mid- to late holocene changes in tropical atlantic temperature seasonality and interannual to...
TRANSCRIPT
Earth and Planetary Science Letters 331-332 (2012) 187–200
Contents lists available at SciVerse ScienceDirect
Earth and Planetary Science Letters
j ourna l homepage: www.e lsev ie r .com/ locate /eps l
Mid- to late Holocene changes in tropical Atlantic temperature seasonality andinterannual to multidecadal variability documented in southern Caribbean corals
Cyril Giry a,⁎, Thomas Felis a, Martin Kölling a, Denis Scholz b,1, Wei Wei c,Gerrit Lohmann c, Sander Scheffers d
a MARUM — Center for Marine Environmental Sciences, University of Bremen, 28359 Bremen, Germanyb Heidelberger Akademie der Wissenschaften, Forschungsstelle Radiometrie, 69120 Heidelberg, Germanyc Alfred Wegener Institute for Polar and Marine Research, 27570 Bremerhaven, Germanyd School of Environmental Science and Management, Southern Cross University, Lismore, NSW 2480, Australia
⁎ Corresponding author.E-mail address: [email protected] (C. Giry).
1 Now at Institute for Geosciences, University of Mainz,
0012-821X/$ – see front matter © 2012 Elsevier B.V. Aldoi:10.1016/j.epsl.2012.03.019
a b s t r a c t
a r t i c l e i n f oArticle history:Received 13 May 2011Received in revised form 20 December 2011Accepted 13 March 2012Available online 10 April 2012
Editor: P. DeMenocal
Keywords:coral Sr/Casouthern Caribbean climatesea surface temperatureseasonalityinterannual to multidecadal variabilityENSO teleconnection
Proxy reconstructions of tropical Atlantic sea surface temperature (SST) that extend beyond the period ofinstrumental observations have primarily focused on centennial tomillennial variability rather than on sea-sonal to multidecadal variability. Here we present monthly-resolved records of Sr/Ca (a proxy of SST) fromfossil annually-banded Diploria strigosa corals from Bonaire (southern Caribbean Sea). The individual coralsprovide time-windows of up to 68 years length, and the total number of 295 years of record allows forassessing the natural range of seasonal to multidecadal SST variability in the western tropical Atlantic dur-ing snapshots of the mid- to late Holocene. Comparable to modern climate, the coral Sr/Ca records revealthat mid- to late Holocene SST was characterised by clear seasonal cycles, persistent quasi-biennial andprominent interannual as well as inter- to multidecadal-scale variability. However, the magnitude of SSTvariations on these timescales has varied over the last 6.2 ka. The coral records show increased seasonalityduring themid-Holocene consistent with climate model simulations indicating that southern Caribbean SSTseasonality is induced by insolation changes on orbital timescales, whereas internal dynamics of the climatesystem play an important role on shorter timescales. Interannual SST variability is linked to ocean–atmosphere interactions of Atlantic and Pacific origin. Pronounced interannual variability in the westerntropical Atlantic is indicated by a 2.35 ka coral, possibly related to a strengthening of the variability of theEl Niño/Southern Oscillation throughout the Holocene. Prominent inter- to multidecadal SST variability isevident in the coral records and slightly more pronounced in the mid-Holocene. We finally argue that ourcoral data provide a target for studying Holocene climate variability on seasonal and interannual to multidecadaltimescales, when using further numerical models and high-resolution proxy data.
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
Improving skills in predicting climate phenomena that have largesocio-economic impacts in the tropical Atlantic region (e.g., hurricaneactivity, Nordeste and Sahel rainfalls) relies on a better understandingof the forcingmechanisms that control regional sea surface temperature(SST) variability on seasonal and interannual to multidecadal time-scales. Unlike the tropical Pacific, the tropical Atlantic is dominated bythe competing influence of modes of climate variability emanatingfrom tropical and extratropical oceans (e.g., Czaja, 2004; Czaja et al.,2002; Hurrell et al., 2006;Marshall et al., 2001)modulating the strengthof the trade winds that affects SST and the distribution and intensity ofrainfall over the surrounding landmasses (Chiang et al., 2002; Enfield
55128 Mainz, Germany.
l rights reserved.
and Mayer, 1997; Giannini et al., 2000, 2001b). Mechanisms involvedin SST variability in the Caribbean region are linked to sea level pressure(SLP) variations in the North Atlantic region (Hastenrath, 1984) mainlythrough changes in surface winds that in turn, affect the dynamics ofthe Atlantic Warm Pool (AWP) (Wang et al., 2007, 2008a). The AWPis a large body of water warmer than 28.5 °C (Fig. 1) which experiencesmodulations on several timescales including seasonal (Wang andEnfield, 2003), interannual (Wang et al., 2006, 2008a) andmultidecadal(Wang et al., 2008b). During itsmaximumextent in boreal summer, theAWP affects summer climate of the western Hemisphere (Wang et al.,2006, 2007) and can induce thermal stress to coral reefs, as observedin the Caribbean in 2005 (Eakin et al., 2010). On interannual andmulti-decadal timescales, the AWP is modulated by the El Niño/SouthernOscillation (ENSO) and the Atlantic Multidecadal Oscillation (AMO)(Wang et al., 2008a), respectively. The knowledge of tropical AtlanticSST variability on these timescales is based on short and sparse instru-mental data. Proxy reconstructions from the tropical Atlantic that
18
20
22
24
26
28
30
Sea
Sur
face
Tem
pera
ture
, SS
T (
°C)
0°10
°N20
°N30
°N
Latit
ude
110°W 100°W 90°W 80°W 70°W 60°W 50°W 40°W 30°W 20°W 10°W
Longitude
0°10
°N20
°N30
°N
Latit
ude
28.5
28.5
28.529
29.5
28.5
28.528.5
29
29Bonaire
Bonaire
March
October
29
Fig. 1. Sea surface temperature (SST) maps for March (top) and October (bottom) representing the months of minimum and maximum SST around Bonaire (white circle) (Smithet al., 2008). Maps are for the year 2010. Thin contour lines reveal SST warmer than 28.5 °C that is characteristic of the Atlantic Warm Pool. Black arrows illustrate the directionand strength of dominant surface winds associated with the North Atlantic subtropical high.
188 C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
extend beyond the observational period have primarily focused on long-term changes in climate variability rather than focusing on short-termvariability (i.e., seasonal to multidecadal) that is more relevant forpredicting hazardous climate phenomena in this region.
We use a unique high-resolution climate archive, fossil annually-banded Diploria strigosa brain corals from the island of Bonaire (south-ern Caribbean Sea), to investigate patterns of short-term SST variabilityrecorded in the strontium (Sr) to calcium (Ca) ratio of their skeletons(Beck et al., 1992; Smith et al., 1979) during the Holocene. Variationsof Sr/Ca in massive corals are an established proxy for SST variabilityon seasonal (Felis et al., 2004; Gagan et al., 1998) and interannual tomultidecadal timescales (Felis et al., 2009; Linsley et al., 2000), evenduring specific seasons of the year (Felis et al., 2010; Hetzinger et al.,2006; Kuhnert et al., 2005). Recent studies have shown that Sr/Ca re-cords from modern D. strigosa corals are excellent proxies for seasonaland interannual to multidecadal variations in local to regional SST inthe Caribbean (Hetzinger et al., 2006, 2010). Our well-preserved fossilD. strigosa corals from Bonaire range in age up to 6220 years before pre-sent (BP) and provide unusually long time-windows of up to 68 years.U-series ages reveal that the individual corals represent snapshotsthroughout the mid- to late Holocene, which allows for evaluatingreconstructed southern Caribbean SST variability under different cli-mate background states.
It has been suggested that Holocene temperature evolution islargely driven by seasonal changes in insolation received in the north-ern and southern Hemisphere (Kim et al., 2004; Lorenz et al., 2006;Wanner et al., 2008). The mid-Holocene was characterised by a stron-ger seasonality of insolation in the Northern Hemisphere which thensteadily decreased towards the present (Berger, 1978). Holocenechanges in insolation may have influenced the spatial distribution ofsurface heating and therefore, may have resulted in direct changesin temperature seasonality or in different oceanic and atmosphericcirculation regimes. In addition, on shorter timescales, further mech-anisms linked to both insolation changes on suborbital timescales(i.e., solar variability) and internal variability of the climate systemmay have played a role in modulating Holocene climate change. Stud-ies of past tropical Atlantic SST variability have essentially relied onforaminiferal δ18O and Mg/Ca as well as alkenones in marine
sediments providing insights into decadal to millennial scale variabil-ity (e.g., Lea et al., 2003; Rühlemann et al., 1999; Tedesco and Thunell,2003). Our Bonaire coral Sr/Ca records allow us to investigate south-ern Caribbean SST variability at subseasonal resolution. We identifyyet non-investigated fundamental features of tropical Atlantic climatevariability and assess the natural range of seasonality and interannualto multidecadal SST variability in the southern Caribbean Sea duringthe mid- to late Holocene.
2. Material and methods
2.1. Climatic setting of the study area
Bonaire (Netherlands) is an open-ocean island in the southern Ca-ribbean Sea (~12°10′N, 68°18′W), located ~100 km north of Venezuela(Fig. 1). The island has a tropical-arid climate characterised by low an-nual rainfall (~500 mm/yr) and is influenced by easterly trade winds.These also referred to as Caribbean low-level jet (CLLJ) (Muñoz et al.,2008; Wang and Lee, 2007) vary with Caribbean SLP anomalies thatare connected to the North Atlantic subtropical high. Wang (2007)found that the correlation between CLLJ and southern Caribbean SSTduring summer and winter is −0.7 and −0.5, respectively. Therefore,cold SST anomalies at Bonaire are associated with high SLP anomaliesthat are consistent with a strong CLLJ and vice versa.
The monthly SST time series representative of a 2°×2° gridboxcentred on Bonaire (12°N; 68°W) (Smith et al., 2008) reveals pro-nounced annual cycles with a mean maximum of 28.5±0.5 °C inSeptember/October and a mean minimum of 25.6±0.6 °C in Febru-ary/March for the period 1910–2000 (Smith et al., 2008), indicatinga SST seasonality of 2.9 °C. Observational studies have shown thatnorthern tropical Atlantic SST variability is remotely affected by thecompeting influence of ENSO and the North Atlantic Oscillation(NAO) on the strength of surface winds inducing changes in latentheating that in turn, generate SST anomalies (Czaja et al., 2002). Notall ENSO events manifest in exactly the same manner in the CaribbeanSea (Wang and Lee, 2007). There are some El Niño events that resultin significant warm SST anomalies, while during other El Niño yearsno significant effects on Caribbean SST are observed. Observational
189C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
studies have suggested that theNAO can play a key role in inhibiting theENSO effect in the Caribbean resulting in a complex signal. For instance,a positive NAOphase implies stronger thannormal surfacewinds enter-ing the Caribbean thus, generating negative SST anomalies (Giannini etal., 2001a). El Niño-related atmospheric anomalies are known to force alagged warming of tropical North Atlantic SST through the weakeningof the northeasterly trade winds and consequent reduction of surfacefluxes (Giannini et al., 2004). Mature phases of ENSO occurring usuallyin December–January–February (DJF) affect remotely, via the so-called“atmospheric-bridge”, northern tropical Atlantic SST in spring(March–April–May: MAM) with a lag of 3–5 months (Enfield andMayer, 1997). As shown in Fig. 2, the high correlation (r=0.61) be-tween the Niño3.4 index (DJF) and MAM Bonaire SST confirms thatunder present-day conditions ENSO teleconnections have a delayedinfluence on southern Caribbean SST. Moreover, given that ENSO af-fects primarily Bonaire SST anomalies in spring (i.e., the coldest sea-son), it is indeed evident that the remote influence of ENSOaffects Bonaire SST seasonality. El Niño events result in reduced am-plitude and La Niña events result in increased amplitude of BonaireSST seasonality (Fig. 2). It is generally believed that interannual SSTvariability in the tropical North Atlantic is strongest (weakest) in bo-real spring (early fall) which reflects primarily the remote forcing byNAO and ENSO, but also the strength of local ocean–atmosphere cou-pling (Czaja, 2004).
2.2. Coral collection and diagenetic investigations
Along the coast of Bonaire, prominent deposits of hurricane and/ortsunami origin occur. Those deposits consist of reef boulders and fossilcoral colonies of Holocene age (Scheffers, 2004; Scheffers and Scheffers,2006; Scheffers et al., 2009). Six fossil D. strigosa colonies recoveredfrom these coastal deposits were drilled along the major growth axis.Similarly, two dead modern D. strigosa colonies from a storm depositformed by hurricane Ivan in 2004 (Scheffers and Scheffers, 2006) were
1930 1940 1950 1960 19
Year
24
25
26
27
28
29
30
SS
T (
°C)
1
2
3
4
5
SS
T s
easo
nal
ity
(°C
)M
ax(0
)-M
in(0
)
Bonaire SST
Niño3.4 (DJ
Fig. 2.Monthly sea surface temperature (SST) for a 2°×2° gridbox centred on Bonaire (12°N;time series (blue line) is positively correlated (r=0.61) with the December–January–Februaseasonality (bottom) calculated as the difference between monthly maximum and minimu(DJF) (r=−0.48). Grey arrows indicate major El Niño/La Niña events that were followed b
drilled. For calibration, a core froma livingD. strigosawas retrieved. Sam-pling sites are displayed in Supplementary Fig. 1. Coral cores were cutalong the growth direction using a circular diamond saw and 6–7 mmthick slabs were sliced from individual half-cores. Slabs were cleanedfor 10 min in an ultrasonic bath, and then placed in a drying ovenat 50 °C overnight and subsequently X-rayed for 7 min at 45 kV. X-radiographs of the coral slabs are shown in Fig. 3.
All corals used in this studywere investigated for possible diageneticalteration of their skeleton using differentmethods. First, the slabswereX-rayed in order to identify both the annual density-banding patternand areas or patches of anomalously high density potentially reflectingdiagenetic alterations such as secondary aragonite or calcite infillings.Slabs that did not show suspicious areas were further investigated forpotential diagenetic textures by powder X-ray diffraction (XRD) analy-sis, Scanning Electron Microscope (SEM) imaging and thin-sectionmicroscopy, using a sample representative for the entire slab thatwas extracted from an area in the centre of the slab. XRDwas appliedin order to quantify the calcite content of the sample. As crystals ofinorganic aragonite cannot be detected with XRD, SEM and thin-section microscopy were used to identify the presence of secondaryaragonite.
2.3. 230Th/U-dating of corals
230Th/U-dating was conducted on individual coral samples at theHeidelberg Academy of Sciences, Heidelberg, Germany. Analyses wereperformed with a Finnigan Thermal Ionization Mass SpectrometerMAT 262 RPQ. Sample preparation and analytical details are similar asdescribed in Scholz et al. (2004). Calibration of the added U and Thspike solutions is described in Hoffmann et al. (2007). Ages were calcu-lated using the half-lives of Cheng et al. (2000). All ages were cor-rected for the effect of detrital contamination assuming a bulkearth 232Th/238Uweight ratio of 3.8 and secular equilibrium between238U, 234U and 230Th. However, this correction is insignificant for allcorals presented here. The reliability of the determined 230Th/U-ages
70 1980 1990 2000 2010
s AD
-2
0
2
Niñ
o3.
4 (D
JF)
SST (MAM)
F)
Bonaire SST seasonality
68°W) (ERSSTv3b; Smith et al., 2008) (top). The March–April–May (MAM) Bonaire SSTry (DJF) Niño3.4 index (Kaplan et al., 1998; Reynolds et al., 2002) (middle). Bonaire SSTm SST values in a single year, displays an inverse relationship with the Niño3.4 indexy anomalous SST seasonality around Bonaire.
BON-3-E6.22 ka coral
BON-6-A2.35 ka coral
BON-9-B1957 AD coral
BON-0-A2009 AD coral
10 cm
10 c
m
Gro
wth
dir
ecti
on
BON-7-A3.79 ka coral
BON-7-B3.83 ka coral
3831
3
4 a
BP
±
BON-4-G4.27 ka coral
4268
2
5 a
BP
±6218
4
6 a
BP
±
3793
2
9 a
BP
±
BON-20-A1.85 ka coral
1837
1
8 a
BP
±
2347
4
9 a
BP
±
Gro
wth
dir
ecti
on
1957
5
AD
±
Jan 2009
Jan 2004
Jan 1999
Jan 1994
10 cm
10 c
m
a
b
BON-9-A1912
1912
8
AD
±
AD coral
Fig. 3. X-radiograph positive prints of slabs of a) modern and b) fossil Bonaire Diploria strigosa corals reveal the annual density banding pattern. Red lines indicate the position of themicrosampling transects along the centre of the thecal walls. Red triangles indicate where samples for U/Th dating and diagenetic screening were taken. Black arrows indicate5-year intervals based on annual high-density bands and grey shading marks changes in growth pattern (i.e., lower annual extension rate and thickening of the thecal walls) atthe top of coral BON-9-A.
190 C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
was checked using established criteria (Scholz and Mangini, 2007),such as initial (234U/238U) in agreement with (234U/238U) of modernseawater (i.e., 1.1466±0.0025, Robinson et al., 2004), U concentrations
comparable to modern corals of the same species and negligible calcitecontent. These criteria are fulfilled for all studied corals except for sam-ple BON-7-A, which shows an initial (234U/238U) slightly lower than the
191C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
seawater value. However, the effect on the 230Th/U-age should benegligible for such a young coral (Scholz and Mangini, 2007). Thus, allages are considered as reliable. U/Th ages are presented in years before1950 AD, with 2σ errors (Supplementary Table 1).
2.4. Microsampling, Sr/Ca analyses, chronology construction, time seriesanalyses
Based on the annual-density banding pattern inferred from X-radiographs, we targeted a sampling resolution of about 12 samplesper year. A combination of naked-eye observations of the slab-surface and density variations inferred from X-radiographs was usedto identify the position and orientation of the dense thecal walls(Fig. 3). Microsamples of skeletal powder were extracted by carefullydrilling with constant sampling depth along the centre of the thecalwall using a 0.6 mm diameter bit. This careful sampling strategy hasbeen successfully deployed and proved to yield the best signal of ambi-ent SST variations (Giry et al., 2010a) and allowing to evaluate Sr/Cavariability at near-monthly resolution.
For Sr/Ca analyses, 0.25–0.30 mg coral powder was dissolved in7 mL 2% suprapure HNO3 containing 1 ppm Sc (Scandium) as internalstandard. Measurements were performed on a Perkin-Elmer Optima3300R simultaneous radial ICP-OES using a CETAC U5000-AT ultra-sonic nebuliser at the Faculty of Geosciences, University of Bremen,Germany. Element wavelengths were detected simultaneously in3 replicates (Ca 317.933 nm, Ca 422.673 nm, Sr 421.552 nm, Sc361.383 nm and Mg 280.271 nm). Ca concentrations measured onan atomic line (422.673 nm) were averaged with the concentrationsfrom an ionic line (317.933 nm) to compensate for possible sensitivitydrift in a radial ICP-OES. Measurements of a laboratory coral standardafter each sample allowed offline correction for instrumental drift.Relative standard deviation of the Sr/Ca determinations was betterthan 0.2%. Twenty five aliquots of the Porites coral powder referencematerial JCp-1 (Okai et al., 2002) were treated like samples and theaverage Sr/Ca value obtained during the course of this study was8.919±0.008 mmol/mol.
Age models for coral Sr/Ca records are based on both annualdensity-banding patterns inferred from X-radiographs and annual cy-cles in Sr/Ca. We assumed that the timing of the annual SST cycle didnot change during the investigated period, which is consistent with aclimate model simulation (see Section 4.2). Therefore, maximum andminimum coral Sr/Ca values in any given year were assigned to the onaverage coldest (February/March) and warmest months (September/October) according to present-day SST (Smith et al., 2008), respectively.Linear interpolation between these tie points resulted in time seriesthat were subsequently interpolated to a monthly resolution followingestablished procedures (Felis et al., 2009; Giry et al., 2010b). Skeletalannual growth rates were calculated as the distance from a maximumcoral Sr/Ca value in a given year to themaximum value of the following
Table 1Details of the coral Sr/Ca records used in this study. Calcite contents were measured with X
Sample ID Age yr BP/AD(± 1σ)
Annual growth rate(cm/yr)b
Record length(yr)
Annual Mean(mmol/mol)
Modern coralsBON-0-A (living) 2009 AD 1.10 15 yr 9.208BON-9-B (dead) 1957±5 AD 0.75 13 yr 9.062BON-9-A (dead) 1912±8 AD 0.66 30 yr 9.079
Fossil coralsBON-20-A 1837±18 yr BP 0.46 7 yr 9.172BON-6-A 2347±49 yr BP 0.61 68 yr 9.183BON-7-A 3793±29 yr BP 0.94 39 yr 9.188BON-7-B 3831±34 yr BP 0.71 33 yr 9.268BON-4-G 4268±25 yr BP 0.72 23 yr 9.232BON-3-E 6218±46 yr BP 0.63 67 yr 9.295
year. Mean coral Sr/Ca values were calculated by averaging monthlydata of each year and subsequently averaging the resulting annualmeans of all years. Coral Sr/Ca seasonality is defined as the differencebetween maximum and minimum monthly coral Sr/Ca values of agiven year corresponding to the coldest and warmest months, respec-tively. Mean coral Sr/Ca seasonality was calculated by averaging thecoral Sr/Ca seasonality of all years on record. We have applied multita-per method (MTM) spectral analysis with a red noise null hypothesis(Ghil et al., 2002) to the detrended and normalised coral Sr/Ca andinstrumental SST records with the mean annual cycle removed.
2.5. Model description and experimental setup
The numerical experiments were performed with the coupled gen-eral circulation model COSMOS consisting of the atmospheric modelECHAM5 (Roeckner et al., 2003), ocean model MPIOM (Marslandet al., 2003), and dynamical vegetation model JSBACH (Raddatz et al.,2007) developed mainly at the Max-Planck-Institute for Meteorology.The atmospheric model has a resolution of T31 (3.75°×3.75°) in hor-izontal and 19 vertical hybrid sigma pressure levels. The ocean modelhas a 3°×1.8° averaged horizontal grid with 40 unevenly spaced ver-tical levels with higher resolution around Greenland and Antartica(Marsland et al., 2003). We carried out two experiments: a pre-industrial (CTL) experiment and a mid-Holocene one (6 ka), by pre-scribing the appropriate orbital parameters and greenhouse gases,identical to those used in the Paleoclimate Modeling IntercomparisonProject (PMIP) (Crucifix et al., 2005). Details of the model experimentsare described in Wei et al. (2012).
3. Results
3.1. Preservation of coral skeletons
The X-radiographs of individual coral slabs (Fig. 3) reveal clearannual-density banding patterns and do not indicate anomalous den-sity patches that are typical for diagenetic alterations of the skeleton(Hendy et al., 2007). XRD analysis performed on fossil corals indicatesb0.5% calcite content similar to that of the modern corals (Table 1).Thin sections indicate an excellent preservation of primary porosityof all coral skeletons, with no evidence for secondary aragonite or cal-cite cements (Fig. 4). SEM analysis reveals that the skeletal elementsof all corals have smooth surfaces free of both secondary overgrowthand dissolution patterns (Fig. 4). Dissepiments and especially thedense thecal walls, the targeted skeletal element for microsampling,are extraordinarilywell preservedwith no obvious differences betweenmodern and fossil corals. Hence, our thorough diagenetic screeningindicates that all corals used in this study have a pristine aragoniticskeleton.
-ray diffraction. Note that modern corals do not temporally overlap.
Sr/Ca STD(2σ)
Standard error of the mean(2SE)
Sampling resolution(samples/yr)
Calcite content(%)
0.052 0.014 14.1 ≤0.250.050 0.014 14.6 ≤0.50.100 0.018 11.8 ≤0.2
0.020 0.008 15.3 ≤0.350.064 0.008 11.2 ≤0.50.072 0.012 11.3 ≤0.20.074 0.012 13.0 ≤0.50.102 0.022 10.8 ≤0.50.108 0.014 11.4 ≤0.5
1 mm 100 µm
1 mm 100 µm
100 µm1 mm
100 µm500 µm 0.2 mm 0.2 mm
BO
N-2
0-A
1.85
ka
cora
l100 µm1 mm
BO
N-0
-A20
09 A
D c
ora
l0.2 mm 0.2 mm
100 µm1 mm
BO
N-9
-B19
57 A
D c
ora
l
0.2 mm 0.2 mm
50 µm200 µm
BO
N-9
-A19
12 A
D c
ora
l
0.2 mm 0.2 mm
100 µm1 mm
BO
N-6
-A2.
35 k
a co
ral
1 mm 1 mm
1 mm 200 µm
BO
N-7
-A3.
79 k
a co
ral
0.2 mm 0.2 mm
BO
N-7
-B3.
83 k
a co
ral
0.2 mm 0.2 mm
BO
N-4
-G4.
27 k
a co
ral
0.2 mm 0.2 mm
BO
N-3
-E6.
22 k
a co
ral
1 mm 1 mm
192 C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
193C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
3.2. Modern coral Sr/Ca-SST relationships
Monthly and annual Sr/Ca-SST regression slopes from theD. strigosacollected live in Bonaire (BON-0-A) are assessed for the period 1993 to2009. Monthly Sr/Ca correlates significantly with satellite SST data(Reynolds et al., 2002). The corresponding regression equation is:
Sr=Ca mmol=molð Þ ¼ 10:15 �0:06ð Þ−0:034 �0:002ð Þ� SST r2 ¼ 0:52; p b0:0001; N ¼ 188
� �
The regression equation for annual mean Sr/Ca values is:
Sr=Ca mmol=molð Þ ¼ 10:6 �0:5ð Þ−0:050 �0:019ð Þ� SST r2 ¼ 0:33; p ¼ 0:018; N ¼ 16
� �:
Similar relationships are assessed for the two dead modern corals.Since U/Th ages of these two corals do not provide enough timeconstraint, we relied on the maximum correlation found betweenSr/Ca and SST data (ERSST, Smith et al., 2008) by moving the coraltime-window within the range given by age uncertainty. The max-imum correlation betweenmonthly coral Sr/Ca of BON-9-B(BON-9-A) and SST of r=−0.72(−0.55) is found for the period 1957–1970(1904–1935). For the same period, correlation for annual meansis r=−0.55(−0.37). The slopes of themonthly (annual) Sr/Ca-SST cal-ibration are −0.040(−0.046) and −0.036(−0.058) mmol/mol/°C forBON-9-B and BON-9-A, respectively. These values are comparable tothe ones previously published from Caribbean D. strigosa (Hetzingeret al., 2006). However, because the Hetzinger et al. (2006) coral Sr/Ca-SST relationships are based on a longer record (~41 years) we considerthem as more robust, and use the corresponding monthly and annualrelationships of −0.042 and −0.066 mmol/mol/°C for the interpreta-tion of our coral Sr/Ca records.
3.3. Coral Sr/Ca-based SST variations throughout the mid- to late Holocene
The monthly resolved coral Sr/Ca records of three modern and sixfossil Bonaire corals provide a total of 295 years of record since themid-Holocene with individual time windows of up to 68 years length(Fig. 5a). The records show clear annual cycles that correlate with thecorresponding annual density-banding pattern in the coral skeleton.Therefore, robust internal chronologies could be established enablingaccurate assessment of seasonal to multidecadal SST variability. De-tails of coral Sr/Ca records used in this study are displayed in Table 1.
3.3.1. Coral Sr/Ca-based changes in mean SSTThe mean Sr/Ca values of the individual corals, based on monthly-
resolved records, are shown in Fig. 6. The three corals which grewduring the last century exhibit between-colony offsets in mean Sr/Ca values ranging from 9.062 to 9.208 mmol/mol. Although the mod-ern coral records do not temporally overlap, such between-colonyoffsets in mean Sr/Ca cannot be explained by corresponding decadalSST changes at Bonaire (Fig. 5a). Similar intercolony offsets havebeen reported for other massive corals at various locations (Abramet al., 2009; Felis et al., 2004) and may be related to biological effects,local effects and other unknown factors influencing the incorporationof Sr into the coral skeleton (e.g., Cohen et al., 2002; de Villiers et al.,1994; de Villiers et al., 1995; Weber, 1973). To address the resultinguncertainties associated with mean SST estimates from individualfossil D. strigosa coral colonies from Bonaire, we applied the combinederror, following established procedures (Abram et al., 2009). Thiserror ascribed to each fossil coral Sr/Ca-SST estimate is derived from
Fig. 4. Scanning Electron Microscope images (left and middle left) and polarised- (middle righcorals. Well-preserved skeletons are observed for both modern and fossil corals.
the combination (root of the sum of the squares) of 1) the standarddeviation (2σ) of the mean Sr/Ca value of the three modern coralsand 2) the standard error (2SE) of the mean Sr/Ca of each coral re-cord. The Bonaire intercolony variability in mean Sr/Ca has a standarddeviation (1σ) of 0.079 mmol/mol. Using the annual relationship of−0.066 mmol/mol/°C (Hetzinger et al., 2006), this corresponds to a1σ uncertainty of ~1.2 °C.
The mean coral Sr/Ca values of all nine corals suggest a warmingtrend in the southern Caribbean since the mid-Holocene (Fig. 6).This is true except for the most recent coral which has a mean Sr/Cavalue similar to the late Holocene corals. However, the overall trendin mean Sr/Ca cannot be attributed to any trend in mean coral growthrates (de Villiers et al., 1995; Goodkin et al., 2007; Goreau, 1977)(Table 1) and thus, is interpreted to reflect SST changes. Consideringthe combined error ascribed to each fossil coral, the cooler conditionsfound at 6.22 ka BP are significant with respect to the average Sr/Cavalue given by three modern corals. This suggests that mid-Holocenesouthern Caribbean SST was at least 0.3 °C cooler than today.
3.3.2. Coral Sr/Ca-based SST seasonalityThe Sr/Ca records of the three modern corals indicate a mean SST
seasonality of 2.8±0.2 °C (1SE) ranging from 2.4 to 3.0 °C (Fig. 7),based on the monthly Sr/Ca-SST relationship of −0.042 mmol/mol/°C(Hetzinger et al., 2006), satisfactorily documenting the present-daySST seasonality around Bonaire of 2.9±0.1 °C (1SE) (1910–2000)(Smith et al., 2008). The between-colony differences in Sr/Ca-SST sea-sonality cannot be attributed to actual differences in SST seasonalityover the last century (Fig. 5) nor to different sampling resolution(Table 1) but are consistent with those reported from other massivecorals (Felis et al., 2004). Consequently, the combined error (Abramet al., 2009) that takes into account modern intercolony differences incoral Sr/Ca seasonality is considered for our estimates of HoloceneSST seasonality (Fig. 7).
The coral Sr/Ca records indicate a mean SST seasonality of 3.1±0.2 °C (1SE) throughout the mid- to late Holocene. The 2.35 ka andthe 6.22 ka corals indicate high SST seasonality that is not associatedwith higher sampling resolution (Table 1). While the SST seasonalityat 6.22 ka BP (3.3±0.8 °C) is not significantly higher than modernvalues when considering our thorough error estimates, the 2.35 kacoral indicates a seasonality of 4.0±0.8 °C that is significantly higherthan both modern and mean mid- to late Holocene values. Spectralanalysis of the coral Sr/Ca-SST seasonality records reveals significantinterannual variability for almost all time intervals (SupplementaryFig. 2).
3.3.3. Coral Sr/Ca-based interannual SST variabilityInstrumental data reveal that SST variations around Bonaire were
characterised by prominent quasi-biennial and interannual variabilityduring the last century, with significant periodicities of 2.2, 3.8, and5.2 years. The monthly coral Sr/Ca records indicate similar SST variabil-ity at quasi-biennial and interannual periodicities during the mid- tolate Holocene which is often superimposed on inter- to multidecadalvariations that are more pronounced than in the modern SST (Fig. 5).Significant spectral peaks in the quasi-biennial (~2-year) band aredetected in almost all coral Sr/Ca records. Significant interannual vari-ability ranging from3 to 6 years is identified inmost of the coral records.However, we note that the variance of interannual variability is relative-ly weak in the 6.22 ka coral record compared to that of 2.35 ka recordand present-day SST variability, suggesting that the magnitude of inter-annual SST variability was reduced during the mid-Holocene (Fig. 5).The 2.35 ka coral record displays significant variability at a prominent
t) and normal-light microscopy (right) of thin sections of Bonaire Diploria strigosa
45*
9.4
9.2
9.0
Co
ral S
r/C
a(m
mo
l/mo
l)
-4-2024
Sr/
Ca-
SS
Tan
om
aly
(°C
)
1837 ± 18 a BP
9.2
9.0
8.8
Co
ral S
r/C
a(m
mo
l/mo
l)
-4-20246
Sr/
Ca-
SS
Tan
om
aly
(°C
)
0.1
1
10
1001912 ± 8 AD 1957 ± 5 AD 2009 AD
1912 AD
9.4
9.2
9.0
Co
ral S
r/C
a(m
mo
l/mo
l)
-4-20246
Sr/
Ca-
SS
Tan
om
aly
(°C
)
0.1
1
10
1002347 ± 49 a BP
2.35 ka BP
0.1
1
10
100
9.4
9.2
9.0
Co
ral S
r/C
a(m
mo
l/mo
l)
-4-20246
Sr/
Ca-
SS
Tan
om
aly
(°C
)3793 ± 29 a BP
3.79 ka BP
9.4
9.2
9.0
Co
ral S
r/C
a(m
mo
l/mo
l)
-4-20246
Sr/
Ca-
SS
Tan
om
aly
(°C
)
0.1
1
10
100
3.83 ka BP
3831 ± 34 a BP
9.4
9.2
9.0
Co
ralS
r/C
a(m
mo
l/mo
l)
-6-4-2024
Sr/
Ca-
SS
Tan
om
aly
(°C
)
0.1
1
10
100
4.27 ka BP
4268 ± 25 a BP
0 0.2 0.4 0.6
Frequency (cycles yr-1)
0.1
1
10
100
6.22 ka BP9.4
9.2
9.0
Co
ral S
r/C
a(m
mo
l/mo
l)
-4-20246
Sr/
Ca-
SS
Tan
om
aly
(°C
)
6218 ± 46 a BP
10 years Growth direction
34*
3.93.02.1 1.2
2.0
3.3
19*1.9
34* 5.7
2.71.8 1.5
15.5*4.1
2.0
a b
1910 1920 1930 1940 1950 1960 1970 1980 1990 2000Years AD
24262830
SS
T(°
C)
0 0.2 0.4 0.60.1
1
10
100SST
ERSST 1910-2000
99 %95 %90 %
Rel
ativ
e va
rian
ceR
elat
ive
vari
ance
Rel
ativ
e va
rian
ceR
elat
ive
vari
ance
Rel
ativ
e va
rian
ceR
elat
ive
vari
ance
Rel
ativ
e va
rian
ce
10.65.2 3.8 2.2
99 % 95 %90 %
99 %95 %90 %
99 %95 %90 %
99 %95 %90 %
99 %95 %90 %
99 %95 %90 %
Fig. 5. a) Monthly Bonaire coral Sr/Ca records and corresponding sea surface temperature (SST) anomalies. Monthly instrumental SST record for Bonaire (2°×2° gridbox centred at12°N; 68°W) (Smith et al., 2008). U-series ages of individual corals are indicated (years before present; a BP). Coral Sr/Ca-based SST anomalies were calculated using the monthlyrelationship of −0.042 mmol/mol/°C (Hetzinger et al., 2006). Grey shading marks changes in growth pattern at the top of coral BON-9-A (cf. Fig. 3 caption). b) Multitaper method(MTM) spectral analysis with a red noise null hypothesis (Ghil et al., 2002) (number of tapers 3; bandwidth parameter 2) of detrended and normalised coral Sr/Ca (of >20 yearslength) and SST records with mean seasonal cycle removed. 90%, 95%, and 99% significance levels and significant spectral peaks (in years) are indicated. Asterisk marks periodicitiesthat are at the limit of detection with respect to the length of the time series.
194 C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
periodicity of 5.7 years that is unprecedented in all other coral records.However, significant variability in the 5–6 year band is evident in theinstrumental SST record.
3.3.4. Coral Sr/Ca-based inter- to multidecadal SST variabilitySignificant variance in the inter- to multidecadal band is detected
in most of the coral Sr/Ca records as well as in instrumental BonaireSST throughout the last century (Fig. 5). Periodicities of 15.5, 19, and
34 years identified in the coral records are, however, at the limit of de-tection with respect to the length of the considered time series, whereoscillatory behaviour becomes indistinguishable from a secular trend.This indicates that inter- tomultidecadal variability is present in the re-cords, but it is not possible to confirm its oscillatory behaviour. Nev-ertheless, spectral analysis of the 6.22 ka coral record revealsenhanced power at inter- to multidecadal timescales relative to thelate Holocene records and present-day SST, suggesting that the
6 5 4 3 2 1 0
Age ka BP
9.5
9.4
9.3
9.2
9.1
9.0
An
nu
al m
ean
Sr/
Ca
(mm
ol/m
ol)
27.5
28.0
28.5
W. A
tlan
tic
Alk
eno
ne-
SS
T (
°C)
25.5
26.0
26.5
27.0
27.5
Car
iaco
Bas
in,
Mg
/Ca-
SS
T (
°C)
-2
0
2
Sr/
Ca-
SS
T(°
C)
Annual mean Sr/Ca
Mg/Ca-SST
Alkenone-SST
Fig. 6. Comparison between Bonaire coral Sr/Ca-derived mean sea surface temperature(SST) estimates and other southern Caribbean Sea temperature reconstructions (Lea etal., 2003; Rühlemann et al., 1999). Error bars represent the combined error (root of thesum of the squares) of the standard deviation (2σ) of the mean Sr/Ca heterogeneity ofthree modern Diploria strigosa colonies and the standard error (2SE) of the mean Sr/Caof each coral record, following established procedures (Abram et al., 2009). Dashed redline indicates the mean Sr/Ca value given by all corals. The dotted black line representsthe mean Sr/Ca value given by the three modern corals. Mean coral Sr/Ca-SST was cal-culated using the annual relationship of −0.066 mmol/mol/°C (Hetzinger et al., 2006).Note that mean Sr/Ca values are based on monthly records of up to 68 years lengthgenerated from individual corals.
195C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
magnitude of SST variability on these timescales was likely larger dur-ing the mid-Holocene. This is further supported by the standard devia-tion of the coral Sr/Ca-based SST records filtered in the inter- tomultidecadal band, which is highest for themid-Holocene record (Sup-plementary Fig. 3).
4. Discussion
4.1. Warming of the western tropical Atlantic over the last 6.2 ka
The Bonaire coral Sr/Ca records suggest that the southern CaribbeanSea was characterised by cooler conditions during the mid-Holocenecompared to today. Although intercolony variability in mean Sr/Caobstructs accurate quantification of the mid-Holocene cooling inthe Caribbean, this result is consistent with alkenone- and foraminif-eral Mg/Ca-based SST estimates frommarine sediments that indicatea warming of the southern Caribbean Sea since the mid-Holocene
0.00
0.04
0.08
0.12
0.16
0.20
Co
ral S
r/C
a se
aso
nal
ity
(mm
ol/m
ol)
6.22
ka
4.27
ka
3.83
ka
3.79
ka
2.35
ka
Fig. 7. Bonaire coral Sr/Ca-derived SST seasonality calculated from individual coral records andseasonality (0.118 mmol/mol corresponding to 2.8 °C) calculated from themean seasonality ofthe squares) of the standarddeviation (2σ) of themean Sr/Ca seasonality of threemodernDiplofollowing established procedures (Abram et al., 2009) but applied to reconstructed seasonalitywas calculated using the monthly relationship of −0.042 mmol/mol/°C (Hetzinger et al., 2006)length generated from individual corals.
(Lea et al., 2003; Rühlemann et al., 1999) (Fig. 6). This warming ofthe western tropical Atlantic has been attributed to a trend in theNAOmean state from amore positive, in the early Holocene, towardsa more negative mean state in the late Holocene possibly linked toinsolation changes (Lorenz and Lohmann, 2004; Rimbu et al., 2003).
4.2. Forcing of SST seasonality in the western tropical Atlantic over thelast 6.2 ka
The few studies that have addressed past SST seasonality in theAtlantic using coral records have either focused on the last centuries(Kilbourne et al., 2010; Kuhnert et al., 2002; Watanabe et al., 2001)or on the last interglacial period (Winter et al., 2003). The presentstudy however, allows for estimating past SST seasonality in the west-ern tropical Atlantic for snapshots of the last 6.2 ka, and indicates no sig-nificant changes throughout this time interval with the exception of aperiod around 2.35 ka BP (Fig. 7). Furthermore, slightly but not signifi-cantly increased SST seasonality is observed in the 6.22 ka Bonairecoral record. This suggests that direct insolation forcing plays, to someextent, a role in controlling the amplitude of the annual SST cycle inthe southern Caribbean, as our result is consistent with a higher ampli-tude of the seasonal insolation cycle in theNorthernHemisphere duringthe mid-Holocene (Berger, 1978). Our climate model simulations areconsistent with coral-based result in indicating an increased SST sea-sonality around Bonaire at 6 ka BP, which is part of a large-scale phe-nomena in the North Atlantic realm (Fig. 8). However, simulated SSTseasonality anomaly at 6 ka (1.2±0.5 °C) is larger compared to theslightly but not significantly increased SST seasonality inferred fromthe 6.22 ka coral. Overall, our results suggest that tropical North AtlanticSST is sensitive to orbitally driven changes in the seasonal insolationcycle, which are, nevertheless, more pronounced in the extratropics.
However, the significantly increased SST seasonality at 2.35 ka BPand the prominent interannual variability of SST seasonality indicatedby all Bonaire coral records (Fig. 5, Supplementary Fig. 2) suggeststhat additional forcing mechanisms are critical in controlling changesin the amplitude of the SST annual cycle in the southern Caribbeanover the last 6.2 ka. Observational studies showed that the ENSO tele-connection on tropical North Atlantic SST is strongest in spring (Czaja,2004; Enfield and Mayer, 1997), as also observed in Bonaire SST(Fig. 2). This seasonal emphasis in ENSO forcing produces significantinterannual SST variability in the coldest season around Bonaire(Fig. 9a) that results in pronounced interannual variability of SST sea-sonality (Fig. 2). Further mechanisms related to the influence of theNAO on spring Atlantic SST (Czaja, 2004) might also play a criticalrole in controlling SST seasonality. In summary, while southern
0
1
2
3
4
5
SS
T s
easo
nal
ity
(°C
)
1.84
ka
1912
AD
1957
AD
2009
AD
ER
SS
Tv3
b
Modern
instrumental SST data (1910–2000). Dashed line represents the meanmodern Sr/Ca-SSTthreemodern corals. Error bars for fossil corals are the combined error (root of the sum ofria strigosa and the standard error (2SE) of the averaged Sr/Ca seasonality of the fossil coral,. 2SE for modern corals and instrumental data are presented. Coral Sr/Ca-SST seasonality. Note that mean Sr/Ca seasonality values are based on monthly records of up to 68 years
SST seasonality anomaly at 6 ka (°C)
M
-2 -1.5 -1 -0.5 -0.2 0.2 0.5 1 1.5 2
J F M A M J J A S O N D
Months
-2
-1
0
1
2
Mo
del
-bas
edS
ST
ano
mal
y(°
C)
J F M A M J J A S O N D
onths
0.08
0.04
0.00
-0.04
-0.08
Co
ralS
r/C
a(m
mo
l/mo
l),
com
po
site
ann
ual
cycl
e
-2
-1
0
1
2
Sr/
Ca-
der
ived
SS
Tan
om
aly
(°C
)
Modern corals6 ka BP coral
CTL6ka
Coral Sr/Ca Model
Fig. 8. Top: Comparison of Bonaire sea surface temperature (SST) composite annual cycles inferred from coral Sr/Ca records (left panel) and numerical simulations from the coupledgeneral circulation model COSMOS (right panel). Top left: Coral Sr/Ca composite annual cycles from 3 modern corals (yellow lines) are compared to the annual Sr/Ca cycle of a mid-Holocene coral (6.22 ka BP) (black line). Corresponding Sr/Ca-SST anomaly was calculated using the monthly Sr/Ca-SST relationship of −0.042 mmol/mol/°C (Hetzinger et al.,2006). Error bars represent the standard error of the mean Sr/Ca value for each month of individual coral records. Top right: Climate model simulations of monthly SST climatologyof the Bonaire gridbox for pre-industrial (CTL) and mid-Holocene (6 ka) time slices. Bottom: SST seasonality anomaly at 6 ka inferred from numerical simulations relative to pre-industrial (CTL) time (coloured contours). Coral record of SST seasonality anomaly at Bonaire is indicated by the coloured circle. Note that model-based SST seasonality anomaly atBonaire (1.2±0.5 °C) is part of a large-scale phenomenon in the tropical and subtropical North Atlantic sector.
196 C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
Caribbean SST seasonality seems to reflect the overall influence oforbitally-induced changes in insolation over the last 6.2 ka, our resultssuggest that internal dynamics of the climate system play an importantrole in controlling short-term changes in SST seasonality.
4.3. Source of prominent quasi-biennial SST variability over the last 6.2 ka
The Bonaire coral records indicate that quasi-biennial SST variationshave been persistent in the southern Caribbean Sea over the last 6.2 ka(Fig. 5). Quasi-biennial variability is found to be a prominent feature inthe Atlantic surface temperature variability (Barnett, 1991; Dima andLohmann, 2004) possibly modulating Atlantic hurricane activity onthese timescales (Gray, 1984). Moreover, southern Caribbean SST vari-ability is influenced by the strength of surfacewinds associatedwith theCLLJ, which has a high-frequency component (1.25- and 2.3-year)that varies in phase with the NAO (Wang, 2007). Quasi-biennial pe-riodicities were also observed in a SST reconstruction from a modernCaribbean coral (Hetzinger et al., 2010), and were coherent with the
quasi-biennial component of the NAO. It has been suggested thatquasi-biennial variability is a global mode with strong projectiononto the tropical Pacific as well as the North Atlantic Oceans for bo-real winter (Barnett, 1991; Dima and Lohmann, 2004). Therefore,the significant quasi-biennial peaks found in any winter (DJF) coraltime series (e.g., Fig. 9) support the persistence of this prominentmode over the last 6.2 ka.
4.4. Sources of interannual SST variability over the last 6.2 ka
The tropical Atlantic is subject to the competing influence of remoteand local forcing. ENSO, the dominant source of global interannual cli-mate variability, plays an important role inmodulatingmodern tropicalAtlantic climate (e.g., Enfield and Mayer, 1997; Giannini et al., 2001b).At interannual timescales, significant spectral power is observed inmost Bonaire coral records indicating that southern Caribbean SSTwas characterised by prominent interannual variability over the last6.2 ka. However, the magnitude and periodicity of this interannual
a b
0.1
1
10
9.3
9.2
9.1
9.0
Monthly dataMAMDJFJJASON
1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
24
26
28
30
0.1
1
10
0 10 20 30 40 50 60 709.5
9.4
9.3
9.2
9.1
9.0
0 0.1 0.2 0.3 0.4 0.50.1
1
10
5.8-5.9
2.63.2
2.92.9
2.1
8.3
3.8
2.2
5.2
2.5
45*C
ora
l Sr/
Ca
(mm
ol/m
ol)
Co
ral S
r/C
a(m
mo
l/mo
l)S
ST
(°C
) Bonaire SST
Internal chronology (years)
Years AD
Sr/
Ca-
SS
T a
no
mal
y (°
C)
Rel
ativ
e va
rian
ceR
elat
ive
vari
ance
Rel
ativ
e va
rian
ce
-1Frequency (cycles yr )
2.35 ka BP
6.22 ka BP
-4
-2
0
2
4
6
4
-4
-2
0
2
4
6
Fig. 9. a) Monthly instrumental SST record for the Bonaire gridbox (12°N; 68°W) (Smith et al., 2008) (top), monthly coral Sr/Ca records for 2.35 ka (middle) and 6.22 ka BP (bottom)and corresponding seasonal coral Sr/Ca time series for boreal winter (DJF), spring (MAM), summer (JJA) and fall (SON). Corresponding Sr/Ca-SST anomalies calculated using themonthly Sr/Ca-SST relationship of −0.042 mmol/mol/°C (Hetzinger et al., 2006) are shown. b) Multitaper method (MTM) spectral analysis with a red noise null hypothesis(Ghil et al., 2002) (number of tapers 3; bandwidth parameter 2) of detrended and normalised seasonal time series. 95% significance level is indicated. Note that the 2.35 ka BP recorddisplays a significant peak at 5.8–5.9 years that is significant in the SON and DJF time series at the 99% level and in the JJA and MAM time series at the 95% level.
197C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
SST variability has changed through time, suggesting changes in theforcing of southern Caribbean SST over the last 6.2 ka. To investigatethis, we use two coral records of similar length (~68 years) that grewduring contrasted periods of reduced/enhanced ENSO activity toassess the role of ENSO forcing on southern Caribbean interannualSST variability.
Periods of weak and enhanced ENSO activity were identified intropical Pacific coral records (McGregor and Gagan, 2004; Tudhopeet al., 2001; Woodroffe et al., 2003). Tudhope et al. (2001) reporteda reduced ENSO activity during the mid-Holocene that was furthersupported by model simulations (Clement et al., 1999). An analysisof Pacific corals from diverse locations (McGregor and Gagan, 2004;Tudhope et al., 2001; Woodroffe et al., 2003) has suggested thatlarge amplitude El Niño events occurred during the time interval2.5–1.7 ka (McGregor and Gagan, 2004). These results are consistentwith a long-term trend towards increased ENSO variability since themid-Holocene as inferred from other proxy records (Haug et al., 2001;Moy et al., 2002).
The 6.22 ka Bonaire coral that grew during a period of weak ENSOactivity indicates significant 3–4 year variability in SST (Fig. 5 and 9).The source of this variability is ambiguous because it can be attributedto air-sea interactions originating from both the Atlantic (Chang et al.,2006; Tourre et al., 1999; Xie and Carton, 2004) and the tropical Pacific(Tourre et al., 1999). In contrast, the 2.35 ka Bonaire coral which grewduring a period of strong ENSO activity indicates enhanced interannualSST variability at periodicities centred at 5.7-years that is unprecedent-ed in any other coral record. Comparable 5–6 year variability is alsoevident in instrumental Bonaire SST, and cross-spectral analysis reveals
coherence with the Niño3.4 index at this timescale (SupplementaryFig. 4). Consequently, one can relate the strong 5–6 year variabilityin southern Caribbean SST at 2.35 ka to pronounced ENSO variabilityand the mid-Holocene 3–4 year variability to ocean–atmosphereinteractions having its origin in the Atlantic realm.
4.5. Change in the seasonal pattern of interannual SST variability at2.35 ka
The ENSO teleconnection to the Caribbean and the tropical NorthAtlantic has been shown to affect climate predominantly during the bo-real spring season (e.g., Czaja, 2004; Enfield and Mayer, 1997; Gianniniet al., 2001b; Sutton et al., 2000). Spectral analyses of instrumentalBonaire spring SST reveals a significant interannual peak centred at5.2-years (Fig. 9) that reflects the seasonal effect of ENSO on south-ern Caribbean SST. To investigate the seasonal pattern of ENSO tele-connection for snapshots of the past 6.2 ka, our monthly coral recordshave been decomposed into seasonal time series and assessed for inter-annual signals during individual seasons (DJF, December–February;MAM, March–May; JJA, June–August; SON, September–November)(Fig. 9).
The 6.22 ka coral indicates significant 3–4 year periodicities in theDJF, MAM and SON time series. Besides, a unique 8.3-year spectralpeak was found in the JJA time series suggesting quasi-decadal vari-ability in Bonaire summer SST at that time. The source of this quasi-decadal signal was also detected in instrumental data (Jury, 2009),and was attributed to modulation of the easterly trade-winds by theAtlantic subtropical high and summer convection (Jury, 2009).
198 C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
Consistent with this, a 8-year mode has been attributed to the meridi-onal movement of the Walker circulation (Dima and Lohmann, 2004).From our data, we can conclude that this mode was most likelystronger at 6.2 ka when the Intertropical Convergence Zone (ITCZ)was located further north (Haug et al., 2001). The 6.22 ka coral re-cord indicates that interannual SST variability was restricted to spe-cific seasons, as observed in most coral records used in this study(not shown).
In contrast, the prominent periodicity centred at 5.7-years evident inthe 2.35 ka coral record is significant in all seasons and not restricted toa particular season (as observed in instrumental data and other coral re-cords). As shown in Fig. 9, warmer (colder) winters are accompaniedwith warmer (colder) summers, thus contributing to the pronouncedinterannual variability significant for all seasons. Giannini et al. (2001a)showed that interdecadal changes in the ENSO teleconnection to the Ca-ribbean region can produce interference between ENSO and NAO affect-ing seasonal anomalies in Caribbean climate. Rogers (1984) found strongcoherence at a periodicity of 5.7-years. Later on, Huang et al. (1998) sug-gested that energy transfer from the tropical Pacific during warm ENSOevents might influences the NAO on these timescales. Consequently,the pronounced 5.7-year variability found in the 2.35 ka coral is possiblyassociated with changed ENSO-NAO interactions at that time.
Warm ENSO events and the negative phase of the NAO are knownfor generating warm tropical Atlantic SST anomalies in winter andspring (Czaja et al., 2002; Giannini et al., 2001a). However, it is notclear from the instrumental records that such constructive interfer-ence between ENSO and NAO can produce SST anomalies that persistuntil the next summer and fall. Therefore, given that modern ENSOaffects tropical North Atlantic SST in a particular season (Czaja, 2004;Enfield andMayer, 1997; Sutton et al., 2000), changes in the seasonalityof ENSO teleconnection or positive ocean–atmosphere feedbacks (i.e.,weakening of surface winds due to warm SST anomalies) might havedriven the pronounced interannual SST variability in all seasons at2.35 ka BP. However, physical mechanisms involved in enhancingboth SST seasonality and interannual variability in the southern Carib-bean are not well understood. Modelling studies are needed to investi-gate the underlying mechanisms and assess their potential role ingenerating such large-scale climate anomalies.
4.6. Inter- to multidecadal SST variability over the last 6.2 ka
Observational studies revealed that Caribbean climate is influencedby the Atlantic Multidecadal Oscillation (AMO) (Enfield et al., 2001;Sutton and Hodson, 2005; Wang et al., 2008a). Given that regional SSTproxy records could be linked to this Atlantic mode of multidecadal cli-mate variability (e.g., Heslop and Paul, 2011; Kilbourne et al., 2008;Poore et al., 2009; Richey et al., 2007; Saenger et al., 2009), recent stud-ies have demonstrated that Caribbean corals are ideal climate archivesfor recording variations of sea surface conditions associated with theAMO (Hetzinger et al., 2008; Kilbourne et al., 2008). Prominent inter-to multidecadal variability found in most of our Bonaire coral Sr/Carecords could suggest a potential influence of the AMO on southernCaribbean SST variability over the last 6.2 ka. Interestingly, SST vari-ability on these timescales was likely more pronounced during themid-Holocene compared to the late Holocene (Fig. 5 and SupplementaryFig. 3).
Coral δ18O records from the Dominican Republic in northern Carib-bean Sea have been interpreted to reflect interdecadal variations in pre-cipitation during the mid-Holocene, possibly related to the latitudinalshift of the ITCZ or increased storm activity at that time (Greer andSwart, 2006). A possible role for tropical Atlantic SST was suggested incontrolling this interdecadal variability in precipitation. In line withthis finding, the 6.22 ka Bonaire coral Sr/Ca record indicates inter- tomultidecadal SST variations during the mid-Holocene. Knudsen et al.(2011) presented evidence that Caribbean precipitation over the last8 ka has experienced multidecadal fluctuations linked to the AMO, yet
the contribution of SST to the underlying climate variability on AMO-like timescales remains poorly constrained due to a lack of high-resolution SST proxy records. Our new coral Sr/Ca records from anAMO-sensitive region provide evidence for inter- to multidecadal SSTvariability in the southern Caribbean Sea over the last 6.2 ka.
5. Conclusions
The natural range of tropical Atlantic SST variabilitywas investigatedwith a long-term perspective for themid- to late Holocene using a totalnumber of 295 years of monthly-resolved coral Sr/Ca records fromBonaire (southern Caribbean Sea). While this study provides addi-tional evidence for cooler conditions in the western tropical Atlanticduring themid-Holocene, patterns of short-term SST variability havebeen investigated which can be summarised as follows:
(1) In combination with climate model simulations, the seasonalcoral Sr/Ca records reveal that insolation changes on orbitaltimescales affect western tropical Atlantic SST seasonality.
(2) The significantly increased SST seasonality at 2.35 ka togetherwith prominent interannual variability of SST seasonality foundin all coral Sr/Ca records suggests a prominent role of internalvariability in modulating short-term changes in western tropicalAtlantic SST seasonality.
(3) Prominent quasi-biennial and interannual SST variability insouthern Caribbean SST during the mid- to late Holocene arelinked to ocean–atmosphere interactions originating from boththe Pacific and Atlantic oceans.
(4) Interannual SST variability at typical ENSO periods was weak at6.22 ka BP, whereas it was enhanced at 2.35 ka BP, consistentwith an increasing influence of ENSO on western tropical At-lantic climate variability throughout the mid- to late Holocene.
(5) Significantly increased SST seasonality accompanied by en-hanced interannual variability is documented by the 2.35 kacoral. Since similar interannual and seasonal SST anomalieslasting for decades were not identified in observational data,it is conceivable that a reorganisation of atmospheric circula-tion possibly linked to changes of teleconnection patterns(e.g., Giannini et al., 2001a) has occurred around 2.35 ka BP.
(6) Inter- to multidecadal SST variability over the last 6.2 ka is indi-cated by coral Sr/Ca records fromanAMO-sensitive region.More-over, our records suggest larger amplitude of SST variability onthese timescales during the mid-Holocene.
As the period of instrumental observations might already containanthropogenic imprints, the total number of 295 years of monthly-resolved coral Sr/Ca records used in this study provides unique insightsinto the natural range of tropical Atlantic SST variability since the mid-Holocene. These records indicate that short-term SST variability overthe last 6.2 ka has arisen from internal dynamics of the climate systemand suggest that a better understanding of changes in climate modesand their interactions are critical for successful understanding of tropi-cal Atlantic climate variability under changing boundary conditions.
Supplementary data to this article can be found online at doi:10.1016/j.epsl.2012.03.019.
Acknowledgements
We thank the Government of the former Island Territory of Bon-aire (Netherlands Antilles) for research and fieldwork permission,E. Beukenboom (STINAPA Bonaire National Parks Foundation) forsupport, J. Pätzold for support with coral drilling and discussion, M.J.A.Vermeij andH. Grobe for supportwith CITES permit, S. Pape for operatingand maintaining the ICP-OES, H. Kuhnert for discussion, M. Zuther for X-ray diffraction analyses, K. Baumann for Scanning Electron Microscopyimaging, C. Fensterer and R. Eichstädter for assistance with U-seriesdating in the lab of A. Mangini, A. Scheffers and C. Maier for initiating
199C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
this collaboration, and G. Wefer for overall support. This work wasfunded by Deutsche Forschungsgemeinschaft (DFG) under the SpecialPriority Programme INTERDYNAMIK (CaribClim project), throughgrants to T.F. (FE 615/3-1, 3-2), A. Mangini and D.S. (MA 821/37-1, 37-2), and G.L. (LO 895/9-1, 9-2). C.G. acknowledges additional supportby GLOMAR (Bremen International Graduate School for Marine Sci-ences) that is funded by DFG.
References
Abram, N.J., McGregor, H.V., Gagan, M.K., Hantoro, W.S., Suwargadi, B.W., 2009. Oscillationsin the southern extent of the Indo-Pacific Warm Pool during the mid-Holocene. Quat.Sci. Rev. 28, 2794–2803.
Barnett, T.P., 1991. The interaction of multiple time scales in the tropical climate system. J.Clim. 4, 269–285.
Beck, J.W., Edwards, R.L., Ito, E., Taylor, F.W., Recy, J., Rougerie, F., Joannot, P., Henin, C.,1992. Sea-surface temperature from coral skeletal strontium/calcium ratios. Sci-ence 257, 644–647.
Berger, A., 1978. Long-term variations of daily insolation and quaternary climatic changes.J. Atmos. Sci. 35, 2362–2367.
Chang, P., Fang, Y., Saravanan, R., Ji, L., Seidel, H., 2006. The cause of the fragile relation-ship between the Pacific El Nino and the Atlantic Nino. Nature 443, 324–328.
Cheng, H., Edwards, R.L., Hoff, J., Gallup, C.D., Richards, D.A., Asmerom, Y., 2000. Thehalf-lives of uranium-234 and thorium-230. Chem. Geol. 169, 17–33.
Chiang, J.C.H., Kushnir, Y., Giannini, A., 2002. Deconstructing Atlantic IntertropicalConvergence Zone variability: influence of the local cross-equatorial sea surfacetemperature gradient and remote forcing from the eastern equatorial Pacific. J.Geophys. Res. 107, 4004.
Clement, A.C., Seager, R., Cane, M.A., 1999. Orbital controls on the El Niño/SouthernOscillation and the tropical climate. Paleoceanography 14, 441–456.
Cohen, A.L., Owens, K.E., Layne, G.D., Shimizu, N., 2002. The effect of algal symbionts onthe accuracy of Sr/Ca paleotemperatures from coral. Science 296, 331–333.
Crucifix, M., Braconnot, P., Harrison, S.P., Otto-Bliesner, B., 2005. Second phase of paleo-climate modelling intercomparison project. Eos Trans. AGU 86 (28). doi:10.1029/2005EO280003.
Czaja, A., 2004. Why is north tropical Atlantic SST variability stronger in Boreal Spring?J. Clim. 17, 3017–3025.
Czaja, A., van der Vaart, P., Marshall, J., 2002. A diagnostic study of the role of remoteforcing in tropical Atlantic variability. J. Clim. 15, 3280–3290.
de Villiers, S., Shen, G.T., Nelson, B.K., 1994. The Sr/Ca-temperature relationship incoralline aragonite: influence of variability in (Sr/Ca)seawater and skeletal growthparameters. Geochim. Cosmochim. Acta 58, 197–208.
de Villiers, S., Nelson, B.K., Chivas, A.R., 1995. Biological controls on coral Sr/Ca and δ18Oreconstructions of sea surface temperatures. Science 269, 1247–1249.
Dima, M., Lohmann, G., 2004. Fundamental and derived modes of climate variability:concept and application to interannual time-scales. Tellus A 56, 229–249.
Eakin, C.M., Morgan, J.A., Heron, S.F., Smith, T.B., Liu, G., Alvarez-Filip, L., Baca, B.,Bartels, E., Bastidas, C., Bouchon, C., Brandt, M., Bruckner, A.W., Bunkley-Williams, L., Cameron, A., Causey, B.D., Chiappone, M., Christensen, T.R.L., Crabbe,M.J.C., Day, O., de la Guardia, E., Díaz-Pulido, G., DiResta, D., Gil-Agudelo, D.L.,Gilliam, D.S., Ginsburg, R.N., Gore, S., Guzmán, H.M., Hendee, J.C., Hernández-Delgado, E.A., Husain, E., Jeffrey, C.F.G., Jones, R.J., Jordán-Dahlgren, E., Kaufman,L.S., Kline, D.I., Kramer, P.A., Lang, J.C., Lirman, D., Mallela, J., Manfrino, C.,Maréchal, J.-P., Marks, K., Mihaly, J., Miller, W.J., Mueller, E.M., Muller, E.M.,Orozco Toro, C.A., Oxenford, H.A., Ponce-Taylor, D., Quinn, N., Ritchie, K.B.,Rodríguez, S., Ramírez, A.R., Romano, S., Samhouri, J.F., Sánchez, J.A., Schmahl,G.P., Shank, B.V., Skirving, W.J., Steiner, S.C.C., Villamizar, E., Walsh, S.M., Walter,C., Weil, E., Williams, E.H., Roberson, K.W., Yusuf, Y., 2010. Caribbean Corals in Cri-sis: Record Thermal Stress, Bleaching, and Mortality in 2005. PLoS One 5, e13969.
Enfield, D.B., Mayer, D.A., 1997. Tropical Atlantic sea surface temperature variabilityand its relation to El Niño-Southern Oscillation. J. Geophys. Res. 102, 929–945.
Enfield, D.B., Mestas Nuñez, A.M., Trimble, P.J., 2001. The Atlantic Multidecadal Oscilla-tion and its relation to rainfall and river flows in the continental U.S. Geophys. Res.Lett. 28, 2077–2080.
Felis, T., Lohmann, G., Kuhnert, H., Lorenz, S.J., Scholz, D., Pätzold, J., Al-Rousan, S.A.,Al-Moghrabi, S.M., 2004. Increased seasonality in Middle East temperatures duringthe last interglacial period. Nature 429, 164–168.
Felis, T., Suzuki, A., Kuhnert, H., Dima, M., Lohmann, G., Kawahata, H., 2009. Subtropicalcoral reveals abrupt early-twentieth-century freshening in the western North PacificOcean. Geology 37, 527–530.
Felis, T., Suzuki, A., Kuhnert, H., Rimbu, N., Kawahata, H., 2010. Pacific Decadal Oscilla-tion documented in a coral record of North Pacific winter temperature since 1873.Geophys. Res. Lett. 37, L14605.
Gagan, M.K., Ayliffe, L.K., Hopley, D., Cali, J.A., Mortimer, G.E., Chappell, J., McCulloch,M.T., Head, M.J., 1998. Temperature and surface-ocean water balance of the mid-Holocene Tropical Western Pacific. Science 279, 1014–1018.
Ghil, M., Allen, M.R., Dettinger, M.D., Ide, K., Kondrashov, D., Mann, M.E., Robertson,A.W., Saunders, A., Tian, Y., Varadi, F., Yiou, P., 2002. Advanced spectral methodsfor climatic time series. Rev. Geophys. 40, 1003.
Giannini, A., Kushnir, Y., Cane, M.A., 2000. Interannual variability of Caribbean rainfall,ENSO, and the Atlantic ocean*. J. Clim. 13, 297–311.
Giannini, A., Cane, M.A., Kushnir, Y., 2001a. Interdecadal changes in the ENSO telecon-nection to the Caribbean Region and the North Atlantic Oscillation*. J. Clim. 14,2867–2879.
Giannini, A., Chiang, J.C.H., Cane, M.A., Kushnir, Y., Seager, R., 2001b. The ENSO telecon-nection to the tropical Atlantic Ocean: contributions of the remote and local SSTs torainfall variability in the tropical Americas*. J. Clim. 14, 4530–4544.
Giannini, A., Saravanan, R., Chang, P., 2004. The preconditioning role of Tropical Atlan-tic variability in the development of the ENSO teleconnection: implications for theprediction of Nordeste rainfall. Clim. Dyn. 22, 839–855.
Giry, C., Felis, T., Kölling, M., Scheffers, S., 2010a. Geochemistry and skeletal structure ofDiploria strigosa, implications for coral-based climate reconstruction. Palaeogeogr.Palaeoclimatol. Palaeoecol. 298, 378–387.
Giry, C., Felis, T., Scheffers, S., Fensterer, C., 2010b. Assessing the potential of SouthernCaribbean corals for reconstructions of Holocene temperature variability. IOP Con-ference Series: Earth and Environmental Science, 9. doi:10.1088/1755-1315/1089/1081/012021.
Goodkin, N.F., Hughen, K.A., Cohen, A.L., 2007. A multicoral calibration method toapproximate a universal equation relating Sr/Ca and growth rate to sea surfacetemperature. Paleoceanography 22, PA1214. doi:10.1029/2006PA001312.
Goreau, T.J., 1977. Coral skeletal chemistry: physiological and environmental regulation ofstable isotopes and trace metals inMontastrea annularis. Proc. R. Soc. Lond. B Biol. Sci.196, 291–315.
Gray, W.M., 1984. Atlantic seasonal hurricane frequency. Part I: El Niño and 30 mbQuasi-Biennial Oscillation influences. Mon. Weather Rev. 112, 1649–1668.
Greer, L., Swart, P.K., 2006. Decadal cyclicity of regional mid-Holocene precipitation:evidence from Dominican coral proxies. Paleoceanography 21, PA2020. doi:10.1029/2005PA001166.
Hastenrath, S., 1984. Interannual variability and annual cycle: mechanisms of circulationand climate in the tropical Atlantic sector. Mon. Weather Rev. 112, 1097–1107.
Haug, G.H., Hughen, K.A., Sigman, D.M., Peterson, L.C., Rohl, U., 2001. Southward migra-tion of the intertropical convergence zone through the Holocene. Science 293,1304–1308.
Hendy, E.J., Gagan, M.K., Lough, J.M., McCulloch, M., deMenocal, P.B., 2007. Impact ofskeletal dissolution and secondary aragonite on trace element and isotopic cli-mate proxies in Porites corals. Paleoceanography 22, PA4101. doi:10.1029/2007PA001462.
Heslop, D., Paul, A., 2011. Can oceanic paleothermometers reconstruct the AtlanticMultidecadal Oscillation? Clim. Past 7, 151–159.
Hetzinger, S., Pfeiffer, M., Dullo, W.-C., Ruprecht, E., Garbe-Schönberg, D., 2006. Sr/Ca andδ18O in a fast-growing Diploria strigosa coral: evaluation of a new climate archivefor the tropical Atlantic. Geochem. Geophys. Geosyst. 7, Q10002. doi:10.1029/2006GC001347.
Hetzinger, S., Pfeiffer, M., Dullo, W.-C., Keenlyside, N., Latif, M., Zinke, J., 2008. Caribbeancoral tracks Atlantic Multidecadal Oscillation and past hurricane activity. Geology36, 11–14.
Hetzinger, S., Pfeiffer, M., Dullo, W.-C., Garbe-Schönberg, D., Halfar, J., 2010. Rapid 20thcentury warming in the Caribbean and impact of remote forcing on climate in thenorthern tropical Atlantic as recorded in a Guadeloupe coral. Palaeogeogr. Palaeo-climatol. Palaeoecol. 296, 111–124.
Hoffmann, D.L., Prytulak, J., Richards, D.A., Elliott, T., Coath, C.D., Smart, P.L., Scholz, D.,2007. Procedures for accurate U and Th isotope measurements by high precisionMC-ICPMS. Int. J. Mass Spectrom. 264, 97–109.
Huang, J., Higuchi, K., Shabbar, A., 1998. The relationship between the North AtlanticOscillation and El Niño/Southern Oscillation. Geophys. Res. Lett. 25, 2707–2710.
Hurrell, J.W., Visbeck, M., Busalacchi, A., Clarke, R.A., Delworth, T.L., Dickson, R.R., Johns,W.E., Koltermann, K.P., Kushnir, Y., Marshall, D., Mauritzen, C., McCartney, M.S.,Piola, A., Reason, C., Reverdin, G., Schott, F., Sutton, R., Wainer, I., Wright, D.,2006. Atlantic climate variability and predictability: a CLIVAR perspective. J. Clim.19, 5100–5121.
Jury, M.R., 2009. A quasi-decadal cycle in Caribbean climate. J. Geophys. Res. 114,D13102.
Kaplan, A., Cane, M.A., Kushnir, Y., Clement, A.C., Blumenthal, M.B., Rajagopalan, B.,1998. Analyses of global sea surface temperature 1856–1991. J. Geophys. Res.103, 18567–18589.
Kilbourne, K.H., Quinn, T.M., Webb, R., Guilderson, T., Nyberg, J., Winter, A., 2008. Paleo-climate proxy perspective on Caribbean climate since the year 1751: evidence ofcooler temperatures and multidecadal variability. Paleoceanography 23, PA3220.doi:10.1029/2008PA001598.
Kilbourne, K.H., Quinn, T.M., Webb, R., Guilderson, T., Nyberg, J., Winter, A., 2010. Coralwindows onto seasonal climate variability in the northern Caribbean since 1479.Geochem. Geophys. Geosyst. 11, Q10006.
Kim, J.-H., Rimbu, N., Lorenz, S.J., Lohmann, G., Nam, S.-I., Schouten, S., Rühlemann, C.,Schneider, R.R., 2004. North Pacific and North Atlantic sea-surface temperaturevariability during the Holocene. Quat. Sci. Rev. 23, 2141–2154.
Knudsen, M.F., Seidenkrantz, M.-S., Jacobsen, B.H., Kuijpers, A., 2011. Tracking theAtlantic Multidecadal Oscillation through the last 8,000 years. Nat. Commun.2, 178.
Kuhnert, H., Pätzold, J., Schnetger, B., Wefer, G., 2002. Sea-surface temperature variabilityin the 16th century at Bermuda inferred from coral records. Palaeogeogr. Palaeocli-matol. Palaeoecol. 179, 159–171.
Kuhnert, H., Crueger, T., Pätzold, J., 2005. NAO signature in a Bermuda coral Sr/Ca record.Geochem. Geophys. Geosyst. 6, Q04004. doi:10.1029/2004GC000786.
Lea, D.W., Pak, D.K., Peterson, L.C., Hughen, K.A., 2003. Synchroneity of tropical andhigh-latitude Atlantic temperatures over the last glacial termination. Science 301,1361–1364.
200 C. Giry et al. / Earth and Planetary Science Letters 331-332 (2012) 187–200
Linsley, B.K., Wellington, G.M., Schrag, D.P., 2000. Decadal sea surface temperaturevariability in the subtropical south Pacific from 1726 to 1997 A.D. Science 290,1145–1148.
Lorenz, S.J., Lohmann, G., 2004. Acceleration technique for Milankovitch type forcing ina coupled atmosphere–ocean circulation model: method and application for theHolocene. Clim. Dyn. 23, 727–743.
Lorenz, S.J., Kim, J.-H., Rimbu, N., Schneider, R.R., Lohmann, G., 2006. Orbitally driveninsolation forcing on Holocene climate trends: evidence from alkenone data andclimate modeling. Paleoceanography 21, PA1002.
Marshall, J., Yochanan, K., David, B., Ping, C., Arnaud, C., Robert, D., James, H., Michael,M., Saravanan, R., Martin, V., 2001. North Atlantic climate variability: phenomena,impacts and mechanisms. Int. J. Climatol. 21, 1863–1898.
Marsland, S., Haak, H., Jungclaus, J., Latif, M., Röske, F., 2003. The Max-Planck-Instituteglobal ocean/sea ice model with orthogonal curvilinear coordinates. Ocean Model.5, 91–127.
McGregor, H.V., Gagan, M.K., 2004. Western Pacific coral δ18O records of anomalousHolocene variability in the El Niño/Southern Oscillation. Geophys. Res. Lett. 31, L11204.
Moy, C.M., Seltzer, G.O., Rodbell, D.T., Anderson, D.M., 2002. Variability of El Nino/Southern Oscillation activity at millennial timescales during the Holocene epoch.Nature 420, 162–165.
Muñoz, E., Busalacchi, A.J., Nigam, S., Ruiz-Barradas, A., 2008. Winter and summerstructure of the Caribbean low-level jet. J. Clim. 21, 1260–1276.
Okai, T., Suzuki, A., Kawahata, H., Terashima, S., Imai, N., 2002. Preparation of a newgeological survey of Japan geochemical reference material: coral JCp-1. Geostand.Newsl. 26, 95–99.
Poore, R., DeLong, K., Richey, J., Quinn, T., 2009. Evidence of multidecadal climate variabil-ity and the Atlantic Multidecadal Oscillation from a Gulf of Mexico sea-surfacetemperature-proxy record. Geo-Mar. Lett. 29, 477–484.
Raddatz, T., Reick, C., Knorr, W., Kattge, J., Roeckner, E., Schnur, R., Schnitzler, K.G.,Wetzel, P., Jungclaus, J., 2007. Will the tropical land biosphere dominate the climate–carbon cycle feedback during the twenty-first century? Clim. Dyn. 29, 565–574.
Reynolds, R.W., Rayner, N.A., Smith, T.M., Stokes, D.C., Wang, W., 2002. An improved insitu and satellite SST analysis for climate. J. Clim. 15, 1609–1625.
Richey, J.N., Poore, R.Z., Flower, B.P., Quinn, T.M., 2007. 1400 yr multiproxy record ofclimate variability from the northern Gulf of Mexico. Geology 35, 423–426.
Rimbu, N., Lohmann, G., Kim, J.H., Arz, H.W., Schneider, R., 2003. Arctic/North AtlanticOscillation signature in Holocene sea surface temperature trends as obtainedfrom alkenone data. Geophys. Res. Lett. 30, 1280.
Robinson, L.F., Belshaw, N.S., Henderson, G.M., 2004. U and Th concentrations and isotoperatios in modern carbonates and waters from the Bahamas. Geochim. Cosmochim.Acta 68, 1777–1789.
Roeckner, E., Bäuml, G., Bonaventura, L., Brokopf, R., Esch, M., Giorgetta, M., Hagemann,S., Kirchner, I., Kornblueh, L., Manzini, E., Rhodin, A., Schlese, U., Schulzweida, U.,Tompkins, A., 2003. The Atmospheric General Circulation Model ECHAM5. PartOne: Model Description Report No.349. Max Planck Institute for Meteorology.
Rogers, J.C., 1984. The association between the North Atlantic oscillation and the Southernoscillation in the Northern Hemisphere. Mon. Weather Rev. 112, 1999–2015.
Rühlemann, C., Mulitza, S., Muller, P.J., Wefer, G., Zahn, R., 1999. Warming of the tropicalAtlantic Ocean and slowdown of thermohaline circulation during the last deglacia-tion. Nature 402, 511–514.
Saenger, C., Cohen, A.L., Oppo, D.W., Halley, R.B., Carilli, J.E., 2009. Surface-temperature trendsand variability in the low-latitude North Atlantic since 1552. Nat. Geosci. 2, 492–495.
Scheffers, A., 2004. Tsunami imprints on the Leeward Netherlands Antilles (Aruba, Curaçao,Bonaire) and their relation to other coastal problems. Quat. Int. 120, 163–172.
Scheffers, A., Scheffers, S., 2006. Documentation of the impact of hurricane Ivan on thecoastline of Bonaire (Netherlands Antilles). J. Coast. Res. 22, 1437–1450.
Scheffers, S.R., Haviser, J., Browne, T., Scheffers, A., 2009. Tsunamis, hurricanes, thedemise of coral reefs and shifts in prehistoric human populations in the Caribbean.Quat. Int. 195, 69–87.
Scholz, D., Mangini, A., 2007. How precise are U-series coral ages? Geochim. Cosmochim.Acta 71, 1935–1948.
Scholz, D., Mangini, A., Felis, T., 2004. U-series dating of diagenetically altered fossil reefcorals. Earth Planet. Sci. Lett. 218, 163–178.
Smith, S.V., Buddemeier, R.W., Redalje, R.C., Houck, J.E., 1979. Strontium-calciumthermometry in coral skeletons. Science 204, 404–407.
Smith, T.M., Reynolds, R.W., Peterson, T.C., Lawrimore, J., 2008. Improvements toNOAA's historical merged land–ocean surface temperature analysis (1880–2006).J. Clim. 21, 2283–2296.
Sutton, R.T., Hodson, D.L.R., 2005. Atlantic Ocean forcing of North American and Europeansummer climate. Science 309, 115–118.
Sutton, R.T., Jewson, S.P., Rowell, D.P., 2000. The elements of climate variability in thetropical Atlantic region. J. Clim. 13, 3261–3284.
Tedesco, K., Thunell, R., 2003. High resolution tropical climate record for the last6,000 years. Geophys. Res. Lett. 30, 1891.
Tourre, Y.M., Rajagopalan, B., Kushnir, Y., 1999. Dominant patterns of climate variability inthe Atlantic Ocean during the last 136 years. J. Clim. 12, 2285–2299.
Tudhope, A.W., Chilcott, C.P., McCulloch, M.T., Cook, E.R., Chappell, J., Ellam, R.M., Lea,D.W., Lough, J.M., Shimmield, G.B., 2001. Variability in the El Niño-Southern oscil-lation through a glacial–interglacial cycle. Science 291, 1511–1517.
Wang, C., 2007. Variability of the Caribbean low-level jet and its relations to climate.Clim. Dyn. 29, 411–422.
Wang, C., Enfield, D.B., 2003. A further study of the tropical western Hemisphere warmpool. J. Clim. 16, 1476–1493.
Wang, C., Lee, S.-k., 2007. Atlantic warm pool, Caribbean low-level jet, and their poten-tial impact on Atlantic hurricanes. Geophys. Res. Lett. 34, L02703.
Wang, C., Enfield, D.B., Lee, S.-K., Landsea, C.W., 2006. Influences of the Atlantic warmpool on western Hemisphere summer rainfall and Atlantic hurricanes. J. Clim. 19,3011–3028.
Wang, C., Lee, S.-K., Enfield, D.B., 2007. Impact of the Atlantic warm pool on the summerclimate of the Western Hemisphere. J. Clim. 20, 5021–5040.
Wang, C., Lee, S.-K., Enfield, D.B., 2008a. Climate response to anomalously large andsmall Atlantic warm pools during the summer. J. Clim. 21, 2437–2450.
Wang, C., Lee, S.-K., Enfield, D.B., 2008b. Atlantic warm pool acting as a link betweenAtlantic multidecadal oscillation and Atlantic tropical cyclone activity. Geochem.Geophys. Geosyst. 9, Q05V03.
Wanner, H., Beer, J., Bütikofer, J., Crowley, T.J., Cubasch, U., Flückiger, J., Goosse, H.,Grosjean, M., Joos, F., Kaplan, J.O., Küttel, M., Müller, S.A., Prentice, I.C., Solomina,O., Stocker, T.F., Tarasov, P., Wagner, M., Widmann, M., 2008. Mid- to late Holoceneclimate change: an overview. Quat. Sci. Rev. 27, 1791–1828.
Watanabe, T., Winter, A., Oba, T., 2001. Seasonal changes in sea surface temperatureand salinity during the Little Ice Age in the Caribbean Sea deduced from Mg/Caand 18O/16O ratios in corals. Mar. Geol. 173, 21–35.
Weber, J.N., 1973. Incorporation of strontium into reef coral skeletal carbonate. Geochim.Cosmochim. Acta 37, 2173–2190.
Wei, W., Lohmann, G., Dima, M., 2012. Distinct modes of internal variability in theGlobal Meridional Overturning Circulation associated with the Southern Hemi-sphere westerly winds. J. Phys, Oceanogr., doi:10.1175/JPO-D-11-038.1.
Winter, A., Paul, A., Nyberg, J., Oba, T., Lundberg, J., Schrag, D., Taggart, B., 2003. Orbitalcontrol of low-latitude seasonality during the Eemian. Geophys. Res. Lett. 30, 1163.doi:10.1029/2002GL016275.
Woodroffe, C.D., Beech, M.R., Gagan, M.K., 2003. Mid-late Holocene El Niño variabilityin the equatorial Pacific from coral microatolls. Geophys. Res. Lett. 30, 1358.
Xie, S.-P., Carton, J.A., 2004. Tropical Atlantic variability: Patterns, mechanisms, andimpacts, Earth Climate: The Ocean–atmosphere Interaction, Geophys. Monogr.,Vol. 147, Amer. Geophys. Union ed, pp. 121–142.