A Paired Apatite and Calcite Clumped Isotope Thermometry Approach to Estimating Cambro-Ordovician Seawater Temperatures and Isotopic Composition

Available online at www.sciencedirect.com ScienceDirect

Geochimica et Cosmochimica Acta 224 (2018) 18–41 www.elsevier.com/locate/gca

A paired apatite and calcite clumped isotope thermometry approach to estimating Cambro-Ordovician seawater temperatures and isotopic composition

Kristin D. Bergmann a,⇑, Seth Finnegan b, Roger Creel a, John M. Eiler c, Nigel C. Hughes d, Leonid E. Popov e, Woodward W. Fischer c

a Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, United States b Department of Integrative Biology, University of California, Berkeley, Berkeley, CA 94720, United States c Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, United States d Department of Earth Sciences, University of California, Riverside, Riverside, CA 92521, United States e Department of Geology, National Museum of Wales, Cardiﬀ CF10 3NP, Wales, UK

Received 8 February 2017; accepted in revised form 10 November 2017; available online 6 December 2017

Abstract

The secular increase in d18O values of both calcitic and phosphatic marine fossils through early Phanerozoic time suggests either that (1) early Paleozoic surface temperatures were high, in excess of 40 °C (tropical MAT), (2) the d18O value of seawater has increased by 7–8‰ VSMOW through Paleozoic time, or (3) diagenesis has altered secular trends in early Paleozoic samples. Carbonate clumped isotope analysis, in combination with petrographic and elemental analysis, can deconvolve fluid composition from temperature effects and therefore determine which of these hypotheses best explain the secular d18O increase. Clumped isotope measurements of a suite of calcitic and phosphatic marine fossils from late Cambrian- to Middle-late Ordovician-aged strata–the first paired fossil study of its kind–document tropical sea surface temperatures near modern temperatures (26–38 °C) and seawater oxygen isotope ratios similar to today’s ratios. Ó 2017 Elsevier Ltd. All rights reserved.

Keywords: Clumped isotope; Apatite; Calcite; Seawater; Oxygen isotope

1. INTRODUCTION when climate transitions are hypothesized to have played a major role in the radiation of many animal lineages Carbonate and phosphatic rocks retain our most com- (e.g. Rasmussen et al., 2016; Trotter et al., 2008). plete evidence of climatic conditions in Earth’s deep past. Efforts to extrapolate from these d18O records broader Determining how temperature and seawater oxygen isotope inferences about climate and biotic evolution have been (d18O) fluctuations influenced carbonate and phosphatic hampered, however, by the difficulty of making meaningful d18O records over geologic time, therefore, is crucial for comparisons among data derived from disparate sources. understanding the coupled evolutions of Earth’s climate For example, d18O measurements of conodont apatite by and biota. This is particularly true in early Paleozoic time, secondary ion mass spectrometry (SIMS) were used to propose high Early Ordovician sea surface temperatures ° ⇑ >40 C that declined to modern-like surface temperatures Corresponding author. (<25 °C) by late Middle Ordovician time; this cooling was E-mail address: [email protected] (K.D. Bergmann). https://doi.org/10.1016/j.gca.2017.11.015 0016-7037/Ó 2017 Elsevier Ltd. All rights reserved. K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 19 postulated to have spurred the radiation of diverse skele- explanations: (1) surface ocean temperatures have tonized calcitic taxa in the Middle Ordovician decreased since Precambrian time; (2) seawater oxygen iso- (Rasmussen et al., 2016; Trotter et al., 2008). In contrast, topic composition has increased through time; or (3) diage- d18O records of Paleozoic brachiopods have been used to nesis has altered older rocks more than younger samples. It argue that early Paleozoic calcitic skeletons grew at extre- has been difficult to evaluate the relative importance of mely low temperatures (e.g. 2 °C) (Giles, 2012), and that these hypotheses because d18O depends both on deposi- the bulk isotopic composition of seawater increased 7‰ tional temperature and on the isotopic composition of the from early Paleozoic values to present levels (Jaffre´s et al., water from which a mineral precipitates. Although prefer- 2007; Kasting et al., 2006; Prokoph et al., 2008; Veizer ential diagenetic alteration of older materials cannot et al., 1999, 1997; Veizer and Prokoph, 2015). These large explain the entire d18O trend in carbonate, phosphate, discrepancies in interpretation arise from different mineral and chert that declines in parallel through the Phanerozoic systems, different dataset approaches, and divergent (Joachimski et al., 2004; Veizer et al., 1999), there is at pre- assumptions about the cause of the observed mineralogical sent no consensus on which of the remaining hypotheses d18O increase over the early Paleozoic (Fig. 1). best explains the secular increase. The secular increase in oxygen isotope values docu- Paired isotopic analyses of co-occurring sample suites of mented in ancient carbonates, apatites, and cherts in the apatite and calcite (Joachimski et al., 2004; Wenzel et al., early Paleozoic and extending though the Phanerozoic era 2000) or apatite and chert (Karhu and Epstein, 1986) pro- (Knauth and Epstein, 1976; Karhu and Epstein, 1986; vide a useful way to test the relative quality of the archives Shields and Veizer, 2002; Jaffre´s et al., 2007; Prokoph held by different phases via their mutual equilibrium frac- et al., 2008; Shemesh et al., 1988; Veizer et al., 1997) (early tions. Previous paired studies, however, have produced Paleozoic increase shown in Fig. 1) has three end-member variable results and yielded conflicting interpretations.

Brachiopod Matrix A 0 Rugose Coral Trilobite (‰) Conodont Linguliform Brachiopod VPDB Phosphate Nodule O

18 -5 δ mineral -10

(‰) 20

VSMOW O 18

δ 15 apatite

10 C 0 (‰)

B D VP O

18 -5 δ mineral -10

5000 48 4060 44 Age

18 18 Fig. 1. Composite mineral d O from Cambrian, Ordovician and early Silurian-aged shells and matrix. (A) d OVPDB from calcitic 18 brachiopods (diamonds) from tropical sites replotted from (Veizer et al., 1999). (B) apatite d OVSMOW from conodonts (circles) and linguliform brachiopods (squares) compiled from (Bassett et al., 2007; Elrick et al., 2011; Lećuyer et al., 1998; Trotter et al., 2008; Wenzel 18 et al., 2000). (C) d OVPDB from calcitic brachiopods (blue diamonds), trilobites (green reverse triangles), rugose corals (pink triangles), matrix (blue circles), and CO2 generated from carbonate groups substituted into linguliform brachiopods (orange squares), and phospatic nodules (brown diamond). The data from this study are overlaid on transparent calcitic brachiopod data from (Veizer et al., 1999). Data from this study that span three significant positive carbon isotope excursions (SPICE, GICE and HICE) are highlighted. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 20 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41

Early work, with imperfect temperature relationships, pos- linguliforms—Lingula and Glottidia—collected from five tulated that equilibrium precipitation of two phases implied locations worldwide: Florida, southern California, Hawaii, exceedingly warm ocean temperatures and no change in the Philippines and Japan. To best calibrate the measure- bulk fluid d18O(Karhu and Epstein, 1986; Shemesh et al., ments in modern samples, shells grown in a range of mean 1988) whereas other studies have argued that one or more average temperatures were selected. Linguliform valves of mineral phases precipitated out of equilibrium with primary Glottidia albida from Terminal Island and Newport Bay, fluids or were subsequently altered (Bassett et al., 2007; CA, and one specimen of the calcitic rhynchonelliform bra- Joachimski et al., 2004; Wenzel et al., 2000). Recent studies chiopod Terebratalia transversa from the San Pedro Jetty, redefining the phosphate-water equilibrium fractionation Los Angeles, CA, were analyzed to compare calcite and relationship (Chang and Blake, 2015; Puceát et al., 2010) phosphate shell materials from nearby localities. Samples have also undermined previous interpretations of paired of Lingula reevii were raised in the Waikiki Aquarium for studies of phosphate-carbonate (Joachimski et al., 2004; >5 years at a known temperature after being harvested Wenzel et al., 2000) and phosphate-chert (Karhu and from Kaneohe Bay, Hawaii. Because the Waikiki Aquar- Epstein, 1986), as well as d18O phosphate records ium used local seawater and temperature ranges, we (Shemesh et al., 1988; Trotter et al., 2008). Additionally, assumed that the move from Kaneohe Bay caused a mini- analytical methods to isolate and measure the d18Oof mal change in temperature. Lingula ungis valves were col- PO4 have varied substantially and yielded inconsistent lected from the Polillo Islands, Philippines, and Glottidia results (Puceát et al., 2010). Recent efforts have also recal- pyramidata valves were collected from Tampa Bay, Florida. ibrated the composition of NBS 120-C, a standard com- Lingula anatina valves were collected from the tidal flats of monly used to calibrate phosphate oxygen isotope ratio Ariake Bay, Japan. A deep-sea calcitic brachiopod was ana- measurements (Puceát et al., 2010). lyzed as a cold temperature end-member. The single valve Carbonate clumped isotope thermometry, a measure of was collected from Monterey Canyon up dip of Shepard’s crystallization temperature not influenced by fluid d18O iso- Meander near studied whale falls, where benthic tempera- tope composition (Eiler, 2007), can substantiate observed tures range from 2 to 7 °C(Lundsten et al., 2010). This d18O trends in O-bearing marine precipitates through early calibration suite spans a temperature range of 2–35 °C Paleozoic time. When paired with classical d18O paleother- (Supplemental Materials, D). mometry (Urey, 1947; Friedman and O’Neil, 1977; Kim and O’Neil, 1997), the technique can also establish the iso- 2.2. Middle to Late Ordovician fossil materials topic composition of the seawater, formation water, or diagenetic fluid with which the mineral last equilibrated and, when Samples of calcitic and phosphatic fossil materials were combined with petrographic and elemental analysis, can collected from six stratigraphic sections of the Late Ordovi- untangle primary isotopic signals from diagenetic artifacts. cian (Sandbian-early Katian) Decorah Formation in Min- Here we use carbonate clumped isotope thermometry to nesota, Iowa and Wisconsin in the summers of 2010 and construct from calcitic and phosphatic biominerals and 2011 (Emerson, 2002; Sell et al., 2015). The unit contains other fabrics a paired phosphate-calcite record for parts four bentonite ash deposits: the Deicke, Millbrig, Elkport of the Cambrian and Ordovician periods—the interval with and Dickeyville beds, which can be correlated across east- the most substantial observed change in global d18O records ern North America (Sell et al., 2015). The Decorah Forma- in the Phanerozoic era (Fig. 1)(Veizer et al., 1997). Bio- tion strata also capture the Guttenberg Carbon Isotope genic phosphates behave differently from calcitic biominer- Excursion (GICE) (Ludvigson et al., 2004; Simo et al., als during diagenesis and serve as an important geochemical 2003). The Decorah Formation sections in Iowa and Min- archive of seawater conditions prior to the diversification of nesota have experienced minimal burial and remain below thick-shelled calcitic skeletons in many groups during the the oil window based on thermal maturity estimates from Ordovician biodiversification (Pruss et al., 2010; Trotter organic material (Tmax < 441 °C, hydrogen indices of et al., 2008). Building on previous phosphatic calibration 278–1024 mg HC/g TOC) and conodonts (CAI = 1) studies of carbonate clumped isotope thermometry in bone (Smith and Clark, 1996; Rohrssen et al., 2015). Northern and teeth (Eagle et al., 2010, 2011; Wacker et al., 2016), we Wisconsin sections not explored in this study have Missis- also measure the clumped isotope values (D47)ofCO2 sippi Valley Type (MVT) mineralization. Biomarker analy- extracted from modern linguliform brachiopods with ses from these MVT-deposits indicate peak burial known growth temperatures. We compare measurements temperatures of 95–118 °C, Ro values of 0.6–0.8 and Tmax 18 18 of the d OofCO3 groups with the d OofPO4 groups of 425 °C(Barker and Pawlewicz, 1994; Rowan and in these phosphatic brachiopods to test whether the Goldhaber, 1996). Maximum burial depth estimates range clumped isotope measurements yield equilibrium fractiona- from 600 to 900 m and assuming a 25 °C/km geothermal tion patterns consistent with primary growth temperatures. gradient and a 22 °C surface temperature, expected burial temperatures are 37–47 °C outside of the zone of MVT 2. SAMPLES AND ANALYTICAL METHODS mineralization (Rowan and Goldhaber, 1996). This is well below the temperatures needed to generate solid state 2.1. Modern brachiopod materials reordering with <1 °C expected reordering for rocks held at 75 °C for 100 million years (Henkes et al., 2014; The modern brachiopods used for calibration Stolper and Eiler, 2015). Brachiopods, rugose corals, trilo- included calcitic brachiopods and both extant genera of bites, crinoids and bryozoa including Prasapora sp. were K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 21 collected for analysis because of their diverse range of mate- collectors, we also purchased exceptionally preserved trilo- rial properties. Large Lingulasma galenaense phosphatic bites contained within a carbonate wackestone matrix from brachiopods and phosphatic nodules were collected from the Russian Aseri and Kunda Formations. the Decorah Bruening Quarry and Pole Line Road localities near Decorah, Iowa (Emerson, 2002; Bergstro¨m and 2.3. late Cambrian fossil materials McKenzie, 2005; Simo et al., 2003)(Fig. 2). The strata of the Volkhov, Kunda and Aseri Stages Phosphatic linguliform brachiopods were collected from (Floian-Darriwilian) of Estonia and adjacent Russia is are five Eau Claire Formation localities in Wisconsin in the among the few successions on earth that preserve abundant summer of 2011. The different localities sampled for this calcitic fossil materials from Middle Ordovician-aged rocks study are all from siliciclastic-dominated, nearshore to mar- with burial history below the oil window (Egerquist and ginal marine environments, with fossils collected by split- Egerquist, 2003; Rohrssen, 2013; Dronov et al., 2005; ting sandstone outcrop hand samples (Fig. 2). Due to the Popov et al., 2005; Rasmussen et al., 2016). The entire sample sizes required for clumped isotope analysis, we were thickness of the Ordovician Strata is 100 m and total burial not able to collect samples from coeval offshore deposits estimates are similarly minimal suggesting solid state because these are known only from drill cores. The collec- reordering is a non-issue (Stolper and Eiler, 2015; Henkes tion sites in Strum, Colfax, and Remington Quarry, Wis- et al., 2014; Rasmussen et al., 2016). Calcitic brachiopods consin, have different dominant species of Lingulepis and from the lower part of the Volkhov Formation were span two trilobite zones, the Cedaria and Crepicephalus collected by P. Fedorov and provided by L. Popov. From zones (Cowan et al., 2005). The Lingulepis specimens from

A B

C D

E F

Fig. 2. Field photographs and shell materials of the Eau Claire and Decorah Formations. (A) Nearshore sandstone lithology of the Eau Claire Formation near Strum, Wisconsin. (B) Mixed poorly lithified mudstone and cemented packstone carbonate lithologies of the Decorah Formation at the Decorah Bruening Quarry, Decorah, IA. (C) Mineralized hardground surfaces with significant burrowing interbedded with poorly lithified mudstone near Rochester, MN. (D) Abundant iron ooids from the middle Decorah Formation near Rochester, MN. (E) Lingulasma galenaense recovered from the St. James Member of the Decorah Formation at the Decorah Bruening Quarry. (F) prepared sample of Lingulepis sp. from the Crepicephalus zone of the Eau Claire Formation near Strum, WI. 22 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 the Cambrian Eau Claire Formation sampled near Colfax, at a working distance between 7–9 mm using a Quadrant WI are relatively thin, white, and in certain cases have lost Back-Scattering Detector (QBSD). all luster in the outer shell. The Lingulepis specimens from Using a JEOL JXA-8200 Electron Microprobe, we Strum, WI display brown growth banding and are generally made quantitative elemental spot analyses and elemental thicker and iridescent. The better preserved shells from maps of the fossil materials on the fossil materials to mea- Strum, WI were analyzed as part of this study (Fig. 2, sure and image micro-scale trace metal abundances in dif- ECIB). The Eau Claire Fm. burial history is minimal and ferent sample textures. For all quantitative results, the broadly similar to the Decorah Formation as it is even clo- accelerating voltage was 15 kV, the beam current was ser to the maximum Paleozoic outcrop extent again sug- 20nA, and the beam size was 1 mm. The CITZAF method gesting solid state reordering is not an issue in this was used for matrix correction. Sample standards for the locality (Henkes et al., 2014; Stolper and Eiler, 2015). The five chemical elements analyzed on calcitic samples total Cambrian-Early Ordovician stratigraphic thickness included: calcite for Ca, dolomite for Mg, siderite for Fe, in the Eau Claire area is 100–150 m thick and no younger rhodochrosite for Mn, strontianite for Sr, and anhydrite stratigraphy is outcropping (Runkel et al., 2007). for S. Ca had an average detection limit of 161 ppm, Mg–294 ppm, Fe–312 ppm, Mn–293 ppm, Sr–507 ppm, 2.4. Sample processing and S–91 ppm. Sample standards for the ten elements mapped on the linguliform brachiopod included: anhydrite We extracted phosphatic linguliform brachiopod shells for S, apatite for P, cerium phosphate for Ce, albite for Na, from the Eau Claire Formation using a Dremel hand drill, fluorine mica for F, apatite for Ca, yttrium phosphate for sonication in DI water and sand abrasion. Phosphatic shells Y, dolomite for Mg, siderite for Fe, and strontianite for Sr. from the Decorah Formation were isolated using dental picks. Once extracted, all taxa were cleaned with a variety 2.6. Bulk powder inductively coupled plasma optical emission of tools including dental picks, toothbrush, sand abrasion, spectroscopy (ICP-OES) and sonication. For isotopic and XRD analyses, calcitic brachiopod For 63 of the 130 calcitic clumped isotope samples, an shells were gently crushed and picked under a dissecting aliquot of powder was analyzed for trace metals on an microscope for flakes preserving shell ultrastructure: ICP-OES. We dissolved between 1 and 3 mg of powder in blocky zones of recrystallization were avoided. Frag- 3 mL of dilute nitric acid for 24 h at 25 °C at the California ments of well-preserved shell ultrastructure were then Institute of Technology. Samples were analyzed at the Jet powdered in a mortar and pestle. Trilobite fragments Propulsion Laboratory using a Thermo iCAP 6300 radial were crushed and powdered. Rugose corals were sec- view ICP-OES with a Cetac ASX260 autosampler with tioned in two directions and a small portion of skeletal solutions aspirated to the Ar plasma using a peristaltic material was drilled, avoiding visible cement-filled bor- pump. Three standard solutions of 10 ppb, 100 ppb and ings and cavities. 1 ppm of Fe, Mn, and Sr were run between every 8 sample A sample of many isolated and cleaned linguliform bra- unknowns. A methods blank was also analyzed during the chiopod fragments from the same horizon was powdered. analytical run. Because modern lingulid skeletons typically contain substantial amounts of organic matter—and organic com- 2.7. Bulk powder x-ray diffraction (XRD) pounds can challenge clumped isotope mass spectrometry—we assessed the precision and accuracy of XRD measurements were made on powders of the lin- several cleaning protocols for modern and ancient samples guliform brachiopod materials used for clumped isotope in order to find the best way to removing organic material measurements on a PANalytical X’Pert Pro within the prior to analysis. The range of protocols that we tested Department of Applied Physics and Materials Science at included: no cleaning, 3% H2O2 for 4 h, 30% H2O2 for 24 the California Institute of Technology. Scans were run from h, 50–50 30% H2O2 and 3% NaOCl for 24 h, and 50–50 5–70° 2h with a step size of 0.008 and a scan step time of 30% H2O2 and 1 M NaOH until no further reaction was 10.16 s. A Cu anode was used at 45 kV and 40 mA. A noticeable (over 36 h). All powders were then treated with zero-background silicon plate was used for all measure- a 0.1 M buffered acetic and acetate solution for 12 h to ments because of sample size limitations. Mineralogical remove labile organic-hosted carbonate groups (Koch phases that most closely matched a given linguliform bra- et al., 1997). chiopod sample were identified using the X’Pert Highscore IDMin function. 2.5. SEM/electron microprobe 2.8. Carbonate clumped isotope thermometry We used a ZEISS 1550 VP Field Emission Scanning Electron Microscope (SEM) equipped with an Oxford Samples were analyzed between December 2010 and INCA Energy 300 X-ray Energy Dispersive Spectrometer January 2013. We weighed 9–12 mg of calcitic powder or (EDS) system in the California Institute of Technology 100–200 mg of phosphatic powder into silver capsules. Geological and Planetary Sciences Division Analytical The powders reacted in a phosphoric common acid bath Facility to characterize the mineralogy and textures of the (103%; 1.90 < q < 1.92) for 20 min at 90 °C. Evolved samples prior to isotopic analysis. Images were collected CO2 was collected and purified with an automated acid K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 23 digestion and gas purification device as described by Passey Dale et al. (2014), Kluge et al. (2014) and John (2015). et al. (2010). This device includes passing the CO2 through Results of pre- and post- processing of standards are pre- multiple dry ice/ethanol and liquid nitrogen cryogenic traps sented in Fig. 3; long-term averages of standard D47CDES as well as through a Porapak-Q 180/20 chromatograph held values with comparisons to D47CDES values measured in at 20 °C with a helium carrier gas. The CO2 was measured three other labs are available in Table 1 (Bonifacie et al., on one of two ThermoFinnigan MAT 253 IRMS housed at 2017; Dennis et al., 2011; Henkes et al., 2013). Values are Caltech. Each measurement consisted of eight acquisitions reported as D47CDES90 with no acid fractionation correction (16 V on m/z = 44) of seven cycles of unknown sample and as two values for D47CDES25 using recent empirically CO2 versus Oztech working gas as outlined by derived acid fractionation factors of 0.082‰ following Huntington et al. (2009). The data collection spans a long (Defliese et al., 2015) and of 0.092‰ following (Henkes period of time over which best practices in the clumped iso- et al., 2013). tope community evolved to better address pressure baseline Following recent clumped isotope studies, we recognize issues, non-linearity in the source and scale compression the importance of larger error bars on samples with few (e.g. Bernasconi et al., 2013; Dennis et al., 2011). The replicate measurements. In the main text the reproducibility unknown samples were collected concurrently with 1000 ° of the measurement is presented as 2SE, twice the com- C heated gases of two different bulk isotopic compositions monly reported 1SE. We present well-replicated measure- and at least two carbonate standards of known composi- ments in the main text (n > 1, with D48 excess <1.5‰ tion. Unknown samples analyzed since 2012 were also ana- (except for the modern organic rich samples where a more lyzed concurrently with 25 °C equilibrated gases of two strict D48 excess <1‰ was required) and 2SE < 15 °C) different bulk isotopic compositions. (Table 2). Table E in the Supplemental Materials, also has single analyses (n = 1), poorly replicated samples and 2.9. Post clumped isotope measurement data processing 95% confidence intervals calculated (two tailed critical value for t-distribution with n 1 degrees of freedom) To address the complexities of the data set— multiple (e.g. Bonifacie et al., 2017; Mu¨ller et al., 2017. mass spectrometers, data processed over many months, We used a temperature calibration of D47CDES90 = samples of multiple mineralogies gathered with and without 0.0422 (±0.0019) * 106/T2 + 0.1262 (±0.0207) (Bonifacie 25 °C equilibrated gases—we input over 1200 raw measure- et al., 2017, Eq. (3)). The D47CDES90 calibration equation ment files of 1000 °C heated and 25 °C equilibrated gases is a composite of both synthetic and biogenic calcite, arag- and carbonate standard data, along with 270 raw measure- onite and dolomite minerals reacted at 70–90 °C in 8 differ- ment files of sample unknowns, into Easotope, an open ent laboratories (Bonifacie et al., 2017; Defliese et al., 2015; source software tool developed specifically for clumped iso- Fernandez et al., 2014; Henkes et al., 2013; Kele et al., 2015; tope data processing (John and Bowen, 2016)(Supplemen- Tang et al., 2014; Wacker et al., 2014)(Fig. 4). This calibra- tal Materials, E and F). The carbonate d18O composition tion includes data garnered at Caltech during the same time was calculated using separate 90 °C acid digestion fraction- interval as our analyses using the same methods outlined ation factors for calcite Kim et al. (2007a) and apatite here (Bonifacie et al., 2017). In addition to the unpublished Passey et al. (2007). Following Schauer et al. (2016) and data presented below we have also reprocessed all of the Dae¨ron et al. (2016), we used 17O correction parameters raw data from Finnegan et al. (2011) using the post- outlined in Brand et al. (2010). Carbonate d13C and d18O measurement data processing approach and statistical eval- values were drift-corrected using a 10 standard moving win- uation outlined above and the temperature calibration from dow to NBS-19 and NBS-19 calibrated internal Carrara Bonifacie et al. (2017). While we analyzed modern bra- marble standards (HCM and CIT). We corrected all of chiopod samples as part of this study, we do not use those the D47 data using a 10 standard moving window for both results to calculate temperature because of recent argu- the linearity correction and empirical transfer function into ments that the best practices for calibration studies must the ‘‘absolute reference frame” or ‘‘Carbon Dioxide Equili- include a large temperature spread, many sample replicates, brated Scale” (CDES) as defined by Dennis et al. (2011). and many samples (Bonifacie et al., 2017). In the Supple- For the linearity correction in Easotope, we used 1000 °C mental Materials, Table E we also present temperatures cal- heated gases of two different bulk isotopic compositions culated using D47CDES25, a 0.082‰ AFF (Defliese et al., and when available 25 °C equilibrated gases of two different 2015), and a calibration from Kelson et al. (2017), the only bulk isotopic compositions. To build an empirical function calibration currently calculated using 17O correction to transfer our data into the CDES reference frame in Easo- parameters following Schauer et al. (2016) and Dae¨ron tope, we first calculated D47CDES values for five in house et al. (2016), as outlined in Brand et al. (2010). standards from periods when both 1000 °C heated and 25 Fluid compositions were calculated using the measured °C equilibrated gases were analyzed (HCM (n = 44), CIT clumped isotope temperatures and an equilibrium fraction- CM (n = 38), TV01 (n = 60), TV03 (n = 32), 102-GC-AZ- ation equation for calcite-water (Kim and O’Neil, 1997). To 01 (n = 30). We then used a 10 standard moving window calculate d18O fluid composition in equilibrium with the of a combination of at least two carbonate standards carbonate in our phosphatic materials, we assumed the (HCM/CIT CM, TV01/TV03, 102-GC-AZ-01) and two equilibrium fractionation of the CO3 groups follows compositions of 1000 °C heated and 25 °C equilibrated calcite-water equilibrium fractionation (Kim and O’Neil, gases (when available) to calculate D47CDES values for the 1997). Error on these calculations uses 2SE on the D47 entire time period, an approach similar to one used in temperature. 24 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41

1000C BOC CIT 25C BOC CIT CARMEL CHALK HCM TV03 1000C EBOC CIT 25C EBOC CIT FAST HAGA NBS19 TV01 102-GC-AZ-01 2_8_E CIT A 0.5

(‰) 0.0 47 RAW

Δ -0.5

-1.0 1.00 B

0.75 (‰) 0.50

0.25 47 CDES90 Δ

0.00

0 250 500 750 1000 1250 #

Fig. 3. D47RAW and D47CDES90 values of 1250 standards and gases run at Caltech between 2009 and 2013. (A) D47RAW values over time show signiﬁcant deviations. (B) D47CDES90 values after a correction using a 10 standard moving window for both the linearity correction (two 1000 °C heated gases (BOC and EBOC) and from 2012 onwards two 25 °C (BOC and EBOC) equilibrated gases) and the empirical transfer function (using at least two carbonate standards (HCM/CIT CM, TV01/TV03, 102-GC-AZ-01), two 1000 °C heated and when available two 25 °C equilibrated gases).

2.10. Phosphate oxygen isotope measurements (Fig. 6). XRD analysis of modern linguliform shell powders indicate that the calcium phosphate has significant amounts 18 The apatite d O of four samples was analyzed by of CO3 and F substituted into the lattice and is most similar P. Quinton in the MacLeod Lab at the University of to the minerals francolite CaF(Ca, C)4[P, C(O,OH,F)4]3, Missouri. The cleaned powders of linguliform brachiopods carbonated fluorapatite Ca10(PO4)5CO3F1.5(OH)0.5 and from the Decorah and Eau Claire formations, as well as hydroxyapatite Ca5(PO4)3(OH) (Fig. 7). two modern locations (southern California and the Polillo After a range of cleaning methods and a 12-h treatment Islands, Philippines), were dissolved in dilute hydrofluoric in a 0.1 M buffered acetic and acetate solution, modern lin- acid. Ag3PO4 was precipitated from the solution using a sil- guliform brachiopods yielded clumped isotope tempera- ver amine solution into which excess Ag+ was added as tures that range from 0 to 64 °C. We attribute the large AgNO3 as described in (Bassett et al., 2007). The pH was range in temperatures to organic interferences during mass then increased using NH4OH buffered with NH4NO3 and spectrometry. The most aggressively cleaned materials with samples were left uncovered at 50 °C. As the pH D48 excess <1.5‰ yield temperatures that generally com- increased, Ag3PO4 precipitated out of solution. The silver pare well to the range of observed sea surface temperatures phosphate crystals were rinsed, dried and stored in a over the last 5–10 years from nearby ocean buoys and fairly desiccator before being weighed into silver capsules. The well to seawater d18O values from GISS monitoring sites capsules were then combusted in a high temperature (Fig. 5)(Schmidt, 1999). We note that this comparison is (1400 °C) TC/EA before being analyzed as CO on an imperfect because some modern linguliform brachiopods IRMS. The isotopic measurements are reported relative often live in tidal flats that can experience significant tem- to VSMOW and analytical precision is ±0.3‰. perature extremes compared to nearby open seawater as evident in the 24 °C range in shallow sediment temperatures 3. RESULTS in Ariake Bay, Japan (Moqsud et al., 2006). The tidal flats may also experience waters that have variably brackish or 3.1. Modern brachiopods evaporative seawater compositions.

Electron microscopy of modern phosphatic shells indi- 3.2. Ordovician samples cates a wide range of phosphatic microstructure morpholo- gies from more massive structure in the mineralized In general, the calcitic brachiopods analyzed display primary layers to elaborate cross-hatching and vertical exceptional preservation of shell ultrastructure. The lami- growth structures in the organic-rich, more chitinous layers nae (Fig. 8) appear very similar to the laminae of a modern Table 1 Isotopic composition of analyzed CIT standards (‰). 13 18 a b c d d,a d,b Standard N d CVPDB SD d OVPDB SD D47CDES90 D47CDES25 D47CDES25 SD 2SE 95% CL D47CIT90 D47CIT25 D47CIT25 HCM 243 2.28 0.06 1.89 0.44 0.301 0.393 0.383 0.020 0.003 0.002 0.263 0.345 0.355 ..Brmn ta./Gohmc tCsohmc ca24(08 84 25 18–41 (2018) 224 Acta Cosmochimica et Geochimica / al. et Bergmann K.D. CIT 40 2.02 0.04 1.70 0.04 0.323 0.415 0.405 0.013 0.004 0.004 0.277 0.359 0.369 NBS19 7 1.89 0.08 2.30 0.10 0.302 0.394 0.384 0.013 0.010 0.012 0.266 0.348 0.358 IPGPe 11 0.314 0.406 0.396 J Hopkinsf 23 0.322 0.414 0.404 J Hopkinsg 12 0.318 0.410 0.400 0.010 Harvardg 7 0.292 0.384 0.374 0.014 TV01 192 2.44 0.09 8.47 0.21 0.627 0.719 0.709 0.019 0.003 0.003 0.572 0.654 0.664 TV03 34 3.29 0.15 8.28 0.13 0.632 0.724 0.714 0.016 0.006 0.006 0.567 0.649 0.659 102-GC-AZ-01 90 0.46 0.08 14.44 0.16 0.615 0.707 0.697 0.017 0.004 0.004 0.578 0.660 0.670 J Hopkinsf 102 0.618 0.710 0.700 IPGPe 20 0.625 0.717 0.710 J Hopkinsg 17 0.620 0.712 0.702 0.006 Harvardg 4 0.625 0.717 0.707 0.010 Caltechg 9 0.643 0.735 0.725 0.014 2_8_E 22 9.19 0.29 6.32 0.11 0.613 0.705 0.695 0.019 0.008 0.009 0.575 0.657 0.667 CARMEL CHALK 42 2.16 0.07 4.11 0.07 0.596 0.688 0.678 0.018 0.006 0.006 0.551 0.633 0.643 FAST HAGA 21 3.06 0.08 5.76 0.08 0.607 0.699 0.689 0.024 0.010 0.011 0.570 0.652 0.662 1000C BOC CIT 221 10.71 0.34 1.27 0.86 0.028 0.120 0.028 0.021 0.003 0.003 0.001 0.081 0.091 1000C EBOC CIT 219 10.75 0.27 24.98 2.95 0.028 0.120 0.028 0.023 0.003 0.003 0.001 0.083 0.093 25C BOC CIT 39 10.60 0.53 0.51 1.32 0.931 1.023 0.931 0.018 0.006 0.006 0.857 0.939 0.949 25C EBOC CIT 34 10.63 0.43 25.18 5.67 0.926 1.018 0.926 0.016 0.006 0.006 0.852 0.934 0.944 a Calculated using a 0.092 Acid Fractionation Factor following Henkes et al. (2013). b Calculated using a 0.082 Acid Fractionation Factor following Defliese et al., 2015. c 95% confidence intervals calculated with a two tailed critical value for t-distribution with n 1 degrees of freedom d D47CIT uses a 10 standard moving window, the HG intercept and slope and stretching factor outlined in Huntington et al., 2009 e Comparison data in italics from Bonifacie et al. (2017). f Comparison data in italics from Henkes et al. (2013). g Comparison data in italics from Defliese et al., 2015. Table 2 18–41 (2018) 224 Acta Cosmochimica et Geochimica / al. et Bergmann K.D. 26 Isotopic composition of analyzed samples (‰). 13 18 a b c 18 Sample Name N Min Age Age Bin d CVPDB SD d OVPDB SD D47CDES90 D47CDES25 D47CDES25 SD 2SE T (°C) 2SE Water d OVSMOW 2SE 916_MF_RC2 2 Ca 440 Rhud. 1.46 0.59 4.40 0.25 0.546 0.628 0.638 0.001 0.002 43.9 0.8 43.9 0.8 901_19_RC_1 2 Ca 444.5 Hirn. 0.88 0.08 3.45 0.44 0.594 0.676 0.686 0.023 0.033 27.4 11.2 0.6 2.3 901_10_RC_1 2 Ca 445 Hirn. 4.11 0.08 0.95 0.12 0.589 0.671 0.681 0.025 0.035 29.0 12.1 2.3 2.5 901_10_RC_2 2 Ca 445 Hirn. 4.84 0.00 1.51 0.01 0.586 0.668 0.678 0.019 0.027 29.0 12.1 2.3 2.5 901_12_RC_1 2 Ca 444.5 Hirn. 4.46 0.12 0.24 0.35 0.580 0.662 0.672 0.000 0.000 31.8 3.7 3.5 0.7 901_12_RC_2 2 Ca 444.5 Hirn. 4.13 0.06 0.95 0.18 0.575 0.657 0.667 0.013 0.019 33.7 6.7 3.2 1.3 901_19_RC_2 2 Ca 444.5 Hirn. 1.07 0.01 3.35 0.04 0.557 0.639 0.649 0.016 0.023 40.0 8.7 1.9 1.7 912_98_2_RC 2 Ca 443.5 Hirn. 2.06 0.01 3.65 0.21 0.535 0.617 0.627 0.006 0.009 48.3 3.6 3.1 0.6 Lbec_F_B 2 Ca 443.5 Hirn. 1.71 0.03 3.87 0.09 0.534 0.616 0.626 0.008 0.012 47.4 5.3 2.8 1.0 908_1_T 3 Ca 445.7 Katian 1.40 2.32 2.57 1.45 0.579 0.661 0.671 0.014 0.016 32.0 5.4 1.2 1.1 MOS_T 2 Ca 447.5 Katian 0.89 0.01 4.80 0.01 0.572 0.654 0.664 0.010 0.014 34.5 4.9 0.5 1.0 MOS_B 2 Ca 447.5 Katian 0.00 0.01 5.02 0.02 0.566 0.648 0.658 0.025 0.036 36.6 13.5 0.4 2.7 WWR1 2 Ca 446.8 Katian 1.05 0.00 4.43 0.01 0.562 0.644 0.654 0.011 0.015 38.2 5.5 0.5 1.0 918_43_2_B2 2 Ca 447 Katian 0.30 0.02 3.77 0.06 0.561 0.643 0.653 0.004 0.006 38.4 2.2 1.2 0.4 DB_SE_S1 4 Ca 451.8 Katian 0.51 0.01 5.40 0.03 0.560 0.642 0.652 0.010 0.010 38.1 5.1 0.5 1.0 ElkR1 3 Ca 446.5 Katian 1.31 0.02 4.36 0.01 0.554 0.636 0.646 0.018 0.021 40.9 8.0 1.1 1.5 904_4_5_RC 3 Ca 446 Katian 2.48 0.00 3.00 0.03 0.554 0.636 0.646 0.007 0.008 41.0 3.1 2.5 0.6 DB_SE_M 3 Ca 451.8 Katian 1.05 0.04 4.99 0.07 0.550 0.632 0.642 0.014 0.017 40.8 4.7 0.4 0.9 DIB 6 Ap 451.8 Katian 1.60 0.07 4.43 0.18 0.543 0.625 0.635 0.007 0.006 44.6 3.0 1.7 0.5 DPL_B 2 Ca 451.8 Katian 0.39 0.01 5.43 0.05 0.542 0.624 0.634 0.011 0.016 45.4 6.3 0.8 1.1 906_1_T 2 Ca 447 Katian 0.24 0.01 4.15 0.11 0.540 0.622 0.632 0.014 0.020 46.2 8.0 2.3 1.5 DB_SE_S6 3 Ca 451.8 Katian 0.08 0.02 5.46 0.09 0.537 0.619 0.629 0.017 0.020 45.4 6.4 0.8 1.2 912_8_2_T 3 Ca 445.7 Katian 0.10 0.06 3.46 0.12 0.534 0.616 0.626 0.012 0.014 48.4 5.6 3.3 1.0 918_43_2_1_B 3 Ca 447 Katian 0.47 0.04 3.52 0.02 0.534 0.616 0.626 0.016 0.018 48.5 7.5 3.3 1.3 DB_SE_S2 3 Ca 451.8 Katian 0.59 0.03 5.80 0.06 0.534 0.616 0.626 0.014 0.016 47.4 4.8 0.8 0.9 DB_SE_R1 2 Ca 451.8 Katian 0.49 0.05 5.46 0.06 0.533 0.615 0.625 0.002 0.003 48.3 0.4 1.3 0.1 DB_SE_S5 3 Ca 451.8 Katian 0.30 0.03 5.80 0.10 0.529 0.611 0.621 0.017 0.019 49.3 5.4 1.1 1.0 DB_SE_B_S1_1 3 Ca 451.8 Katian 0.42 0.01 5.82 0.06 0.525 0.607 0.617 0.016 0.018 52.3 7.8 1.6 1.4 DB_PN 2 Ap 453 Sand. 5.77 0.04 2.77 1.44 0.596 0.678 0.688 0.018 0.026 26.6 8.7 0.1 1.8 Dec_B_1 2 Ca 453 Sand. 0.43 0.01 4.91 0.18 0.550 0.632 0.642 0.001 0.001 42.6 0.4 0.8 0.1 DCQ_m0_15 2 Ca 453 Sand. 0.63 0.02 4.50 0.50 0.539 0.621 0.631 0.023 0.026 46.7 10.6 2.0 1.9 Dec_T_2 2 Ca 453 Sand. 0.53 0.01 5.16 0.21 0.533 0.615 0.625 0.004 0.006 48.9 2.4 1.7 0.4 Kunda_M 4 Ca 467 Darw. 0.42 0.02 4.99 0.06 0.578 0.660 0.670 0.015 0.015 33.2 6.5 1.0 1.3 Asery_M 4 Ca 465 Darw. 0.45 0.04 5.97 0.05 0.573 0.655 0.665 0.016 0.016 34.9 6.8 1.7 1.3 VkB4 5 Ca 468 Darw. 0.21 0.02 5.38 0.04 0.572 0.654 0.664 0.015 0.014 34.6 4.8 1.1 0.9 VkB14 2 Ca 468 Darw. 0.68 0.00 5.36 0.03 0.566 0.648 0.658 0.010 0.014 36.6 5.1 0.7 1.0 Kunda_T 5 Ca 467 Darw. 0.08 0.02 6.01 0.04 0.564 0.646 0.656 0.019 0.017 37.5 6.2 1.2 1.2 Asery_T1 5 Ca 465 Darw. 0.58 0.01 5.63 0.04 0.563 0.645 0.655 0.017 0.015 37.6 5.4 0.8 1.0 VkB13 3 Ca 468 Darw. 0.17 0.05 5.34 0.08 0.555 0.637 0.647 0.003 0.003 39.3 1.5 0.2 0.3 VkB3 3 Ca 468 Darw. 0.14 0.00 5.70 0.03 0.554 0.636 0.646 0.011 0.013 40.9 4.9 0.2 0.9 VkB2 2 Ca 468 Dap. 0.38 0.01 5.13 0.06 0.568 0.650 0.660 0.023 0.032 35.9 11.8 0.6 2.3 (continued on next page) K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 27

calcitic brachiopod Terebratalia transversa, recovered from

2SE the San Pedro Jetty, CA (Fig. 8A and B). In some of the calcite brachiopod shells examined on the SEM, the ultrastructure has a recrystallized appearance although

VSMOW Rasmussen et al. (2016) attribute similar features to dust O

18 created during mounting and not recrystallization (Fig. 8). d Electron microprobe elemental maps and spot analyses 0.2 1.0 0.2 1.6 1.4 0.2 2.5 1.2 of the diﬀerent Ordovician fossil materials highlight a range of post-depositional diagenetic modiﬁcations of shell material. Transects of spot analyses across a calcitic bra-

2SE Water chiopod shell from the Decorah Formation indicate local-

c ized zones of Fe enrichment. The punctae are in places C)

° ﬁlled with pyrite, and a zone of dissolution and recrystallization of the shell calcite spreads outwards from the punctae. In the region of alteration Fe concentrations can be as high as 5500 ppm. Outside of the zone of alteration, Sr concentrations are high, while Mn and Fe concentrations are low and more similar to younger brachiopod materials (Fig. 9B, F, J, N, R, RSQ_2_1_B). SD 2SE T ( A brachiopod (Paurorthis parva) from the Volkhov For- b mation is enriched in Mn in speciﬁc horizons within the skeleton that follow the orientation of original growth

47CDES25 laminae in the shell ultrastructure. Mn enrichments can D be as high as 8000 ppm (Fig. 9A, E, I, M, Q, VkB2).

a An asaphid trilobite (Asaphus cornutus) from the Aseri Stage shows two features suggesting that the original skeletal calcite has been diagenetically altered. The ﬁrst 47CDES25

D is even, high enrichment in Fe across the entire thickness of the cuticle to values over 3000 ppm. The second is large dolomite rhombs that are ingrown at the sediment-shell interface (Fig. 9C, G, K, O, S, Aseri_T1). 47CDES90

D The linguliform brachiopod Lingulasma galenaense from the Decorah Formation show microstructure similar to . .

, Eq. (3)). modern linguliform brachiopods, including massive miner- SD alized layers and ﬁne-scale cross-hatched phosphatic precipitates in the chitinous layers (Fig. 6). The globular

VPDB nature of the skeleton ultrastructure in both modern and O 5.24 0.09 0.560 0.642 0.652 0.013 0.014 38.6 5.3 5.28 0.00 0.560 0.642 0.652 0.016 0.022 38.7 8.2 5.31 0.01 0.556 0.638 0.648 0.004 0.006 40.2 2.2 0.0 0.4 5.39 0.03 0.544 0.626 0.636 0.012 0.014 44.8 5.5 0.8 1.0 5.86 0.08 0.569 0.651 0.661 0.003 0.003 35.7 1.2 6.21 0.03 0.580 0.662 0.672 0.020 0.018 31.7 6.1 18 fossil linguliform brachiopods makes it more challenging d to identify recrystallization than in their calcitic relatives.

Henkes et al. (2013) Deﬂiese et al. (2015) XRD analysis of the Lingulasma galenaense samples from SD Bonifacie et al., 2017 the Decorah Formation yield spectra similar to the modern linguliform brachiopod spectra and are most similar to car-

VPDB bonated hydroxyapatite and francolite (Fig. 7). C 0.09 0.01 0.71 0.03 0.14 0.01 0.34 0.01 5.23 0.43

13 Recalculated values of Late Ordovician samples from d the Cincinnati Series and Anticosti Island preserve the previously described trends (Finnegan et al., 2011). The Hir-

+ 0.1262 (±0.0207) ( nantian Isotope Excursion (HICE) has a minimum 2

T temperature of 29.0 ± 12 °C (2SE, 901_10_RC_2, n =2) / 6 as well as two other similar temperatures of 31 ± 4 °C (901_12_RC_1, n = 2) and 33.7 ± 6.7 °C (901_12_RC_2, n = 2). The d18O ﬂuids of these three samples are 2.3 ± 2.5‰, 3.5 ± 0.7‰, and 3.2 ±1.3‰ VSMOW respectively. The minimum Katian temperature is 32 ± 5.4 °C (908_1_T, n = 3) with another nearby at 34.5 ± 4.9 °C

) (MOS_T, n =2)(Fig. 10). All Katian brachiopod tempera- °

= 0.0422 (±0.0019) * 10 tures are warmer with a cluster around 37 C (918_43_2_B2, n = 2, WWR1, n = 2, DB_SE_S1, n = 4). continued The d18O ﬂuid compositions of the two lowest temperatures

47CDES90 from the Katian are 1.2 ± 1‰, and 0.5 ± 1‰ VSMOW Calculated using a 0.092Calculated Acid using Fractionation a Factor 0.082 following D Acid Fractionation Factor following c a b respectively (Table 2). Table 2 ( Sample Name N Min Age Age Bin VkB18VkB12 3 2 Ca Ca 468 468 Dap. Dap. 0.21 0.01 VkB8 2 Ca 468 Dap. VkB16 3 Ca 468 Dap. VkB5 3 Ca 472 Floian ECIB 5 Ap 499 Marj. 28 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41

0.7 Biogenic calc. & arag. (Henkes 13) Dolomite (Bonifacie 16) Inorg. aragonite (Defliese 15) Inorg. calcite (Defliese 15) 0.6 Inorg. siderite (Fernandez 14) Travertine aragonite & calcite (Kele 15) Inorg. calcite (Tang 14) Biogenic calcite (Wacker 14) Biogenic apatite (Wacker 16) 0.5 Biogenic apatite (this study) (‰) Biogenic calcite (this study) 47 CDES90

Δ 0.4

0.3

0.2

4681012 106/T2 (K)

Fig. 4. D47CDES90 values from modern calcitic and linguliform brachiopods (D48 excess <1.5) compared to recently published D47CDES90 values from calcite, dolomite and aragonite calibration studies (Bonifacie et al., 2017) and references therein) and a recently published dataset of

D47CDES90 values of modern apatite enamel measurements (Wacker et al., 2016).

Southern Southern ABPhillipines Japan Phillipines Japan California California Hawaii Hawaii Florida Florida Dp. Sea Dp. Sea

Annual T range and mean δ 18 60 Local Oseawater (VSMOW)

) 50 C ° (‰) ( 40

30 VSMOW 5 O 18

20 δ Temperature 10 water 0 0

51015 51015 Site Site

Fig. 5. All modern linguliform and calcitic brachiopod clumped isotope calibration results. (A) clumped isotope temperatures calculated using D47CDES90 values compared to min, max and mean temperature from the closest buoy in NODC except for Japan which has in-situ tidal ﬂat temperature data (±2SE) (Moqsud et al., 2006). Multiple cleaning methods were tried as poorly cleaned samples produced depleted D47 values relative to expected and had D48 excesses >1 (smaller transparent symbols). Calcitic brachiopods are plotted as circles. (B) calculated 18 water d OVSMOW compositions compared to the closest GISS measurements (±2SE) (Schmidt, 1999; Kim et al., 2007b).

The lowest recorded temperature from Decorah Forma- a mean temperature of 44.6 ± 3.0 °C (DIB, n = 6). The tion is 26.6 ± 8.7 °C from a phosphatic nodule found calculated mean d18O fluid, assuming an equilibrium rela- less than a meter below the onset of the GICE (Figs. 10 tionship that follows calcite, is 1.7 ± 0.5‰ VSMOW (Kim and 11). The calculated d18O fluid composition is and O’Neil, 1997)(Fig. 11, Table 2). 0.1 ± 1.8‰ VSMOW (DB_PN, n = 2). The lowest tem- The brachiopod clumped isotope temperatures from the perature calcitic brachiopod from above the GICE is Floian-Darriwilian Volkhov Formation, trilobites and 38.1 ± 5.1 °Cwithad18O fluid composition of 0.5 ± 1.0‰ matrix from the Darriwilian Aseri and Kunda formations (DB_SE_S1, n = 4). Results from clumped isotope analyses are similar with a minimum temperature of 33.2 ± 6.5 °C of the CO3 groups substituted into the phosphate lattice of (Kunda_M, n = 4) with eight other samples between 33 linguliform brachiopods from the Decorah Formation yield and 37 °C (VkB4, n = 5, VkB5, n = 3, VkB12, n =2, K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 29

Fig. 6. SEM images of modern and ancient linguliform brachiopods. (A & C) shell structure of the secondary layer of a modern linguliform brachiopod, Glottidia albida from the San Pedro shelf oﬀ Newport Bay, CA. (E) shell ultrastructure of a modern linguliform brachiopod, Lingula reevii from the Waikiki Aquarium. (B) ultrastructure of the secondary layers in Lingulepis sp. recovered from the Crepicephalus zone of the Eau Claire Formation near Strum, WI. Scale bar is 20 mm. (D) ultrastructure of the secondary layers in Lingulepis sp. recovered from the Crepicephalus zone of the Eau Claire Formation near Colfax, WI. Scale bar is 40 mm. (F) shell ultrastructure of Lingulasma galenaense from the Decorah Formation collected from the Decorah Bruening Quarry, IA.

A 4

8 B 6 4 2 Intensity 6 C 4 quartz sand

0 30 40 50 60 2θ (°)

Fig. 7. XRD data for modern and ancient linguliform brachiopod samples. (A) Modern cleaned and powdered samples of Glottidia albida from the San Pedro shelf oﬀ Newport Bay, CA (green) and Lingula unguis from Polillo Islands, Philippines (blue). (B) two separate cleaned and powdered samples of Lingulasma galenaense from the Decorah Formation (purple and pink). (C) cleaned and powdered samples of Lingulepis sp. from the Crepicephalus zone of the Eau Claire Formation near Strum, WI (light blue and dark blue) and Colfax, WI (red). (For interpretation of the references to color in this ﬁgure legend, the reader is referred to the web version of this article.) 30 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41

A B mineralogy (francolite, hydroxyapatite, and carbonate hydroxyapatite). The spectra also indicate that while not all of the quartz sand was removed from each powder prior to analysis, no calcite or dolomite remained in the analyzed samples (Fig. 7). The Eau Claire linguliform brachiopod sample yields a temperature of 31.7 ± 6.1 °C (ECIB, n = 5). The average calculated d18O water, assuming an equilibrium relation- C D ship that follows calcite, is 2.5 ± 1.2‰ VSMOW (Kim and O’Neil, 1997)(Table 2).

4. DISCUSSION

The geological context and materials examined here provide several insights into the nature of clumped isotope E F records in deep time. Below we discuss three case studies that illustrate the behavior of clumped isotopes in diﬀerent materials and from diﬀerent paleoenvironments that inform interpretation. We discuss how these new data constrain interpretations of Cambro-Ordovician climate, seawater d18O values, and the trend in d18O values of minerals through early Paleozoic time. G H 4.1. Case study: Organic-rich modern linguliform brachiopods can yield anomalously high clumped isotope temperatures

The modern linguliform brachiopod calibration data highlight the challenge with analyzing organic-rich materials. Part of the pre-processing clean up method for making carbonate clumped isotope measurements involves running Fig. 8. SEM images of modern and ancient articulate brachiopods. the CO2 sample through a poropak trap held at 20 °C. (A and B) shell ultrastructure of Terebratalia transversa from the The goal of this treatment is to remove organic materials San Pedro Jetty, CA. Note the abundant punctae. (C–H) calcitic that may create isobaric interferences with different iso- brachiopods from the Decorah Formation in varying states of topologues of CO2. Different laboratories have slight varia- ultrastructure preservation and recrystallization (C = tions on this cleaning step that range from a freeze down DB_SE_S5_1, D = DB_SE_S2_1, E = DB_SE_S6_1, F = finger filled with poropak to a diffusive flow though poro- DB_SE_S3_1, G = DB_SE_S4_1, H = DB_SE_S1_1). They are pak packed 1/400 tube, to a He flow through poropak organized from highest (C) to lowest (H) clumped isotope 00 packed 1/8 stainless steel column. CO2 is moved through temperature. In some of the calcite brachiopod shells examined 00 on the SEM, the ultrastructure has a recrystallized appearance the 1/8 column in a He carrier gas stream. While the lab (shown with arrows) although Rasmussen et al. (2016) attribute at Caltech uses one of the more stringent organic cleaning similar features to dust created during mounting and not protocols currently applied, the high amount of organic recrystallization. phases present in modern linguliform brachiopod skeletons (approaching 40% by weight) was not effectively removed in VkB14, n = 2 VkB18, n = 3, Aseri_T1, n = 5, Aseri_M, modern linguliform brachiopod samples that had no chem- n = 4 Kunda_T1, n = 5). These temperatures are slightly ical pre-treatment (e.g. PIB_none). Clumped isotope analy- 13 18 lower than the temperatures from the Decorah Formation ses are depleted in d C, d O, and D47 and elevated in D48, and other Late Ordovician sections but it has a higher pale- suggesting hydrolysable carboxylic groups are sourced dur- olatitude (30°S) (excepting the samples from the HICE ing digestion and contributed isobaric interferences to the and GICE). The calculated d18O fluid composition for the measurements. The apparent clumped isotope temperature lowest temperature sample is 1.0 ± 1.3‰ (Kunda_M, of modern samples contaminated with organic molecules is n =4)(Table 2). higher than growth temperatures by as much as 40 °C. Interferences on Mass 45 and 48 appear to be the most sig- 13 3.3. Cambrian samples nificant with clear offsets in both D48 and d C between uncleaned and cleaned samples. Additionally, the organic The Lingulepis specimens have very similar microstruc- material is sometimes ‘sticky’ and can remain in the mass tures to both the Decorah and modern linguliform bra- spectrometer system resulting in high standard residuals. chiopod species. The more porous organic-rich layers The organic fragments are visible on a full mass scan as have cross-hatching and vertical phosphate precipitates low-resolution peaks. (Fig. 6B and D). The XRD spectra are similar to the Attempts to react away the 40% organic material modern linguliform brachiopod spectra and have the same with a range of chemical pre-treatment methods met with K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 31

Fig. 9. Electron microprobe maps and spot analyses of ancient brachiopods. (A) Calcitic brachiopod from the Volkhov Formation with overlain elemental maps generated on an electron microprobe of Ca (E), Fe (I), & Mn (M) and quantitative spot measurements (Q) along a transect shown in (A) (VkB8). Scale bar is 110 mm. (B) Calcitic brachiopod from the Decorah Formation with overlain elemental maps of Ca (F), Fe (J), & Mn (N) and quantitative spot measurements (R) along a transect shown in (B) (RSQ_8_B). Scale bar is 60 mm. (C) Calcitic trilobite from the Aseri Formation, Russia with overlain elemental maps of Ca (G), Fe (K), & Mn (O) and quantitative spot measurements (S) along a transect shown in (C) (Aseri_T1). Note the ingrown dolomite rhombs along the base of the shell. Scale bar is 110 mm. (D) Phosphatic brachiopod from the Eau Claire Formation with overlain elemental maps of P (H), F (L), Na (P), & Ce (T) (ECIB). Scale bar is 200 mm.

variable success. Variability in sample shell thicknesses 4.2. Case study: Samples with elevated Fe and Mn sometimes and total organic material affected the success of the yield low clumped isotope temperatures cleaning treatment. More aggressive cleaning methods yield results consistent with or slightly lower than nearby Trace metal abundance has long been used to assess dia- buoy temperatures (Fig. 5). Ineffective cleaning treat- genesis and screen samples before isotopic analysis. During ments were identified with D48 excess >1 in these materi- burial, calcitic brachiopod shells may recrystallize in the als. Symbols in Fig. 5 with a D48 excess >1 are presence of high-temperature, anoxic metal-rich fluids and transparent, and are not subsequently plotted in Fig. 4. become enriched in Fe and Mn (Brand and Veizer, 1980). The temperature results of the modern phosphatic sam- Carbonate clumped isotope thermometry provides an inde- ples without D48 excesses and the two measured calcitic pendent measure of diagenesis for carbonate or apatite brachiopods are in good agreement with recently pub- materials, yielding deeper insight into alteration processes lished calibration datasets (Fig. 4)(Bonifacie et al., associated with a range of environments. The combined 2017 and references therein). application of microscopy, trace metal analysis and 32 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41

60 Brachiopod Matrix A Rugose Coral Trilobite 50 Linguliform Brachiopod

) Phosphate Nodule C ° ( 40

30 Temperature 20 36 ± 12 3 38 ± 5 4 48 ± 4 2 44 ± 1 2 32 ± 6 5 36 ± 1 3 35 ± 5 5 27 ± 9 2 32 ± 5 3 27 ± 11 2

5 B (‰)

0 VSMOW O 18 δ water

-5

500 480 460 440 Age (Ma)

18 Fig. 10. Compilation of clumped isotope temperature and calculated water d OVSMOW for Cambrian, Ordovician and earliest Silurian time. (A) Clumped isotope temperatures calculated using D47CDES90 values with text detailing the sample with the minimum temperature from each 18 time bin (±2SE, # of replicates in italics). (B) Calculated water d OVSMOW (±2SE). Time bins from this study that span three signiﬁcant positive carbon isotope excursions (the SPICE, GICE and HICE) are highlighted in blue. (For interpretation of the references to color in this ﬁgure legend, the reader is referred to the web version of this article.)

clumped isotopes for diagenetic screening within a more expected diagenetic screening metrics (e.g. Brand and nuanced assessment of sample context is required. Veizer, 1980; Grossman, 2012a; Brand et al., 2012). The clumped isotope data show that trace metal concen- No such pattern is apparent for the older Decorah, Vol- trations in calcitic taxa can become altered across a range of khov, Aseri, and Kunda formations despite similar and temperatures. This effect is more evident in Middle and even lower clumped isotope temperatures (Fig. 12). The cal- Late Ordovician materials from the Decorah, Volkhov, citic taxa from these formations show metal enrichments Aseri and Kunda materials than in previous analyses of significantly outside of documented patterns of metal incor- end-Ordovician calcitic taxa from Anticosti Island, Quebec poration into modern biogenic materials (Veizer et al., and the Cincinnati Series of Kentucky and Ohio (Fig. 12) 1999). This high Fe and Mn incorporation appears both (Came et al., 2007; Finnegan et al., 2011). The end- along individual horizons, as seen with Mn enrichment in Ordovician taxa, including brachiopods, rugose corals the Aseri trilobite sample, and in pyrite-filled punctae, as and trilobites, are enriched in Fe and Mn only when they seen in the Fe enriched Decorah samples (Fig. 9). Despite also are visibly recrystallized and have elevated clumped elevated Fe and Mn concentrations, the lowest clumped iso- isotope temperatures, and even then their trace metal tope results in these Mid- and Late Ordovician samples enrichments are not extreme. Petrography and elemental resemble those seen directly before the Hirnantian, and analysis, combined with metal and isotope data from Late the maximum temperature measured in the Volkhov, Aseri Ordovician and early Silurian samples, suggest that and Kunda materials, 44.8 ± 5.5 °C (VkB16, n = 3), is sig- increases in clumped isotope temperature accompany trace nificantly lower than maximum temperatures previously metal incorporation into shell materials during diagenesis measured in Late Ordovician and Silurian sample locales and that clumped isotope variability at a single location (Fig. 12)(Came et al., 2007; Cummins et al., 2014; directly relates to degree of recrystallization (Fig. 12) Finnegan et al., 2011). (Came et al., 2007; Cummins et al., 2014; Finnegan et al., Depositional context helps explain these trace metal and 2011). In this respect, the younger suite of samples follows clumped isotope patterns. The Decorah Formation in the K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 33

the largest transgressions of the Phanerozoic era 65 A A (Emerson, 2002); evidence for these low sedimentation rates 60 includes phosphatized grains in the lower Decorah Forma- 55 tion, iron ooids which appear in the more northerly sections C in Minnesota, and abundant hardgrounds (Emerson, 2002;

) 50

C D Simo et al., 2003)(Fig. 2). ° ( 45 E Fossils from the Baltic-Ladoga Klint, a cratonic escarp- 40 ment on the European platform exposed in Estonia, Russia, H and Sweden, are found within limestone wackestone and 35 packstone beds as well as in the unlithiﬁed clay-rich inter- 30 Temperature layers (Egerquist and Egerquist, 2003; Tolmacheva et al., 25 2003). Very low sedimentation rates also characterize these Brachiopod Matrix deposits. The Volkhov Formation, carbonate-rich, con- 20 Rugose Coral Trilobite formably overlies predominantly quartz sand and clay 15 Linguliform Brachiopod Phosphate Nodule lithologies of Cambrian and Early Ordovician age and 10 was deposited on a broad epicontinental shelf (Dronov B et al., 1996, 2005). Hardgrounds in the Volkhov are abundant, the sediments predominantly glauconiferous, and the 2 ooids ferruginous (Dronov et al., 1996, 2005; Popov et al., 2005). The Kunda and Aseri stages of the Baltic-Ladoga Klint (12 meters total thickness) also contain iron ooids C (‰) E and glauconite grains as well as abundant, well-preserved D trilobite and brachiopod fauna (Dronov et al., 1996,

VSMOW 2005; Rasmussen et al., 2016). O 0

18 The stratigraphies sampled for this study contain abun- δ H dant sedimentological evidence—including hardgrounds, phosphatized grains and iron ooids—for early seaﬂoor water

Anticosti Island 0.992 -2 Cincinnati Series 0.988 Decorah Formation Volkov Formation 0.984 Aseri & Kunda formations 0.980 B M RC T IB PN 60 Taxa ) C

Fig. 11. Comparison of calcitic and phosphatic taxa from the ° ( Decorah Formation. (A) Clumped isotope temperatures calculated using D47CDES90 values for calcitic brachiopods (blue diamonds), matrix (blue circles), rugose corals (pink triangles), trilobites (green 40 reverse triangles), and CO2 generated from carbonate groups

substituted into linguliform brachiopods (orange squares), and Temperature phospatic nodules (brown diamond) (± 2SE). The labeled calcitic brachiopods (C, D, E and H) reference SEM images in Fig. 8. The coldest sample, a phosphatic nodule is from directly below the onset of the GICE in the uppermost Spects Ferry Member. (B) 18 calculated water d OVSMOW compositions with the dashed line at 20 ‰ predicted modern ice-free conditions ( 1.2 )(Lear, 2000) -5 -2 0 (± 2SE). (For interpretation of the references to color in this 18 mineral δ O (‰) figure legend, the reader is referred to the web version of this VPDB article.) Fig. 12. Cross plot of clumped isotope temperature and mineral 18 d OVPDB. Clumped isotope temperatures calculated using D47CDES90 values (±2SE). Grey lines are constant water 18 Upper Mississippi Valley region is a mixed carbonate- d OVSMOW. The PCA1 of trace metals abundances of log Fe siliciclastic unit deposited in the Hollandale Embayment, (0.703), log Mn (0.680) and log Sr (0.207) is plotted as a colormap a broad shallow depression across Minnesota and Iowa on calcitic taxa from units spanning Ordovician (Volkov Forma- (Emerson, 2002; Simo et al., 2003). Lithologies in the Dec- tion (reverse triangle), Aseri and Kunda Formations (square), orah vary from poorly lithified mixed carbonate muds and Decorah Formation (triangle), Cincinnati Series (diamond), and d18 illite clays to fossiliferous amalgamated packstone beds Anticosti Island (circle). The more OVPDB depleted Volkov, Aseri, Kunda and Decorah formations have similar to lower reworked by storms and riddled with well-developed hard- clumped isotope temperatures than the Cincinnati and Anticosti grounds and Chondrites burrows (Emerson, 2002). The samples but more enriched trace metals. Evidence suggests this is Decorah Formation is characterized by notably low aver- due to low oxygen conditions at or near the seafloor. (For age sedimentation rates, due to its position at the geody- interpretation of the references to color in this figure legend, the namic center of the North American craton during one of reader is referred to the web version of this article.) 34 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 lithification in an environment with low sedimentation rates different interpretation. The Upper Cambrian-aged Eau and enhanced metal cycling (Fig. 2). Early lithification Claire Formation of Wisconsin and Minnesota was one likely led to syndepositional mineralization of fossil materi- of the first Cambrian sandstones deposited on the craton als, which elevated trace metal concentrations. Despite at these localities (Runkel et al., 1998, 2007). The Eau trace metal enrichments, however, the entire suite of Claire Formation is composed of well-rounded, medium- clumped isotope temperatures (fossils and cemented car- grained quartzose sandstones intercalated with fine felds- bonate mud) remains low, a result consistent with low bur- pathic sandstones and reflects nearshore paleoenvironments ial estimates from organic maturity data (Rohrssen, 2013). (Cowan et al., 2005; Runkel et al., 2012). The nearshore sili- Each of these examples underwent scant subsequent sedi- ciclastics of the Eau Claire transition to carbonates down mentary burial diagenesis due to the geodynamic settings stratigraphic dip in Iowa (Runkel et al., 2012, 2007). of these continental basins, for a substantial transgression Interpretation of sedimentation patterns indicates a during this time interval generated little accommodation well-developed ’moat’ at this time across Minnesota and space for sediment accumulation (Peters and Gaines, Wisconsin that allowed incoming freshwater to mix with 2012). This finding reinforces the need to couple sedimen- marine water in an estuary-like setting (Cowan et al., tology, petrography and elemental analyses, particularly 2005; Myrow et al., 2012; Runkel et al., 2012, 2007). Previ- on the micro-scale, with carbonate clumped isotope ther- ous geochemical study of the Eau Claire Formation also mometry in order to fully characterize diagenetic alteration, supports a marginal marine environment setting (Cowan for even materials with high Fe or Mn abundances may be et al., 2005). Trilobite fauna from the ’inner detrital belt’ well preserved if they formed in low O2 conditions. This is is also restricted and distinct from more offshore environ- particularly true for older carbonate successions when ments (Ludvigsen and Westrop, 1983; Westrop, 1996). lower benthic oxygen concentrations (and higher degrees The Eau Claire Formation preserves the Steptoean Pos- of metal cycling) may have been more frequent (e.g. itive Carbon Isotope Excursion (SPICE) in carbonate Creveling et al., 2014). lithologies in Iowa (Elrick et al., 2011; Saltzman et al., 13 This case study demonstrates that both the trace metal 2004) and in the d CofCO3 groups substituted into the distribution and the range of clumped isotope temperatures phosphate lattice of linguliform brachiopods (Cowan in a sample suite are influenced by sedimentation rate and et al., 2005). Based on their statigraphic position and total overburden. Fossil materials preserved in environ- d13C composition, the samples analyzed for this study pre- ments with high sedimentation rates, and environments that serve the onset of the SPICE excursion (Cowan et al., 2005). have seen significant tectonic loading, usually exceed the The linguliform brachiopods from Iowa record the SPICE blocking temperature for calcite, and the fossils experience with d13C values identical to the surrounding carbonate partial or complete reordering (Finnegan et al., 2011; matrix and other global sections recording the SPICE 13 Henkes et al., 2014). In intermediate burial environments, (Cowan et al., 2005). Interestingly, the d C of the CO3 fossil materials have some components with elevated from linguliform brachiopods recovered from nearshore clumped isotope temperatures, including sparry calcite sand-dominated sediments in the Eau Claire Formation is cements with higher trace metal concentrations (Fig. 12). consistently depleted by about 5‰ compared to down-dip Fossil materials from depositional environments with very marine environments, a reasonable offset if the nearshore low sedimentation rates and the thinnest overburden, like sandy environment in Wisconsin was estuarine, with DIC the cratonic Volkhov and Decorah formations, have low isotopically distinct from a well-mixed marine DIC pool clumped isotope temperatures and little evidence for coarse (Mook, 1971; Cowan et al., 2005). This explanation is also 18 recrystallization. Although the trace metal enrichments and supported by the d O of the CO3 groups measured in the depleted d18O values in these formations suggest that these two populations: the d18O in the Wisconsin brachiopods fossil materials are poorly preserved, we contend that the is 4‰ lighter than that of the Iowa brachiopods d18O depletion in Decorah and Volkov samples relative to (Cowan et al., 2005). This mineral offset is similar to our younger fossil materials is primary and reflects temperature calculated water d18O offset from nominal ’seawater,’ change (Fig. 12). implying that deposition from different fluids but in similar temperatures could explain the offset between Wisconsin 4.3. Case study: Estuarine waters and Iowa. There is an additional test to determine whether the lin- The phosphatic brachiopods from the Cambrian-aged guliform brachiopods from the Eau Claire Formation grew Eau Claire Formation are the only samples we examined in brackish waters or, alternatively, whether global seawa- that yield both low clumped isotope temperatures and cal- ter d18O composition evolved between Late Cambrian culated fluid compositions more depleted than estimates for and Middle Ordovician time: pairing clumped isotope and ice-free seawater (estimated between 0.89 and 1.2‰) conventional d18O records of carbonate groups and apatite. (Lear, 2000; Cramer et al., 2011). In principle, these results Each record is imperfect. Carbonate minerals are more may imply a change in the seawater d18O value from Cam- likely to dissolve and recrystallize in burial environments. brian to Ordovician time, though they are certainly not as For apatite, the phosphate-water equilibrium relationship 18O-depleted as some hypotheses for Cambrian-aged sea- is long debated (Longinelli and Nuti, 1973, 1996; Chang water (2.5 ± 1.2‰ VSMOW, ECIB, n = 5 versus 7‰) and Blake, 2015; Kolodny et al., 1983; Lećuyer et al., (e.g. Prokoph et al., 2008; Veizer et al., 1999). However, 2013; Puceát et al., 2010). Additionally, Rodland et al. the sedimentological context of the samples suggests a (2003) recover high-resolution variability in apatite mineral Table 3 Water d18 O comparison of apatite samples (‰). 13 18 d e 18 Sample Name N Min Age d CVPDB SD d O SD D47CDES90 SD 2SE T (°C) 2SE Water d OVSMOW 2SE ..Brmn ta./Gohmc tCsohmc ca24(08 84 35 18–41 (2018) 224 Acta Cosmochimica et Geochimica / al. et Bergmann K.D. DIB 6 Ap 451.8 1.60 0.07 4.43 0.18 0.543 0.007 0.006 44.6 3.0 1.7 0.5 Chang and Blake (2015) 17.5 44.6 1.5 Puceát et al. (2010)c 17.5 44.6 0.8 Longinelli and Nuti (1973) 17.5 44.6 2.0 Lećuyer et al. (2013) 17.5 44.6 1.3 ECIB 5 Ap 499 5.23 0.43 6.21 0.03 0.580 0.020 0.018 31.7 6.1 2.5 1.2 Chang and Blake (2015) 16.1 31.7 4.8 Puceát et al. (2010)c 16.1 31.7 3.6 Longinelli and Nuti (1973) 16.1 31.7 2.4 Lećuyer et al. (2013) 16.1 31.7 3.0 TIIBa 1 Ap 0 1.09 0.01 1.08 0.01 0.608 0.017 0.004 24.0 1.3 3.3 0.3 Chang and Blake (2015) 21.4 22.2 0.8 Puceát et al. (2010)c 21.4 22.2 0.1 Longinelli and Nuti (1973) 21.4 22.2 1.1 Lećuyer et al. (2013) 21.4 22.2 0.7 PIBb 4Ap0 4.37 0.11 0.09 0.09 0.579 0.014 0.007 33.1 1.6 3.0 0.4 Chang and Blake (2015) 20.5 33.1 0.3 Puceát et al. (2010)c 20.5 33.1 1.1 Longinelli and Nuti (1973) 20.5 33.1 2.3 Lećuyer et al. (2013) 20.5 33.1 1.8 a Only TIIB with no D48 excess (TIIB_NaOH_HP). b Average of three cleaned samples of PIB that do not show a D48 excess (PIB_4_12_POW, PIB_4_12_250, and PIB_24_12_POW). c With the corrected value for NBS120c from Joachimski and Lambert (2015). d Relative to VPDB for the CO3 group analyses (grey), relative to VSMOW for the apatite analyses. e 6 2 D47CDES90 = 0.0422 (±0.0019) * 10 /T + 0.1262 (±0.0207) (Bonifacie et al., 2017, Eq. (3)). 36 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 d18O (>5‰) across linguliform brachiopod shells and sug- average values from the compilation are unlikely to provide 18 gest that CO3 groups provide a better record of equilibrium accurate estimates of mineral d OVPDB of this age conditions. Despite the uncertainties, combined measure- (Grossman, 2012a, 2012b; Grossman et al., 2008). Though ments of clumped isotope temperatures in modern, Ordovi- the lowest temperatures from the Cambrian-aged Eau cian and Cambrian samples can be used to test whether the Claire Formation, measured in Wisconsin linguliform bra- oxygen isotopes of these two mineral phases precipitated at chiopods imply fluid compositions depleted in 18O or near equilibrium from the same fluid. Using the mea- (2.6‰), the sedimentological, geochemical and fossil evi- sured clumped isotope temperature from the CO3 group dence suggest that these nearshore sandstones were depos- and measured d18O mineral composition (either on the apa- ited in a brackish environment lower in d18O than marine 18 tite or CO3 group)—we calculated water d O compositions waters due to riverine influx. Pairing the Wisconsin bra- using four phosphate-water d18O equilibrium fractionation chiopod temperatures with carbonate d18O values from equations and one calcite-water d18O equilibrium fraction- contemporaneous linguliform brachiopods from offshore ation equation (Longinelli and Nuti, 1973, 1996; Kim and Iowa sections yields marine seawater d18O estimates O’Neil, 1997; Chang and Blake, 2015; Kolodny et al., (Cowan et al., 2005). 1983; Lećuyer et al., 2013; Puceát et al., 2010) with an These estimates for seawater d18O accord with ophiolite updated value for NBS120c for Puceát et al. (2010) studies and records of hydrothermal alteration of mid- (Joachimski and Lambert, 2015). Notably, water ocean ridge basalts that show little change in seawater 18 18 d OVSMOW values calculated from apatite measurements d O since the Archean Eon (Muehlenbachs and Clayton, using two PO4-water equilibrium equations (Longinelli 1976; Holmden and Muehlenbachs, 1993; Muehlenbachs, and Nuti, 1973; Lećuyer et al., 2013) provided a better 1998; Turchyn et al., 2013). The ophiolite data evince a con- 18 18 match to carbonate water d OVSMOW than other equations stant gradient in d O of oceanic crusts through time, with 18 (Chang and Blake, 2015; Puceát et al., 2010) for the pillow basalt enrichment (d OVSMOW 6–10‰) relative to 18 Cambro-Ordovician samples while the modern CO3 group concurrent gabbros (d OVSMOW 3–6‰) indicating waters are high relative to GISS measurements from the exchange with 0‰ VSMOW seawater at different tempera- sites suggesting the need for better modern data of a range tures (e.g. Holmden and Muehlenbachs, 1993). The of apatite materials (Table 3). Further work is needed to constancy of the oceanic d18O reservoir is supported by fully address why some equations agree better in the past. mass-balance models of seawater isotope exchange, which suggest an equilibrium between isotopic depletion from 4.4. Climate of the Cambrian and Ordovician and seawater low-temperature pillow basalt weathering and hydrother- d18O through time mal enrichment of gabbroic mid-ocean ridge deposits (Muehlenbachs and Clayton, 1976). Assuming past spread- The decline in apatite and calcite mineral d18O values ing rates to be at least 50% of present rates, these exchange through deep time has historically had three explanations: reactions, along with subduction and recycling of saturated climate and baseline temperatures changed; seawater d18O sediments, are thought to have maintained seawater com- evolved; or burial diagenesis had a greater impact in the positions of 0‰ (Muehlenbachs and Clayton, 1976; past than it currently has. We find no evidence for evolution Holmden and Muehlenbachs, 1993; Muehlenbachs, 1998). of seawater d18O beyond stratigraphic trends likely associ- Reconstructing sea surface temperatures in the ancient ated with cycles of continental ice sheet growth, but signif- past with accuracy remains challenging because the icant evidence that both climate change (albeit milder than clumped isotope thermometer is particularly sensitive to some hypotheses have suggested; e.g. Trotter et al., 2008) diagenetic alteration. Each study site has a complex history and diagenesis contributed to Paleozoic d18O trends. of burial, recrystallization, trace metal migration, and During the time interval that the Cambrian- and changes in permeability and porosity. The petrography, Ordovician-aged strata we sampled represent, calcite bra- SEM and EPMA work that complements our clumped chiopod d18O increased from 11‰ to 4‰ and apatite isotope studies indicates that reprecipitated carbonate mineral d18O increased from 14‰ to 18‰ (Fig. 1). The influences even those brachiopods typically regarded as d18O water compositions of these strata, calculated from well-preserved. At a single sample site, inclusion of repre- clumped isotope measurements, indicate that the oxygen cipitated material into samples with similar d18O mineral isotopic fluctuations over the Ordovician largely reflects compositions can cause a range of clumped isotope temper- changes in ice volume related to the end-Ordovician glacia- atures (Fig. 11, Fig. 8). The full range of brachiopod tion (e.g. Finnegan et al. 2011). Middle Ordovician-aged clumped isotope temperatures from the Decorah Forma- samples from Russia and Late Ordovician-aged samples tion can be explained by connecting those temperatures from the Decorah Formation yield seawater d18O values with textural differences between skeletal materials viewed predicted for ice-free conditions in oceans with modern by SEM: samples with higher clumped isotope tempera- d18O water composition (between 0.89 and 1.2‰) tures have more recrystallized shell material and commonly 18 (Lear, 2000; Cramer et al., 2011). The mineral d OVPDB have thin laminae susceptible to dissolution (Fig. 11, compositions of these samples are at the high end of the Fig. 8). Fig. 10A includes the minimum, well-replicated Phanerozoic d18O dataset compiled from tropical bra- clumped isotope temperature from each time bin as our best chiopods (Veizer et al., 1999), as are bulk oxygen isotope temperature estimate from the least recrystallized material. values from Lower Ordovician brachiopods from Baltica We find no evidence supporting recent claims that clumped (Rasmussen et al., 2016). These difference suggests that isotope temperature variability is independent of texture, K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 37 trace metal or chemical/isotopic signals in study sites with in the past. Using clumped isotope thermometry, coupled low temperature burial histories (Veizer and Prokoph, with micro-analytical screening, we evaluate the origins of 2015). d18O records of carbonate and apatite by deconvolving In the late Cambrian – middle Ordovician time interval fluid composition from temperature. Results from Ordovi- our clumped isotope dataset brackets previously published cian and Cambrian calcitic and phosphatic taxa indicate conodont phosphate analyses. Unlike Trotter et al. (2008), that seawater d18O did not approach 7‰. Instead, by con- we observe no monatonic trend towards warmer tempera- straining the effect of diagenetic recrystallization on the tures into the Cambrian however we lack the sample reso- materials analyzed, we find equatorial temperatures of lution to fully reconstruct earliest Ordovician Ordovician and Cambrian epeiric seas to be between 29 temperatures. Rather, temperature differences between and 38 °C, a temperature range within modern equatorial samples within positive carbon isotope excursions (the temperatures and estimates from proxies and models for SPICE, GICE and HICE) and those outside excursions the Cretaceous and Eocene (Huber and Sloan, 2001; suggest that these positive carbon isotope excursions may Douglas et al., 2014; Lunt et al., 2012; Upchurch et al., each represent a short interval associated with cooler tem- 2015). Finally we hypothesize that d18O values trend from peratures (Buggisch et al., 2010; Elrick et al., 2011; low to high during late Cambrian – Early Ordovician time Finnegan et al., 2011). The minimum temperatures from because of (1) decreased diagenesis due to the taphonomic other time bins equal or slightly exceed temperatures in change in skeletal materials that accompanied the Ordovi- the modern West Pacific warm pool and are warm relative cian radiation, and (2) the deposition of thin, thermally to modern tropical seawater (34–38 °C vs. 30). They also immature strata in the middle of continents during global match estimates by climate models and proxies for equato- eustatic highstand, which contain fossil materials better rial sea surface temperatures during the ice-free Eocene Cli- preserved than their Cambrian-age counterparts. matic Optimum and the Cretaceous, suggesting that past Phanerozoic greenhouse conditions may have reached sim- ACKNOWLEDGEMENTS ilar maxima (Huber and Sloan, 2001; Douglas et al., 2014; Lunt et al., 2012; Upchurch et al., 2015). Skeletal growth We thank Ethan Grossman, Cedric John and an anonymous temperatures may also reflect seasonal biases (e.g. Butler reviewer for helpful comments that improved the manuscript. et al., 2015) or differ systematically from a global seawater The Los Angeles County Museum of Natural History, the Waikiki average due to the unique hydrography of tropical epeiric Aquarium and Norene Tuross provided samples of modern linguli- seaways. form and calcitic brachiopods. Victoria Orphan provided the cal- Two aspects of the fossil record may magnify the impor- citic brachiopod from Monterey Canyon. Matthew Robles and tance of diagenesis in Cambrian and Early Ordovician cal- Gerry Gunderson provided Eau Claire materials and collaborated citic fossils relative to its importance in younger materials. on fieldwork in the Eau Claire Formation. Norlene Emerson, Clint Cowan, and Pat McLaughlin provided invaluable assistance on the First, the lowest values in brachiopod calcite d18O predates Decorah Formation in the field. Nami Kitchen, Joel Hurowitz, Chi the Ordovician radiation of calcifying taxa, when many Ma and Katie Eiler assisted with analytical measurements and groups of thick-shelled taxa emerged. Those thick shells method development. Page Quinton and Ken MacLeod are respon- are likely better able to withstand diagenesis. A skeletal sible for phosphate d18O measurements on splits of the ancient lin- control on diagenesis is apparent in the divergence between guliform brachiopod samples and two of our modern linguliform calcite d18O and phosphate d18O values in the Cambrian samples. A National Science Foundation Graduate Research Fel- and Early Ordovician (Fig. 1, light points in upper panel). lowship was provided to K. Bergmann. Support for analyses was Second, the Ordovician intracratonic seas that developed provided by a National Science Foundation award (EAR- following the Sauk Transgression—a eustatic sea level high 1053523) to WWF and JME. WWF acknowledges support from over the past five hundred million years (Peters and Gaines, the Agouron Institute and a David and Lucile Packard Fellowship for Science and Engineering. LEP acknowledges logistical support 2012)—deposited strata on the geodynamically stable con- from the National Museum of Wales, Cardiff. tinental interior, creating a unique sedimentary archive with low burial and superior fossil preservation. We postulate that the taphonomic opportunities associated with concur- APPENDIX A. SUPPLEMENTARY MATERIAL rent development of thick-shelled calcitic skeletons and deposition of gently buried midcratonic deposits in the Supplementary data associated with this article can be mid- to late Ordovician drove global trends in d18O mineral found, in the online version, at https://doi.org/10.1016/j. records that naturally reflect time-varying diagenetic gca.2017.11.015. signals.

REFERENCES 5. CONCLUSIONS

d18 Barker C. E. and Pawlewicz M. J. (1994) Calculation of vitrinite The long-term increase in O of carbonate, apatite and reﬂectance from thermal histiries amd peak temperatures. A d18 chert has three end-member explanations: (1) the O comparison of methods. Vitr. reﬂectance as a Matur. Param. record was controlled by cooling of surface environments Apl. limitations, 216–229. https://doi.org/10.1021/bk-1994- towards the present, (2) the bulk isotopic concentration of 0570.ch014. seawater increased 7‰ from the Cambrian period to the Bassett D., Macleod K. G., Miller J. F. and Ethington R. L. (2007) present, and (3) diagenetic alteration was more prevalent Oxygen isotopic composition of biogenic phosphate and the 38 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41

temperature of early Ordovician seawater. Palaios 22, 98–103. Cosmochim. Acta 140, 241–258. https://doi.org/10.1016/j. https://doi.org/10.2110/palo.2005.p05-089r. gca.2014.05.024. Bergstro¨m S. M. and McKenzie P. (2005) Biostratigraphic and Dae¨ron M., Blamart D., Peral M. and Affek H. P. (2016) Absolute paleoceanographic relations between the type Richmondian isotopic abundance ratios and the accuracy of D47 measure- (Upper Ordovician) in the Cincinnati region and the Upper ments. Chem. Geol. 442, 83–96. https://doi.org/10.1016/j. Mississippi Valley succession. Iowa Geol. Surv. Guideb. Ser., chemgeo.2016.08.014. 35–37. Dale A., John C. M., Mozley P. S., Smalley P. C. and Muggeridge Bernasconi S. M., Hu B., Wacker U., Fiebig J., Breitenbach S. F. A. H. (2014) Time-capsule concretions: unlocking burial M. and Rutz T. (2013) Background effects on Faraday diagenetic processes in the Mancos shale using carbonate collectors in gas-source mass spectrometry and implications clumped isotopes. Earth Planet. Sci. Lett.. https://doi.org/ for clumped isotope measurements. Rapid Commun. Mass 10.1016/j.epsl.2014.03.004. Spectrom. 27, 603–612. https://doi.org/10.1002/rcm.6490. Defliese W. F., Hren M. T. and Lohmann K. C. (2015) Compo- Bonifacie M., Calmels D., Eiler J. M., Horita J., Chaduteau C., sitional and temperature effects of phosphoric acid fractiona-

Vasconcelos C., Agrinier P., Katz A., Passey B. H., Ferry J. M. tion on D47 analysis and implications for discrepant and Bourrand J. J. (2017) Calibration of the dolomite clumped calibrations. Chem. Geol. 396, 51–60. https://doi.org/10.1016/ isotope thermometer from 25 to 350 °C, and implications for a j.chemgeo.2014.12.018.

universal calibration for all (Ca, Mg, Fe)CO3 carbonates. Dennis K. J., Aﬀek H. P., Passey B. H., Schrag D. P. and Eiler J. Geochim. Cosmochim. Acta 200, 255–279. https://doi.org/ M. (2011) Deﬁning an absolute reference frame for ‘‘clumped”

10.1016/j.gca.2016.11.028. isotope studies of CO2. Geochim. Cosmochim. Acta 75, 7117– Brand U., Jiang G., Azmy K., Bishop J. and Montan˜ez I. P. (2012) 7131. https://doi.org/10.1016/j.gca.2011.09.025. Diagenetic evaluation of a Pennsylvanian carbonate succession Douglas P. M. J., Aﬀek H. P., Ivany L. C., Houben A. J. P., Sijp (Bird Spring Formation, Arrow Canyon, Nevada, U.S.A.) – 1: W. P., Sluijs A., Schouten S. and Pagani M. (2014) Pronounced Brachiopod and whole rock comparison. Chem. Geol. 308–309, zonal heterogeneity in Eocene southern high-latitude sea 26–39. https://doi.org/10.1016/j.chemgeo.2012.03.017. surface temperatures. Proc. Natl. Acad. Sci. U. S. A. 111, 1–6. Brand U. and Veizer J. (1980) Chemical diagenesis of a multicom- https://doi.org/10.1073/pnas.1321441111. ponent carbonate system-1: trace elements. J. Sediment. Petrol.. Dronov A. V., Tolmacheva T., Raevskaya E. and Nestell M. (2005) Brand W. A., Assonov S. S. and Coplen T. B. (2010) Correction for Cambrian and Ordovician of St. Petersburg Region, Fossils. 17 13 the O interference in d C measurements when analyzing CO2 Dronov A. V., Savitsky J. V., Fedorov P. V. and Tsyganova E. A. with stable isotope mass spectrometry (IUPAC Technical (1996) Detailed lithostratigraphy of the Ordovician lower Report). Pure Appl. Chem. 82, 1719–1733. https://doi.org/ Volkhovian limestone along the eastern part of the Baltic- 10.1351/PAC-REP-09-01-05. Ladoga Glint, northwestern Russia. Geol. Foeren. Stockholm Buggisch W., Joachimski M. M., Lehnert O., Bergstro¨m S. M., Foerh. 118, 19–24. Repetski J. E. and Webers G. F. (2010) Did intense volcanism Eagle R. A., Schauble E. A., Tripati A. K., Tu¨tken T., Hulbert R. trigger the ﬁrst late Ordovician icehouse? Geology 38, 327–330. C. and Eiler J. M. (2010) Body temperatures of modern and https://doi.org/10.1130/G30577.1. extinct vertebrates from 13C–18O bond abundances in bioap- Butler S., Bailey T. R., Lear C. H., Curry G. B., Cherns L. and atite. Proc. Natl. Acad. Sci. U. S. A. 107, 10377–10382. https:// McDonald I. (2015) The Mg/Ca-temperature relationship in doi.org/10.1073/pnas.0911115107. brachiopod shells: calibrating a potential palaeoseasonality Eagle R. A., Tutken T., Martin T. S., Tripati A. K., Fricke H. C., proxy. Chem. Geol. 397, 106–117. https://doi.org/10.1016/j. Connely M., Cifelli R. L. and Eiler J. M. (2011) Dinosaur body chemgeo.2015.01.009. temperatures determined from isotopic (13C-18O) ordering in Came R. E., Eiler J. M., Veizer J., Azmy K., Brand U. and fossil biominerals. Science 333, 443–445. https://doi.org/ Weidman C. R. (2007) Coupling of surface temperatures and 10.1126/science.1206196.

atmospheric CO2 concentrations during the Palaeozoic era. Egerquist E. and Egerquist E. (2003) New brachiopods from the Nature 449, 198–201. https://doi.org/10.1038/nature06085. Lower-Middle Ordovician (Billingen–Volkhov stages) of the Chang S. J. and Blake R. E. (2015) Precise calibration of East Baltic. Acta Palaeontol. Pol. 48, 31–38. equilibrium oxygen isotope fractionations between dissolved Eiler J. M. (2007) ‘‘Clumped-isotope” geochemistry–the study of phosphate and water from 3 to 37 °C. Geochim. Cos- naturally-occurring, multiply-substituted isotopologues. Earth mochim. Acta 150, 314–329. https://doi.org/10.1016/j. Planet. Sci. Lett. 262, 309–327. https://doi.org/10.1016/j. gca.2014.10.030. epsl.2007.08.020. Cowan C. A., Fox D. L., Runkel A. C. and Saltzman M. R. (2005) Elrick M., Rieboldt S., Saltzman M. and McKay R. M. (2011) Terrestrial-marine carbon cycle coupling in 500-m.y.-old Oxygen-isotope trends and seawater temperature changes phosphatic brachiopods. Geology 33, 661–664. https://doi.org/ across the Late Cambrian Steptoean positive carbon-isotope 10.1130/G21434.1. excursion (SPICE event). Geology 39, 987–990. https://doi.org/ Cramer B. S., Miller K. G., Barrett P. J. and Wright J. D. (2011) 10.1130/G32109.1. Late Cretaceous-Neogene trends in deep ocean temperature Emerson N. R. (2002) Sedimentology, sequence stratigraphy, and and continental ice volume: Reconciling records of benthic brachiopod biostratigraphy of the Ordovician (Mohawkian) foraminiferal geochemistry (d18O and Mg/Ca) with sea level Decorah Formation, Midcontinent, USA. history. J. Geophys. Res. Ocean. 116, C12023. https://doi.org/ Fernandez A., Tang J. and Rosenheim B. E. (2014) Siderite 10.1029/2011JC007255. ‘‘clumped” isotope thermometry: a new paleoclimate proxy for Creveling J. R., Johnston D. T., Poulton S. W., Kotrc B., Ma¨rz C., humid continental environments. Geochim. Cosmochim. Acta Schrag D. P. and Knoll A. H. (2014) Phosphorus sources for 126, 411–421. https://doi.org/10.1016/j.gca.2013.11.006. phosphatic Cambrian carbonates. Bull. Geol. Soc. Am. 126, Finnegan S., Bergmann K., Eiler J. M., Jones D. S., Fike D. A., 145–163. https://doi.org/10.1130/B30819.1. Eisenman I., Hughes N. C., Tripati A. K. and Fischer W. W. Cummins R. C., Finnegan S., Fike D. A., Eiler J. M. and Fischer (2011) The magnitude and duration of Late Ordovician-Early W. W. (2014) Carbonate clumped isotope constraints on Silurian glaciation. Science 331, 903–906. https://doi.org/ Silurian ocean temperature and seawater d18O. Geochim. 10.1126/science.1200803. K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 39

Friedman I. and O’Neil J. R. (1977) Compilation of stable isotope Karhu J. and Epstein S. (1986) The implication of the oxygen fractionation factors of geochemical interest. Geol. Surv. Prof. isotope records in coexisting cherts and phosphates. Geochim. Pap. 440. https://doi.org/10.1017/CBO9781107415324.004. Cosmochim. Acta 50, 1745–1756. https://doi.org/10.1016/0016- Giles P. S. (2012) Low-latitude Ordovician to Triassic brachiopod 7037(86)90136-5. habitat temperatures (BHTs) determined from d18O bra- Kasting J. F., Howard M. T., Wallmann K., Veizer J., Shields G. chiopod calcite: a cold hard look at ice-house tropical oceans. and Jaffre´s J. B. D. (2006) Paleoclimates, ocean depth, and the Palaeogeogr. Palaeoclimatol. Palaeoecol. 317–318, 134–152. oxygen isotopic composition of seawater. Earth Planet. Sci. https://doi.org/10.1016/j.palaeo.2012.01.002. Lett. 252, 82–93. https://doi.org/10.1016/j.epsl.2006.09.029. Grossman E. L. (2012a) Applying oxygen isotope paleothermom- Kele S., Breitenbach S. F. M., Capezzuoli E., Meckler A. N., etry in deep time. Paleontol. Soc. Pap. 18, 39–68. https://doi. Ziegler M., Millan I. M., Kluge T., Dea´k J., Hanselmann K., org/10.1017/s1089332600002540. John C. M., Yan H., Liu Z. and Bernasconi S. M. (2015) Grossman E. L. (2012b) Oxygen isotope stratigraphy. Geol. Time Temperature dependence of oxygen- and clumped isotope Scale 181–206. https://doi.org/10.1016/B978-0-444-59425- fractionation in carbonates: a study of travertines and tufas in 9.00010-X. the 6–95°C temperature range. Geochim. Cosmochim. Acta 168, Grossman E. L., Yancey T. E., Jones T. E., Bruckschen P., 172–192. https://doi.org/10.1016/j.gca.2015.06.032. Chuvashov B., Mazzullo S. J. and Mii H. sheng. (2008) Kelson J. R., Huntington K. W., Schauer A. J., Saenger C. and Glaciation, aridification, and carbon sequestration in the Lechler A. R. (2017) Toward a universal carbonate clumped Permo-Carboniferous: the isotopic record from low latitudes. isotope calibration: diverse synthesis and preparatory methods Palaeogeogr. Palaeoclimatol. Palaeoecol. 268, 222–233. https:// suggest a single temperature relationship. Geochim. Cosmochim. doi.org/10.1016/j.palaeo.2008.03.053. Acta 197, 104–131. https://doi.org/10.1016/j.gca.2016.10.010. Henkes G. A., Passey B. H., Grossman E. L., Shenton B. J., Perez- Kim S.-T. and O’Neil J. R. (1997) Equilibrium and nonequilibrium Huerta A. and Yancey T. E. (2014) Temperature limits for oxygen isotope effects in synthetic carbonates. Geochim. Cos- preservation of primary calcite clumped isotope paleotemper- mochim. Acta 61, 3461–3475. https://doi.org/10.1016/S0016- atures. Geochim. Cosmochim. Acta 139, 362–382. https://doi. 7037(97)00169-5. org/10.1016/j.gca.2014.04.040. Kim S. T., Mucci A. and Taylor B. E. (2007a) Phosphoric acid Henkes G. A., Passey B. H., Wanamaker A. D., Grossman E. L., fractionation factors for calcite and aragonite between 25 and Ambrose W. G. and Carroll M. L. (2013) Carbonate clumped 75 °C: revisited. Chem. Geol. 246, 135–146. https://doi.org/ isotope compositions of modern marine mollusk and bra- 10.1016/j.chemgeo.2007.08.005. chiopod shells. Geochim. Cosmochim. Acta 106, 307–325. Kim S. T., O’Neil J. R., Hillaire-Marcel C. and Mucci A. (2007b) https://doi.org/10.1016/j.gca.2012.12.020. Oxygen isotope fractionation between synthetic aragonite and Holmden C. and Muehlenbachs K. (1993) The 18O/16O ratio of 2- water: influence of temperature and Mg2+ concentration. billion-year-old seawater inferred from ancient oceanic crust. Geochim. Cosmochim. Acta 71, 4704–4715. https://doi.org/ Science 259, 1733–1736. https://doi.org/ 10.1016/j.gca.2007.04.019. 10.1126/science.259.5102.1733. Kluge T., Affek H. P., Zhang Y. G., Dublyansky Y., Spo¨tl C., Huber M. and Sloan L. C. (2001) Heat transport, deep waters, and Immenhauser A. and Richter D. K. (2014) Clumped isotope thermal gradients: coupled simulation of an Eocene greenhouse thermometry of cryogenic cave carbonates. Geochim. Cos- climate. Geophys. Res. Lett. 28, 3481–3484. https://doi.org/ mochim. Acta 126, 541–554. https://doi.org/10.1016/j. 10.1029/2001GL012943. gca.2013.11.011. Huntington K. W., Eiler J. M., Affek H. P., Guo W., Bonifacie M., Knauth L. P. and Epstein S. (1976) Hydrogen and oxygen isotope Yeung L. Y., Thiagarajan N., Passey B., Tripati A., Daeron M. ratios in nodular and bedded cherts. Geochim. Cosmochim. Acta and Came R. (2009) Methods and limitations of ‘‘clumped” 40, 1095–1108. https://doi.org/10.1016/0016-7037(76)90051-X.

CO2 isotope (D47) analysis by gas-source isotope ratiomass Koch P. P. L., Tuross N. and Fogel M. L. M. (1997) The effects of spectrometry. J. Mass Spectrom. 44, 1318–1329. https://doi. sample treatment and diagenesis on the isotopic integrity of org/10.1002/jms.1614. carbonate in biogenic hydroxylapatite. J. Archaeol. Sci. 24, Jaffre´s J. B. D., Shields G. A. and Wallmann K. (2007) The oxygen 417–429. https://doi.org/10.1006/jasc.1996.0126. isotope evolution of seawater: a critical review of a long- Kolodny Y., Luz B. and Navon O. (1983) Oxygen isotope standing controversy and an improved geological water cycle variations in phosphate of biogenic apatites, I. Fish bone model for the past 3.4 billion years. Earth-Sci. Rev. 83, 83–122. apatite—rechecking the rules of the game. Earth Planet. Sci. https://doi.org/10.1016/j.earscirev.2007.04.002. Lett. 64, 398–404. Joachimski M. M. and Lambert L. L. (2015) Salinity contrast in Lear C. H. (2000) Cenozoic deep-sea temperatures and global ice the US Midcontinent Sea during Pennsylvanian glacio-eustatic volumes from Mg/Ca in benthic foraminiferal calcite. Science highstands: evidence from conodont apatite d18O. Palaeogeogr. 287, 269–272. https://doi.org/10.1126/science.287.5451.269. Palaeoclimatol. Palaeoecol. 433, 71–80. https://doi.org/10.1016/ Lećuyer C., Amiot R., Touzeau A. and Trotter J. (2013) Calibra- j.palaeo.2015.05.014. tion of the phosphate d18O thermometer with carbonate-water Joachimski M. M., van Geldern R., Breisig S., Buggisch W. and oxygen isotope fractionation equations. Chem. Geol. 347, 217– Day J. (2004) Oxygen isotope evolution of biogenic calcite and 226. https://doi.org/10.1016/j.chemgeo.2013.03.008. apatite during the Middle and Late Devonian. Int. J. Earth Sci. Lećuyer C., Grandjean P., Barrat J.-A. A., Nolvak J., Emig C., 93, 542–553. https://doi.org/10.1007/s00531-004-0405-8. Paris F. and Robardet M. (1998) D18O and REE contents of John C. M. (2015) Burial estimates constrained by clumped isotope phosphatic brachiopods: a comparison between modern and thermometry: example of the Lower Cretaceous Qishn Forma- lower Paleozoic populations. Geochim. Cosmochim. Acta 62, tion (Haushi-Huqf High, Oman). Geol. Soc. London, Spec. 2429–2436. https://doi.org/10.1016/S0016-7037(98)00170-7. Publ. 435. https://doi.org/10.1144/SP435.5. Lećuyer C., Grandjean P. and Emig C. C. (1996) Determination of John C. M. and Bowen D. (2016) Community software for oxygen isotope fractionation between water and phosphate challenging isotope analysis: First applications of ‘‘Easotope” from living lingulids: potential application to palaeoenviron- to clumped isotopes. Rapid Commun. Mass Spectrom. 30, 2285– mental studies. Palaeogeogr. Palaeoclimatol. Palaeoecol. 126, 2300. https://doi.org/10.1002/rcm.7720. 101–108. 40 K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41

Longinelli A. and Nuti S. (1973) Revised phosphate-water isotopic and d34S database through Earth history. Earth-Sci. Rev. 87, temperature scale. Earth Planet. Sci. Lett. 19, 373–376. https:// 113–133. https://doi.org/10.1016/j.earscirev.2007.12.003. doi.org/10.1016/0012-821X(73)90088-5. Pruss S. B., Finnegan S., Fischer W. W. and Knoll A. H. (2010) Ludvigsen R. and Westrop S. R. (1983) Trilobite biofacies of the Carbonates in skeleton-poor seas: new insights from Cambrian cambrian-ordovician boundary interval in northern North and Ordovician strata of Laurentia. Palaios 25, 73–84. https:// America. Alcheringa 7, 301–319. https://doi.org/10.1080/ doi.org/10.2110/palo.2009.p09-101r. 03115518308619614. Puceát E., Joachimski M. M., Bouilloux A., Monna F., Bonin A., Ludvigson G. A., Witzke B. J., Gonza´lez L. A., Carpenter S. J., Motreuil S., Morinie`re P., Heńard S., Mourin J., Dera G. and Schneider C. L. and Hasiuk F. (2004) Late Ordovician Quesne D. (2010) Revised phosphate–water fractionation (Turinian-Chatfieldian) carbon isotope excursions and their equation reassessing paleotemperatures derived from biogenic stratigraphic and paleoceanographic significance. Palaeogeogr. apatite. Earth Planet. Sci. Lett. 298, 135–142. Palaeoclimatol. Palaeoecol. 210, 187–214. https://doi.org/ Rasmussen C. M. Ø., Ullmann C. V., Jakobsen K. G., Lindskog 10.1016/j.palaeo.2004.02.043. A., Hansen J., Hansen T., Eriksson M. E., Dronov A., Frei R., Lundsten L., Schlining K. L., Frasier K., Johnson S. B., Kuhnz L. Korte C., Nielsen A. T. and Harper D. A. T. (2016) Onset of A., Harvey J. B. J., Clague G. and Vrijenhoek R. C. (2010) main Phanerozoic marine radiation sparked by emerging Mid Time-series analysis of six whale-fall communities in Monterey Ordovician icehouse. Sci. Rep. 6, 18884. https://doi.org/ Canyon, California, USA. Deep Sea Res. Part I: Ocean. Res. 10.1038/srep18884. Pap. 57(12). https://doi.org/10.1016/j.dsr.2010.09.003. Rodland D. L., Kowalewski M., Dettman D. L., Flessa K. W., Lunt D. J., Jones T. D., Heinemann M., Huber M., LeGrande A., Atudorei V. and Sharp Z. D. (2003) High-resolution analysis of Winguth A., Loptson C., Marotzke J., Roberts C. D., Tindall d18O in the biogenic phosphate of modern and fossil lingulid J., Valdes P. and Winguth C. (2012) A model-data comparison brachiopods. J. Geol. 111, 441–453. https://doi.org/10.1086/ for a multi-model ensemble of early Eocene atmosphere-ocean 375283. simulations: EoMIP. Clim. Past 8, 1717–1736. https://doi.org/ Rohrssen M. (2013) Molecular organic geochemical records of late 10.5194/cp-8-1717-2012. Ordovician biospheric evolution. Mook W. G. (1971) Paleotemperatures and chlorinities from stable Rohrssen M., Gill B. C. and Love G. D. (2015) Scarcity of the C 30 carbon and oxygen isotopes in shell carbonate. Palaeogeogr. sterane biomarker, 24-n-propylcholestane, in Lower Paleozoic Palaeoclimatol. Palaeoecol. 9, 245–263. https://doi.org/10.1016/ marine paleoenvironments. Org. Geochem. 80, 1–7. https://doi. 0031-0182(71)90002-2. org/10.1016/j.orggeochem.2014.11.008. Moqsud M. A., Hayashi S., Du Y. J. and Suetsugu D. (2006) Rowan E. L., Goldhaber M. B. (1996). Fluid Inclusions and Evaluation of temperature trend in contaminated tidal flat in Biomarkers in the Upper Mississippi Valley Zinc-lead District; the Ariake Sea, Japan. Am. J. Environ. Sci. 2, 104–108. Implications for the Fluid-flow and Thermal History of the Muehlenbachs K. (1998) The oxygen isotopic composition of the Illinois Basin. Bulletin. oceans, sediments and the seafloor. Chem. Geol. 145, 263–273. Runkel A., McKay R., Cowan C., Miller J., Taylor J. (2012). The Muehlenbachs K. and Clayton R. N. (1976) Oxygen isotope Sauk Megasequence in the cratonic interior of North America: composition of the oceanic crust and its bearing on seawater. J. interplay between a fully developed inner detrital belt and the Geophys. Res. 81, 4365–4369. https://doi.org/10.1029/ central Great American Carbonate Bank. AAPG Mem. 98 Gt. JB081i023p04365. Am. Carbonate Bank. pp. 1001–1011. http://doi.org/10.1306/ Mu¨ller I. A., Fernandez A., Radke J., van Dijk J., Bowen D., 13331526M983522. Schwieters J. and Bernasconi S. M. (2017) Carbonate clumped Runkel A. C., McKay R. M. and Palmer A. R. (1998) Origin of a isotope analyses with the long-integration dual-inlet (LIDI) classic cratonic sheet sandstone: stratigraphy across the Sauk workflow: scratching at the lower sample weight boundaries. II–Sauk III boundary in the Upper Mississippi Valley. GSA Rapid Commun. Mass Spectrom. 31, 1057–1066. https://doi.org/ Bull. 110. 10.1002/rcm.7878. Runkel A. C., Miller J. F., McKay R. M., Palmer A. R. and Taylor Myrow P. M., Taylor J. F., Runkel A. C. and Ripperdan R. L. J. F. (2007) High-resolution sequence stratigraphy of lower (2012) Mixed siliciclastic–carbonate upward-deepening cycles Paleozoic sheet sandstones in central North America: the role of the upper Cambrian inner detrital belt of Laurentia. J. of special conditions of cratonic interiors in development of Sediment. Res. 82, 216–231. https://doi.org/10.2110/jsr.2012.20. stratal architecture. Bull. Geol. Soc. Am. 119, 860–881. https:// Passey B. H., Cerling T. E. and Levin N. E. (2007) Temperature doi.org/10.1130/B26117.1. dependence of oxygen isotope acid fractionation for modern Saltzman M. R., Cowan C. A., Runkel A. C., Runnegar B., and fossil tooth enamels. Rapid Commun. Mass Spectrom. 21, Stewart M. C. and Palmer A. R. (2004) The Late Cambrian 2853–2859. https://doi.org/10.1002/rcm.3149. Spice (13C) event and the Sauk II-Sauk III regression: new Passey B. H., Levin N. E., Cerling T. E., Brown F. H., Eiler J. M. evidence from Laurentian basins in Utah, Iowa, and New- and Turekian K. K. (2010) High-temperature environments of foundland. J. Sediment. Res. 74, 366–377. https://doi.org/ human evolution in East Africa based on bond ordering in 10.1306/120203740366. paleosol carbonates. Proc. Natl. Acad. Sci. U. S. A. 107, 11245– Schauer A. J., Kelson J., Saenger C. and Huntington K. W. (2016) 17 11249. https://doi.org/10.1073/pnas.1001824107. Choice of O correction affects clumped isotope (D47) values of Peters S. E. and Gaines R. R. (2012) Formation of the ‘‘Great CO2 measured with mass spectrometry. Rapid Commun. Mass Unconformity” as a trigger for the Cambrian explosion. Nature Spectrom. 30, 2607–2616. https://doi.org/10.1002/rcm.7743. 484, 363–366. https://doi.org/10.1038/nature10969. Schmidt G. A. (1999) Forward modeling of carbonate proxy data Popov L. E., Egerquist E. and Zuykov M. A. (2005) Ordovician from planktonic foraminifera using oxygen isotope tracers in a (Arenig-Caradoc) syntrophiidine brachiopods from the East global ocean model. Paleoceanography 14, 482–497. https://doi. Baltic region. Palaeontology 48, 739–761. https://doi.org/ org/10.1029/1999PA900025. 10.1111/j.1475-4983.2005.00487.x. Sell B. K., Samson S. D., Mitchell C. E., McLaughlin P. I., Koenig Prokoph A., Shields G. A. and Veizer J. (2008) Compilation and A. E. and Leslie S. A. (2015) Stratigraphic correlations using time-series analysis of a marine carbonate d18O, d13C, 87Sr/86Sr trace elements in apatite from late Ordovician (Sandbian- K.D. Bergmann et al. / Geochimica et Cosmochimica Acta 224 (2018) 18–41 41

Katian) K-bentonites of eastern North America. Bull. Geol. temperatures during the latest Cretaceous: congruence of Soc. Am. 127, 1259–1274. https://doi.org/10.1130/B31194.1. geologic data and climate models. Geology 43, 683–686. Shemesh A., Kolodny Y. and Luz B. (1988) Isotope geochemistry https://doi.org/10.1130/G36802.1. of oxygen and carbon in phosphate and carbonate of phos- Urey H. C. (1947) The thermodynamic properties of isotopic phorite francolite. Geochim. Cosmochim. Acta 52, 2565–2572. substances. J. Chem. Soc. 562. https://doi.org/10.1039/ https://doi.org/10.1016/0016-7037(88)90027-0. jr9470000562. Shields G. and Veizer J. (2002) The Precambrian marine carbonate Veizer J., Ala D., Azmy K., Bruckschen P., Buhl D., Bruhn F., isotope database: version 1.1. Geochem. Geophys. Geosyst. 3,1– Carden G., Diener A., Ebneth S., Godderis Y., Jasper T., Korte 12. https://doi.org/10.1029/2001GC000266. C., Pawellek F., Podlaha O. and Strauss H. (1999) 87Sr/86Sr, Simo J. A., Emerson N. R., Byers C. W. and Ludvigson G. A. d13C and d18O evolution of Phanerozoic seawater. Chem. Geol. (2003) Anatomy of an embayment in an Ordovician epeiric sea, 161, 59–88. https://doi.org/10.1016/S0009-2541(99)00081-9. Upper Mississippi Valley, USA. Geology 31, 545–548. https:// Veizer J., Bruckschen P., Pawellek F., Diener A., Podlaha O. G., doi.org/10.1130/0091-7613(2003)0310545:AOAEIA2.0.CO;2. Carden G. A. F., Jasper T., Korte C., Strauss H., Azmy K. and Smith G. L. and Clark D. L. (1996) Condonts of the Lower Ala D. (1997) Oxygen isotope evolution of Phanerozoic Ordovician Prairie du Chien Group of Wisconsin and Min- seawater. Palaeogeogr. Palaeoclimatol. Palaeoecol. 132, 159– nesota. Micropaleontology 42, 363. https://doi.org/10.2307/ 172. https://doi.org/10.1016/S0031-0182(97)00052-7. 1485958. Veizer J. and Prokoph A. (2015) Temperatures and oxygen isotopic Stolper D. A. and Eiler J. M. (2015) The kinetics of solid-state composition of Phanerozoic oceans. Earth-Sci. Rev. 146, 92– isotope-exchange reactions for clumped isotopes: a study of 104. https://doi.org/10.1016/j.earscirev.2015.03.008. inorganic calcites and apatites from natural and experimental Wacker U., Fiebig J., To¨dter J., Scho¨ne B. R., Bahr A., Friedrich samples. Am. J. Sci. 315, 363–411. https://doi.org/10.2475/ O., Tu¨tken T., Gischler E. and Joachimski M. M. (2014) 05.2015.01. Empirical calibration of the clumped isotope paleothermometer Tang J., Dietzel M., Fernandez A., Tripati A. K. and Rosenheim using calcites of various origins. Geochim. Cosmochim. Acta B. E. (2014) Evaluation of kinetic eﬀects on clumped isotope 141, 127–144. https://doi.org/10.1016/j.gca.2014.06.004.

fractionation (D47) during inorganic calcite precipitation. Wacker U., Rutz T., Lo¨ffler N., Conrad A. C., Tu¨tken T., Bo¨ttcher Geochim. Cosmochim. Acta 134, 120–136. https://doi.org/ M. E. and Fiebig J. (2016) Clumped isotope thermometry of 10.1016/j.gca.2014.03.005. carbonate-bearing apatite: revised sample pre-treatment, acid Tolmacheva T., Egerquist E., Meidla T., Tinn O. and Holmer L. digestion, and temperature calibration. Chem. Geol. 443, 97– (2003) Faunal composition and dynamics in unconsolidated 110. https://doi.org/10.1016/j.chemgeo.2016.09.009. sediments: a case study from the Middle Ordovician of the East Wenzel B., Lećuyer C. and Joachimski M. M. (2000) Comparing Baltic. Geol. Mag. 140, 31–44. https://doi.org/10.1017/ oxygen isotope records of silurian calcite and phosphate—d18O S001675680200701X. compositions of brachiopods and conodonts. Geochim. Cos- Trotter J. A., Williams I. S., Barnes C. R., Lećuyer C. and Nicoll mochim. Acta 64, 1859–1872. R. S. (2008) Did cooling oceans trigger Ordovician biodiver- Westrop S. R. (1996) Temporal persistence and stability of sification? Evidence from conodont thermometry. Science 321, Cambrian biofacies: Sunwaptan (Upper Cambrian) trilobite 550–554. https://doi.org/10.1126/science.1155814. faunas of North America. Palaeogeogr. Palaeoclimatol. Turchyn A. V., Alt J. C., Brown S. T., DePaolo D. J., Coggon R. Palaeoecol. 127, 33–46. https://doi.org/10.1016/S0031-0182 M., Chi G., Be´dard J. H. and Skulski T. (2013) Reconstructing (96)00087-9. the oxygen isotope composition of late Cambrian and Creta- ceous hydrothermal vent fluid. Geochim. Cosmochim. Acta 123, Associate editor: F. McDermott 440–458. https://doi.org/10.1016/j.gca.2013.08.015. Upchurch G. R., Kiehl J., Shields C., Scherer J. and Scotese C. (2015) Latitudinal temperature gradients and high-latitude