Tropical seasonality in the late Campanian (Late Cretaceous): Comparison between multiproxy records from three bivalve taxa from Oman

de Winter, Niels J.1, Goderis, Steven1,2, Dehairs, Frank1, Jagt, John W.M.3, Fraaije, René H.B.4, Van Malderen, Stijn J.M.2, Vanhaecke, Frank2, Claeys, Philippe1

1Analytical, Environmental and Geochemistry (AMGC) Department, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussels, Belgium

2Department of Analytical Chemistry (A&MS), Ghent University, Krijgslaan 281, Campus Sterre - S12, 9000 Ghent, Belgium

3Natuurhistorisch Museum Maastricht (NHMM), Bosquetplein 7, 6211 KJ, Maastricht, the Netherlands

4Oertijdmuseum De Groene Poort, Bosscheweg 80, 5283 WB, Boxtel, the Netherlands

Corresponding author: niels.de.winter@vub.be

Keywords: Palaeotemperature, mollusc, rudist, trace element, stable isotope, µXRF

Abstract

Geochemical proxy records from calcite shells of bivalves constitute an important archive for the reconstruction of palaeoenvironmental conditions on sub-annual timescales. However, the incorporation of these trace element and stable isotope proxies into the shell is influenced by a multitude of physiological and environmental factors that need to be disentangled to enable reliable reconstruction of palaeoclimate and palaeoenvironment. In this study, records of multiple proxies in three bivalve taxa from the same late Campanian locality in Oman are used to study the expression of various geochemical proxies in relation to each other and to the palaeoenvironment. Micro-X-Ray Fluorescence mapping allows the localization, discussion and evasion of diagenetically altered parts of the fossil shells. X-Ray Fluorescence line scanning calibrated with Laser Ablation Inductively Coupled Plasma Mass Spectrometry is used to measure trace element profiles through well-preserved calcitic parts of the shells. Records of stable carbon and oxygen isotope ratios of shell calcite are combined with these high-resolution trace element concentration profiles to study sub-annual variations in shell chemistry and reconstruct changes in the palaeoenvironment of the bivalves on a seasonal scale. Spectral analysis routines are used to detect cyclicity in stable isotope (δ18O and δ13C) and trace element (Mg/Ca, Sr/Ca, S/Ca and Zn/Ca) records. Differences in seasonal expression between these chemical proxies and between individual shells are discussed in terms of the relative influence of palaeoenvironment and potential species-specific physiological effects. Stable oxygen isotope ratios between shells suggest a local palaeotemperature seasonality of 8°C around an annual mean of 28°C, with the shell of the rudistid Torreites sanchezi milovanovici yielding slightly higher average temperatures. The discussion of the application of various Mg/Ca palaeotemperature calibrations on Mg/Ca records in these bivalve species emphasizes the complexity of using trace element proxies in extinct bivalve species. It shows that long-term changes

12

34

5

67

89

1011

12

13

14

15

16

17

18

19

20212223242526272829303132333435363738394041

in Mg/Ca ratios in ocean water need to be taken into account and that Mg/Ca ratios in bivalves might be influenced by vital effects. Sr/Ca and S/Ca ratios in these fossil taxa are likely controlled by growth and metabolic rates of the shell, although an influence of local salinity on strontium-to-calcium ratios cannot be excluded. Sub-annual variations in zinc concentrations in shell calcite may reflect seasonal variations in palaeoproductivity and redox conditions in the water column.

1. Introduction

1.1 Late Cretaceous climate reconstruction

The Late Cretaceous is a well-studied interval of geological time, in part due to its hothouse climate punctuated by episodes of rapid climatic and environmental change (Jenkyns, 1980; Huber et al., 1995; Li and Keller, 1999; Jenkyns et al., 2004). During the Late Cretaceous, atmospheric pCO2 levels are estimated to have been two to four times higher (i.e. ~400 to ~1000 ppmV) compared to the pre-industrial level (280 ppmV; e.g., Berner, 1990; Andrews et al., 1995; Ekart et al., 1999; Quan et al., 2009). The values projected by the Intergovernmental Panel on Climate Change (IPCC, 2013) indicate that such high values may be reached within the 21st century – the current global mean pCO2 value is already >400 ppmV and rises at a rate of 2-3 ppmV/year (Dlugokencky, 2017). This makes the Late Cretaceous an important geological time period to study the effects of increased pCO2 on Earth’s climate. As such, long-term changes in climate and environment during this time period have received much attention (e.g., Eshet and Almogi-Labin, 1996; Clarke and Jenkyns, 1999; Huber et al., 2002; Faul et al., 2003; Friedrich et al., 2012). However, climatic and environmental changes in the Cretaceous greenhouse world should also be understood on a much shorter timescale to put long term climate models and reconstructions in their proper context. Therefore, there is a need to supplement widely available, long-term palaeoclimate records, such as those obtained through deep-sea drilling expeditions and continuous sedimentary records (e.g. Li and Keller, 1999; Huber et al., 2002), with more rare data sets that record sub-annual variations.

1.2 Bivalve shells as climate recorders

Fossil bivalve shells are well-studied records that track these sub-annual climate fluctuations. Marine bivalves record changes in their environment, and thus local climate, as a result of environment-dependent chemical and isotopic fractionation that takes place during the incremental growth of their carbonate (aragonitic/calcitic) shell (Epstein et al., 1953; Grossman and Ku, 1986; Jones, 1980; 1983; Klein et al., 1996a; Lazareth et al., 2003). Chemical variations of pristinely preserved shell material can record changes in the animal’s environment and local climate down to daily resolution, much higher than can be retrieved through conventional long timescale climate archives. Thick-shelled rudistid (Mollusca: Hippuritida, see Newell, 1965) and ostreid bivalves (Mollusca: Ostreoida, see Bouchet et al., 2010) are suitable taxa for this type of research because their thick calcitic shells are more resistant to diagenesis than those of their aragonitic relatives (Al-Aasm and Veizer, 1986a; b; Steuber, 1999; Ullman and Korte, 2015). The use of such shells to reconstruct climate on a sub-seasonal scale has been amply demonstrated by earlier work on both rudistid (Steuber, 1996; 1999; Steuber et al., 2005; Immenhauser et al., 2005; Hennhöfer et al., 2012) and ostreid shells (Kirby et al., 1998; Brigaud et al., 2008; Ullman et al., 2010; Fan et al., 2011). However, there are complications with the use of shells of extinct bivalve species to reconstruct climate because of potential species-specific differences in physiological (“vital”) effects, such as the effects of varying growth and metabolic rate on palaeoclimate proxies (Fastovsky et al., 1993;

4243444546

47

48

49

50

5152535455565758596061626364656667

68

69

7071727374757677787980818283848586

Steuber, 1999; Lorrain et al., 2005; Gillikin et al., 2005; Immenhauser et al., 2005; Schöne et al., 2011). This requires the construction of species-specific transfer functions for the relationship between climatic and environmental parameters and geochemical proxies in the shell, which requires culture experiments (e.g. Wanamaker et al., 2007; Freitas et al., 2008). For extinct bivalve species, proxy data sets can be cross calibrated using related extant species (e.g. Armendáriz et al., 2008). For species without close extant relatives, such as rudistids, the use of modern bivalve transfer functions can, however, severely reduce the reliability of reconstructions.

1.3 Multi-species approach

The reliability of proxy interpretation in these bivalves can be improved by comparing records from multiple species and by using multiple proxies. This approach allows isolation of the effect of different parameters on the measured proxies when culture experiments cannot be carried out (Klein et al., 1996b; Carroll et al., 2009). The present work discusses a multi-proxy analysis of a single ostreid shell and two associated rudistid taxa from the Campanian Samhan Formation from the Saiwan area in the Sultanate of Oman, which is characterized by long horizontally-continuous successions of reef-like deposits of bivalves, stromatoporoids and corals. This shallow marine Tethys Ocean succession contains exceptionally well-preserved rudistids and other bivalves in life position, allowing multi-species paleoenvironmental reconstructions to be placed in a wider framework of Late Cretaceous reconstructions in the Tethys Ocean. The use of three different species in a multi-proxy approach enables an assessment of potential species-specific differences in proxy relationships and offers a unique opportunity to isolate seasonal fluctuations of climatic parameters recorded in extinct bivalve species. In addition to reconstructing past seasonality, the use of multiple species and multiple proxies permits a more general discussion of the value of trace element and stable isotope records from fossil bivalve shells.

2. Background

2.1 Palaeoenvironmental proxies in calcite skeletons

A wide range of marine calcifying organisms has been the subject of studies aiming to reconstruct sub-annual climate. Detailed studies of stable isotope ratios and trace element concentrations in corals (Goreau, 1977; Swart, 1983; Sinclair et al., 1998), coralline algae (Halfar et al., 2000; Hetzinger et al., 2011), sclerosponges (Lazareth et al., 2000; Swart et al., 2002), molluscs (e.g. Jacob et al., 2008; McConnaughey and Gillikin, 2008; Elliot et al., 2009; Goodwin et al., 2013; Ullmann et al., 2010; 2013; 2015), brachiopods (Veizer et al., 1986; Grossman et al., 1996) and stromatolites (Chafetz et al., 1991; Andrews and Brasier, 2005) have shown that many of these organisms and colonies record changes in the chemistry of their carbonate skeleton in response to seasonal changes in their environment. It is reported that most of these marine organisms precipitate their carbonate skeletons in equilibrium with ambient seawater, allowing sub-annual reconstruction of palaeotemperature from their stable isotope ratios. However, several other environmental parameters, such as salinity, nutrient availability and sea water chemistry, complicate the reconstruction of palaeoseasonality in shallow marine records (Klein et al., 1996b; Purton and Brasier, 1997; Surge and Lohmann, 2002). In response to these complications and in an attempt to obtain a more complete picture of sub-annual variations in the local environment, several trace element proxies for changes in sea water chemistry have been developed for calcifying organisms. Magnesium-to-calcium (Mg/Ca) ratios in carbonates have been proposed as proxies for

87888990919293

94

95

96979899

100101102103104105106107108109110

111

112

113

114

115116117118119120121122123124125126127128129130131

palaeotemperature, potentially allowing independent control on temperature reconstructions and the determination of salinity changes in combination with stable oxygen isotope ratios (Mitsuguchi et al., 1996; Klein et al., 1996a; Richardson et al., 2004; Wanamaker et al., 2008; Schöne et al., 2011). Additionally, strontium-to-calcium (Sr/Ca) ratios have been shown to vary with calcification temperature and salinity in some bivalves and corals (Dodd and Crisp, 1982; de Villiers et al., 1995; Richardson et al., 2004; Freitas et al., 2006). Several trace element proxies have been suggested to record changes in palaeoproductivity and redox conditions in ambient sea water. Examples include redox sensitive elements such as manganese (Mn; Freitas et al., 2006), elements that are enriched in the skeletons of primary producers such as barium (Ba, Gillikin et al., 2008; Hatch et al., 2012; Marali and Schöne, 2015) and micronutrients such as zinc (Zn) and cadmium (Cd) which are readily taken up into the shells of bivalves (Carriker et al., 1980; Calmano et al., 1993; Jackson et al., 1993; Wang and Fisher, 1996; Guo et al., 1997). Other studies have demonstrated the reproducibility of trace element profiles between specimens living in the same environment, showing that these records are likely to be governed by environmental factors (Gillikin et al., 2008; Marali et al., 2017).

2.2 Vital effects

However, it is now recognized that trace elements are often not taken up in equilibrium with the environment and that vital effects influence many of these proxies (see Weiner and Dove, 2003 and references therein). Examples include studies that show relationships between trace element concentrations and mineral growth in bivalves (Lorrain et al., 2005; Gillikin et al., 2005; Freitas et al., 2006; Carré et al., 2006; Schöne et al., 2011), evidence of the breakdown of proxy relationships during periods of growth stress (Lorens and Bender, 1980; Hendry et al., 2001; Takesue and van Geen, 2004) and the effect of changing redox conditions in the sea water (Vander Putten et al., 2000; Hendry et al., 2001; Freitas et al., 2006; Beck et al., 2017). Independent chronologies and proxies for the amount of organic matrix in the shell (e.g. sulphur (S) concentrations) may be used to better constrain vital effects in biogenic carbonate records by reconstructing the physiological cycle of the organism (de Villiers et al., 1994; Sinclair et al., 1998; Rosenberg et al., 2001; Dauphin et al., 2003; Marali and Schöne, 2015). An example of how the integration of proxies for physiological changes enables the discussion of proxy interpretation, is the observation of covarying magnesium (Mg) and sulphur (S) concentrations in bivalve records as evidence for physiological control on the incorporation of trace elements into bivalve shell calcite (Lorens and Bender, 1980; Dauphin et al., 2003; England et al., 2007; Perrin et al., 2017). Recent studies have therefore focused on the integration of multi-proxy records from multiple sub-annual climate records in order to isolate the effects of environmental parameters (e.g. Black et al., 2009; Wanamaker et al., 2011; Marali et al., 2017).

2.3 Geological setting

The three bivalve specimens described in this study originate from the upper Campanian middle member of the Samhan Formation in the Saiwan area of the Huqf Desert in east-central Oman (Schumann, 1995). The two rudistid specimens were retrieved in life position from a 200 m wide Vaccinites-dominated rudist biostrome that contains many large Torreites rudists (Unit V1 in profile 1 in Schumann 1995; biostrome unit 2 in Philip and Platel, 1996), while the thick-shelled oyster (Oscillopha figari) originated from the top of the marly limestone unit directly below the rudistid horizon (top of Unit 1 in profile 1 in Schumann 1995). The Samhan Formation records the development of a shallow marine carbonate platform during a gradual transgression (Platel et al.,

132133134135136137138139140141142143144145

146

147

148149150151152153154155156157158159160161162163164165166

167

168

169170171172173174175176

1994). Previously established as early Campanian (Platel et al., 1994), the middle and upper Samhan Formation were later dated as late Campanian using ammonite biostratigraphy by Kennedy et al. (2000). This is in congruence with the dating of the Orbitoides tissoti zone, in which the Samhan Formation is found (Philip and Platel, 1995), to the early late Campanian (Korbar et al., 2010). Schumann (1995) described the Samhan Formation as a shelf setting far from the coast, with no evidence of coastal sedimentation. The palaeoenvironment was frequently disturbed by strong currents and storm events and was characterized by high input of weathering products from Omani ophiolites. The Vaccinites-dominated biostromes are presumably formed during periods of relatively lower hydro-energy and were very short-lived (100s of years; Schumann, 1995). However, it is hypothesized that the low diversity of these biostromes suggests an adaptation of Vaccinites and Torreites to a highly turbulent environment. The O. figari specimen was found 1-2m below the biostrome and is therefore assumed to have lived before the closely associated rudistid species.

2.4 Late Campanian palaeoclimate

The late Campanian period is characterized by a climatic optimum with mean annual sea surface temperatures in northern Europe rising to 19-20°C (Thibault et al., 2016). Modelling results indicate that mean annual temperatures were approximately 10°C warmer compared to today (DeConto, 1996), but meridional temperature gradients were probably lower, meaning that tropical temperatures were likely less different from the present day situation, with tropical surface temperatures estimated at 30°C (Huber et al., 1995; Brady et al., 1998; Alsenz et al., 2013). At the same time, a higher eustatic sea level caused continental shelves to be submerged and continents were likely surrounded by extensive shallow seas (Brady et al., 1998). The geographical setting of this study’s locality fits this description, with shallow marine conditions above the storm wave base occurring without signs of deltaic sedimentation. Palaeoceanographic studies in the area have demonstrated that the Late Cretaceous Tethys Ocean was prone to rapid changes in palaeoproductivity and redox conditions as the basin became progressively more confined (Dercourt et al., 1986; Almogi-Labin et al., 1993; Steuber and Löser, 2000; Jarvis et al., 2002; Fleurance et al., 2013). Besides these local changes, long-term trends in global ocean water chemistry also significantly influenced the environment in which the bivalves in this study precipitated their shells. Marine magnesium-to-calcium (Mg/Ca) and strontium-to-calcium (Sr/Ca) ratios were much lower (1-2 mol/mol and 4-6 mmol/mol, respectively, compared to 5 mol/mol and 8-9 mmol/mol in modern oceans; Stanley and Hardie, 1998; Coggon et al., 2010; Rausch et al., 2013), which is likely to have influenced the trace element concentrations recorded in shell calcite (e.g. Lear et al., 2015).

3. Methods

3.1 Specimen acquisition and preparation

The shells of two rudistid bivalves (Torreites sanchezi milovanovici Douvillé, 1927; Natuurhistorisch Museum Maastricht collections, NHMM 2014 052, hereafter TS and Vaccinites vesiculosus Woodward, 1855; Oertijdmuseum De Groene Poort collections, MAB D.R.81 5/19078, hereafter VV) and a thick-shelled oyster (Oscillopha figari Fourtau, 1904, NHMM 2014 050a, hereafter OF; Malchus, 1990; Aqrabawi, 1993) were acquired for the present study. The VV shell was subject of an earlier publication by de Winter and Claeys (2016). All shells were collected in life position in the upper Campanian Samhan Formation in the Saiwan area of the Huqf Desert in east-central Oman (Unit 1 of Schumann 1995, 30°39’ N, 57°31’ E, see also Platel et al., 1994; Philip and Platel, 1996; Kennedy et

177178179180181182183184185186187188

189

190

191192193194195196197198199200201202203204205206207208209

210

211

212

213

214215216217218219220221

al., 2000 and section above). Shells were cut longitudinally along their central growth axis and high-grade polished by increasingly fine-grained silicon carbide polishing discs (up to P2400) at the Vrije Universiteit Brussel (VUB, Belgium). High-resolution (6400 dpi) colour scans were produced of the shell surfaces to track all sample locations in an overview figure (Fig. 1).

3.2 µXRF element mapping

High-resolution elemental abundance maps of the polished surface of shells were created using a Tornado M4 micro X-ray fluorescence (µXRF) scanner (Bruker nano GmbH, Berlin, Germany). µXRF mapping was done along a 2D grid with 50 µm spacing, a spot size of 25 µm and an integration time of 1 ms per pixel. The X-ray source was operated under maximum energy settings (600 µA, 50 kV) and no source filters were applied. Details of the setup of the Tornado M4 scanner for the mapping of bivalve calcite can be found in de Winter and Claeys (2016). Higher resolution (25 µm spacing) maps were made of smaller areas of the shells. Mapping by µXRF permitted qualitative element abundance distributions to be visualized (see Fig. 1).

3.3 µXRF line scanning

To determine the best approach for µXRF scanning, trace element profiles were measured on the outer calcite layer along the length of the VV shell using three methods: 1) conventional µXRF line scanning according to de Winter and Claeys (2016), 2) a µXRF point-by-point line scanning method with longer integration time (60s) after de Winter et al. (2017), and 3) two different transects quantified by Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry (LA-ICP-MS). To optimize the results, XRF maps were used to guide sampling of well-preserved parts of the shells in all transient line scanning methods. The method for conventional XRF line scanning on bivalve calcite is detailed in de Winter and Claeys (2016). For point-by-point line scanning, the X-ray beam is allowed to dwell on each point for 60 s at maximized energy settings (50 kV, 600 µA, no source filter) with a sampling resolution of 100 µm for a total of 978 and 1051 points respectively for VV and TS and a total of 431 points for the OF shell. This allows more time for each XRF spectrum to be accumulated and results in more reproducible results (de Winter et al. 2017). All XRF line scans were quantified according to the Fundamental Parameters method of the Esprit software calibrated to BAS CRM393 limestone trace element reference material (Bureau of Analyzed Samples Ltd, UK). XRF point-by-point scans were carried out on the outer calcite layer in the growth direction of the rudistid shells (VV and TS) and through a polished section of the hinge of the OF shell perpendicular to the growth layers (see Fig. 1).

3.4 LA-ICP-MS line scanning

LA-ICP-MS measurements were carried out at the Atomic and Mass Spectrometry research unit of Ghent University (Ghent, Belgium). A detailed account of LA-ICP-MS methodology is found in supplementary data 1. XRF line scans were calibrated by using the bulk concentration of one of the LA-ICP-MS transects as an additional standard and establishing a calibration curve. The results of the application of these methods and a schematic overview of the calibration of XRF results with LA-ICP-MS results are shown in Figure 2. A detailed explanation of the calibration calculations is given in supplementary data 1. After this correction, the average offset of strontium (Sr) concentrations between corrected XRF measurements and results of the second LA-ICP-MS line scan (not used in the correction) was 9.1% for point-by-point analysis and 12% for conventional line scan analysis. All trace

222223224225

226

227

228229230231232233234235

236

237

238239240241242243244245246247248249250251252253254

255

256

257258259260261262263264265

element profiles discussed further in this study are measured by point-by-point XRF line scanning and calibrated using the LA-ICP-MS calibration. Reproducibility standard deviations and instrument errors are reported in Table 1 and were calculated according to de Winter et al. (2017). Data of all calibrated trace element and stable isotope measurements used in this study are provided in supplementary data 2-5.

3.5 Stable isotopic analysis

Pristine calcite in cross sections of all three bivalve shells was sampled for stable isotope analysis using a microscope guided drill (Merchantek/Electro Scientific Industries Inc. (ESI), Portland (OR), USA, coupled to Leica GZ6, Leica Microsystems GmbH, Wetzlar, Germany) equipped with a 300 µm wide tungsten carbide drill bit. Samples for stable isotope analysis were taken at a resolution of 500 µm for TS (243 measurements) and 250 µm for VV and OF (328 and 92 measurements respectively). The spatial sampling resolution was decided upon based on the presence of annual growth bands and recognized annual cycles in trace element ratios measured by µXRF, and was chosen such that each interpreted year contained at least four samples. Samples for stable isotope analysis were taken along the same transects as XRF and LA-ICP-MS measurements (see Fig. 1). In the OF shell, all samples for stable isotope analysis were taken in the dense foliated calcite layers which are concentrated in the hinge of the left valve of the shell (Fig. 1). Sample aliquots with a typical weight of 30-50 µg were reacted with phosphoric acid (H3PO4.H2O) at 70°C in a Nu Carb carbon preparation device coupled via dual inlet to a Nu Perspective isotope ratio mass spectrometer (Nu Instruments, UK) and the resulting CO2 was analysed for stable carbon and oxygen isotope ratios. Analytical uncertainty was determined by repeated measurement (N = 194) of the in-house reference material MAR2 (Marbella marble, δ13C: 3.41 ± 0.10‰VPDB; δ18O: 0.13 ± 0.20‰ VPDB; 1 standard deviation, SD) and found to be 0.05‰ and 0.09‰ for δ13C and δ18O values (1 SD), respectively. This MAR2 reference material was previously calibrated using the international NBS-19 stable isotope standard (Friedman et al., 1982). All stable isotope values are reported in permille relative to the Vienna Pee Dee Belemnite standard (‰VPDB).

3.6 Statistical analyses

Statistical analyses were performed to assess the possible presence of seasonal cyclicity in the records. To test for periodic behaviour in proxy records, periodograms were generated by R 3.1.2 software package (R Development Core Team, 2008) using three methods of spectral analysis in a frequency range of 10-5 and 10-3 µm-1: 1) A Lomb-Scargle Periodogram (LSP, see Lomb 1976; Scargle 1982) was calculated from untreated datasets using the R “lomb” package by Ruf (1999) with an oversampling factor of 100 and a confidence level of 90%. 2) A Multi Taper Method Spectral Analysis (MTM, see Thomson, 1982) was carried out on uniformly resampled and linearly detrended datasets using the R “astrochron” package by Meyers (2014). 3) A simple periodogram (SP, see Bloomfield, 2004) was calculated from uniformly resampled and linearly detrended datasets using R core functions. Confidence levels for detected frequencies were calculated via “red noise” estimation and harmonic F-test calculation (Thomson, 1982) using functions in the R “astrochron” package by Meyers (2012) and significant periodicities were extracted from the records by bandpass filtering (Hays et al., 1976) using the “bandpass” function in the R “astrochron” package (Meyers, 2012) on untreated datasets using a Cosine-tapered window and a zero-padding factor of 5. The script used for these spectral analysis routines including the parameters that were used and the resultant data of filters, powerspectra and confidence levels is given in supplementary data 6-9. In addition to spectral

266267268269270

271

272

273274275276277278279280281282283284285286287288289290291292

293

294

295296297298299300301302303304305306307308309310

analysis, p-values and Pearson’s correlation coefficients were calculated for correlations between all palaeoenvironmental records from the bivalve shells to provide a statistical basis for the visually observed correlations between records.

4. Results

4.1 XRF mapping

An overview of all XRF maps is given in Figure 1. The element distribution maps render a qualitative overview of the presence of elements in a sample and may therefore be used to detect heterogeneity within the sample. Comparison of the manganese (Mn) and iron (Fe) heatmap with the colour scan of VV indicates that brown spots visible on the polished surface of the shell are associated with increased concentrations of Fe, Mn and Si. The Fe-Si-Ca map of the VV shell (Fig. 1A) illustrates how various comparatively Fe- and Mn-rich spots are found within the more Ca-rich outer shell surface. It also shows that the innermost area of the shell as well as the left valve are characterized by higher concentrations of Fe and Mn. XRF maps of the second rudistid shell (TS) indicate that a larger part of this shell shows enrichment in elements other than Ca compared to VV. The Fe-Si-Ca distribution map (Fig. 1B) indicates that the nature of these enrichments is different in the sense that a major section of area is enriched in silicon (Si) rather than Fe. A large area of the inside of the outer shell layer of TS has lower concentrations of Si, Fe and Mn. Enrichments in Fe, Mn and Si only occur in the outer ±10 mm of the shell and a corona of several millimetres around the more porous inner parts of the shell. The central part of the outer calcite layer is characterized by the presence of fine growth layers, which are shown in the close-up of Figure 1B. The oyster shell (OF; fractured during sample preparation) contains the lowest concentrations of Fe, Mn and Si (Fig. 1C). Thin coronas of carbonate enriched in Si, Fe and Mn surround the OF shell, while the interior shows no signs of enrichment in these elements. A close-up of a cross section through the hinge of the oyster (Figure 1C) indicates a thin rim of Fe- and Mn-rich carbonate, while the inside of the shell is characterized by low concentrations of these elements and well-preserved dark growth layers.

4.2 Trace element concentrations

Figure 3 provides an overview of the range in concentrations of major and trace elements measured in all three shells by LA-ICP-MS-calibrated XRF point-by-point line scans. Obtained concentrations of all measured elements fall within the same order of magnitude between different shells, with the exception of Fe. Even though an effort was made to avoid inclusions of Mn and Fe based on their location in Figure 1 (see above), the VV record contains higher values for Fe, Mn and Si. Concentration ranges of other elements are mostly similar between shells except for Mg, which is significantly lower in the OF record. Concentrations of Mg, Sr, S and Zn (±1100, ±900, ±200 and ±100 µg/g, respectively) are similar to concentrations of these elements found in other studies of well-preserved fossil and modern bivalve calcite (Carriker et al., 1980; Al-Aasm and Veizer, 1986a; b; Steuber, 1999; Vander Putten et al., 2000; Holmden and Hudson, 2003).

Figures 4, 5 and 6 present an overview of a selection of trace element and stable isotope records of the three characterized shells. Running averages of 11 points of the trace element records illustrate mm-scale variations in the records and are favoured over individual point measurements, which may be influenced by the irregularities on the sample surface (see de Winter and Claeys, 2017; de Winter et al., 2017). A comparison of Figure 4 with Figure 5 and 6 confirms that concentrations of Mn, Fe,

311312313

314

315

316

317

318319320321322323324325326327328329330331332333334335336337

338

339

340341342343344345346347348349

350351352353354

and Si in VV are overall higher than in the TS and OF shells. Further inspection of the records of these elements shows that the measurements that yield high Mn, Fe, and Si abundances are highly localized and are often associated with lower than average Ca concentrations. As evident from the XRF maps (Figure 1A), the lower 22 mm of the VV shell is characterized by areas with high concentrations of Mn, Fe and Si. In contrast, concentrations of Mn, Fe and Si in TS and OF are an order of magnitude lower (Mn < 150 µg/g, Fe < 40 µg/g, Si < 200 µg/g) and Ca concentrations in these shells do not fall below 38 weight per cent (wt%). Concentrations of other trace elements occur in the same order of magnitude in all shells, with the exception of Mg, which shows consistently lower concentrations in OF than in the rudistid shells (see also Fig. 3). Records of Mg/Ca, Sr/Ca, Zn/Ca and S/Ca show regular fluctuations with an amplitude of 50 to 75 per cent of the average value in all three shells. Regular fluctuations occur on a scale of 10 mm in VV (Fig. 4), 15-20 mm in TS (Fig. 5) and 2-3 mm in the oyster record (OF; Fig. 6). The expression of these fluctuations in trace element records, especially those of Zn and S, is in general not sinusoidal but rather shows a punctuated character with rapid shifts in concentration and longer episodes of relative stability. Trace element records often co-vary (most notably Mg/Ca and Zn/Ca) or show anti-correlations (e.g. Zn/Ca and S/Ca). The Sr/Ca record of OF is characterized by a pronounced ontogenetic increase and shows a large drop at the end of the record. Fluctuations in the Sr record of OF are smaller than in the rudistid shells. Another ontogenetic trend in OF is observed in the S/Ca record, which shows much less variation in the initial ±6 mm of the record followed by an increase in fluctuation for the remainder of the record. A similar, though less pronounced, intensification of the trace element fluctuations is observed in the records of Mg/Ca and Zn/Ca.

4.3 Stable isotope ratios

Figures 4-6 show that stable carbon and oxygen isotope ratios (δ13C and δ18O) fluctuate periodically around a stable background in all bivalve shells. An exception to this is the last part of the δ13C record in OF (Fig. 6) where a decreasing trend is observed. While δ13C values are similar between shells (±0.5‰ to 2.5‰ with a mean value of ±1.5‰ in VV and TS and ±2‰ in OF), large inter-shell differences are observed in δ18O values. Oxygen isotope ratios are highest in OF (roughly between -4‰ and -3‰ with a mean value around -3.5‰) and significantly lower in TS (-6‰ to -5‰, mean value around -5‰) with intermediate values for VV (between -5‰ and -3‰ with a mean value of -4‰). Fluctuations in stable isotope values also vary between shells. Measured δ13C values fluctuate with an amplitude of approximately 1‰ in VV and OF, while fluctuations in TS are about 1.5‰. Oxygen isotope values fluctuate from 1 to 1.5‰ in TS and OF, but are closer to 2‰ in VV. Stable isotope fluctuations seem to follow the same periodicity as the trace element records (10-20 mm in rudistid shells, 2-3 mm in OF) and relationships between these records are described in paragraph 4.5. Fluctuations in the stable isotope records are more regular and have a more constant amplitude compared to trace element records.

4.4 Seasonal cyclicity

Relatively scaled powerspectra of three methods of spectral analysis (MTM, Lomb-Scargle and simple periodogram) are plotted in Figure 7, 8 and 9. Confidence levels for detected frequencies show that frequencies with high spectral power according to all spectral analysis methods (black rectangle in Figs. 7-9) are generally associated with high (>80%) estimated confidence levels. These are the frequency ranges that are filtered from the records using bandpass filtering. Resulting bandpass filters in the frequency domain of seasonality pick up fluctuations in the records that were

355356357358359360361362363364365366367368369370371372373374375

376

377

378379380381382383384385386387388389390391

392

393

394395396397398399

identified in Figures 4-6. Figures 7-9 show that the period best describing the data is ±8 mm long in VV, ±20 mm in TS and ±3 mm in OF. Results of spectral analysis show that all records show periodicity in the seasonal range, but cyclicity is, in general, best represented in Sr/Ca, Mg/Ca and stable isotope records, where bandpass filters most successfully pick up the cyclicity in the records. Results from the three different methods of spectral analysis are not similar in all cases but agree, in general, in the frequency domain where the seasonal cycle is found. Bandpass filters of the proxy records reflect the periodic variation in stable isotope records, but fail to explain the majority of the variation in trace element records, which have a less sinusoidal character. Furthermore, interpretation of growth seasons in Figures 7-9 show that the amount of calcite deposition in the shells is not equal between seasons, resulting in larger parts of the TS shell being deposited under high-δ18O conditions and slightly larger parts of the VV shell deposited under low-δ18O conditions. There is no observable difference in the length of seasons in the OF shell. Interpretation of seasonal cyclicity in the three shells yields an age of 9 years for OF, 6 years for TS and 8-9 years for VV. Age of VV is uncertain due to the poorly preserved lower ±20 mm which precluded isotope analysis, forcing the growth period to be estimated to 2-3 years for this part of the shell (see Fig. 7).

4.5 Phase relationships

As indicated in Figures 4-9, periodic fluctuations of stable isotope and trace element records can be traced between different proxy records. An overview of visually observed and statistically significant correlations between proxy records of the bivalves is given in Table 2. Figures 7-9 show that many proxies only show clear phase relationships in parts of the record and correlations between total shell records are generally low (see Table 2), as proxies shift in and out of phase. One example of this is found in the VV shell, where periodicity in Zn/Ca, δ13C and Sr/Ca is most pronounced in the central part of the record (30 – 70 mm; Fig. 7). Another case is in the OF record where the last two millimetres are influenced by a strong trend in δ13C and Sr/Ca, whereas the part of the record that immediately precedes this drop (i.e., 14 – 22 mm) shows clear periodic variation. A close inspection of the phase relationships between proxies (Figs 7-9) demonstrates that periodicities in δ18O records are generally in phase with Mg/Ca. However, this relationship is only statistically significant in OF (Table 2). On the other hand, the positive relationships between Mg/Ca and S/Ca that are visually observed in all shells do have a statistical basis. Positive correlation of Sr/Ca with S/Ca is both visually observed and statistically significant in the rudistid shells, but not in the oyster shell. Another difference between the rudistids and the oyster is a negative correlation of δ13C with S/Ca and Sr/Ca in the rudistid shells, which in the OF shell is positive. The latter is presumably caused in large by the sharp drop in Sr/Ca and δ13C near the end of the record. Sr/Ca ratios seem to be in antiphase with Mg/Ca ratios in most shell records, but this correlation is only weakly significant in TS and OF. In the TS shell, contrary to the other two shells, both stable isotope records are strongly correlated. The relationship of Zn/Ca with other records is complicated, as not many correlations have statistically significant p-values (p < 0.05). Pearson’s coefficients (r) indicate that Zn/Ca ratios often exhibit weak positive correlations with Mg/Ca and negative correlations with Sr/Ca. However, these relationships are not always supported by visual inspection, mostly because the phase relationship of these records tends to change across the record (see Figs. 4-9). Interestingly, visual inspection suggests a negative relationship between Zn/Ca and S/Ca, especially in the parts of VV and OF that show well-expressed periodicity. However, this relationship is only significant in the bulk statistics of the TS record. This again indicates that phase relationships between these proxies are complex, change throughout the records and cannot easily by summarized by a statistical value.

400401402403404405406407408409410411412413414

415

416

417418419420421422423424425426427428429430431432433434435436437438439440441442443444

445

5. Discussion

5.1 Shell preservation

5.1.1 Spatial distribution

In the case of fossil bivalve taxa, a map of Fe, Si and calcium (Ca) concentrations highlights zones in the sample, where the calcite underwent diagenetic overprinting. As such, Figure 1 provides an initial survey of the state of preservation of the bivalve shells. In addition, a heatmap of the combined intensities of Fe and Mn (elements both associated with diagenesis, see Al-Aasm and Veizer, 1986a; b; Steuber, 1999), highlights those spots where diagenetic carbonates, (re-)crystallized post-mortem, are most common. XRF maps indicate that the best-preserved calcite is found in the outer shell layers of rudistid shells. As discussed by de Winter and Claeys (2016), this layer is punctuated by zones of diagenetic material that should be avoided during sampling. In the VV shell, the left part of the shell cross-section contains a higher amount of these spots affected by diagenesis than the right part and spots are present in greater concentration in the lower part of the shell (as oriented in Fig. 1). The concentrations of Ca, Si, Fe and Mn in the sampled parts of shells (Figs. 3-6) indicate that the preservation state of the calcite sampled in TS and OF is better than that of VV. Trace element records together with the XRF maps in Figure 1 indicate that diagenesis, although present to some extent in all shells, is more localized and therefore easier to avoid during sampling of TS and OF. This explains why the VV record contains more intervals with increased concentrations of Si, Fe and Mn (see also Fig. 3). Another difference between the VV shell and TS and OF is the relative concentrations of Si, Fe and Mn in altered parts of the shell (Compare Fig. 4 with Figs 5 and 6). This difference may be explained by two different types of diagenesis that affected the fossils.

5.1.2 Boring

The first is represented by the highly Fe and Mn enriched spots in VV, which can be explained by boring into the shell by predators (most likely sponges; Sanders and Baron-Szabo, 1997). Bore holes are later filled in by detrital material rich in Fe, Mn and Si (see de Winter and Claeys, 2017). These bore holes explain the drops (by several weight percent) in concentration of Ca observed in Figure 4, where calcite is replaced by (Fe-, Mn- and Si-rich) inclusions of detrital material that filled up the bore holes. This boring into the bivalve shells most likely took place during the life time of the bivalve or early post-mortem and has most strongly affected the VV shell. The relatively thin outer shell layers of VV probably made it more susceptible to boring predation than the significantly more thick-shelled TS and OF (Sanders and Baron-Szabo, 1997), explaining more frequent bore holes in the VV shell and higher local Fe and Mn concentrations (see also Fig. 3).

5.1.3 Silicification

The second type of diagenesis is silicification, which is evident from the corona of Si-rich material that is present on the outside and in the cavity of the TS shell and to a lesser extent on the OF shell (Figure 1). This silicification took place after burial of the fossils, is strongly controlled by the local concentration of Si in the pore fluid (Maliva and Siever, 1988) and has left large (mm- to cm-scale) euhedral quartz crystals in the porous cavity of the TS shell (Fig. 1). The fact that VV was relatively unaffected by silicification seems counter-intuitive, as VV and TS shells are stratigraphically more

446

447

448

449

450

451452453454455456457458459460461462463464465466467468

469

470

471472473474475476477478479480

481

482

483484485486487488

closely associated with each other than with OF, which was found slightly lower in the stratigraphy. However, since the V1-member of the Samhan Formation, from which VV and TS originate, is laterally continuous for 200m (Schumann, 1995), it cannot be excluded that local differences in Si content of the pore fluid in this member were large enough to cause the observed difference in the extent of silicification between the shells. While traces of silicification are also visible in OF, it is hypothesized that the calcite shell of this bivalve was likely more resistant to replacement by quartz than the partly aragonitic TS shell (see Schubert et al., 1997), preventing the extensive (cm-deep) invasion of Si-rich pore fluid observed in the TS shell from affecting OF. The last 2 mm of the Sr/Ca record in the OF shell (Fig. 6) were probably affected by diagenesis as drops in Sr concentration are much more severe than those observed in pristine modern bivalves. The sharp decrease in Sr/Ca is matched by a decrease in δ13C, potentially indicating incorporation of 12C-enriched carbonate during diagenesis. The statistically significant positive correlation between Sr/Ca and δ13C (Table 2) is strongly controlled by this diagenetic drop at the end of the record.

5.1.4 Data filtering

In Figure 4 (VV), parts of the trace element record that show peaks in the concentrations of these elements were removed from records of Mg/Ca, Zn/Ca, Sr/Ca and S/Ca to avoid interpretation of parts of the record that were affected by diagenesis. This was not necessary in the records of TS and OF as the Mn and Fe concentration in these shells mostly remained well below 200 µg/g, which has been proposed as a threshold for pristine shell material (Brand and Veizer, 1980; Al-Aasm and Veizer, 1986; Steuber, 1999). However, as is evidenced by the presence of silicification in these shells, thresholds of Mn, Fe and Sr concentrations as indicators for pristine carbonate preservation are highly simplistic and can only be used to trace calcite recrystallization from reducing pore fluids (Aller, 1980; Brand and Veizer, 1980). Furthermore, fossilization of the bivalves has probably limited the preservation of organic matter in the shells, as indicated by the lower S concentrations found in the present study (100-200 µg/g) compared to higher (~500 µg/g) S concentrations found in modern bivalve studies (Lorens and Bender, 1980; Dalbeck et al., 2006). Microscope-guided drilling for stable isotope samples guided by XRF mapping allowed bore holes and silicified shell parts to be avoided, and the resulting filtered trace element and stable isotope records (Figure 4-6) are indicative for well-preserved bivalve calcite.

5.2 Multi-proxy records

5.2.1 Comparison with modern shells

Element-to-Ca ratios of Mg, Zn and Sr of fossil shells in this study are in agreement with ratios found in modern bivalves (e.g. Carriker et al., 1980; Al-Aasm and Veizer, 1986a; b; Steuber, 1999; Vander Putten et al., 2000; Holmden and Hudson, 2003). Mg/Ca and Sr/Ca in OF are similar to those found in modern oyster species (Carriker et al., 1980; Ullmann et al., 2013; Mouchi et al., 2013), though with lower Mg/Ca than found in the estuarine oyster Crassostrea virginica (Surge and Lohmann, 2008). Concentrations of Zn, Mn and Fe are also in close agreement with those found in modern oyster shells (Carriker et al., 1980; Ullmann et al., 2013), and S concentrations are similar to those found in foliated calcite structures of Crassostrea gigas by Dauphin et al. (2013), though slightly lower than those found in other marine bivalves such as Mytilus edulis (Lorens and Bender, 1980; Dalbeck et al., 2006). Trace element concentrations are generally higher in VV and TS, and are in closer agreement with those in calcite shells of modern marine bivalves (Vander Putten et al.,

489490491492493494495496497498499500501

502

503

504505506507508509510511512513514515516517518

519

520

521

522

523524525526527528529530531532533

2000; Gillikin et al., 2005; Freitas et al., 2009), with the exception of S, which is lower than found in calcite of marine bivalves (Lorens and Bender, 1980; Dalbeck et al., 2006) and low-Mg calcite shells of brachiopods (England et al., 2007), but in closer agreement with concentrations in the aragonitic tridacnid bivalve Hippopus hippopus (Yoshimura et al., 2013). Higher concentrations of trace elements in rudistid bivalves compared to the hinge of the oyster are in agreement with similar offsets between prismatic and foliated calcite structures in modern oysters (Carriker et al., 1980). The fact that Mg/Ca ratios are low compared to modern shells is likely a result of the lower Mg/Ca ratios in Late Cretaceous ocean water, in which the bivalves grew (Stanley and Hardie, 1998). The fact that Mg/Ca ratios in these bivalves approach modern values if the (3.3 times) lower Mg/Ca ratios of seawater in which they grew where taken into account indicates that these bivalves lived in full marine conditions and that they incorporated Mg into their shell proportional to ambient Mg/Ca ratios similar to modern bivalves (Surge and Lohmann, 2008).

5.2.2 Phase relationships

The complexity of phase relationships in this multi-proxy study is illustrated by the lack of significant, strong (p<0.05; r>0.60) correlations between proxy records (Table 2). This lack of linear relationships is caused by different parameters acting on the proxies at different times and because the expression of seasonality in the shells is not always equally strong (Figs. 7-9). A close examination of Figures 7-9 shows that this is indeed the case, as filtered sinusoidal trends only explain a part of the variation. The amplitude of seasonality observed in proxy records is also highly variable along the length of the shell records, and on many occasions this seasonality (especially in trace element records) is overprinted by short-lived variations, obscuring phase relationships between proxies. Finally, the amount of shell material deposited in a year is variable, generally becoming shorter with shell age, an observation that is in agreement with observations in modern bivalves (e.g. Richardson et al., 2004; Schöne et al., 2005; 2011).

5.3 Bivalve physiology

5.3.1 Ontogenetic trends

Despite the fact that bivalves in this study were relatively short-lived, some ontogenetic trends are observed in the record. The gradually decreasing trend in Sr/Ca in the OF shell is similar to ontogenetic trends in Sr/Ca found in modern oysters (Ullmann et al., 2013), suggesting an effect of growth rate on the incorporation of Sr into OF. Discarding the final drop in δ13C, no significant ontogenetic trend in δ13C values is observed in OF. On a seasonal scale, lows in Sr/Ca coincide with high δ13C (Figs. 7 and 9), indicating that differences in metabolic rate may have controlled Sr/Ca ratios, as observed in culture studies on modern bivalves (Lorrain et al., 2005; Freitas et al., 2006). Since Sr/Ca seasonality in OF is weak, its correlation is controlled by the stronger ontogenetic trend. This illustrates that it remains important to visually inspect and compare seasonal records to complement bulk statistical analysis of the proxy records. Interpretation of seasonal cycles in these figures further indicate that shell growth varies seasonally. The length of the interpreted warm/low-salinity season (lower δ18O) and cold/higher-salinity season (higher δ18O) are not equal, showing differential growth rates between the seasons or even complete growth halts during extreme conditions (Wanamaker et al., 2011). The difference in the amount of shell material deposited during different seasons indicates either that the records of TS was more strongly affected by extreme conditions in the warm/low-salinity season than the other shells, or that its growth conditions were

534535536537538539540541542543544545

546

547

548549550551552553554555556557558

559

560

561

562

563564565566567568569570571572573574575576577578

more extreme in terms of high temperatures and/or low salinities than those of VV and OF. The former hypothesis might indicate that TS was more resistant to extreme conditions, but the latter hypothesis cannot be excluded because the length of these short-lived bivalve records means that the probability of these organisms living at the same moment in time is small. Thickness of annual deposits in the shells decreases with age, suggesting ontogenetically decreasing growth rates, in agreement with other studies of rudistids (Steuber, 1995; 1996) and oysters (Ullmann et al., 2010). Periodic fluctuations in the stable isotope records of the OF shell are paced to the growth increments observed in Figure 1, showing that these growth increments are likely annual in Oscillopha figari, contrary to those observed in modern Crassostrea gigas (Ullmann et al., 2013).

5.3.2 Physiological control on trace element incorporation

The correlation between Mg/Ca and S/Ca found by Lorens and Bender (1980) is also observed in the bivalve records in the present study (Table 2), and is especially clearly visible in the OF record. S is present in the organic matrices of bivalve calcite in the form of ester sulphates and comprises approximately 10 per cent of the organic matter in the shell, meaning that S/Ca ratios of up to 10 mmol/mol can be explained by the occurrence of S in organic matter in the shell (Crenshaw, 1972; Lorens and Bender, 1980). However, more recent studies have concluded that a large fraction of S in bivalve shells is present in inorganic form (Yoshimura et al., 2014). Annual fluctuations in S contents in extant aragonitic and calcitic bivalves have been linked to growth rate and concentrations in the calcifying fluid, with lower S concentrations during more rapid growth (Yoshimura et al., 2013, 2014). These fluctuations become progressively more pronounced with age in the OF shell, showing that perhaps these short episodes of rapid growth become more common with age in the oyster shell. This increase in fluctuation of the S/Ca record is mirrored in the Mg/Ca record. The covariations of S/Ca with Mg/Ca can be explained by local increases in the amount of organic matrix in the carbonate structure of the shell. It has been shown that bivalves secrete additional sulphate-bearing organic matter under the influence of crystal growth-inhibiting concentrations of Mg in order to seed calcite growth by locally increasing Ca concentrations, causing local enrichments of Mg and S in the shell (Lorens and Bender, 1980). This physiological control on Mg-incorporation would mean that Mg concentrations, whenever correlated to S concentrations, are associated with the organic matrix of the shell rather than the calcite, implying that caution must be exercised when interpreting Mg/Ca ratios in terms of palaeotemperature (Rosenberg and Hughes, 1991; England et al., 2007). Alternatively, if part of the Mg in the shells is bound to the organic matrix rather than the calcite lattice, increases in S/Ca and Mg/Ca ratios may indicate calcite deposited during times of stress, when more organic matter is secreted (Lorens and Bender, 1980; Rosenberg et al., 1989; 2001). Such periods of increased stress may also be controlled by growth rate or environmental stress, as episodes of rapid growth or outside stress may rupture the periostracal rim resulting in increased organic matrix production and allowing sulphate-rich sea water to enter the extra-pallial fluid (Carriker et al., 1980; Skelton et al., 1990; Rosenberg et al., 2001). This hypothesis is consistent with the occurrence of storm events in the turbulent environmental setting of the bivalves, which could have caused disruption of the bivalve’s life environment on a regular basis.

5.4 Sea water chemistry

5.4.1 Zn/Ca ratios and palaeoproductivity

579580581582583584585586587

588

589

590591592593594595596597598599600601602603604605606607608609610611612613614615616617618

619

620

621

622

Parts of all shell records show periodic spikes in Zn/Ca, indicating that there were seasonal changes in productivity in the environment. The punctuated character of the Zn/Ca record, with sharp increases rather than more sinusoidal fluctuation, can be explained by episodes of release of Zn into the environment after productivity blooms. The fact that Zn concentrations are lower during the remainder of the record is ascribed to its role as a micronutrient. It is continuously taken up by primary producers and bound to organic ligands without being replenished fast enough unless during times when respiration increases dramatically, i.e. after productivity blooms (Bruland, 1989; Calvert and Pedersen, 1993; Vance et al., 2017). If Zn concentrations in the shell increase after episodes of increased productivity, as shown in modern environment studies (Calvert and Pedersen, 1993; Jackson et al., 1993; Guo et al., 2002), Zn/Ca fluctuations record blooms of palaeoproductivity in the shallow marine environment of the shells. This also explains why often multiple peaks in Zn/Ca ratios per year are observed, especially in the OF and TS shell records, which display double peaks in Zn/Ca paced to seasonal variation in δ18O records. These increases in palaeoproductivity likely strongly affected the local surface ocean chemistry, especially in the progressively more confined Late Cretaceous Tethys Ocean (Dercourt et al., 1986; Almogi-Labin et al., 1993; Steuber and Löser, 2000; Jarvis et al., 2002; Fleurance et al., 2013). Indeed, Zn/Ca peaks in VV and OF seem to be paced to intervals with higher δ13C values, which may reflect increased δ13C values in the water indicative of increased productivity (Grossman, 1984; Jarvis et al., 2002; Takesue and van Geen, 2004) or decreased metabolic activity (Klein et al., 1996b; Owen et al., 2002; Gillikin et al., 2006). The latter may be explained by changes in the redox state of the shallow ocean as a consequence of the reduction of organic matter after a productivity bloom. It has been proposed that this relationship between metabolic rate and δ13C values is especially strong during periods of relatively low respiration and low ambient CO2/O2 ratios, as is the case in well-oxygenated waters (McConnaughey et al., 1997; Gillikin et al., 2006). The observed anti-phase relationship between Zn/Ca and Sr/Ca in VV and OF supports this hypothesis of reduced metabolic rate in the shells during periods corresponding to peaks in Zn/Ca ratios. Such seasonal changes in redox conditions in the surface ocean are common in present-day low-latitude settings (Morrison et al., 1998). Under more oxygen-poor conditions, the mobility and bioavailability of redox sensitive elements in the water column increases significantly (Zwolsman et al., 1997; Gerringa et al., 2001). Seasonal cycles in Zn/Ca and S/Ca are especially evident in the OF shell, showing that changes in Zn availability are readily recorded in oyster shells, as shown by Carriker et al. (1980) and Guo et al. (2002). Large S/Ca shifts show that seasonal changes in the organic matter fraction incorporated in the shell of OF are more pronounced than in the rudistid shells.

5.4.2 S/Ca ratios and sea water chemistry

In parts of VV and OF where Zn/Ca shows well-expressed periodicity, the record is in antiphase with S/Ca ratios, which might be interpreted as evidence for an environmental control mechanism on S concentrations in the shells (Figs. 4, 6, 7, 9). One hypothesis might be that more reducing conditions allow S to be precipitated in sulphides while more oxidizing conditions can remobilize S from the sediment, increasing S concentrations (Ku et al., 1999; Kah et al., 2004). However, since the sulphate concentration in ocean water is several orders of magnitude higher than the concentrations of S found in the present study, it is unlikely that changes in redox conditions in this shallow, open-marine system could have brought about conditions where this source would be limited (Chester, 2009). Evidence of a highly turbulent environment in which these bivalves lived renders severe redox stratification unlikely (Schumann, 1995). Furthermore, a redox shift that allows the precipitation of sulphides would also reduce the bioavailability of Zn by means of precipitation of Zn sulphides, and this is not consistent with the observed antiphase relationship between Zn/Ca and S/Ca (Calvert and

623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655

656

657

658659660661662663664665666667668669

Pedersen, 1993; Guo et al., 1997). Furthermore, the precipitation of ZnS requires more reducing conditions than may be expected in the shallow marine, turbulent environment of the bivalves in this study (Guo et al., 1997; Schumann, 1995). Instead, it is more likely that a change in growth and metabolic rate caused by increases in productivity, when Zn concentrations are highest, resulted in an increase in S/Ca ratios.

5.5 Palaeotemperature seasonality

5.5.1 Stable oxygen isotopes

In several studies, relationships have been found between chemical proxies (Mg/Ca, Sr/Ca and δ18O) in biogenic carbonates and the calcification temperature of these carbonates (e.g. Epstein et al., 1953; Craig, 1965; O’Neil et al., 1969; Tarutani et al., 1969; Hays and Grossman, 1991; Takesue and van Geen, 2004; Wanamaker et al., 2008). Since other factors (e.g., salinity, growth rate and water chemistry) also appear to control these proxies in molluscan shells, different proxies need to be compared to isolate the effect of temperature. Temperature transfer functions were either obtained through controlled inorganic calcium carbonate precipitation (e.g., Epstein et al, 1953; O’Neil et al., 1969; Kim and O’Neil, 1997) or from culture experiments and comparison of proxy data with historical temperature records in bivalves as well as in other calcifying organisms (e.g., de Villiers et al., 1995; Freitas et al., 2005; Takesue and van Geen, 2004; Surge and Lohmann, 2008; Wanamaker et al., 2008). For palaeotemperature reconstructions based on δ18O values, a value of -1‰ is assumed for Late Cretaceous sea water (Savin and Yeh, 1981; Steuber, 1995). Temperature estimates of minima and maxima of the seasonal cycle based on stable oxygen isotope ratios from shell OF and VV yield temperatures of 25°C and 32°C respectively within a 1-2°C range depending on the calibration used (e.g. Epstein et al., 1953; Craig, 1965; O’Neil et al., 1969; Tarutani et al., 1969; Hays and Grossman, 1991). Minimum seasonal temperatures based on δ18O values in the OF and VV shells are in agreement, while maximum seasonal temperatures recorded in VV (~34°C) exceed those determined for OF (~30°C). Temperatures for the TS shell using the same parameters turn out higher, ranging between 34°C and 43°C. Temperatures determined based on OF and VV are in agreement with temperatures found for Vaccinites cornuvaccinum, Vaccinites ultimus, Vaccinites oppeli, all of which are in the same family as VV, as well as to temperatures derived from other rudistid taxa from the Upper Cretaceous (Steuber, 1995, 1996, 1999). Temperatures found in the seasonal cycle of TS overlap with those documented for other rudistids, but are still exceptionally high (Steuber et al., 2005), suggesting that δ18O values in this rudistid may have been controlled by other factors besides temperature. Seasonal changes in salinity could account for a change in stable oxygen isotope ratios (Klein et al., 1996a). The distal, open marine environmental setting of these shells, however, precludes a strong riverine component to explain large changes in salinity. Because of the relatively high δ18O values of rainfall in tropical climates, changes in salinity in the order of 10 g/kg would be required to account for a 1‰ difference in δ18O in the TS shell (Ravelo and Hillaire-Marcel, 2007). However, it cannot be excluded that seasonal salinity changes did influence the stable isotope records of the shells differently, primarily because it is not certain exactly how much time existed between the lifetimes of the bivalves. Furthermore, as speculated above, it is likely that extreme parts of the seasonal cycle were not recorded in one species while they were recorded in another. The preferential exclusion of part of the seasonal cycle from the TS shell (in this case the 18O-depleted warm/low-salinity part of the cycle) would explain the difference in seasonality between the δ18O record of this shell and that of the closely associated VV, but does not explain the 1‰ difference in mean value between the records. Differences in seasonality between the rudistid shells and the OF

670671672673674

675

676

677

678

679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715

are probably explained by actual changes in climate between their lifetimes, as a significant amount of time may separate their formation. The fact that δ18O records in modern oysters seem to faithfully record water temperatures (if δ18O of seawater is well-constrained; see Surge and Lohmann, 2008; Ullman et al., 2010) puts slightly more confidence in the temperatures reconstructed from the VV shell. Although isotope ratios from Torreites shells from the same region have been interpreted to fractionate in disequilibrium with respect to ambient sea water (Steuber, 1999), the data in this study cannot confirm or disprove this statement. However, since disequilibrium fractionation of isotopes in marine calcitic bivalve shells is rare, it cannot be excluded that TS precipitated its calcite under significantly lower-salinity (-10 g/kg) or higher temperature (+8°C) with respect to the other two bivalves. Furthermore, it must be stated that the use of one specimen for each species in this study precludes the discussion of specimen-specific effects and therefore adds a certain uncertainty to the reconstructions. This uncertainty is somewhat alleviated by the agreement of the temperature reconstructions in this study with other bivalve-based (Steuber, 1995, 1996, 1999) and other reconstructions (e.g. Huber et al., 1995; Brady et al., 1998; Alsenz et al., 2013).

5.5.2 Mg/Ca ratios

Even though Mg/Ca ratios have been proposed as a proxy for calcification temperature in a range of calcifying organisms (e.g. Klein et al., 1996a; Vander Putten et al., 2000; Elderfield and Ganssen, 2000; Takesue and van Geen, 2004; Freitas et al., 2005; Wanamaker et al., 2008; Surge and Lohmann, 2008; Cléroux et al., 2008; Mouchi et al., 2013), empirically determined dependencies of Mg concentrations on sea water temperature as well as mean concentrations of Mg in these biogenic calcites differ strongly, causing widely different transfer functions between these studies, as illustrated in Figure 10. The fact that Mg/Ca ratios in the OF shell are systematically lower than those from the rudistid shells shows that applying the same transfer function for these shells would lead to different palaeotemperature reconstructions. Using oyster-based calibrations to calculate temperatures from Mg/Ca ratios in the OF shells yields unrealistically cold mean annual temperature estimates of 2.9°C (Surge and Lohmann, 2008) and 8.8°C (Mouchi et al., 2013) with a seasonal temperature variability of ±2.5°C and ±6°C, respectively. A probable reason for these underestimations of temperature with respect to the temperature reconstructions based on δ18O is that the Mg/Ca ratio of ocean water in the Late Cretaceous (1-2 mol/mol; Stanley and Hardie, 1998; Coggon et al., 2010) was lower than it is today (5 mol/mol; Stanley and Hardie, 1998). While the incorporation of trace elements into bivalve calcite is not yet fully understood, it seems likely that concentrations in ambient sea water exert a dominant control on the concentrations in the shell. Taking this estimated decrease in sea water Mg/Ca ratios by a factor of 3.3 with respect to present-day values into account and recalculating temperatures based on Mg/Ca ratios in the OF shell yields mean annual palaeotemperature estimates of 9°C with a temperature seasonality of 10°C using the Surge and Lohmann (2008) calibration and 25°C with a temperature seasonality of 14°C for the Mouchi et al. (2013) calibration. These mean annual temperature estimates, especially those based on Surge and Lohman (2008), are still much lower than those based on δ18O values, and temperature seasonalities of 10°C or more seem unrealistic for tropical ocean water. When this correction for different sea water Mg/Ca ratios is applied on the Mg/Ca records of VV and TS, reconstructed seasonalities based on bivalve calibrations (Klein et al., 1996a; Vander Putten et al., 2000; Takesue and van Geen, 2004; Wanamaker et al., 2008) are unrealistically high as well (>10°C), while mean annual temperatures vary widely depending on the calibration that is used (see Fig. 10). It is possible that the temperature dependence of the incorporation of Mg into bivalve calcite was different in low-Mg/Ca oceans or in the much hotter climate of the Late Cretaceous, or that a part of the Mg in bivalve shells is associated with the organic matrix in the shell and controlled by physiological

716717718719720721722723724725726727728729

730

731

732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762

processes rather than ambient temperature (Freitas et al., 2009; see discussion in section 5.3). In any case, the present exercise shows that palaeotemperature reconstructions based on the Mg/Ca proxy need to be treated with care until the mechanisms controlling the incorporation of Mg into the shell is better understood.

5.5.3 Sr/Ca ratios

Palaeotemperature estimates based on Sr/Ca ratios in bivalve calcite also need to be treated with care, since previous studies have shown that changes in Sr/Ca ratios correlate with changes in growth and metabolic rate (Klein et al., 1996b; Lorrain et al., 2005; Freitas et al., 2006) and salinity (Klein et al., 1996b; Wanamaker et al., 2008). Nevertheless, some authors have proposed relationships between Sr/Ca ratios and sea water temperature (Takesue and van Geen, 2004; Wanamaker et al., 2008). Calibrations from both these studies show a positive correlation of Sr/Ca with temperature, while a negative relationship would be expected based on experimental data and thermodynamic considerations (Rimstidt et al., 1998). Applying these calibrations on this study’s Sr/Ca records leads to severely inconsistent estimates and consistent underestimation of sea water temperatures with respect to temperatures based on δ18O values (-13°C to 30°C mean annual temperatures according to the calibration of Wanamaker et al., 2008 assuming 32 g/kg salinity and 5°C to 8°C mean annual temperatures according to the Takesue and van Geen (2004) intertidal aragonite shell calibration). Furthermore, the Wanamaker et al. (2008) calibrations yield unrealistically large seasonal temperature ranges. Correction for lower Sr/Ca ratios in Late Cretaceous ocean water or use of a low salinity model from Wanamaker et al. (2008) yields impossibly high temperature estimates (>100°C). Using the calibration from Takesue and van Geen (2004) with sea water-corrected Sr/Ca values yields temperatures closer to temperatures based on δ18O, though much lower (mean annual temperatures range from 12°C for OF to 14°C for TS with temperature seasonalities of 1.5°C for OF and 10°C for the rudistids). Nevertheless, reconstructions based on Sr/Ca ratios vary significantly between specimens and are much lower than would be expected of this locality. These results illustrate the complexity of Sr/Ca ratios in bivalves and shows that they are not necessarily temperature dependent (e.g. Lorens and Bender, 1980; Klein et al., 1996b, Steuber, 1999). It is possible that the Sr/Ca is in part controlled by temperature, while much of the larger variation in Sr/Ca is caused by other processes such as the above-mentioned growth rate and salinity effects.

6. Conclusions

The study of three co-occurring extinct bivalve taxa sheds light on differences in the response of these bivalves to seasonally changing environmental conditions. µXRF line scanning on polished shell surfaces provides a method for simultaneous acquisition of multiple trace element records, including S, along the growth axis of the shells with a strong control on possible effects by diagenesis. A combination of these trace element records with stable carbon and oxygen isotope measurements illustrates the complex interplay of environmental and physiological processes that control the incorporation of trace elements and stable isotopes into bivalve shells in this low-latitude setting in the late Campanian south eastern Tethys Ocean on the seasonal scale. Our results indicate that δ18O-based temperature reconstructions yield a seasonal temperature cycle of ±8°C around a mean annual sea surface temperature of ±28°C, which is comparable to the present-day situation in the oceans near Oman and in agreement with other reconstructions. Temperature reconstructions based on Mg/Ca or Sr/Ca ratios remain problematic in fossil bivalves, because of the vast differences in the

763764765766

767

768

769770771772773774775776777778779780781782783784785786787788789790791792

793

794

795

796797798799800801802803804805806807

transfer functions found for modern bivalves and the uncertainty in long-term fluctuations of trace element concentrations in the oceans. Correction of Mg/Ca ratios for reconstructed differences in sea water Mg/Ca ratios between the present and the Late Campanian yields temperatures that are close to those reconstructed via stable oxygen isotopes, but such reconstructions exhibit much larger seasonal differences, indicating that Mg/Ca ratios in these bivalves are controlled by more than just sea water temperature. The existence of physiological effects on Mg incorporation into these bivalve shells is evidenced by covarying Mg/Ca and S/Ca ratios. Palaeotemperature reconstructions using Sr/Ca ratios are different from those based on stable oxygen isotope ratios, leading to the conclusion that temperature dependence of this proxy is weak at best and that other factors such as growth rate, metabolism and salinity probably have strong effects on the Sr/Ca ratios. Zn records show that these Late Campanian shallow seas were characterized by productivity blooms, which control Zn concentrations in the water column analogous to present-day shallow-marine systems. After these blooms, enhanced respiration remobilized Zn in the surface ocean and decreased growth rates of the bivalves, leading to the decreased incorporation of S in their shells. The oyster species Oscillopha figari appears to have been especially sensitive to these changing conditions and a clear periodic record of Zn and S is recorded in its shell. Since some of the studied proxies are likely to have been controlled by the animal’s growth rate and metabolism, further research into the quantification of growth rates as recorded in extinct bivalve shells could reflect a valuable tool in the interpretation of these multi-proxy records.

Acknowledgments

The authors thank Prof. Dr. Alan Wanamaker and one anonymous reviewer for their thorough and insightful review comments that helped to improve this manuscript. We also thank Prof. Dr. Thomas Algeo for guiding the review process. This work is made possible by the loan of fossil bivalve shells from the Oertijdmuseum De Groene Poort in Boxtel, the Netherlands and the Natuurhistorisch Museum Maastricht. NJW is financed by a personal PhD fellowship from IWT Flanders (IWT700). SG is a Postdoctoral Fellow of the Research Foundation Flanders (FWO). PC thanks the Hercules foundation Flanders for an upgrade of the stable isotope laboratory (grant HERC9) and the acquisition of XRF instrumentation (grant HERC1309). The authors acknowledge financial and logistic support from the Flemish Research Foundation (FWO, research project G017217N) and Teledyne CETAC Technologies as well as support from VUB Strategic Research. SVM is a Ph.D. Fellow of the FWO. We thank Lieven Bekaert and Quintin Dierckens for their help with the sample preparation.

References

Al-Aasm, I.S., Veizer, J., 1986a. Diagenetic stabilization of aragonite and low-Mg calcite, II. Stable isotopes in rudists. Journal of Sedimentary Research 56.

Al-Aasm, I.S., Veizer, J., 1986b. Diagenetic Stabilization of Aragonite and Low-mg Calcite, I. Trace Elements in Rudists. Journal of Sedimentary Research 56.

Aller, R.C., 1980. Diagenetic processes near the sediment-water interface of Long Island Sound. II. Fe and Mn. Advances in Geophysics 22, 351–415.

Almogi-Labin, A., Bein, A., Sass, E., 1993. Late Cretaceous upwelling system along the southern Tethys margin (Israel): interrelationship between productivity, bottom water environments, and organic matter preservation. Paleoceanography 8, 671–690.

Alsenz, H., Regnery, J., Ashckenazi-Polivoda, S., Meilijson, A., Ron-Yankovich, L., Abramovich, S., Illner, P., Almogi-Labin, A., Feinstein, S., Berner, Z., others, 2013. Sea surface temperature record of a Late Cretaceous tropical southern Tethys upwelling system. Palaeogeography, Palaeoclimatology, Palaeoecology 392, 350–358.

Andrews, J.E., Brasier, A.T., 2005. Seasonal records of climatic change in annually laminated tufas: short review and future prospects. Journal of Quaternary Science 20, 411–421.

Andrews, J.E., Tandon, S.K., Dennis, P.F., 1995. Concentration of carbon dioxide in the Late Cretaceous atmosphere. Journal of the Geological Society 152, 1–3.

Aqrabawi, M., 1993. Oysters (Bivalvia–Pteriomorphia) of the Upper Cretaceous rocks of Jordan. Palaeontology, stratigraphy and comparison with the Upper Cretaceous oysters of northwest Europe. Mitteilungen aus dem Geologisch-Paläontologischen Institut der Universität Hamburg 75, 1–135.

808809810811812813814815816817818819820821822823824825826

827

828

829830831832833834835836837838839840

841

842843844845846847848849850851852853854855856857858859860

Armendáriz, M., Rosales, I., Quesada, C., 2008. Oxygen isotope and Mg/Ca composition of Late Viséan (Mississippian) brachiopod shells from SW Iberia: palaeoclimatic and palaeogeographic implications in northern Gondwana. Palaeogeography, Palaeoclimatology, Palaeoecology 268, 65–79.

Beck, M., Reckhardt, A., Amelsberg, J., Bartholomä, A., Brumsack, H.-J., Cypionka, H., Dittmar, T., Engelen, B., Greskowiak, J., Hillebrand, H., others, 2017. The drivers of biogeochemistry in beach ecosystems: A cross-shore transect from the dunes to the low water line. Marine Chemistry.

Berner, R., 1990. Atmospheric carbon dioxide levels over Phanerozoic time. Science 249, 1382–1386.Black, B.A., Copenheaver, C.A., Frank, D.C., Stuckey, M.J., Kormanyos, R.E., 2009. Multi-proxy reconstructions of northeastern Pacific sea

surface temperature data from trees and Pacific geoduck. Palaeogeography, Palaeoclimatology, Palaeoecology 278, 40–47.Bloomfield, P., 2004. Fourier analysis of time series: an introduction. John Wiley & Sons.Bouchet, P., Rocroi, J.-P., Bieler, R., Carter, J.G., Coan, E.V., 2010. Nomenclator of bivalve families with a classification of bivalve

families. Malacologia 52, 1–184.Brady, E.C., DeConto, R.M., Thompson, S.L., 1998. Deep water formation and poleward ocean heat transport in the warm climate extreme

of the Cretaceous (80 Ma). Geophysical Research Letters 25, 4205–4208.Brand, U., Veizer, J., 1980. Chemical diagenesis of a multicomponent carbonate system–1: Trace elements. Journal of Sedimentary Research

50.Brigaud, B., Pucéat, E., Pellenard, P., Vincent, B., Joachimski, M.M., 2008. Climatic fluctuations and seasonality during the Late Jurassic

(Oxfordian–Early Kimmeridgian) inferred from δ 18 O of Paris Basin oyster shells. Earth and Planetary Science Letters 273, 58–67.

Bruland, K.W., 1989. Complexation of zinc by natural organic ligands in the central North Pacific. Limnology and Oceanography 34, 269–285.

Calmano, W., Hong, J., Förstner, U., 1993. Binding and mobilization of heavy metals in contaminated sediments affected by pH and redox potential. Water science and technology 28, 223–235.

Calvert, S.E., Pedersen, T.F., 1993. Geochemistry of Recent oxic and anoxic marine sediments: Implications for the geological record. Marine Geology, Marine Sediments, Burial, Pore Water Chemistry, Microbiology and Diagenesis 113, 67–88. doi:10.1016/0025-3227(93)90150-T

Carré, M., Bentaleb, I., Bruguier, O., Ordinola, E., Barrett, N.T., Fontugne, M., 2006. Calcification rate influence on trace element concentrations in aragonitic bivalve shells: evidences and mechanisms. Geochimica et Cosmochimica Acta 70, 4906–4920.

Carriker, M.R., Palmer, R.E., Sick, L.V., Johnson, C.C., 1980. Interaction of mineral elements in sea water and shell of oysters (Crassostrea virginica (Gmelin)) cultured in controlled and natural systems. Journal of experimental marine biology and ecology 46, 279–296.

Carroll, M.L., Johnson, B.J., Henkes, G.A., McMahon, K.W., Voronkov, A., Ambrose, W.G., Denisenko, S.G., 2009. Bivalves as indicators of environmental variation and potential anthropogenic impacts in the southern Barents Sea. Marine pollution bulletin 59, 193–206.

Chafetz, H.S., Utech, N.M., Fitzmaurice, S.P., 1991. Differences in the 18O and 13C signatures of seasonal laminae comprising travertine stromatolites. Journal of Sedimentary Research 61.

Chester, R., 2009. Marine geochemistry. John Wiley & Sons.Clarke, L.J., Jenkyns, H.C., 1999. New oxygen isotope evidence for long-term Cretaceous climatic change in the Southern Hemisphere.

Geology 27, 699–702.Cléroux, C., Cortijo, E., Anand, P., Labeyrie, L., Bassinot, F., Caillon, N., Duplessy, J.-C., 2008. Mg/Ca and Sr/Ca ratios in planktonic

foraminifera: Proxies for upper water column temperature reconstruction. Paleoceanography 23.Coggon, R.M., Teagle, D.A., Smith-Duque, C.E., Alt, J.C., Cooper, M.J., 2010. Reconstructing past seawater Mg/Ca and Sr/Ca from mid-

ocean ridge flank calcium carbonate veins. Science 327, 1114–1117.Craig, H., 1965. The measurement of oxygen isotope paleotemperatures. Stable isotopes in oceanographic studies and paleotemperatures 23.Crenshaw, M.A., 1972. The soluble matrix from Mercenaria mercenaria shell. Biomineralization 6, 6–11.Dalbeck, P., England, J., Cusack, M., Lee, M.R., Fallick, A.E., 2006. Crystallography and chemistry of the calcium carbonate polymorph

switch in M. edulis shells. European Journal of Mineralogy 18, 601–609. doi:10.1127/0935-1221/2006/0018-0601Dauphin, Y., Ball, A.D., Castillo-Michel, H., Chevallard, C., Cuif, J.-P., Farre, B., Pouvreau, S., Salomé, M., 2013. In situ distribution and

characterization of the organic content of the oyster shell Crassostrea gigas (Mollusca, Bivalvia). Micron 44, 373–383.Dauphin, Y., Cuif, J., Doucet, J., Salomé, M., Susini, J., Williams, C., 2003. In situ mapping of growth lines in the calcitic prismatic layers of

mollusc shells using X-ray absorption near-edge structure (XANES) spectroscopy at the sulphur K-edge. Marine Biology 142, 299–304.

de Villiers, S., Nelson, B.K., Chivas, A.R., 1995. Biological controls on coral Sr/Ca and delta18O reconstructions of sea surface temperatures. Science 269, 1247.

de Villiers, S., Shen, G.T., Nelson, B.K., 1994. The SrCa-temperature relationship in coralline aragonite: Influence of variability in (SrCa) seawater and skeletal growth parameters. Geochimica et Cosmochimica Acta 58, 197–208.

de Winter, N., Sinnesael, M., Makarona, C., Vansteenberge, S., Claeys, P., 2017. Trace element analyses of carbonates using portable and micro-X-ray fluorescence: Performance and optimization of measurement parameters and strategies. Journal of Analytical Atomic Spectrometry 32, 1211-1224.

de Winter, N.J., Claeys, P., 2016. Micro X-ray fluorescence (μXRF) line scanning on Cretaceous rudist bivalves: A new method for reproducible trace element profiles in bivalve calcite. Sedimentology. doi:10.1111/sed.12299

Deconto, R.M., 1996. Late Cretaceous climate, vegetation and ocean interactions: an earth system approach to modeling an extreme climate.Dercourt, J. et al, Zonenshain, L.P., Ricou, L.-E., Kazmin, V.G., Le Pichon, X., Knipper, A.L., Grandjacquet, C., Sbortshikov, I.M.,

Geyssant, J., Lepvrier, C., others, 1986. Geological evolution of the Tethys belt from the Atlantic to the Pamirs since the Lias. Tectonophysics 123, 241–315.

Dlugokencky, E., Tans P., 2017, NOAA/ESRL (www.esrl.noaa.gov/gmd/ccgg/trends/), accessed 31/01/2017.Douvillé, H., 1927. Nouveaux rudistes du Crétacé de Cuba. Bulletin de la Société géologique de France 4, 49–56.Ekart, D.D., Cerling, T.E., Montanez, I.P., Tabor, N.J., 1999. A 400 million year carbon isotope record of pedogenic carbonate: implications

for paleoatmospheric carbon dioxide. American Journal of Science 299, 805–827.Elderfield, H., Ganssen, G., 2000. Past temperature and δ18O of surface ocean waters inferred from foraminiferal Mg/Ca ratios. Nature 405,

442–445.Elliot, M., Welsh, K., Chilcott, C., McCulloch, M., Chappell, J., Ayling, B., 2009. Profiles of trace elements and stable isotopes derived from

giant long-lived Tridacna gigas bivalves: potential applications in paleoclimate studies. Palaeogeography, Palaeoclimatology, Palaeoecology 280, 132–142.

England, J., Cusack, M., Lee, M.R., 2007. Magnesium and sulphur in the calcite shells of two brachiopods, Terebratulina retusa and Novocrania anomala. Lethaia 40, 2–10.

Epstein, S., Buchsbaum, R., Lowenstam, H.A., Urey, H.C., 1953. Revised carbonate-water isotopic temperature scale. Geological Society of

861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936

America Bulletin 64, 1315–1326.Eshet, Y., Almogi-Labin, A., 1996. Calcareous nannofossils as paleoproductivity indicators in Upper Cretaceous organic-rich sequences in

Israel. Marine Micropaleontology 29, 37–61.Fan, C., Koeniger, P., Wang, H., Frechen, M., 2011. Ligamental increments of the mid-Holocene Pacific oyster Crassostrea gigas are reliable

independent proxies for seasonality in the western Bohai Sea, China. Palaeogeography, Palaeoclimatology, Palaeoecology 299, 437–448.

Fastovsky, D.E., Arthur, M.A., Strater, N.H., Foss, A., 1993. Freshwater bivalves (Unionidae), disequilibrium isotopic fractionation, and temperatures. Palaios 602–608.

Faul, K.L., Anderson, L.D., Delaney, M.L., 2003. Late Cretaceous and early Paleogene nutrient and paleoproductivity records from Blake Nose, western North Atlantic Ocean. Paleoceanography 18.

Fleurance, S., Cuney, M., Malartre, F., Reyx, J., 2013. Origin of the extreme polymetallic enrichment (Cd, Cr, Mo, Ni, U, V, Zn) of the Late Cretaceous–Early Tertiary Belqa Group, central Jordan. Palaeogeography, Palaeoclimatology, Palaeoecology 369, 201–219.

Fourtau, R., 1904. Contribution a L’Étude de la faune crétacique d’Egypte. Imprimerie nationale.Freitas, P., Clarke, L.J., Kennedy, H., Richardson, C., Abrantes, F., 2005. Mg/Ca, Sr/Ca, and stable-isotope (δ18O and δ13C) ratio profiles

from the fan mussel Pinna nobilis: Seasonal records and temperature relationships. Geochemistry, Geophysics, Geosystems 6.Freitas, P.S., Clarke, L.J., Kennedy, H., Richardson, C.A., Abrantes, F., 2006. Environmental and biological controls on elemental (Mg/Ca,

Sr/Ca and Mn/Ca) ratios in shells of the king scallop Pecten maximus. Geochimica et Cosmochimica Acta 70, 5119–5133.Freitas, P.S., Clarke, L.J., Kennedy, H., Richardson, C.A., others, 2009. Ion microprobe assessment of the heterogeneity of Mg/Ca, Sr/Ca and

Mn/Ca ratios in Pecten maximus and Mytilus edulis (bivalvia) shell calcite precipitated at constant temperature. Biogeosciences Discussions 6, 1267.

Freitas, P.S., Clarke, L.J., Kennedy, H.A., Richardson, C.A., 2008. Inter-and intra-specimen variability masks reliable temperature control on shell Mg/Ca ratios in laboratory and field cultured Mytilus edulis and Pecten maximus (bivalvia). Biogeosciences Discussions 5, 531–572.

Friedman, I., O’NEIL, J., CEBULA, G., 1982. Two New Carbonate Stable-Isotope Standards. Geostandards Newsletter 6, 11–12.Friedrich, O., Norris, R.D., Erbacher, J., 2012. Evolution of middle to Late Cretaceous oceans—a 55 my record of Earth’s temperature and

carbon cycle. Geology 40, 107–110.Gerringa, L.J.A., De Baar, H.J.W., Nolting, R.F., Paucot, H., 2001. The influence of salinity on the solubility of Zn and Cd sulphides in the

Scheldt estuary. Journal of Sea Research 46, 201–211.Gillikin, D.P., Lorrain, A., Bouillon, S., Willenz, P., Dehairs, F., 2006. Stable carbon isotopic composition of Mytilus edulis shells: relation

to metabolism, salinity, δ 13 C DIC and phytoplankton. Organic Geochemistry 37, 1371–1382.Gillikin, D.P., Lorrain, A., Navez, J., Taylor, J.W., André, L., Keppens, E., Baeyens, W., Dehairs, F., 2005. Strong biological controls on

Sr/Ca ratios in aragonitic marine bivalve shells. Geochemistry, Geophysics, Geosystems 6.Goodwin, D.H., Gillikin, D.P., Roopnarine, P.D., 2013. Preliminary evaluation of potential stable isotope and trace element productivity

proxies in the oyster Crassostrea gigas. Palaeogeography, Palaeoclimatology, Palaeoecology 373, 88–97.Goreau, T.J., 1977. Coral skeletal chemistry: physiological and environmental regulation of stable isotopes and trace metals in Montastrea

annularis. Proceedings of the Royal Society of London B: Biological Sciences 196, 291–315.Grossman, E.L., 1984. Stable isotope fractionation in live benthic foraminifera from the Southern California Borderland. Palaeogeography,

Palaeoclimatology, Palaeoecology 47, 301–327.Grossman, E.L., Ku, T.-L., 1986. Oxygen and carbon isotope fractionation in biogenic aragonite: temperature effects. Chemical Geology:

Isotope Geoscience section 59, 59–74.Grossman, E.L., Mii, H.-S., Zhang, C., Yancey, T.E., 1996. Chemical variation in Pennsylvanian brachiopod shells; diagenetic, taxonomic,

microstructural, and seasonal effects. Journal of Sedimentary Research 66, 1011–1022.Guo, L., Santschi, P.H., Ray, S.M., 2002. Metal partitioning between colloidal and dissolved phases and its relation with bioavailability to

American oysters. Marine Environmental Research 54, 49–64.Guo, T., DeLaune, R.D., Patrick, W.H., 1997. The influence of sediment redox chemistry on chemically active forms of arsenic, cadmium,

chromium, and zinc in estuarine sediment. Environment International 23, 305–316.Halfar, J., Zack, T., Kronz, A., Zachos, J.C., 2000. Growth and high-resolution paleoenvironmental signals of rhodoliths(coralline red algae)-

A new biogenic archive. Journal of Geophysical Research 105, 22–107.Hays, J.D., Imbrie, J., Shackleton, N.J., others, 1976. Variations in the Earth’s orbit: pacemaker of the ice ages. American Association for the

Advancement of Science.Hays, P.D., Grossman, E.L., 1991. Oxygen isotopes in meteoric calcite cements as indicators of continental paleoclimate. Geology 19, 441–

444.Hendry, J.P., Perkins, W.T., Bane, T., 2001. Short-term environmental change in a Jurassic lagoon deduced from geochemical trends in

aragonite bivalve shells. Geological Society of America Bulletin 113, 790–798.Hetzinger, S., Halfar, J., Zack, T., Gamboa, G., Jacob, D.E., Kunz, B.E., Kronz, A., Adey, W., Lebednik, P.A., Steneck, R.S., 2011. High-

resolution analysis of trace elements in crustose coralline algae from the North Atlantic and North Pacific by laser ablation ICP-MS. Palaeogeography, Palaeoclimatology, Palaeoecology 302, 81–94.

Holmden, C., Hudson, J.D., 2003. 87Sr/86Sr and Sr/Ca Investigation of Jurassic mollusks from Scotland: Implications for paleosalinities and the Sr/Ca ratio of seawater. Geological Society of America Bulletin 115, 1249–1264.

Huber, B.T., Hodell, D.A., Hamilton, C.P., 1995. Middle–Late Cretaceous climate of the southern high latitudes: stable isotopic evidence for minimal equator-to-pole thermal gradients. Geological Society of America Bulletin 107, 1164–1191.

Huber, B.T., Norris, R.D., MacLeod, K.G., 2002. Deep-sea paleotemperature record of extreme warmth during the Cretaceous. Geology 30, 123. doi:10.1130/0091-7613(2002)030<0123:DSPROE>2.0.CO;2

Immenhauser, A., Nägler, T.F., Steuber, T., Hippler, D., 2005. A critical assessment of mollusk 18 O/16 O, Mg/Ca, and 44 Ca/40 Ca ratios as proxies for Cretaceous seawater temperature seasonality. Palaeogeography, Palaeoclimatology, Palaeoecology 215, 221–237.

IPCC, 2013. Climate change 2013: the physical science basis: Working Group I contribution to the Fifth assessment report of the Intergovernmental Panel on Climate Change. Cambridge University Press.

Jackson, L.J., Kalff, J., Rasnnussen, J.B., 1993. Sediment pH and redox potential affect the bioavailability of Al, Cu, Fe, Mn, and Zn to rooted aquatic macrophytes. Canadian Journal of Fisheries and Aquatic Sciences 50, 143–148.

Jacob, D.E., Soldati, A.L., Wirth, R., Huth, J., Wehrmeister, U., Hofmeister, W., 2008. Nanostructure, composition and mechanisms of bivalve shell growth. Geochimica et Cosmochimica Acta 72, 5401–5415.

Jarvis, I., Mabrouk, A., Moody, R.T., de Cabrera, S., 2002. Late Cretaceous (Campanian) carbon isotope events, sea-level change and correlation of the Tethyan and Boreal realms. Palaeogeography, Palaeoclimatology, Palaeoecology 188, 215–248.

Jenkyns, H.C., 1980. Cretaceous anoxic events: from continents to oceans. Journal of the Geological Society 137, 171–188.Jenkyns, H.C., Forster, A., Schouten, S., Damsté, J.S.S., 2004. High temperatures in the late Cretaceous Arctic Ocean. Nature 432, 888–892.Jones, D.S., 1980. Annual cycle of shell growth increment formation in two continental shelf bivalves and its paleoecologic significance.

937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999

1000100110021003100410051006100710081009101010111012

Paleobiology 6, 331–340.Kah, L.C., Lyons, T.W., Frank, T.D., 2004. Low marine sulphate and protracted oxygenation of the Proterozoic biosphere. Nature 431, 834–

838.Kennedy, W.J., Jagt, J.W., Hanna, S.S., Schulp, A.S., 2000. Late Campanian ammonites from the Saiwan area (Huqf Desert, Sultanate of

Oman). Cretaceous Research 21, 553–562.Kim, S.-T., O’Neil, J.R., 1997. Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates. Geochimica et

Cosmochimica Acta 61, 3461–3475.Kirby, M.X., Soniat, T.M., Spero, H.J., 1998. Stable Isotope Sclerochronology of Pleistocene and Recent Oyster Shells (Crassostrea

virginica). PALAIOS 13, 560. doi:10.2307/3515347Klein, R.T., Lohmann, K.C., Thayer, C.W., 1996a. Bivalve skeletons record sea-surface temperature and δ18O via Mg/Ca and 18O/16O

ratios. Geology 24, 415–418.Klein, R.T., Lohmann, K.C., Thayer, C.W., 1996b. SrCa and 13C12C ratios in skeletal calcite of Mytilus trossulus: Covariation with

metabolic rate, salinity, and carbon isotopic composition of seawater. Geochimica et Cosmochimica Acta 60, 4207–4221.Korbar, T., Tesovic, B.C., Radovanovic, I., Krizmanic, K., Steuber, T., Skelton, P.W., 2010. Campanian Pseudosabinia from the Pučišća

Formation on the island of Hvar (Adriatic Sea, Croatia). Turkish Journal of Earth Sciences 19, 721–731.Ku, T.C.W., Walter, L.M., Coleman, M.L., Blake, R.E., Martini, A.M., 1999. Coupling between sulfur recycling and syndepositional

carbonate dissolution: evidence from oxygen and sulfur isotope composition of pore water sulfate, South Florida Platform, USA. Geochimica et Cosmochimica Acta 63, 2529–2546.

Lazareth, C.E., Vander Putten, E., André, L., Dehairs, F., 2003. High-resolution trace element profiles in shells of the mangrove bivalve Isognomon ephippium: a record of environmental spatio-temporal variations? Estuarine, Coastal and Shelf Science 57, 1103–1114.

Lazareth, C.E., Willenz, P., Navez, J., Keppens, E., Dehairs, F., André, L., 2000. Sclerosponges as a new potential recorder of environmental changes: Lead in Ceratoporella nicholsoni. Geology 28, 515–518.

Lear, C.H., Coxall, H.K., Foster, G.L., Lunt, D.J., Mawbey, E.M., Rosenthal, Y., Sosdian, S.M., Thomas, E., Wilson, P.A., 2015. Neogene ice volume and ocean temperatures: Insights from infaunal foraminiferal Mg/Ca paleothermometry. Paleoceanography 30, 1437–1454.

Li, L., Keller, G., 1999. Variability in Late Cretaceous climate and deep waters: evidence from stable isotopes. Marine Geology 161, 171–190.

Lomb, N.R., 1976. Least-squares frequency analysis of unequally spaced data. Astrophysics and space science 39, 447–462.Lorens, R.B., Bender, M.L., 1980. The impact of solution chemistry on Mytilus edulis calcite and aragonite. Geochimica et Cosmochimica

Acta 44, 1265–1278.Lorrain, A., Gillikin, D.P., Paulet, Y.-M., Chauvaud, L., Le Mercier, A., Navez, J., André, L., 2005. Strong kinetic effects on Sr/Ca ratios in

the calcitic bivalve Pecten maximus. Geology 33, 965–968.Malchus, N., 1990. Revision der Kreide-Austern (Bivalvia: Pteriomorphia) Ägyptens (Biostratigraphie, Systematik). Berliner

geowissenschaftliche Abhandlungen A 125, 1–231.Maliva, R.G., Siever, R., 1988. Mechanism and controls of silicification of fossils in limestones. The Journal of Geology 96, 387–398.Marali, S., Schöne, B.R., 2015. Oceanographic control on shell growth of Arctica islandica (Bivalvia) in surface waters of Northeast Iceland

—Implications for paleoclimate reconstructions. Palaeogeography, Palaeoclimatology, Palaeoecology 420, 138–149.Marali, S., Schöne, B.R., Mertz-Kraus, R., Griffin, S.M., Wanamaker, A.D., Matras, U., Butler, P.G., 2017. Ba/Ca ratios in shells of Arctica

islandica—Potential environmental proxy and crossdating tool. Palaeogeography, Palaeoclimatology, Palaeoecology 465, 347–361.

McConnaughey, T.A., Burdett, J., Whelan, J.F., Paull, C.K., 1997. Carbon isotopes in biological carbonates: Respiration and photosynthesis. Geochimica et Cosmochimica Acta 61, 611–622. doi:10.1016/S0016-7037(96)00361-4

McConnaughey, T.A., Gillikin, D.P., 2008. Carbon isotopes in mollusk shell carbonates. Geo-Marine Letters 28, 287–299.Meyers, S.R., 2014. Astrochron: An R package for astrochronology. http://cran.r-project.org/package=astrochron.Meyers, S.R., 2012. Seeing red in cyclic stratigraphy: Spectral noise estimation for astrochronology. Paleoceanography 27.Mitsuguchi, T., Matsumoto, E., Abe, O., Uchida, T., Isdale, P.J., 1996. Mg/Ca thermometry in coral skeletons. Science 274, 961.Morrison, J.M., Codispoti, L.A., Gaurin, S., Jones, B., Manghnani, V., Zheng, Z., 1998. Seasonal variation of hydrographic and nutrient

fields during the US JGOFS Arabian Sea Process Study. Deep Sea Research Part II: Topical Studies in Oceanography 45, 2053–2101.

Mouchi, V., De Rafélis, M., Lartaud, F., Fialin, M., Verrecchia, E., 2013. Chemical labelling of oyster shells used for time-calibrated high-resolution Mg/Ca ratios: a tool for estimation of past seasonal temperature variations. Palaeogeography, Palaeoclimatology, Palaeoecology 373, 66–74.

Newell, N.D., 1965. Classification of the Bivalvia. American Museum novitates; no. 2206.O’Neil, J.R., Clayton, R.N., Mayeda, T.K., 1969. Oxygen isotope fractionation in divalent metal carbonates. Univ. of Chicago.Owen, R., Kennedy, H., Richardson, C., 2002. Experimental investigation into partitioning of stable isotopes between scallop (Pecten

maximus) shell calcite and sea water. Palaeogeography, Palaeoclimatology, Palaeoecology 185, 163–174.Perrin, J., Rivard, C., Vielzeuf, D., Laporte, D., Fonquernie, C., Ricolleau, A., Cotte, M., Floquet, N., 2017. The coordination of sulfur in

synthetic and biogenic Mg calcites: The red coral case. Geochimica et Cosmochimica Acta 197, 226–244.Philip, J.M., Platel, J.-P., 1995. Stratigraphy and rudist biozonation of the Campanian and the Maastrichtian of Eastern Oman. Revista

mexicana de ciencias geológicas 12, 15.Platel, J.P., Philip, J., Bourdillon-de-Grissac, C., Babinot, J.F., Roger, J., Mercadier, C., 1994. Modalites de la transgression campanienne sur

le massif du Haushi-Huqf (Oman); stratigraphie, contexte geodynamique et paleoenvironnements. Bulletin de la Societe Geologique de France 165, 147–161.

Purton, L., Brasier, M., 1997. Gastropod carbonate δ18O and δ13C values record strong seasonal productivity and stratification shifts during the late Eocene in England. Geology 25, 871–874.

Quan, C., Sun, C., Sun, Y., Sun, G., 2009. High resolution estimates of paleo-CO 2 levels through the Campanian (Late Cretaceous) based on Ginkgo cuticles. Cretaceous Research 30, 424–428.

Rausch, S., Böhm, F., Bach, W., Klügel, A., Eisenhauer, A., 2013. Calcium carbonate veins in ocean crust record a threefold increase of seawater Mg/Ca in the past 30 million years. Earth and Planetary Science Letters 362, 215–224.

Richardson, C.A., Peharda, M., Kennedy, H., Kennedy, P., Onofri, V., 2004. Age, growth rate and season of recruitment of Pinna nobilis (L) in the Croatian Adriatic determined from Mg: Ca and Sr: Ca shell profiles. Journal of Experimental Marine Biology and Ecology 299, 1–16.

Rimstidt, J.D., Balog, A., Webb, J., 1998. Distribution of trace elements between carbonate minerals and aqueous solutions. Geochimica et Cosmochimica Acta 62, 1851–1863.

ROSENBERG, G.D., HUGHES, W.W., 1991. A metabolic model for the determination of shell composition in the bivalve mollusc, Mytilus

1013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088

edulis. Lethaia 24, 83–96.Rosenberg, G.D., Hughes, W.W., Parker, D.L., Ray, B.D., 2001. The geometry of bivalve shell chemistry and mantle metabolism. American

Malacological Bulletin 16, 251–261.ROSENBERG, G.D., HUGHES, W.W., TKACHUCK, R.D., 1989. Shell form and metabolic gradients in the mantle of Mytilus edulis.

Lethaia 22, 343–344.Ruf, T., 1999. The Lomb-Scargle periodogram in biological rhythm research: analysis of incomplete and unequally spaced time-series.

Biological Rhythm Research 30, 178–201.Sanders, D., Baron-Szabo, R.C., 1997. Coral-rudist bioconstructions in the Upper Cretaceous Haidach section (Gosau Group; Northern

Calcareous Alps, Austria). Facies 36, 69–89.Savin, S.M., Yeh, H.-W., 1981. Stable isotopes in ocean sediments. The sea 7, 1521–1554.Scargle, J.D., 1982. Studies in astronomical time series analysis. II-Statistical aspects of spectral analysis of unevenly spaced data. The

Astrophysical Journal 263, 835–853.Schöne, B.R., Houk, S.D., Castro, A.D.F., Fiebig, J., Oschmann, W., Kröncke, I., Dreyer, W., Gosselck, F., 2005. Daily growth rates in

shells of Arctica islandica: assessing sub-seasonal environmental controls on a long-lived bivalve mollusk. Palaios 20, 78–92.Schöne, B.R., Zhang, Z., Radermacher, P., Thébault, J., Jacob, D.E., Nunn, E.V., Maurer, A.-F., 2011. Sr/Ca and Mg/Ca ratios of

ontogenetically old, long-lived bivalve shells (Arctica islandica) and their function as paleotemperature proxies. Palaeogeography, Palaeoclimatology, Palaeoecology 302, 52–64.

Schubert, J.K., Kidder, D.L., Erwin, D.H., 1997. Silica-replaced fossils through the Phanerozoic. Geology 25, 1031–1034.Schumann, P.D.D., 1995. Upper cretaceous rudist and stromatoporid associations of Central Oman (Arabian Peninsula). Facies 32, 189–202.

doi:10.1007/BF02536868Sinclair, D.J., Kinsley, L.P., McCulloch, M.T., 1998. High resolution analysis of trace elements in corals by laser ablation ICP-MS.

Geochimica et cosmochimica acta 62, 1889–1901.Skelton, P.W., Nolan, S.C., Scott, R.W., 1990. The Maastrichtian transgression onto the northwestern flank of the Proto-Oman Mountains:

sequences of rudist-bearing beach to open shelf facies. Geological Society, London, Special Publications 49, 521–547.Stanley, S.M., Hardie, L.A., 1998. Secular oscillations in the carbonate mineralogy of reef-building and sediment-producing organisms

driven by tectonically forced shifts in seawater chemistry. Palaeogeography, Palaeoclimatology, Palaeoecology 144, 3–19.Steuber, T., 1999. Isotopic and chemical intra-shell variations in low-Mg calcite of rudist bivalves (Mollusca-Hippuritacea): disequilibrium

fractionations and late Cretaceous seasonality. International Journal of Earth Sciences 88, 551–570.Steuber, T., 1996. Stable isotope sclerochronology of rudist bivalves: Growth rates and Late Cretaceous seasonality. Geology 24, 315.

doi:10.1130/0091-7613(1996)024<0315:SISORB>2.3.CO;2Steuber, T., 1995. Stable isotope sclerochronology of Vaccinites cornuvaccinum (Hippuritidae) from Beotia (Greece). Revista mexicana de

ciencias geológicas 12, 20.Steuber, T., Löser, H., 2000. Species richness and abundance patterns of Tethyan Cretaceous rudist bivalves (Mollusca: Hippuritacea) in the

central-eastern Mediterranean and Middle East, analysed from a palaeontological database. Palaeogeography, Palaeoclimatology, Palaeoecology 162, 75–104.

Steuber, T., Rauch, M., Masse, J.-P., Graaf, J., Malkoč, M., 2005. Low-latitude seasonality of Cretaceous temperatures in warm and cold episodes. Nature 437, 1341–1344. doi:10.1038/nature04096

Surge, D., Lohmann, K.C., 2008. Evaluating Mg/Ca ratios as a temperature proxy in the estuarine oyster, Crassostrea virginica. Journal of Geophysical Research: Biogeosciences 113.

Surge, D.M., Lohmann, K.C., 2002. Temporal and spatial differences in salinity and water chemistry in SW Florida estuaries: effects of human-impacted watersheds. Estuaries and Coasts 25, 393–408.

Swart, P.K., 1983. Carbon and oxygen isotope fractionation in scleractinian corals: a review. Earth-Science Reviews 19, 51–80.Swart, P.K., Thorrold, S., Rosenheim, B., Eisenhauer, A., Harrison, C.G.A., Grammer, M., Latkoczy, C., 2002. Intra-annual variation in the

stable oxygen and carbon and trace element composition of sclerosponges. Paleoceanography 17.Takesue, R.K., van Geen, A., 2004. Mg/Ca, Sr/Ca, and stable isotopes in modern and Holocene Protothaca staminea shells from a northern

California coastal upwelling region. Geochimica et Cosmochimica Acta 68, 3845–3861.Tarutani, T., Clayton, R.N., Mayeda, T.K., 1969. The effect of polymorphism and magnesium substitution on oxygen isotope fractionation

between calcium carbonate and water. Geochimica et Cosmochimica Acta 33, 987–996.Thomson, D.J., 1982. Spectrum estimation and harmonic analysis. Proceedings of the IEEE 70, 1055–1096.Ullmann, C.V., Böhm, F., Rickaby, R.E., Wiechert, U., Korte, C., 2013. The Giant Pacific Oyster (Crassostrea gigas) as a modern analog for

fossil ostreoids: isotopic (Ca, O, C) and elemental (Mg/Ca, Sr/Ca, Mn/Ca) proxies. Geochemistry, Geophysics, Geosystems 14, 4109–4120.

Ullmann, C.V., Frei, R., Korte, C., Hesselbo, S.P., 2015. Chemical and isotopic architecture of the belemnite rostrum. Geochimica et Cosmochimica Acta 159, 231–243.

Ullmann, C.V., Wiechert, U., Korte, C., 2010. Oxygen isotope fluctuations in a modern North Sea oyster (Crassostrea gigas) compared with annual variations in seawater temperature: Implications for palaeoclimate studies. Chemical Geology 277, 160–166.

Vance, D., Little, S.H., de Souza, G.F., Khatiwala, S., Lohan, M.C., Middag, R., 2017. Silicon and zinc biogeochemical cycles coupled through the Southern Ocean. Nature Geoscience.

Vander Putten, E., Dehairs, F., Keppens, E., Baeyens, W., 2000. High resolution distribution of trace elements in the calcite shell layer of modern Mytilus edulis: Environmental and biological controls. Geochimica et Cosmochimica Acta 64, 997–1011.

Veizer, J., Fritz, P., Jones, B., 1986. Geochemistry of brachiopods: oxygen and carbon isotopic records of Paleozoic oceans. Geochimica et Cosmochimica Acta 50, 1679–1696.

Wanamaker, A.D., Kreutz, K.J., Borns, H.W., Introne, D.S., Feindel, S., Funder, S., Rawson, P.D., Barber, B.J., 2007. Experimental determination of salinity, temperature, growth, and metabolic effects on shell isotope chemistry of Mytilus edulis collected from Maine and Greenland. Paleoceanography 22.

Wanamaker, A.D., Kreutz, K.J., Schöne, B.R., Introne, D.S., 2011. Gulf of Maine shells reveal changes in seawater temperature seasonality during the Medieval Climate Anomaly and the Little Ice Age. Palaeogeography, Palaeoclimatology, Palaeoecology 302, 43–51.

Wanamaker Jr, A.D., Kreutz, K.J., Wilson, T., Borns Jr, H.W., Introne, D.S., Feindel, S., 2008. Experimentally determined Mg/Ca and Sr/Ca ratios in juvenile bivalve calcite for Mytilus edulis: implications for paleotemperature reconstructions. Geo-Marine Letters 28, 359–368.

Wang, W.-X., Fisher, N.S., 1996. Assimilation of trace elements and carbon by the mussel Mytilus edulis: effects of food composition. Limnology and Oceanography 4, l.

Woodward, S.P., 1855. On the structure and affinities of the Hippuritidae. Quarterly Journal of the Geological Society 11, 40–61.Yoshimura, T., Suzuki, A., Tamenori, Y., Kawahata, H., 2014. Micro-X-ray fluorescence-based comparison of skeletal structure and P, Mg,

Sr, O and Fe in a fossil of the cold-water coral Desmophyllum sp., NW Pacific. Geo-Marine Letters 34, 1–9.Yoshimura, T., Tamenori, Y., Suzuki, A., Nakashima, R., Iwasaki, N., Hasegawa, H., Kawahata, H., 2013. Element profile and chemical

1089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164

environment of sulfur in a giant clam shell: insights from μ-XRF and X-ray absorption near-edge structure. Chemical Geology 352, 170–175.

Zwolsman, J.J., Van Eck, B.T., Van Der Weijden, C.H., 1997. Geochemistry of dissolved trace metals (cadmium, copper, zinc) in the Scheldt estuary, southwestern Netherlands: impact of seasonal variability. Geochimica et Cosmochimica Acta 61, 1635–1652.

Figure 1: Overview of colour scans and XRF maps of Vaccinites vesiculosus (VV, 1A), Torreites sanchezi milovanovici (TS, 1B) and Oscillopha figari (OF, 1C) sections. Images on the left show colour scans of the polished surface of the bivalves with locations of the lines of measurement in yellow. Central images show XRF maps of relative intensities of Fe (red), Si (green) and Ca (blue). Images on the right-hand side show XRF heatmaps of the relative intensity of Fe and Mn, with red colours indicating higher concentrations and blue colours indicating lower concentrations (see scale bar in 1A). Figures 1B and 1C contain close-ups showing growth microstructure of the calcite in the outer shell layer and the hinge of the TS and OF shells, respectively. Background of the VV maps (1A) shows glass beads used to stabilize the shell during measurement. This was not necessary for the measurement of the heavier TS and OF shells (1B and 1C).

11651166116711681169

1170

1171117211731174117511761177117811791180

Figure 2: A) Plots of Sr concentrations measured along the growth axis of Vaccinites vesiculosus (VV) using LA-ICP-MS (blue and green) and two µXRF methods; conventional line scanning (linescan) in red and point-by-point (pt-by-pt) scanning in purple. XRF data in 2A are corrected using conventional single-standard correction with CRM393. Dashed lines marked “A”, “B” and “C” correspond to marks on the sampling lines in Figure 1A. B) Schematic representation of the correction of XRF data from single-standard calibration (solid gray line) to two-standard calibration (solid black line). This re-correction allows any single-standard calibrated value [M]old to be corrected to a two-standard calibrated value ([M]new; see formula and explanation in supplementary data). C) Linear regression plots of Sr concentrations of the first LA-ICP-MS line (blue line in 2A) against strontium concentrations measured in the other three line scans, both before (left) and after the calibration with the second LA-ICP-MS linescan.

1181

11821183118411851186118711881189119011911192

Figure 3: Average concentrations of all trace elements measured in the VV (purple), TS (blue) and OF (orange) shells using LA-ICP-MS calibrated point-by-point XRF measurements. Bars indicate ranges of concentrations measured in the shell calcite and black dots indicate the averages of these ranges. Values below 1 ppm are dashed because these are below the detection limit of the method (de Winter and Claeys, 2016). Elements of interest for this study are highlighted.

1193

11941195119611971198

Figure 4: Vaccinites vesiculosus (VV) shell. Overview of stable isotope (‰VPDB), elemental abundance (wt.% and µg/g) and relative abundance records (mmol/mol, 11 point running average). Shaded vertical red bars indicate intervals with high concentrations of Si, Fe and Mn. Dashed lines marked “I”, “II” and “III” correspond to marks on the sampling lines in Figure 1A.

1199

1200120112021203

Figure 5: Torreites sanchezi milovanovici (TS) shell. Overview of stable isotope (‰VPDB), elemental abundance (wt.% and µg/g) and relative abundance records (mmol/mol, 11 point running average). Dashed lines marked “I”, “II” and “III” correspond to marks on the sampling lines in Figure 1B.

1204

120512061207

Figure 6: Oscillopha figari (OF) shell. Overview of stable isotope (‰VPDB) and elemental abundance records (wt.% and µg/g) and relative abundance records (mmol/mol, 11 point running

1208

12091210

average). Dashed lines marked “I”, “II” and “III” correspond to marks on the sampling lines in Figure 1C.

Figure 7: Records of stable isotopes and four element-to-calcium ratios from Vaccinites vesiculosus (VV) are plotted together with cyclical components filtered from the data using bandpass filtering. Spectrograms show powerspectra from the three methods of spectral analysis (coloured lines) as well as confidence levels calculated using red noise estimation and harmonic F-tests (grey shaded areas) of every record. The black rectangle on the powerspectra indicates the range of frequencies with significant periodicity used as the window for bandpass filtering. Yellow and white bars represent interpreted seasons in the δ18O record. Dashed lines marked “I”, “II” and “III” correspond to marks on the sampling lines in Figure 1A.

12111212

1213

12141215121612171218121912201221

Figure 8: Records of stable isotopes and four element-to-calcium ratios from Torreites sanchezi milovanovici (TS) are plotted together with cyclical components filtered from the data using bandpass filtering. Spectrograms show powerspectra from the three methods of spectral analysis (coloured lines) as well as confidence levels calculated using red noise estimation and harmonic F-tests (grey shaded areas) of every record. The black rectangle on the powerspectra indicates the range of frequencies with significant periodicity used as the window for bandpass filtering. Yellow and white bars represent interpreted seasons in the δ18O record. Dashed lines marked “I”, “II” and “III” correspond to marks on the sampling lines in Figure 1B.

1222

12231224122512261227122812291230

Figure 9: Records of stable isotopes and four element-to-calcium ratios from Oscillopha figari (OF) are plotted together with cyclical components filtered from the data using bandpass filtering. Spectrograms show powerspectra from the three methods of spectral analysis (coloured lines) as well as confidence levels calculated using red noise estimation and harmonic F-tests (grey shaded areas) of every record. The black rectangle on the powerspectra indicates the range of frequencies with significant periodicity used as the window for bandpass filtering. Yellow and white bars represent interpreted seasons in the δ18O record. Dashed lines marked “I”, “II” and “III” correspond to marks on the sampling lines in Figure 1C.

1231

12321233123412351236123712381239

Figure 10: Crossplot showing the Mg/Ca ratios in biogenic calcite and its relationship to calcification temperature. Dashed lines indicate palaeothermometer calibrations of foraminifera (blue), oysters (green) and other bivalves (purple). Shaded boxes indicate the ranges of Mg/Ca ratios and δ18O-inferred palaeotemperatures in this study. The arrow points towards Mg/Ca values that were corrected for Mg/Ca values of that sea water, estimated to be ±3.3 times lower than today. Horizontal and vertical bars indicate the seasonal amplitude of the records.

1240

124112421243124412451246

Table 1Overview of measured average concentrations, measurement reproducibility (1 standard deviation, SD) and instrumental uncertainty (1 SD) based on XRF deconvolution for major and trace element measurements on the Vaccinites vesiculosus (VV) shell using the two XRF methods (conventional line scanning and point-by-point analysis) and LA-ICP-MS. For LA-ICP-MS, the coefficient of determination (R2) of the calibration curve based on standard measurements is given.

1247

12481249125012511252

Table 2: Overview matrix of correlations between proxy records in the three bivalves. Upper matrices show correlations made on the basis of on visual inspection. Red fills indicate antiphase correlations while green fills signify in-phase correlations. Matrices in the second row show p-values of correlations between these records. Green fills indicate statistically significant relationships (p < 0.05), while red fills indicate no significant relationship. The bottom row of matrices show Pearson’s coefficients (r) of correlations between records. Green fills indicate positive correlations between records (r > 0) and red and orange fills indicate negative correlations (r < 0). White fills indicate that there is no significant correlation between the records.

Supplementary file 1: Detailed description of LA-ICP-MS methodology and calibration of XRF data.Supplementary file 2: Trace element results of VV shellSupplementary file 3: Trace element results of TS shellSupplementary file 4: Trace element results of OF shellSupplementary file 5: Compilation of stable isotope resultsSupplementary file 6: Spectral analysis routine used for data in figures 7, 8 and 9Supplementary file 7: Spectral analysis results for VV shellSupplementary file 8: Spectral analysis results for TS shellSupplementary file 9: Spectral analysis results for OF shell

1253

12541255125612571258125912601261

1262

1263126412651266126712681269127012711272