Silurian strontium isotope stratigraphy Karem Azmy* Ottawa-Carleton Geoscience Centre, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada Jan Veizer Ottawa-Carleton Geoscience Centre, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada, and Institut für Geologie, Ruhr Universität, 44780 Bochum, Germany Bernd Wenzel Institut für Geologie der Universität Erlangen-Nürnberg, 91054 Erlangen, Germany Michael G. Bassett Department of Geology, National Museum of Wales, Cardiff CF1 3NP, United Kingdom Paul Copper Department of Earth Sciences, Laurentian University, Sudbury, Ontario P3E 2C6, Canada ABSTRACT Burke et al., 1982; Veizer, 1989; McArthur, 1994), condition is that the temporal 87Sr/86Sr trends are to understand continental weathering processes characterized by steep slopes. This was the case A sample set of 164 calcitic brachiopod and mid-oceanic ridge hydrothermal circulation for several intervals of Phanerozoic time, and shells, covering the entire Silurian Period (~ 30 (cf. Hodell et al., 1990; Richter et al., 1992; Farrell particularly in Cenozoic time (e.g., Mead and m.y.) with a resolution of about 0.7 m.y., was et al., 1995), and to correlate and date marine sed- Hodell, 1995). collected from stratotype sections at Anticosti imentary rocks (e.g., Elderfield, 1986; Quinn et al., The main objectives of this study are to (1) re- Island (Canada), Wales (United Kingdom), 1991; McArthur, 1994). The dominant driving fine the Sr-isotope curve for the Silurian seawa- Gotland (Sweden), Podolia (Ukraine), Latvia, forces causing the changes in seawater isotopic ra- ter, (2) utilize such a refined curve for high-reso- and Lithuania. They show 87Sr/86Sr values tios are suggested to be (1) continental runoff and lution stratigraphic correlation, and (3) improve ranging from 0.707930 to 0.708792 that pro- ground-water runout, both of which supply radi- understanding of geochemical cycling for Sr dur- gressively increase with time. This may indicate ogenic strontium to the oceans, and (2) sea- ing Silurian time. an increasing riverine flux of radiogenic Sr into water–oceanic crust interaction, particularly hy- the ocean from weathering of continental sialic drothermal rift–related activities, supplying a GEOLOGICAL SETTING rocks due to progressive warming of the cli- less-radiogenic strontium (Palmer and Elderfield, mate. Exceptionally high increases in 87Sr/86Sr 1985). Other factors, such as diagenetic flux and The samples for this study were selected from values were observed in early Llandovery carbonate recycling, may account for a minor con- diverse depositional settings on different paleo- (Rhuddanian), late Llandovery (Telychian), tribution (Elderfield, 1986; Veizer, 1989). continents. Paleogeographic reconstructions (cf. and late Ludlow (Gorstian-Ludfordian bound- Low-Mg calcite brachiopod shells, particularly McKerrow et al., 1991) place all these basins ary) samples. Partial linear regressions, based if nonluminescent, have been documented to fre- within the tropical paleolatitudes and they in- on a stepwise climbing pattern, with local drops quently retain the primary Sr-isotope signals of clude Anticosti Island, Québec, Canada (Lauren- around the Llandovery-Wenlock boundary ambient seawater (Popp et al., 1986; Banner and tia), England, Sweden, Lithuania, Latvia, and and in latest Ludlow time, were used to esti- Kaufman, 1994; Diener et al., 1996). When bio- Podolia in the Ukraine (Baltica). The lithology mate relative ages with a resolution of about ±2 genic marine carbonate forms, the 87Sr/86Sr of of the studied sequences comprised mainly lime- biozones (~1.5–2 m.y.). The Sr-isotope curve ocean water is incorporated into its structure with- stones of shallow shelf environments, frequently shows distinct inflection points in earliest Wen- out fractionation. Oceanic uniformity of 87Sr/86Sr associated with reefs. The stratigraphic assign- lock and mid-PÍídolí time. These may be used at any given time is expected, because the resi- ment of these sections (Fig. 1) follows the global to correlate the Llandovery-Wenlock boundary dence time of Sr in the oceans (~ 106 yr) is much Silurian standard time scale. For further details in the United Kingdom, Gotland, and Lithua- longer than the time it takes for currents to mix the of geology and samples see Azmy (1996), Azmy nia, and the Kaugatuma-Ohesaare boundary oceans (Faure, 1986). However, for highly strati- et al. (1998), Wenzel and Joachimski (1996), and in the Baltic states and Podolia. fied oceans, the response may be different due to Wenzel (1997). possible mixing rates of bottom waters approach- INTRODUCTION ing the residence times for Sr in seawater METHODOLOGY (McArthur, 1994). Variations in seawater 87Sr/86Sr over geologic Variations in 87Sr/86Sr composition of past sea- The selected brachiopods were identified and time, particularly for Phanerozoic time, have been water, resolvable on a short-time scale of 107 to two identical slabs (~1.5 mm thick) were cut lon- used to reconstruct the evolutionary history of an- 106 yr, may be utilized for high-resolution strati- gitudinally through the umbo zone, using an cient seawater (e.g., Veizer and Compston, 1974; graphic correlations with an accuracy compara- ISOMET low-speed saw. The slabs were gently ble to, and perhaps higher than, that of biostratig- polished on a glass plate using Al2O3 powder *E-mail: [email protected]. raphy, the latter usually being 1 to 5 m.y. The (size 9.5 µm). A thin section of the sample, made Data Repository item 9933 contains additional material related to this article. GSA Bulletin; April 1999; v. 111; no. 4; p. 475–483; 7 figures; 3 tables. 475 AZMY ET AL. Graptolite Conodont Anticosti Is. Wales Gotland Podolia Kolka 54 Biozones Biozones CANADA BRITAIN SWEDEN UKRAINE LATVIA LITHUANIA 408.5 Ma 41 transgrediens ~ 40perneri detorta Dzwinogorod Ohesaare Jura 39bouceki 38lochkovensis ^ 37pridoliensis Rashkov Kaugatuma Minija PRIDOLI 36ultimus eosteinhornensis 411 Ma 35parultimus Sundre Prigorodok 34balticus/codatus Hamra 33kozlowski snajdri Isakovtsy 32inexpectatus 31auriculatus 30 cornutus Burgsvik siluricus Grinchuk 29bohemicus Ludfordian 415Ma 28 leintwardinensis Eke ploeckensis LUDLOW 27 hemiaversus Sokol 26invertus Hemse 25 scanicus 24 progenitor Konovka Gorstian crassa 424 Ma 23 nilssoni Much Wenlock Klinteberg G 22 ludensis Limestone 21 nassa stauros Halla Mulde Gèluva W 20 lundgreni Homerian Slite 19 ellesae 18 flexilis amsdeni Coalbrookdale 17 rigidus Tofta Riga SILURIAN 16 riccartonensis WENLOCK Hogklint 15 murchisoni ranuliformis 14 430.5 Ma Sheinwoodian centrifugus Buildwas U. Visby 13 crenulata amorphognathoides Chicotte L. Visby 12 griestoniensis 11 crispus celloni 10 turriculatus Jupiter Telychian 9 sedgwickii 8 convolutus staurognathoides 7 leptotheca 6 magnus Gun River 5 Aeronian triangulatus kentuckensis LLANDOVERY 4 cyphus Merrimack 3 acinaces 2 atavus nathani Becscie 1 acuminatus 439 Ma Rhuddanian ORDOVICIAN Ellis Bay Figure 1. Sampled Silurian sections and their stratigraphic assignments (modified from Basset et al., 1989; Siveter et al., 1989; Jin et al., 1990; Kaminskas and Musteikis, 1994; Long and Copper, 1994). from one of the slabs, was studied under a polar- at the University of Ottawa, to test for shell chem- on a tantalum filament. The strontium isotope ra- izing microscope to examine the preservation of ical preservation (Azmy, 1996). tio was measured using the Finnigan MAT 261 the calcite fibers. The thin section and the other For Sr-isotope analysis, about 1 mg of the multicollector thermal ionization mass spectro- polished slab were viewed under cathodolumi- powdered sample was dissolved for 30 min in 1.5 meter at Carleton University. The laboratory stan- nescence, with the operating conditions at ~10 to ml of 2.5N suprapure HCl at room temperature, dard used was NBS 987 (87Sr/86Sr = 0.710249) 11 kv, gun current of 350 to 400 mA, and vacuum and Sr was extracted via a clean 10 ml column with a (2σ) precision calculated from 30 mea- of ~ 0.03 Torr. filled with Dowex AG50-X8 cation resin. The surements of ± 0.000017. The blanks were 0.4 to Carbonate material from the nonluminescent eluent was dried at 125 °C for 2 hr. The dried 0.8 ng. The measured Sr isotope data are listed in parts of the secondary layer was microsampled sample was dissolved in 0.01N HCl for a few Table DR1, GSA Data Repository.1 For further from the slab under a binocular microscope by minutes and passed through a clean 10 ml sepa- details of geology, samples, and analytical and se- smashing the shell with a stainless steel dental ration column filled with Teflon resin to trap Ca lection procedures, see Azmy (1996) and Azmy pick. The fragments were cleaned in an ultra- that may not have been separated from Sr by first et al. (1998). sonic bath. column. The collected sample was evaporated at Another set of brachiopod samples, from Got- A fragment from each sample was studied un- 125 °C for 2 hr to be ready for running on the land, was prepared and measured independently at der a scanning electron microscope to examine mass spectrometer. Blank samples were fre- the laboratory of Ruhr Universität in Bochum fol- the preservation of the calcite crystals. The rest of quently run and spiked using 84Sr to measure any the sample was ground in an acid-washed agate contamination that might occur during the 1GSA Data Repository item 9933, Table DR1, is mortar and ~ 3 mg were used for trace element process of separation. available on request from Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301. E-mail: analysis by a Thermo Jarrell Ash-AtomScan 25 The sample was dissolved in 0.4 ml of 1M [email protected]. Web: http://www.geosociety inductively coupled plasma source spectrometer, H3PO4 for 2 min and about half of it was loaded .org/pubs/ftpyrs.htm. 476 Geological Society of America Bulletin, April 1999 SILURIAN STRONTIUM ISOTOPE STRATIGRAPHY Graptolite 87 86 Biozones Sr/ Sr 0.7079 0.7081 0.7083 0.7085 0.7087 0.7089 0.7091 408.5 Ma 37 Í P ídolí 410.7Ma 31 Ludfordian Ludlow 25 Gorstian 424Ma Homerian 19 SILURIAN Wenlock 430.4 Ma Sheinwoodian 13 Telychian 7 This study Aeronian Bertram et al.