dated coral microatolls from four sites in western Sarawak. The record spans 450 years of RSL from 7450 to 7000 yr BP. Our data suggest that RSL was higher than present and rapid RSL rise had ceased by 7450 yr BP. We compare these RSL reconstructions with a regional model of glacial-isostatic adjustment (GIA). The RSL reconstructions from three sites off the coast of Sarawak show a spatial gradient opposite to that predicted by the GIA model. This disagreement can best be explained by tectonic deformation since 7000 yr BP, which was previously unrecognized. We propose vertical land motion of 0.7–1.45 m due to slip on the Serabang fault, which runs between our four sites. This slip may have occurred in response to the loading of the Sunda Shelf by rising sea level.