86Sr fractionation in inorganic aragonite and in corals

Available online at www.sciencedirect.com

ScienceDirect Geochimica et Cosmochimica Acta 178 (2016) 268–280 www.elsevier.com/locate/gca

88

Sr/86Sr fractionation in inorganic aragonite and in corals Noa Fruchter a,⇑, Anton Eisenhauer a, Martin Dietzel b, Jan Fietzke a, Florian Bo¨hm a, Paolo Montagna c, Moti Stein d, Boaz Lazar e, Riccardo Rodolfo-Metalpa f, Jonathan Erez e a

GEOMAR, Helmholtz-Zentrum fu¨r Ozeanforschung Kiel, Wischhofstr. 1-3, 24148 Kiel, Germany b Institute of Applied Geosciences, Graz University of Technology, 8010 Graz, Austria c CNR – ISMAR – U.O.S. di Bologna, Via Gobetti, 101 – 40129 Bologna, Italy d Geological Survey of Israel, Malkhei Israel 30, Jerusalem, Israel e Institute of Earth Sciences, The Hebrew University of Jerusalem, 91904, Jerusalem, Israel f Institut de Recherche pour le De´veloppement, Unite 227 CoReus 2, Noumea, New Caledonia Received 5 February 2015; accepted in revised form 31 January 2016;

Abstract Conflicting results have been reported for the stable Sr isotope fractionation, specifically with respect to the influence of temperature. In an experimental study we have investigated the stable Sr isotope systematics for inorganically precipitated and biogenic (coral) aragonite (natural and laboratory-cultured). Inorganic aragonite precipitation experiments were performed from natural seawater using the CO2 diffusion technique. The experiments were performed at different temperatures and different carbonate ion concentrations. 88Sr/86Sr of the inorganic aragonite precipitated in the experiments are 0.2‰ lighter than seawater, but showed no correlation to the water temperature or to CO2
3 concentration. Similar observations are made in different coral species (Cladocora caespitosa, Porites sp. and Acropora sp.), with identical fractionation from the bulk solution 88 86 and no correlation to temperature or CO2
3 concentration. The lack of Sr/ Sr variability in corals at different environmental 88 86 parameters and the similarity to the Sr/ Sr fractionation in inorganic aragonite may indicate a similar Sr incorporation mechanism in corals skeleton and inorganic aragonite, and therefore the previously proposed Rayleigh-based multi element model (Gaetani et al., 2011) cannot explain the process of Sr incorporation in the coral skeletal material. It is proposed that the relatively constant 88Sr/86Sr fractionation in aragonite can be used for paleo reconstruction of seawater 88Sr/86Sr composition. The seawater 88Sr/86Sr ratio reconstruction can be further used in calcite samples to reconstruct paleo precipitation rates. Ó 2016 Elsevier Ltd. All rights reserved.

1. INTRODUCTION Scleractinian coral skeletons are one of the main archives of paleoclimate information, such as sea surface temperature (SST), pH and ocean chemistry. A temperature dependency for the partitioning between seawater and ⇑ Corresponding author at: Geological Survey of Israel, Malkhei

Israel 30, Jerusalem, Israel. E-mail address: [email protected] (N. Fruchter). http://dx.doi.org/10.1016/j.gca.2016.01.039 0016-7037/Ó 2016 Elsevier Ltd. All rights reserved.

corals was found for several elemental ratios such as Mg/ Ca (e.g. Mitsuguchi et al., 1996), Sr/Ca (e.g. Beck et al., 1992) and B/Ca (e.g. Fallon et al., 1999; Montagna et al., 2007), and recently also Li/Mg (e.g. Case et al., 2010; Montagna et al., 2014). The isotope fractionation of oxygen, d18O (e.g. Weber and Woodhead, 1972), is also being used for SST reconstruction from coral skeletons. Among all geochemical proxies, Sr/Ca and d18O are the most widely used and established paleothermometers (Corre`ge, 2006). However, temperature may not be the only

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

controlling factor for trace element partitioning and isotope fractionation (c.f. Gagnon et al., 2007). Sr/Ca ratios, for example, have been shown to differ significantly in neighboring colonies of the same species exposed to the same environmental conditions (Marshall and McCulloch, 2002). Similarly, the applicability of Mg/Ca as temperature proxy is undermined by recent studies showing a weak correlation between skeletal Mg/Ca and SST (Quinn and Sampson, 2002) and the fact that Mg is strongly correlated to organic materials in coral skeletons (Meibom et al., 2004; Finch and Allison, 2008). In addition to temperature d18O is known to be influenced by seawater salinity changes (e.g. Schmidt, 1999). A new approach to extract temperature records from the geochemical signal in corals was recently proposed by Gaetani and Cohen (2006) and Gaetani et al. (2011) based on a Rayleigh-based multi element partitioning model. The Rayleigh-based model defines calcification in corals as a precipitation process from a renewable finite solution. Elements are depleted from the solution during precipitation according to their inorganic partitioning coefficient. Since the solution is finite, the coral’s apparent fractionation factor is influenced by the Rayleigh effect. Temperature is strongly correlated to the Rayleigh-based mechanism (Gaetani et al., 2011), as the depletion or enrichment of elements is a function of increased precipitation rate as a consequence of increasing temperature. By combining information from two elemental ratios (Sr/Ca and Mg/Ca) Gaetani et al. (2011) were able to reconstruct the remaining fraction in the bulk solution and the temperature at which the Acropora sp. specimens were cultured. The increasing interest in non-traditional stable isotopes (Johnson et al., 2004) and the search for a reliable paleo proxy record has led to several studies of 88Sr/86Sr fractionation as a possible paleo-temperature proxy. Fietzke and Eisenhauer (2006) first reported a measurable 88Sr/86Sr fractionation in the tropical coral Pavona clavus and in inorganic aragonites and showed a d88/86Sr dependency on temperature. In a follow up study Ru¨ggeberg et al. (2008) found a similar trend between temperature and d88/86Sr in the cold water coral, Lophelia pertusa. Evidence of significantly different d88/86Sr between the precipitating mineral and the bulk solution were also found in terrestrial environments by Halicz et al. (2008). While the earlier work of d88/86Sr used the samplestandard bracketing technique to reliably correct for instrumental mass fractionation using Multiple-Collector ICP-MS (MC-ICPMS) (Fietzke and Eisenhauer, 2006; Halicz et al., 2008; Ru¨ggeberg et al., 2008), later studies applied the 87 Sr/84Sr double spike method that was developed for Thermal Ionization Mass Spectrometry (TIMS) by Krabbenho¨ft et al. (2009). Measurements using the TIMS method did not confirm the significant positive d88/86Sr-temperature trend in the tropical coral Acropora sp. (Krabbenho¨ft et al., 2010) and in the cold water coral L. pertusa (Raddatz et al., 2013). The discrepancies between the results of these studies may be due to methodological differences that produced erratic data in one set of measurements. In the current study we measured d88/86Sr using the double spike method in modern natural Porites sp. from

269

two different locations, cultured Acropora sp. and Cladocora caespitosa as well as inorganically-precipitated aragonite. A large range of temperatures and precipitation rates (obtained by different CO2
concentrations) were 3 applied to both the inorganic aragonite and the coral culturing experiments in order to detect environmental drivers of the Sr isotopes fractionation process. In addition, the comparison between inorganic aragonite and coralline aragonite through 88Sr/86Sr fractionation sheds some light on the enigmatic coral calcification process. 2. MATERIAL AND METHODS 2.1. Inorganic aragonite precipitation experiments Inorganic aragonite precipitation experiments were conducted at the Graz University of Technology, Austria using the CO2 diffusion technique as described in detail in Dietzel et al. (2004) and Tang et al. (2008). Briefly, CO2 diffuses from an inner solution to an outer solution via a polyethylene (PE) membrane. For the experiments two different membranes thicknesses, 0.02 mm or 2 mm were used. The inner solution has a higher pCO2 than the outer solution and consists of 35 g of NaHCO3 dissolved in 0.5 L of double deionized water acidified to pH 7.5. In some experiments 10 ml 1 N HCl was added to the inner solution in order to increase the pCO2 even more (Table 1). The outer solution (or bulk solution), from which the precipitation occurs, consists of 5 L of filtered natural seawater collected at 1000 m water depth from the Gulf of Aqaba using a CTD-Rosette system (19.09°N 39.49°E, salinity 40.7). Filtration was done using a Sartobram P. sterile MidiCap double filter (0.45 lm and 0.2 lm). The pH of the outer solution was kept constant at 8.30 ± 0.03 using 0.5 N NaOH titration, controlled using a Schott Blueline 28 pH meter combined with an automatic titration burette. The pCO2 gradient between the inner and the outer solutions controls the CO2 diffusion rate and the saturation state for CaCO3 in the outer solution. Nine different experiments (Table 1), which varied in temperature (15 °C, 25 °C and 30 °C) and CO2 uptake rates (using different PE membranes of 0.02 mm or 2 mm thickness and different pCO2 in the inner solution) were run. Water samples from the outer solution were collected at the beginning and at the end of the experiments for the analyses of Sr/Ca and 88Sr/86Sr. The entire precipitated solid material was collected and dried at 80 °C for 24 h, and then analyzed with XRD and FTIR (PerkinElmer spectrum 100 FTIR) to determine its mineralogy and detect the presence of the poorly crystallized mineral brucite. The precipitation rate (R in lmol h
1) of aragonite was determined according the equation R ¼ ð½Cai
½Caf Þ  V =Dt

ð2:1Þ

where [Ca]i and [Ca]f are the Ca concentration in the initial and the final solutions, respectively, V is the volume of the solution and Dt is the duration of the precipitation. Even though precipitation rates are influenced by the surface area and surface reactions (Wiechers et al., 1975), we could not include the surface area measurements in the calculation

270

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

Table 1 Inorganic aragonite experiments setup conditions and elemental ratio results in the solid. Standard deviation from the mean (2SD) are presented in parentheses. Sample Label

Temp. (°C)

Diffusion membrane thickness (mm)

d88/86Srb (‰

IA-1 IA-2 IA-3 IA-4 IA-5 IA-6 IA-7 IA-9 IA-10

25 25 30 30 30 30 15 15 15

0.02 0.02 2 0.02 0.02a 2a 2 2a 0.02a

0.214 0.208 0.193 0.181 0.207 0.230 0.225 0.261 0.212

a b *

SRM987)

Sr/Ca (mmol/mol)

Mg/Ca (mmol/mol)

log R (mol/h)

9.80(0.48) 9.54(0.31) 9.48(0.31) 9.10(0.13) 9.37(0.40) 9.57(0.34) 10.82(0.41) 11.06(0.13) 10.72(0.40)

5.44(0.37) 105.47(4.79)* 2.91(0.17) 47.57(0.47)* 63.13(0.35)* 4.10(0.28) 6.54(0.14) 4.96(0.07) 4.75(0.14)

3.29 2.95 1.84 3.01 3.20 2.55 2.33 1.71 2.51

Addition of 10 ml 1 N HCl to the inner solution to increase pCO2. Reproducibility of 0.020‰ (2SD) according to JCp-1 measurements (n = 10). Brucite, Mg(OH)2 precipitation in addition to aragonite.

since in some experiments up to 10% brucite (Mg(OH)2) precipitated. Freshly precipitated brucite has a significantly higher surface area of 75–90 m2 g
1 (current study surface area measurements of pure brucite), than the 0.81–2.41 m2 g
1 surface area of aragonite (Niedermayr, 2011; Romanek et al., 2011). Thus, even a small amount of brucite in the precipitated solid material would result in large errors in the surface normalized rate. 2.2. Cultured corals Small colonies of Acropora sp. (21 cm3) were cultured at the department of Earth Sciences at the Hebrew University of Jerusalem, Israel. The specimens were collected from one mother colony several weeks prior to the beginning of the experiments. They were attached to an epoxy base and were placed back in the main laboratory aquarium (i.e. holding aquarium) for recovery. Before the beginning of the experiments the specimens were stained using a 15 ppm solution of Alizarin-Red S (Dodge et al., 1984) for seven hours. Alizarin-Red S is incorporated into a thin layer of skeleton and provides a marker for measuring skeletal growth. The incubation experiments lasted for seven weeks, and were placed in a 220 ml clear chamber supplied with 19 ml h
1 of diluted Red-Sea water. The seawater was collected from the pier at the Interuniversity Institute for Marine Sciences in Eilat (IUI) and diluted, for the experiment, to salinity of 37. The flow-through system had a residence time of ca. 12 h and the solution was constantly stirred in the chamber using a magnetic stirrer. Light was supplied at an illumination of 200 ± 10 lmol photonsm
2 s
1 through metal-halide lamps in 12 h dark/ light cycles. Temperature was kept constant in all tanks using heaters or a refrigerating system connected to electronic controllers. The seven different incubation conditions varied in temperature (19.0 ± 0.3 °C, 22.0 ± 0.3 °C, 25.0 ± 0.3 °C and 28.0 ± 0.3 °C) and CO2
3 concentration (171 ± 13, 212 ± 12, 279 ± 15 and 352 ± 28 lmol kg
1). CO2
concentrations were modified by acid addition to 3 the experimental solutions and were calculated from pH and alkalinity (AT) measurements using the CO2Sys program (Lewis et al., 1998).

The carbonate chemistry of the solutions was monitored on a weekly basis through pH and AT of the water source solutions (reservoirs), the reaction vessels (both at light and dark conditions) and the outflow solutions. In addition, 50 ml of water samples from the reservoirs and the outflow solutions (representing the total duration of the experiments) were collected for geochemical analyses. Calcification rates of the Acropora specimens were calculated using the alkalinity anomaly technique (Smith and Kinsey, 1978; Chisholm and Gattuso, 1991). Calcification rates were normalized to the weight in air of the colonies at the end of the experiment using the buoyant weight technique. Every 10 days, during the experiments, the corals specimens were fed with Artemia sp. in an external 5 l vessel for 1 h. While feeding, algae growth was removed from the experiment chambers. The experiments were terminated after seven weeks. The coral skeletons were separated from their tissue using an air-brush followed by bleaching in 1% NaClO and washing in double deionized water (DDW). C. caespitosa specimens were collected in the Gulf of La Spezia (Ligurian Sea, 44°030 N, 9°550 E) at 25 m depth in June 2006. The corals were cultured in two separate experiments at the Scientific Centre of Monaco (CSM). In the first experiment 20 coral nubbins (5 for each condition) were maintained for 87 days at a salinity of 38.5 and four constant temperatures, 15.00 ± 0.05 °C, 18.00 ± 0.05 °C, 21.00 ± 0.05 °C and 23.00 ± 0.05 °C. In the second experiment 14 corals (7 for each condition) were cultured for 10 months under controlled pCO2 conditions, 390 ± 48 and 701 ± 78 ppm pCO2 at a temperature of 16.4 ± 2.6 °C, salinity of 37.9–38.4 and constant AT of 2545 ± 15 leq kg
1. This resulted in two different pH (8.09 ± 0.02, 7.87 ± 0.02) and carbonate concentrations (CO2
is 3 219 ± 17, 145 ± 15 lmol mol
1). Before the initiation of the experiment, the corals were stained for 24 h with a solution of 10 mg L
1 of Alizarin Red S (Dodge et al., 1984). Light was provided by hydrargyrum quartz iodide lamps (60 lmol photon m
2 s
1; photoperiod: 12 h dark/12 h light). Fresh seawater was constantly added to the experiment tanks from the local seawater in front of the Monaco Scientific Center. The linear extension rate was measured

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

for each specimen from the Alizarin Red S marking line to the skeleton tip. The specimens subject to each temperature were then divided into three sample powders according to three statistical groups; average extension rates (Average ± 1SE), low extension rates (Average + 1SE) groups. This statistical separation was not done in the pCO2 experiments specimens. 2.3. Coral cores Two cores from living colonies of Porites sp. were collected from two different locations (see below for details). Each core was cut to a 2–3 mm thick slab, which was subsampled using a micro-drill of 0.5 mm in diameter along the growth axis of the coral for geochemical analyses. The powders were split for d18O analyses, together with trace element and isotope measurements. The powders not designated for d18O analysis were bleached with 1% NaClO and washed with alkaline water (at pH 9 achieved by the addition of ammonium hydroxide to double deionized water (DDW)) prior to dissolution. 2.3.1. Core TH1 This Porites sp. coral core was drilled in July 1995 from a massive coral colony (TH1) located in the south-eastern part of the lagoon in Teahupoo, Tahiti French Polynesia at 2.4 m (c.f. Cahyarini et al. (2008)). A total of 36 samples (1 mg each) (Table 2) were drilled along the growth axis of the slab following a 9.6 cm long transect. The same coral core was used by Cahyarini et al. (2008). We therefore sampled the core comparing the radiograph pictures of both slabs. This enabled us to compare our d18O results with Cahyarini et al. (2009). Cahyarini et al. (2009) calibrated their d18O results to sea surface temperature (SST) records in Tahiti that show variations between 26 °C (minimum in the dry season) and 29 °C (maximum in the wet season) (Cahyarini et al., 2009). This calibration was also used in the current study. 2.3.2. Core SOT-1 This Porites lutea coral core was drilled in July 2007 from the Gulf of Aqaba 29.50°N 34.92°E (coral colony SOT-1). The coral grew at 5 m water depth near the underwater observatory, south of the Nature Reserve Reef, Eilat Israel. The seawater temperature was recorded at 29.47°N 34.92°E at depths of 20–40 m to avoid the diurnal thermocline. Temperature measured at these depths can be up to 2 °C lower than the sea surface in the summer season from June to September. The total of 18 powder samples (ca. 2 mg each) (Table 2) were drilled from SOT-1 core slab from a 3 cm long transect along the main growth axis of the coral. The clear density variations on the slab were used to estimate the period of time that is covered by the transect samples. By counting the total bands from the upper part of the core to the sampled bands, we estimate that the sampled bands cover the years 1996–2000. 2.4. Element/Ca ratios and isotope ratio measurements Part of each powder that was sampled from the two Porites core slabs (TH1 and SOT-1) was measured for d18O on

271

a Finnigan MAT252 at GEOMAR, Kiel. The data are presented in the standard d-notation versus the VPDB standard. The analytical uncertainty was smaller than 0.06‰ (2SD). Powders from all coral samples (Porites sp. Acropora sp. and C. caespitosa) and from the inorganic experiments were dissolved in 2% HNO3 and analyzed for elemental ratios on an Agilent 7500 Quadrupole ICP-MS at the massspectrometer facilities of GEOMAR, Kiel. Drift correction was done using the bracketing standard method, in which a standard, JCp-1 (Okai et al., 2001) was measured once in every five samples measurements. In addition, all samples and standards were prepared at 10 ppm Ca concentration. The precision of the measurements was better than 2% depending on the element measured. Water samples were analyzed for elemental ratios on a Varian 720-ES ICPOES at GEOMAR, Kiel. An internal 10 ppm Yttrium (Y) standard was added to each sample to correct for analytical drift. Strontium isotope (d88/86Sr) measurements of both water samples and aragonite powders were done on a Finnigan Triton TIMS at GEOMAR in Kiel using the 87 Sr/84Sr double spike method (Krabbenho¨ft et al., 2009). Each sample was separated into two aliquots, the double spike was added to one aliquot (spiked) and the other aliquot was left unspiked. Sr was separated from the matrix solution of each aliquot using the strontium-selective chromatographic resin Eichrom Sr-spec (50–100 mesh) that is loaded in a 650 ll BIO-RAD column tube. The recovery for Sr separation was better than 90%. In each batch of carbonate sample measurements, the JCp-1 standard was also prepared and analyzed repeatedly (d88/86Sr = 0.20 ± 0.02‰, 2SD, n = 10) using the same protocol as used for the unknown samples. The IAPSO Seawater Standard (batch ID. P152) was measured in batches of seawater sample measurements (d88/86Sr = 0.39 ± 0.02‰, 2SD, n = 4). No blank correction was needed since the total procedural Sr blanks were about 0.04 ng, which is insignificant compared to the amount of Sr in the measured samples (300–700 ng). Sr isotopic values are presented in the standard d-notation relative to SRM987 as the standard (at value of 88Sr/86Sr = 8.375209). 88

d88=86 Sr ¼ ð88

Sr=86 Srsample
1Þ  1000 Sr=86 SrSRM987

ð2:2Þ

For comparison, the data is presented as the 88Sr/86Sr fractionation (D88/86Sr), which is the d88/86Sr-difference between the sample and a seawater solution that was measured separately (Table A.1 in Appendix A). D88=86 Srð‰Þ ¼ d88=86 Srsolid
d88=86 Srsolution 88/86

ð2:3Þ

The d Sr is presented with an uncertainty of 2SD, as obtained from the external long term reproducibility of measurements of the JCp-1 coral standard, all seawater samples (of approximately constant 88Sr/86Sr composition), and the IAPSO standard. The measurements of both JCp-1 and seawater (including the IAPSO seawater standard), yielded an identical external reproducibility of 0.02‰ (2SD) (Appendix A).

272

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

Table 2 d88/86Sr results of Porites sp. corals. Porites lutea from the Gulf of Eilat and Porites sp. from French Polynesia. Sample label

Sr/Ca (mmol/mol)

d18O (‰ vPDB)

Correlateda SST (°C)

d88/86Srb (‰

Porites lutea – SOT-1 P07L00A P07L00B P07H00A P07H00B P07H00C P07L99A P07L99B P07H99A P07H99B P07H99c P07L98A P07L98B P07H97A P07H97B P07H97C P07L97A P07L97B P07H96A

9.00 9.26 8.97 8.86 8.96 8.91 8.96 9.23 9.15 8.75 9.07 9.32 9.23 8.96 8.92 9.09 9.33 9.14


2.977
2.716
2.784
3.177
3.318
3.334
2.939
2.780
2.942
3.365
2.859
2.687
2.816
3.032
2.910
2.833
2.568
2.802

24.1 21.0 21.9 26.2 26.2 23.8 22.8 21.2 22.0 26.5 23.8 21.5 21.7 24.5 25.0 23.6 22.8 22.4

0.170 0.159 0.203 0.179 0.202 0.166 0.181 0.179 0.204 0.190 0.181 0.165 0.196 0.166 0.188 0.212 0.204 0.205

8.57 8.73


4.791
4.765
4.779
4.439
4.273
4.202
4.466
4.652
4.756
4.547
4.415
4.444
4.334
4.390
4.462
4.609
4.539
4.555
4.532
4.448
4.614
4.332
4.272
4.445
4.493
4.656
4.923
4.812
4.704
4.566
4.474
4.264
4.487
4.657
4.751
4.699

28.7 28.6 28.7 26.9 26.0 25.6 27.0 28.0 28.6 27.5 26.8 26.9 26.3 26.6 27.0 27.8 27.4 27.5 27.4 26.9 27.8 26.3 26.0 26.9 27.2 28.0 29.4 28.9 28.3 27.6 27.1 26.0 27.1 28.0 28.5 28.3

0.232 0.194

Porites sp. – TH1 TH-1 TH-2 TH-3 TH-4 TH-5 TH-6 TH-7 TH-8 TH-9 TH-10 TH-11 TH-12 TH-13 TH-14 TH-15 TH-16 TH-17 TH-18 TH-19 TH-20 TH-21 TH-22 TH-23 TH-24 TH-25 TH-26 TH-27 TH-28 TH-29 TH-30 TH-31 TH-32 TH-33 TH-34 TH-35 TH-36 a b

8.69

8.64

8.54 8.56 8.70 8.58 8.59 8.56 8.72 8.66 8.61 8.74 8.68 8.93 8.73 8.51 8.68 8.62

SST as reconstructed from Sr/Ca in SOT-1 and from d18O in TH-1. Reproducibility of 0.020‰ (2SD) according to JCP-1 measurements (n = 10).

0.194

0.230

0.217

0.217 0.190 0.224 0.200

0.199

SRM987)

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

3. RESULTS 3.1. Inorganic aragonite experiments at different temperatures and saturation states Inorganic experiments data were corrected for Rayleigh distillation effects, as the experiments were done in a closed system and a significant amount of Ca and Sr (up to 52%) were depleted from the bulk solution. The details of the corrections are explained in Appendix B. The data in the results are presented as the corrected results. The inorganic aragonite d88/86Sr seems to be inversely correlated to temperature (Fig. 1a). Yet, the variations between different samples precipitated at the same temperature (but with different CO2 diffusion rates) were as large as the variations between temperatures (Anova-1 P = 0.2). The current study’s inorganic results do not support any significant temperature dependency. This observation is in contrast to the inorganic aragonite results reported in Fietzke and Eisenhauer (2006) who reported a slightly positive temperature trend (with a slope of 0.0054‰ °C
1) for d88/86Sr in inorganic aragonite precipitated under similar experimental conditions. The high Mg concentration of the seawater (the experiment solution) and the addition of NaOH to control the pH caused the precipitation of brucite, Mg(OH)2, as a byproduct in some inorganic aragonite precipitation experiments. While in inorganic aragonite Sr concentration is at least 8000 ppm, the Sr concentration in brucite is approximately 5 ppm. In the current study brucite precipitation was minor compared to the total solid precipitation. We observed up to 10% of brucite in the solid from FTIR measurements. Similar observations have been made using the Mg/Ca ratio of the solid, assuming that samples with Mg/Ca > 4 mmol mol
1 (Gabitov et al., 2008) contain brucite in addition to aragonite. According to these data, the Sr that was incorporated into the brucite should be less than 0.007% of the total Sr in the precipitated solid. Therefore, we can assume that Sr in the solution was not reduced significantly due to the brucite formation, and the Sr isotope fractionation in inorganic aragonite is represented also in the experiments where brucite was formed.

273

No significant correlation is observed between d88/86Sr and precipitation rate (Fig. 1b). The apparent d88/86Srprecipitation rate relationship in the 15 °C experiments is insignificant (d88/86Sr =
0.057(±0.010)log (rate)
2 0.36(±0.02); r = 0.97; P = 0.11). In addition, d88/86Sr in the 30 °C experiments is devoid of any trend with precipitation rate. 3.2. Variations in d88/86Sr in cultured Acropora sp The geochemical data of the cultured Acropora sp. experiments was corrected due to mixing of preexperimental and the experiment skeleton in the samples (Fig. 2). The mixing correction explanation is provided in detail in Appendix C. Briefly, the binary correction was done according to the pre-experimental skeleton Sr/Ca ratio and the temperature related skeleton Sr/Ca from previous studies (Gallup et al., 2006; Reynaud et al., 2007). The correction was verified using 87Sr/86Sr values in the measured skeleton (Fig. 3). The corrected d88/86Sr results show a positive correlation between d88/86Sr and temperature (Fig. 4a), where d88/86Sr = 0.004(±0.001)T(°C) + 0.10(±0.03) (r2 = 0.84,

Fig. 2. Sr/Ca results of cultured Acropora sp. compared to the temperature of precipitation in the current study and in previous studies. Three experiments display significant differences from previous studies. The lower Sr/Ca is a result of contamination with pre-experimental skeleton, which grew in a solution that was depleted in Sr/Ca ratio.

Fig. 1. d88/86Sr in inorganic aragonite. The apparent inverse trend with temperature (a) is insignificant compared to the variations at constant temperature (P = 0.2). The variation within each temperature experiment represent different CO2 diffusion rates, which within the conditions of the experiments, are causing different precipitation rates, as shown in plot (b). Plot (b) displays D88/86Sr variations versus precipitation rates. No relationship was detected between d88/86Sr and precipitation rate. The apparent negative trend in the 15 °C experiments (white circles) is insignificant (P = 0.11).

274

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

P = 0.084). Nevertheless, this correlation is insignificant, as the method’s uncertainty is 0.02‰ (2SD). In the CO2
experiments, CO2
varied significantly 3 3 from 171 ± 13 to 352 ± 28 lmol kg
1 (Table 3). However, D88/86Sr did not vary significantly, and no d88/86Sr–CO2
3 relationship is observed (Fig. 4b). 3.3. Temperature calibrations of the cores and correlation to d88/86Sr 3.3.1. Core TH1 We obtained a good correlation in d18O between the current study results and the results of the adjacent slab that was measured by Cahyarini et al. (2008). Radiography of the slab did not show distinct density bands. Therefore, the duration covered by the transect was determined from

Fig. 3. 87Sr/86Sr in the Acropora sp. skeletons versus the amount of mixture between the pre-experiment and the experiment skeleton. The value at 0% was obtained from a sample containing only preexperiment skeleton. Mixing ratios were calculated using Sr/Ca ratios (Table 3). The good correlation (R2 = 0.98) between the extent of mixing and 87Sr/86Sr is due to the significant differences in the pre-experiment and the experiment solution composition. It confirms that the mixing correction done on Sr/Ca through Reynaud et al. (2007) is accurate enough also for correcting the isotopic composition.

the seasonal variations of the measured d18O. However, since the definite period of the transect could not be reconstructed, we used the d18O-temperature calibration that was obtained by Cahyarini et al. (2008) to calculate temperature from d18O (d18O = 0.67–0.19T(°C)). We observed four high d18O peaks (low temperatures) and five low d18O peaks (high temperatures), which represent 4.5 years of growth along the transect (Fig. 5). The d88/86Sr was measured mainly on samples with high and low d18O values (Fig. 5). The d88/86Sr values vary from 0.190‰ to 0.232‰ (0.04‰), which is close to the measurement uncertainty, and do not show any correlation with d18O and temperature. 3.3.2. Core SOT-1 All samples of Po. lutea were collected along the main growth axis on a transect, which includes several seasonal cycles with three low-density and four high-density bands. One low-density/high-density band pair represents a full annual cycle. According to the seawater temperature data, the temperature in the Gulf of Aqaba ranges between 21 °C in winter and 27 °C in summer. The extreme values of Sr/Ca were assigned to the corresponding yearly seasonal extreme temperatures (Fig. 6). The intermediate Sr/Ca values were distributed between the seasonal extremes as a linear function of their distance in the transect and the dates of the extreme temperature (Fig. 6a). The resulting correlation is Sr/Ca(mmol mol
1) = 10.8 ± 0.3–0.08 ± 0.01T(°C) (r2 = 0.6, P = 5.5  10
5). Low Sr/Ca ratios were obtained from the low density bands, whereas high Sr/Ca ratios were obtained from the high density bands. Therefore, low density bands were precipitated at high temperatures (summer season) and the high density bands were precipitated at low temperatures (winter season). This is in agreement with the timing of the high and low density bands formed in Porites sp. in this area at shallow depths (e.g. Klein and Loya, 1991; Rosenfeld et al., 2003). The corresponding d88/86Sr values vary from 0.159‰ to 0.212‰ (0.05‰), but show no relationship to Sr/Ca and temperature (Fig. 6b) (r2  0.001).

Fig. 4. d88/86Sr variations of cultured Acropora sp. with temperature (a) and with CO
2 3 (b). The black squares are the values of the samples from the temperature experiment and the gray squares represent the values of the samples in constant temperature (25 °C) but different CO
2 3 experiments. The d88/86Sr appears to be positively correlated to temperature, yet not significant (p = 0.08). In the CO
2 concentration 3 experiments three out of the four samples showed no significant variation, only one sample (212 lmol/mol of CO
2 3 ) was significantly less fractionated than the other samples.

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

275

Table 3 Acropora sp. setup conditions and results. Standard deviation from the mean (2SD) are presented in parentheses in terms of least units cited. CO
2a 3 (lmol/kg)

Sample Label

Temp. (°C)

N01 N02 N03 N04 N05 N06 N07 N08

25 171(13) 25 212(12) 25 279(15) 25 352(28) 19 293(25) 22 309(31) 28 357(34) Pre-experiment skeleton

Coral growth compared to final weight (%)

Meas. Sr/Ca (mmol/mol)

% of experiment skeleton from Gallup et al. (2006)b

% of experiment skeleton from Reynaud et al. (2007)b

87 Sr/86Sr normalizedc

Corrected d88/86Sr (‰SRM987)d

7.7 9.1 12.1 13.0 1.4 4.7 8.6

8.33 9.32 9.27 9.33 8.40 8.63 9.33 2.99

82(2) 97(3) 96(3) 97(3) 78(2) 83(2) 100(3) 0

84(2) 100(3) 100(3) 100(3) 78(2) 85(2) 100(3) 0

0.709164 0.709183 0.709206 0.70914 0.709162 0.709168 0.709182 0.708998

0.191 0.227 0.198 0.199 0.165 0.184 0.197 0.302

a

CO2
3 was calculated from the measured TA and pH using CO2Sys. % of experiment skeleton in the sample was obtained from a mixing line calculation. Sr/Ca values from previous studies were used as the experiment Sr/Ca end member and N08 Sr/Ca as the pre-experiment end member (explanation in Appendix C). c Reproducibility of 3E
5 (2SD) according to JCP-1 measurements (n = 10). d reproducibility of 0.020‰ (2SD) according to JCP-1 measurements (n = 10). b

3.4. Temperature, extension rate and pCO2 differences in C. caespitosa

Fig. 5. Results of d18O and d88/86Sr on a Porites sp. core from Tahiti (n = 10). The results are presented versus sample number from the top of the slab transect. The sampling was fairly equidistant. While d18O results display 4.5 annual cycles and correlate to the SST record from the area of the colony growth (Cahyarini et al., 2008), d88/86Sr values show no seasonal cycling or systematic relationship to the d18O values.

Samples of C. caespitosa with the high extension rates of the 21 °C and 23 °C experiments have lower 88Sr/86Sr fractionation compared to the low extension rate samples of the same temperature experiments (Table 4). Although the lower fractionation with low extension rate is observed in both temperature experiments, the differences in d88/86Sr are very small, and are significant only at the higher temperature (Fig. 7). d88/86Sr variations between temperatures are smaller than the samples’ reproducibility (2SD), and therefore show no significant correlation to temperature. Similarly, d88/86Sr did not vary significantly between the 390 ppm 701 ppm pCO2 experiments and the d88/86Sr is within the sample’s reproducibility (2SD). However, comparable to the results observed for the different extension rates, a decrease in pCO2 (which means an increase in pH and in the aragonite saturation state) causes a small increase in 88Sr/86Sr fractionation (Fig. 7). This may suggest a small decrease in Sr isotope fractionation with the increase of calcification rates.

Fig. 6. (a) In the Porites lutea core from the Gulf of Eilat (n = 18) an inverse correlation was found between SST (black circles) and the measured Sr/Ca (gray circles), (Sr/Ca(mmol/mol) = 10.8–0.8 T(°C) and R2 = 0.6). No inferences can be made on the d88/86Sr–Sr/Ca relationship (b), since the measured variation of d88/86Sr values (black circles in (b)) is only slightly above reproducibility, i.e. all values are identical within statistical uncertainy.

276

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

Table 4 Results of Cladocora caespitosa experiments. Temperature and pCO2 results are reported with the standard deviation of the mean (2SD). The linear extension is reported with the standard error (1SE). Sample Label

Temp. (°C)

pCO2 (ppm)

Linear extension (mm)

Sr/Ca (mmol/mol)

d88/86Sr (‰SRM987)a

Clad Clad Clad Clad Clad Clad Clad Clad

15.00(0.05) 18.00(0.05) 21.00(0.05) 23.00(0.05) 21.00(0.05) 23.00(0.05) 16.4(2.6) 16.3(2.6)

– – – – – – 390(48) 701(78)

0.728(0.045) 1.305(0.060) 1.984(0.069) 1.283(0.055) >2.053 >1.338 N.A N.A

9.20 9.14 8.90 8.86 8.86 8.78 9.08 9.04

0.191 0.195 0.180 0.160 0.205 0.209 0.193 0.182

a

15 18 21 23 21L 23L 400 700

Reproducibility of 0.020‰ (2SD) according to JCP-1 measurements (n = 10).

aragonite and the different coral species (Acropora sp., Porites sp. and C. caespitosa) display no significant differences (Fig. 9). 4.1. 88Sr/86Sr at varying environmental parameters in inorganic and corals aragonite

Fig. 7. d88/86Sr results of Cladocora ceaspitosa that was cultured at different temperatures and different pCO2. The variations between the different experiment conditions fall within the 2SD. At 23 °C we observe a significant difference between the specimens with low linear extension (more fractionated) and the specimens with high linear extention (less fractionated). Smaller fractionation is observed as the linear extention increases (in 21 and 23 °C experiments). d88/86Sr in the two pCO2 experiments (black circles) are similar within the 2SD.

4. DISCUSSION In the current study we examined several different aragonite samples with temperature variations of up to 15 °C. The 88Sr/86Sr isotope ratios in all aragonite samples analyzed in this study show a significant fractionation from their respective bulk solutions (seawater). This fractionation is observed for both inorganic and coral aragonite (from laboratory-cultured experiments and natural specimens), and results in an average isotopic difference of 0.2‰ relative to the bulk solution (Fig. 9). Since the Red-Sea water that was used in all experiments has a similar composition to IAPSO (see Appendix A), the D88/86Sr was a subtraction of the measured solid sample from the average of all the measured Red-Sea water and IAPSO measurements. The D88/86Sr values of the inorganic

Early studies of d88/86Sr in corals and inorganic aragonite have reported a variable fractionation in d88/86Sr correlated to temperature. Specifically, Fietzke and Eisenhauer (2006) found a positive correlation with temperature in both P. clavus and inorganic aragonite with slopes of 0.033‰ °C
1 and 0.0054‰ °C–1, respectively. A positive d88/86Sr-temperature correlation (d88/86Sr = 0.026T(°C)–0. 059) was also observed by Ru¨ggeberg et al. (2008) for the cold water coral, L. pertusa. In contrast to the early studies, later studies could not confirm either the large variations in d88/86Sr in coral aragonite, or their temperature correlation (Krabbenho¨ft et al., 2010; Raddatz et al., 2013). The current study consists of a large dataset including several coral species and inorganic aragonite. The d88/86Sr data of the current study displays only a minor range of variations. Most of the observed variations are close to the analytical uncertainty of TIMS-based double spike measurement, despite the larger temperature and precipitation rate range covered. For comparison, Fietzke and Eisenhauer (2006) observed a 0.17‰ variation over a 5 °C temperature change in P. clavus and the in current study a 9 °C variation in Acropora sp. experiments resulted in only 0.03‰ of variability in d88/86Sr. The entire set of data result in only 0.08‰ change in total. The total D88/86Sr variation is indeed larger than the 2SD of the measurements, but shows no correlation to either temperature or precipitation rates. There is no significant difference between inorganic aragonite D88/86Sr and coral aragonite D88/86Sr as the average inorganic aragonite D88/86Sr is
0.171 ± 0.046‰ and the average coral D88/86Sr is
0.190 ± 0.049‰. The inorganic aragonite precipitation experiments were conducted over large range of precipitation rates (the highest rate is 40 times larger than the lowest rate) and a 15 °C temperature range. However, d88/86Sr results varied by 0.08‰ and show no significant correlation to either temperature or to precipitation rate. The apparent slightly negative trends observed in the D88/86Sr vs. temperature plot (R2 = 0.41 Anova-1 P = 0.2) and also in the D88/86Sr vs.

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

Fig. 8. Application of the Rayleigh-based-multi-element model to the Acropora sp. experiment of the current study. Mg/Ca vs. Sr/Ca as obtained by the model (dashed line) shows a polynomial fit that follows the trend of the measured Mg/Ca and Sr/Ca of the Acropora sp. samples. The good fit between the model and the measured data demonstrates that the Rayleigh-based model can be applied to the subject experiment and can further be used to test the model on Sr isotopic system.

precipitation rate plot (for the 15 °C experiment; R2 = 0.99, P = 0.11) are insignificant (Fig. 1). 4.2. Can the Rayleigh-based multi element model (Gaetani et al., 2011) explain the Sr incorporation in corals skeleton? The Rayleigh-based multi element model was applied to the data from the Acropora sp. temperature experiment using the equations from Gaetani et al. (2011) model

277

and are explained in Appendix D. The calculated Mg/Ca–Sr/Ca data of the Acropora sp. specimens are inversely correlated and are in accordance with the measured Mg/Ca–Sr/Ca (Fig. 8). This good fit shows that the Acropora sp. experiment complies with the Rayleigh-based multi element model, as suggested by Gaetani et al. (2011). According to the model calculation, the proportion of Sr that remains in the calcifying fluid is between 0.22 and 0.09 from its initial concentration (Eq. B1 in Appendix B), for the 19 °C and 28 °C experiments, respectively. This means that as the temperature increases more Sr is removed from the calcifying fluid and the Rayleigh distillation effect increases. Because of a near quantitative removal the element and isotopic ratios tend to shift closer towards the initial solution composition as the temperature increases. When applying the Rayleigh-based model to the D88/86Sr results (Appendix D) for the same Acropora sp. experiments, the results predicted by the model conflict with the measured data. The model suggests that D88/86Sr in coral aragonite should be up to 0.15‰ less fractionated than open-system inorganic aragonite (Fig. 9, Table 5). Since the inorganic aragonite and all coral samples, including the Acropora sp. experiment, show the same narrow range of variation in D88/86Sr, it appears that the model does not predict the fractionation of Sr observed in the corals skeleton. In our calculations we considered that the initial d88/86Sr is similar to the seawater composition (0.39 ± 0.02‰). Gaetani et al. (2011), however, suggested that the initial composition of the bulk solution can be slightly modified from the seawater composition. In order to test the possibility that the initial d88/86Sr within the coral calcifying space is modified from the normal seawater value, we calculated the initial d88/86Sr of the bulk solution that would result

Fig. 9. D88/86Sr and d88/86Sr versus temperature in biogenic and inorganic aragonite from the current study. Fractionation from the seawater composition with preference for the light isotope is observed in both inorganic and biogenic aragonites (
0.2‰). The D88/86Sr values vary within a narrow range of 0.1‰, where different coral species and inorganic aragonite overlap. There is no observed temperature effect on D88/86Sr. The Rayleigh based model (Gaetani et al., 2011) predicts a smaller fractionation in corals compared to the inorganic aragonite (marked as two dashed gray lines). The two lines were calculated from the Rayleigh model using the maximum and minimum 88Sr/86Sr fractionation factors (aA/aq) in the inorganic aragonite experiments. Our results show that in contrast to the Rayleigh model there no significant difference between the inorganic and coral aragonite 88Sr/86Sr fractionation.

278

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

Table 5 Rayleigh based model calculation for D88/86Sr in corals. Standard deviation of the mean (2SD) is reported in the parentheses in terms of least units cited. a

Temp. (°C)

f

15 20 25 30

0.99945 0.99933 0.99923 0.99922

a

a = 0.9997875b calculated D88/86Sr

a = 0.99986847b calculated D88/86Sr

Average corals’ D88/86Sr from the current studyc


0.058(10)
0.044(10)
0.028(20)
0.029(40)


0.101(10)
0.076(10)
0.049(20)
0.051(40)


0.201(20)
0.209(18)
0.191(20)
0.176(20)

n=1 n=5 n=6 n=6

The mass fraction of the initial fluid that remained at the end of the experiment (Eq.D.2). Sr/86Sr fractionation factors, a, were calculated form the inorganic aragonite experiments. Average for each temperature of all corals samples measured in the current study Porites sp., Acropora sp. and Cladocora caespitosa.

b 88 c

in coral aragonite having similar d88/86Sr to inorganic aragonite (Appendix D.2). We obtained a large range of d88/86Sr in the calcifying solution which depends on the seawater temperature (0.317‰ at 15 °C and 0.232‰ at 30 °C). The different initial conditions for each temperature suggest a temperature dependent isotopic fractionation that takes place when seawater enters the calcifying space and contradicts the simple Rayleigh-based model, which assumes that the calcifying fluid can be treated as constant throughout the year for a given coral. A similar observation was discussed by Rollion-Bard et al. (2009) who found comparable fractionation in Li isotopes between inorganic and coral aragonite. While the Li/Ca ratio follows the Rayleigh-based model, Li isotopes are unaffected by the Rayleigh process. The authors explained the disagreement of the Li isotopes data with the Rayleigh-based model by having a large reservoir of Li in solution compared to the small amount incorporated into the coral skeleton. Case et al. (2010) proposed that Ca concentration variations in the calcifying fluid dominate the Rayleigh fractionation effect. For this reason, highly discriminated elements such as Mg and Li vary according to the Rayleigh fractionation of Mg/Ca and Li/Ca, but Li isotopes are not affected (Case et al., 2010). For Sr a similar argument is not reasonable. Sr, in contrast to Li, is not highly discriminated in the aragonite lattice as the partitioning coefficient of Sr/Ca is close to unity. This implies that if Sr/Ca in corals is strongly affected by the Rayleigh process (Gagnon et al., 2007; Gaetani et al., 2011), Sr isotopes should be strongly affected, too. Our data do not show such behavior which in our opinion challenges the application of the Rayleigh-based model to coral calcification.

aragonite due to the finite solution conditions during coral calcification. However, our data show that inorganic aragonite precipitated in an open system has similar D88/86Sr as coral aragonite. 88 Sr/86Sr fractionation (D88/86Sr) of
0.20‰ is shown by both inorganic aragonite and biogenic (coralline) aragonite. Since the 88Sr/86Sr fractionation in aragonite is not influenced by precipitation rates (either caused by temperature or saturation state variations), it is possible to use the relatively constant fractionation for bulk solution 88Sr/86Sr composition reconstruction. Furthermore, combining the reconstructed 88Sr/86Sr composition in the bulk solution can therefore be used to estimate calcite precipitation rates (as in calcite 88Sr/86Sr fractionation is influenced by precipitation rates). ACKNOWLEDGMENTS We wish to thank Ana Kolevica, Regina Surberg, Andrea Wolf and Maria Hierz for supporting the lab work of this study. C. caespitosa coral samples were cultured during two experiments funded by the Centre Scientifique de Monaco with the contribution of the European Project on Ocean Acidification (EPOCA). We thank C. Ferrier-Page`s, D. Allemand, S. Reynaud, J.-P. Gattuso and S. Martin for their help and contribution during these experiments. Funding support was provided through the European Marie Curie Initial Training Network ‘‘Calcification by Marine Organisms” (CalMarO) and the European Community’s Seventh Framework Programme (FP7/2007-2013). Finally, the authors would like to thank M. S. Fantle and the two anonymous reviewers for their valuable comments and suggestions, which significantly improved that paper. The authors would also like to thank A. E. Claudine Stirling for the editorial handling and the detailed helpful advices.

5. CONCLUSIONS APPENDIX A. SUPPLEMENTARY DATA The aragonitic corals analyzed in this study show a total range of 0.08‰ in D88/86Sr values. While this range has significant variability (being larger than the analytical uncertainty of 0.02‰), it is neither related to the ambient temperature nor to precipitation rates. This result stands in contrast to previous reports of temperature dependency in d88/86Sr in corals (Fietzke and Eisenhauer, 2006; Ru¨ggeberg et al., 2008). The d88/86Sr values from corals contradict the Rayleighbased model. The model predicts smaller 88Sr/86Sr fractionation in corals compared to open system inorganic

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.gca.2016.01.039. REFERENCES Beck J. W., Edwards R. L., Ito E., Taylor F. W., Recy J., Rougerie F., Joannot P. and Henin C. (1992) Sea-surface temperature from coral skeletal strontium/calcium ratios. Science 257(80), 644–647.

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280 Cahyarini S. Y., Pfeiffer M., Timm O., Dullo W. C. and Schoenberg D. G. (2008) Reconstructing seawater d18O from paired coral d18O and Sr/Ca ratios: methods, error analysis and problems, with examples from Tahiti (French Polynesia) and Timor (Indonesia). Geochim. Cosmochim. Acta 72, 2841–2853. Cahyarini S. Y., Pfeiffer M. and Dullo W. C. (2009) Improving SST reconstructions from coral Sr/Ca records : multiple corals from Tahiti (French Polynesia). Int. J. Earth Sci., 31–40. Case D. H., Robinson L. F., Auro M. E. and Gagnon A. C. (2010) Environmental and biological controls on Mg and Li in deepsea scleractinian corals. Earth Planet. Sci. Lett. 300, 215–225. Chisholm J. R. M. and Gattuso J. P. (1991) Validation of the alkalinity anomaly technique for investigating calcification and photosynthesis in coral reef communities. Limnol. Oceanogr. 36, 1232–1239. Corre`ge T. (2006) Sea surface temperature and salinity reconstruction from coral geochemical tracers. Palaeogeogr. Palaeoclimatol. Palaeoecol. 232, 408–428. Dietzel M., Gussone N. and Eisenhauer A. (2004) Co-precipitation of Sr2+ and Ba2+ with aragonite by membrane diffusion of CO2 between 10 and 50 °C. Chem. Geol. 203, 139–151. Dodge R. E., Wyers S. C., Frith H. R., Knap A. H., Smith S. R., Cook C. B. and Sleeter T. D. (1984) Corals calcification rates by the buoyant weight technique: effect of alizarin staining. J. Exp. Mar. Biol. Ecol. 15, 217–232. Fallon S. J., McCulloch M. T., van Woesik R. and Sinclair D. J. (1999) Corals at their latitudinal limits: laser ablation trace element systematics in Porites from Shirigai Bay, Japan. Earth Planet. Sci. Lett. 172, 221–238. Fietzke J. and Eisenhauer A. (2006) Determination of temperaturedependent stable strontium isotope (88Sr/86Sr) fractionation via bracketing standard MC-ICP-MS. Geochemistry Geophys. Geosystems 7. Finch A. A. and Allison N. (2008) Mg structural state in coral aragonite and implications for the paleoenvironmental proxy. Geophys. Res. Lett. 35, 1–5. Gabitov R. I., Gaetani G. A., Watson E. B., Cohen A. L. and Ehrlich H. L. (2008) Experimental determination of growth rate effect on U6+ and Mg2+ partitioning between aragonite and fluid at elevated U6+ concentration. Geochim. Cosmochim. Acta 72, 4058–4068. Gaetani G. A. and Cohen A. L. (2006) Element partitioning during precipitation of aragonite from seawater: a framework for understanding paleoproxies. Geochim. Cosmochim. Acta 70, 4617–4634. Gaetani G. A., Cohen A. L., Wang Z. and Crusius J. (2011) Rayleigh-based, multi-element coral thermometry: A biomineralization approach to developing climate proxies. Geochim. Cosmochim. Acta 75, 1920–1932. Gagnon A. C., Adkins J. F., Fernandez D. and Robinson L. (2007) Sr/Ca and Mg/Ca vital effects correlated with skeletal architecture in a scleractinian deep-sea coral and the role of Rayleigh fractionation. Earth Planet. Sci. Lett. 261, 280–295. Gallup C. D., Olson D. M., Edwards R. L., Gruhn L. M., Winter A. and Taylor F. W. (2006) Sr/Ca-Sea surface temperature calibration in the branching Caribbean coral Acropora palmata. Geophys. Res. Lett. 33, 2–5. Halicz L., Segal I., Fruchter N., Stein M. and Lazar B. (2008) Strontium stable isotopes fractionate in the soil environments? Earth Planet. Sci. Lett. 272, 406–411. Johnson, C. M., Beard, B. and Albarede, F. (eds.) (2004) Geochemistry of non-traditional stable isotopesRosso, J. J. (ed.) (2004)Reviews in Mineralogy and Geochemistry. Mineralogical Society of America, pp. 1–24, vol. 55. Klein R. and Loya Y. (1991) Skeletal growth and density patterns of two Porites corals from the Gulf of Eilat, Red Sea. Mar. Ecol. Prog. Ser. 77, 253–259.

279

Krabbenho¨ft A., Fietzke J., Eisenhauer A., Liebetrau V., Bo¨hm F. and Vollstaedt H. (2009) Determination of radiogenic and stable strontium isotope ratios (87Sr/86Sr; d88/86Sr) by thermal ionization mass spectrometry applying an 87Sr/84Sr double spike. J. Anal. At. Spectrom. 24, 1267. Krabbenho¨ft A., Eisenhauer A., Bo¨hm F., Vollstaedt H., Fietzke J., Liebetrau V., Augustin N., Peucker-Ehrenbrink B., Mu¨ller M. N., Horn C., Hansen B. T., Nolte N. and Wallmann K. (2010) Constraining the marine strontium budget with natural strontium isotope fractionations (87Sr/86Sr*, d88/86Sr) of carbonates, hydrothermal solutions and river waters. Geochim. Cosmochim. Acta 74, 4097–4109. Lewis E., Wallace D. and Allison L. J. (1998) Program Developed for CO2 System Calculations. Marshall J. F. and McCulloch M. T. (2002) An assessment of the Sr/Ca ratio in shallow water hermatypic corals as a proxy for sea surface temperature. Geochim. Cosmochim. Acta 66, 3263– 3280. Meibom A., Cuif J., Constantz B. R. and Juillet A. (2004) Distribution of magnesium in coral skeleton. Geophys. Res. Lett. 31, 2–5. Mitsuguchi T., Matsumoto E., Abe O., Uchida T. and Isdale P. J. (1996) Mg/Ca thermometry in coral skeletons. Science 274(80), 962–963. Montagna P., McCulloch M. T. and Mazzoli C. (2007) The nontropical coral Cladocora caespitosa as the new climate archive for the Mediterranean: high-resolution (weekly) trace element systematics. Quat. Sci. Rev. 26, 441–462. Montagna P., McCulloch M., Douville E., Lo´pez Correa M., Trotter J., Rodolfo-Metalpa R., Dissard D., Ferrier-Page`s C., Frank N., Freiwald A., Goldstein S., Mazzoli C., Reynaud S., Ru¨ggeberg A., Russo S. and Taviani M. (2014) Li/Mg systematics in scleractinian corals: calibration of the thermometer. Geochim. Cosmochim. Acta 132, 288–310. Niedermayr A. (2011) Effects of magnesium, polyaspartic acid, carbonate accumulation rate and temperature on the crystallization, morphology, elemental incorporation and isotopic fractionation of calcium carbonate phases (Dissertation). TU Graz. Okai T., Suzuki A. and Kawahata H. (2001) Preparation of a New Geological Survey of Japan Geochemical Reference Material: coral JCp-1. J. Geostand. Geoanal. 26, 95–99. Quinn T. M. and Sampson D. E. (2002) A multiproxy approach to reconstructing sea surface conditions using coral skeleton geochemistry. Paleoceanography 17. Raddatz J., Liebetrau V., Ru¨ggeberg A., Hathorne E. C., Krabbenho¨ft A., Eisenhauer A., Bo¨hm F., Vollstaedt H., Fietzke J., Lo´pez Correa M., Freiwald A. and Dullo W.-C. (2013) Stable Sr-isotope, Sr/Ca, Mg/Ca, Li/Ca and Mg/Li ratios in the scleractinian cold-water coral Lophelia pertusa. Chem. Geol. 352, 143–152. Reynaud S., Ferrier-Page`s C., Meibom A., Mostefaoui S., Mortlock R., Fairbanks R. and Allemand D. (2007) Light and temperature effects on Sr/Ca and Mg/Ca ratios in the scleractinian coral Acropora sp. Geochim. Cosmochim. Acta 71, 354–362. Rollion-Bard C., Vigier N., Meibom A., Blamart D., Reynaud S., Rodolfo-Metalpa R., Martin S. and Gattuso J.-P. (2009) Effect of environmental conditions and skeletal ultrastructure on the Li isotopic composition of scleractinian corals. Earth Planet. Sci. Lett. 286, 63–70. Romanek C. S., Morse J. W. and Grossman E. L. (2011) Aragonite kinetics in dilute solutions. Aquat. Geochem. 17, 339–356. Rosenfeld M., Yam R., Shemesh A. and Loya Y. (2003) Implication of water depth on stable isotope composition and skeletal density banding patterns in a Porites lutea colony: results from a long-term translocation experiment. Coral Reefs 22, 337–345.

280

N. Fruchter et al. / Geochimica et Cosmochimica Acta 178 (2016) 268–280

Ru¨ggeberg A., Fietzke J., Liebetrau V., Eisenhauer A., Dullo W. C. and Freiwald A. (2008) Stable strontium isotopes (d88/86Sr) in coldwater corals – A new proxy for reconstruction of intermediate ocean water temperatures. Earth Planet. Sci. Lett. 269, 570–575. Schmidt G. A. (1999) Error analysis of paleosalinity calculations. Paleoceanography 14, 422–429. Smith S. V. and Kinsey D. W. (1978) Calcification and organic carbon metabolism as indicated by carbon dioxide. Coral Reefs Res. Methods. Tang J., Dietzel M., Bo¨hm F., Ko¨hler S. J. and Eisenhauer A. (2008) Sr2+/Ca2+ and 44Ca/40Ca fractionation during inorganic

calcite formation: II. Ca isotopes. Geochim. Cosmochim. Acta 72, 3733–3745. Weber J. N. and Woodhead P. M. J. (1972) Temperature dependence of oxygen-18 concentration in reef coral carbonates. J. Geophys. Res. 77, 463. Wiechers H. N. S., Sturrock P. and Marais G. V. R. (1975) Calcium carbonate crystallization kinetics. Water Res. 9, 835– 845. Associate editor: Claudine Stirling