A Chemostratigraphic Investigation of the Late Ordovician Greenhouse to Icehouse Transition: Oceanographic, Climatic, and Tectonic Implications

Home , Allen Young, Belonechitina, Cryptospore, Hanson Creek Formation, Reedsville Formation, Viola Formation

A CHEMOSTRATIGRAPHIC INVESTIGATION OF THE LATE ORDOVICIAN GREENHOUSE TO ICEHOUSE TRANSITION: OCEANOGRAPHIC, CLIMATIC, AND TECTONIC IMPLICATIONS

DISSERTATION

Presented in Partial Fulfillment of the Requirements

For the Degree Doctor of Philosophy in the

Graduate School of The Ohio State University

Seth Allen Young, M.S.

*****

The Ohio State University 2008

Dissertation Committee:

Dr. Matthew R. Saltzman, Adviser

Dr. Kenneth A. Foland

Dr. William I. Ausich

Dr. Andrea G. Grottoli

ABSTRACT

The latest Ordovician (444 million years ago) was a critical period in Earth history. This was a time of significant climatic global change with large-scale continental glaciation. Moreover, the end-Ordovician mass extinction is recognized as the second- most devastating mass extinction to have affected the Earth. The anomalous Late

Ordovician icehouse period has perplexed many researchers because all previous model and proxy climate evidence suggest high levels of atmospheric CO2 during the Late

Ordovician glaciation. Also associated with this period is a large positive carbon isotope

(δ13C) excursion (up to +7‰) that represents a global perturbation of the carbon cycle.

Additionally, a large decrease (0.001) in seawater 87Sr/86Sr occurs several million years prior (~460 million years ago); this could reflect an increase in atmospheric CO2 uptake due to weathering of volcanic rocks involved in uplift of the early Appalachian

Mountains.

To address these Ordovician anomalies, well-studied, thick, and continuous Late

Ordovician limestone sequences from eastern West Virginia, south-central Oklahoma, central Nevada, Quebec (Canada), Estonia, and China have been sampled. Carbon and strontium isotopic ratios have been measured on samples from these localities of which

Estonian and Chinese sample sites represent separate paleocontinents (Baltica and South ii China) and are compared with other data sets from North America. These data test previous interpretations that the well-documented latest Ordovician carbon isotope excursion coincides with maximum glaciation. They support a hypothesis that the large positive carbonate carbon isotope excursion was coincident with a warm interglacial

(high CO2 levels) period that separated two major glacial advances (with lowered CO2 levels). There are clear parallels between the Late Ordovician and the Late Cenozoic (the most recent) greenhouse to icehouse transitions, with silicate weathering providing the initiator and positive feedback on changes in atmospheric CO2 levels. The results lead to a more complete understanding of climatic and biotic events of this critical interval, which will certainly help the understanding of the period of global climatic and biotic change affecting Earth today.

iii ACKNOWLEDGMENTS

I would like to thank my advisor, Dr. Matthew R. Saltzman and for all of his advice, stimulating discussions, encouragement, and supervision throughout the duration of my doctoral degree. His time and involvement was given freely, and I am very grateful for his guidance. Thanks to Dr. Kenneth A. Foland, Dr. William I. Ausich, and Dr.

Andrea G. Grottoli for serving on my committee and their support and advice throughout the dissertation process. I also thank Dr. Stig M. Bergström for his very helpful support, guidance, discussions, and collaborations over the last four years. I am grateful for stimulating discussions, assistance in the field, and collaborations with Dr. Stephen A.

Leslie (James Madison University), Dr. Dimitri Kaljo (Tallinn Technical University), Dr.

Mats Eriksson (Lund University), Dr. Birger Schmitz (Lund University), Dr. Lee R.

Kump (Pennsylvania State University), Dr. Berry Lyons (The Ohio State University), and

Dr. Loren E. Babcock (The Ohio State University). Many thanks to laboratory technicians Jeff Linder, Yohei Matsui, and Frtiz Hubacher for their assistance with sample preparation and mass spectrometer operations in the Radiogenic Isotopes and

Stable Isotope Biogeochemistry Laboratories here at The Ohio State University. I thank my group members and the following graduate students at OSU for useful discussions, field and lab assistance: Alyssa Bancroft, Brad Cramer, Abbey Chrystal, Scott Drew,

iv Steve Goldsmith, Jeremy Gouldey, Alex Lin, Ryan Moyer, Alexa Sedlacek, Kate

Tierney, and Branwen Williams. I would like to express sincere thanks to my fiancé

Jessica R. Lucas for her unyielding academic, personal, and financial support of me throughout this process. I also thank my parents Deborah Poe and Doyle Young for their encouragement and support during my life as a graduate student.

My dissertation research was funded by grants from Friends of Orton Hall Fund

(OSU), Office of International Affairs Grant (OSU), Alumni Research Grant (OSU), and a GraduateStudent Research Grant from The Geological Society of America. Additional funding was provided by the National Science Foundation Grants to M.R. Saltzman, K.

A. Foland, and W. I. Ausich. Thanks to the School of Earth Sciences for providing me with a Graduate Teaching Assistantship through the first three years of my graduate studies. The last year of my Ph.D. program I was supported by a University Presidential

Fellowship from The Ohio State University.

v VITA

March 27, 1978 ……………………………………..Born-Little Rock, Arkansas

Fall, 2000………….………….………….………….B.S. Geology, University of Arkansas at Little Rock

Fall, 2003…………...…………...…………...………M.S. Geological Sciences, The Ohio State University

Fall, 2001-Fall, 2003……………..………………….Graduate Teaching Associate, The Ohio State University (M.S.)

Winter, 2004-Spring, 2007…………...……………...Graduate Teaching Associate, The Ohio State University (Ph.D.)

Summer, 2007-present……………..………………...Presidential Fellowship, The Graduate School, The Ohio State University (Ph.D.)

PUBLICATIONS

Research Publications

Young, S.A., Saltzman, M.R., Ausich, W.I., & Kaljo, D., 2007, A global change in δ13C of organic matter during the Late Ordovician (Hirnantian): Implications for

atmospheric CO2 levels and glaciation. Geological Society of America, Abstracts with Programs, 39 (6): 84.

Bergström, S.M., Xu, Chen, Young, S.A., Schmitz, B., & Saltzman, M.R., 2007, The First Record of The Ordovician Guttenberg δ13C excursion (GICE) in Asia: Chemostratigraphy of the Pagoda Limestone and Yanwashan Formations in South-eastern China. Geological Society of America, Abstracts with Programs, 39 (6): 144.

Saltzman, M. R., Young, S.A., Gill, B., Lyons, T., Kump, L., & Runnegar, B., 2007, Increased Carbon Isotopic Fractionation during the Late Cambrian SPICE Event

vi and a Pulse of Atmospheric Oxygen. Geological Society of America, Abstracts with Programs, 39 (6): 24.

Bergström, S. M., Young, S.A., Schmitz, B., Barta, N., & Saltzman, M.R., 2007, Katian and Hirnantian (Upper Ordovician) δ13C chemostratigraphy in North America and Baltoscandia: A trans-Atlantic comparison. WOGOGOB 2007. SGU Rapporter och meddelanden . No. 128, 80-81.

Goldman, D., Leslie, S. A., Nõlvak, J., Young, S., Bergström, S.M., and Huff, W.D., 2007, The Global Stratotype Section and Point (GSSP) for the base of the Katian Stage of the Upper Ordovician Series at Black Knob Ridge, Southeastern Oklahoma, USA. Episodes, v. 30, p. 258-270

Bergström, S. M., Young, S.A., Schmitz, and Saltzman, M.R., 2007, Upper Ordovician δ13C chemostratigraphy: A trans-Atlantic comparison. Acta Palaeontologica Sinica 46: 37-39.

Young, S.A., Saltzman, M.R., Bergström, S.M., 2005, Upper Ordovician (Mohawkian) carbon isotope (δ13C) stratigraphy in eastern and central North America: Regional expression a perturbation of the global carbon cycle. Palaeogeography, Palaeoclimatology, Palaeoecology 222: 53-76.

Young, S. A., Saltzman, M. R., Foland, K. A., Linder, J.S., 2006, A large, rapid 87Sr/86Sr drop across the Middle to Late Ordovician boundary: Implications for paleoceanography & weathering of the Taconic highlands. Geological Society of America, Abstracts with Programs, 38 (7): 478.

Bergström, S. M., Young, S.A., Schmitz, B., & Saltzman, M.R., 2006, A detailed δ13C curve through the Cincinnatian Series (Upper Ordovician) in its type area: significance for regional chemostratigraphic correlations. Geological Society of America, Abstracts with Programs, 38 (7): 146.

Young, S.A., Bergström, S. M., Schmitz, B., & Saltzman, M.R., 2006, Trans-atlantic correlation of Upper Ordovician δ13C excursions. Geological Society of America, Abstracts with Programs, 38 (4): 73.

Saltzman, M.R., Young, S.A., 2005, Long-lived glaciation in the Late Ordovician? Isotopic and sequence-stratigraphic evidence from western Laurentia. Geology 33 (2): 109-112. vii

Young, S.A., Saltzman, M.R., Ausich, W.I., 2005, Hirnantian-Llandovery (uppermost Ordovician-lowermost Silurian) carbon isotope (δ13C) stratigraphy from Anticosti Island, Quebec: implications for oceanography, glaciation, and organic carbon burial. The Dynamic Silurian Earth: Subcommision on Silurian Stratigraphy Field Meeting 2005. SGU Rapporter och meddelanden . No. 121, 97-98.

Young, S.A., Saltzman, M.R., Ausich, W.I., 2005, Richmondian-Hirnantian (Upper Ordovician) carbon isotope (δ13C) stratigraphy from Anticosti Island, Quebec: implications for oceanography, glaciation, and organic carbon burial. Abstracts for the Second International Symposium of IGCP503 on Ordovician Palaeogeography and Palaeoclimate. Insight, A Milwaukee Public Museum Series in Natural History. (2) 40.

Young, S.A., 2004, Carbon isotope stratigraphy; a gateway to the past, present, and future. Sedimentary Record 2 (2): 11 p.

Young, S. A., Saltzman, M. R., Foland, K. A., Linder, J.S., 2004, A large seawater 87Sr/86Sr drop (~0.0008) in the Middle to Late Ordovician (Late Whiterockian): Evidence from central Nevada. Geological Society of America Abstracts with Programs,160 p. Young, S. A., Saltzman, M. R., Bergström, S. M., Holmden, C., and Patterson, W. P., 2003, Paleooceanographic Aspects of the Early Chatfieldian (Upper Middle Ordovician) Positive δ13C Excursion (GICE). Geological Society of America, Abstracts with Programs, 43 (7): 600.

Saltzman, M. R., Young, S. A., Bergström, S. M., Holmden, C., and Patterson, W. P, 2003, Age and Significance of the Sequence Boundary at the Base of the Eureka Quartzite in Central Nevada. Geological Society of America, Abstracts with Programs, 43 (7): 473.

Young, S. A., Bergström, S. M., and Saltzman, M. R, 2003, Regional Carbon Isotope (δ13C) Stratigraphy of Chatfieldian (Upper Middle Ordovician) Carbonates in Central and Eastern North America. Geological Society of America, Abstracts with Programs, 35 (2): 5.

Young, S.A., 2000, Upper Middle Ordovician conodont biostratigraphy of the upper

viii Carters limestone and lower Hermitage formation in central Tennessee. Geological Society of America, Abstracts with Programs, 32(3): A-45.

Young, S.A. & Leslie, S.A.. 1999, The search for K-bentonite beds in Arkansas and their potential for regional correlation. Proceeding Journal of the Arkansas Undergraduate Research Conference and Space Grant Symposium: 231-234.

Young, S.A. and S.A. Leslie, 1999. The search for K-bentonite beds in Arkansas and their potential for regional correlation. Program and Abstracts of the Arkansas Undergraduate Research Conference and Space Grant Symposium, p. 82.

FIELDS OF STUDY

Major Field: Geological Sciences

Studies in carbon and strontium isotope chemostratigraphy, paleoceanography, paleoclimatology, biostratigraphy, stratigraphy, paleobiology, and sedimentology.

ix TABLE OF CONTENTS

Page Abstract……………………………………………………………………………………ii

Acknowledgements……………………………………………………………………….iv

Vita………………………………………………………………………………………..vi

List of Tables……………………………………………………………………………xiii

List of Figures……………………………………………………………………….…..xiv

Chapters:

1. Introduction……………………………………………………………………...... 1

Background……………………………………………………………...... 1 Organization of the dissertation……………….………………………...... 5

2. A Volcanic-Weathering Hypothesis for the Late Ordovician Ice Age…………..10

Abstract.……………………………………………………………...... 10 Introduction ……………………………………………………………...10 Background and Geologic Setting.……………………………………....12 Methods and Results……………………………………………………..12 Discussion………………………………………………………………..15 Controls on Middle to Late Ordovician seawater 87Sr/86Sr……...16 Weathering of juvenile volcanic rocks…………………………..17 Implications and Conclusions…………………………………………....18

13 13 3. Paired δ Ccarb and δ Corg records of Upper Ordovican (Sandbian-Katian) carbonates: Implications for paleoceanographic and environmental change…………………………………………...………………...……………..29

x Abstract……………………………………………………………...... 29 Introduction……………………...…………………………………….....29 Geologic Background……………….…………………………………...31 Methods…………….……………….…………………………………....33 Results……………………………….…………………………………...34 Puxihe, Hubei Province, southeastern China…………………….34 Fittstown, Arbuckle Mountains, southern Oklahoma……………35 Dolly Ridge, Central Appalachians, West Virginia……………...36 Discussion…………..……………….…………………………………...37 13 Factors affecting the isotopic difference between δ Ccarb and 13 δ Corg………..……………….…………………………………..38 13 Photosynthetic fractionation εp and its influence on δ C trends through the GICE……...... 41 εp in the Pennsylvania reference section…………………42 South China.……...………………….…………………...43 Oklahoma.………………………………………………..43 West Virginia…..………………….……………………..44 Implications and Conclusions………………………………………...... 45

4. A Global Change in δ13C of Organic Matter During the Late Ordovician Glaciation: Implications for Atmospheric Composition………………………....66

Abstract…..…………………………………………………………...... 66 Introduction…………………………………………………………...... 67 Background and Geologic Setting……………………………………….68 Methods and Results…………….……………….………………………69 Discussion…………..……………….…………………………………...72 13 13 Paired analysis of δ Ccarb and δ Corg……………………………72 The timing of the Hirnantian glaciation………………………….73 Seawater 87Sr/86Sr and climate………………….………………..75 Implications and Conclusions…………………………………………....76

5. Conclusions……...…………………………………………………………….....96

References………………………………………………………………………………100

Appendicies……………………………………………………………………………..115 xi Appendix A Additional Strontium Isotope Data used in Chapter 4………….115 Appendix B Additional Stable Carbon Isotope Data used in Chapter 4……..119 13 13 18 Appendix C Cross plots of δ Ccarb, δ Corg, δ O, and total organic carbon (TOC) data from Chapters 3 and 4………...... 123

xii LIST OF TABLES

Table Page

2.1 Data table for 87Sr/86Sr values used in Figures 2.2 through 2.5..…………………27

2.2 Seawater Sr Model Parameters.………...... 28

3.1 Stable carbon isotope data from Fittstown, Oklahoma section..…………………62

3.2 Stable carbon isotope data from the Dolly Ridge, West Virginia section.………63

3.3 Stable carbon isotope data from the Puxihe, China section……...………………65

4.1 Stable carbon isotope data from Anticosti Island, Quebec, Canada.………….....93

4.2 Stable carbon isotope data from the Kardla drill core, Estonia……………….....95

A.1 Additional Strontium Isotope data from central Nevada, USA.………………..116

A.2 Additional Strontium Isotope data from the Kardla drill core, Estonia………...117

A.3 Additional Strontium Isotope data from the Anticosti Island Canada …………118

B.1 Additional stable data from the Anticosti Island Canada ……………………...120

xiii LIST OF FIGURES

Figure Page

1.1 Modeled atmospheric CO2 levels throughout the Phanerozoic and major glacial and/or cool climate intervals of time………………………………………6

1.2 Generalized curve showing the relationship between Δ13C and atmospheric pCO2 levels ……………………………………………………………………….7

1.3 The Strontium cycle……………………………………………………………….8

1.4 Middle and Upper Ordovician stratigraphic correlation chart between the diagnostic graptolite, conodont zones, and global and North American Series and Stage Names …………………………………………………………………9

2.1 Middle to Late Ordovician paleogeographic map……………………………….20

2.2 Plot of 87Sr/86Sr vs. Sr concentration…………………………………………….21

2.3 Plot of 87Sr/86Sr vs. δ18O (‰ VPDB)..…………………………………………...22

2.4 Generalized 87Sr/86Sr curve and lithostratigrahic units from Nevada (Fig. 2.2) replotted on Figure 8 of Shields et al. (2003)…………………………23

2.5 Plots of 87Sr/86Sr, δ13C, and a simplified stratigraphic column for the Ordovician Antelope-Monitor Range composite……………....……………..…24

2.6 Calculated Middle to Late Ordovician seawater 87Sr/86Sr curve plotted along with actual measured values.……………....…………………………………….25

2.7 Tectonic cross sections through Laurentia in the Ordovician showing different fluxes of Sr to the oceans that produced the large seawater 87Sr/86Sr drop recorded in central Nevada)…………………...………………………...….26

3.1 Katian (Upper Ordovician) paleogeographic reconstruction…………………….47

xiv 3.2 Geologic Map of south-central China, near Yichang City in Hubei Province. Marked on this map is the Puxihe Quarry sample locality (triangle), northeast of Yichang City…………………….…………………………………………….48

3.3 Paired carbon isotope data from the Puxihe Quarry, near Yichang, Hubei Province, China, with diagnostic conodonts and graptolites ranges plotted for the Pagoda and Miaopo Shale Formations plotted ……...…………………...50

3.4 A, Map of south-central North America showing state outlines and Ordovician outcrops in Oklahoma (OK). B, Inset map showing the location of the roadcut along Highway 99……………………………...…………………51

3.5 Paired carbon isotope data from the Fittstown section, Arbuckle Mountains, Oklahoma…………………………………………………..………..53

3.6 Map of eastern North America showing Ordovician outcrops in the southern and central Appalachian Basin Region and state outlines. B, Inset map showing the location of the type-section for the Dolly Ridge Formation ……………………………...………………………………………...54

3.7 Paired carbon isotope data from the Dolly Ridge section, central Appalachians, West Virginia …………………………………………………... 56

13 3.8 Comparison of three point running average δ Corg curves from our three sections presented here, and the two previously studied sections from Iowa (SS-9 drill core) and Pennsylvania (roadcut near Reedsville) ……...…………...57

3.9 Comparison of three point running average Δ13C curves from Oklahoma, south China, and the previously studied Pennsylvania section.……..58

3.10 Generalized relationship between the photosynthetic fractionation effect (εp), atmospheric CO2 levels, and modern marine phytoplankton growth rates (µm/day) ……...…………...... 59

3.11 Comparison of three point running average Δ13C curves from West Virginia, south China, and the previously studied Pennsylvania section ……...…………..60

xv 3.12 Generalized relationship between the photosynthetic fractionation effect (εp), atmospheric CO2 levels, and modern marine phytoplankton growth rates (µm/day)..……...…………...... 61

4.1 Latest Ordovician (Hirnantian) paleogeographic map…………………………...78

13 13 4.2 Upper Ordovician to Lower Silurian δ Ccarb and δ Corg trends for the Kardla drill core, Estonia …………………………...... 80

13 13 4.3 A composite Upper Ordovician to Lower Silurian δ Ccarb and δ Corg trends for the Vaureal, Ellis Bay, and lower Becscie Formations on Anticosti Island, Quebec, Canada. …………………………...... 82

13 4.4 Correlation of Upper Ordovician to Silurian δ Ccarb (filled circles) 13 and δ Corg (open circles) trends from Estonia to a section at Point Laframboise, western Anticosti Island, Quebec. …………………………...... 84

13 4.5 Correlation of Upper Ordovician to Silurian δ Ccarb (filled circles) 13 and δ Corg (open circles) trends from Estonia to a section at Point Laframboise, western Anticosti Island, Quebec. …………………………...... 86

87 86 13 4.6 Upper Ordovician to Lower Silurian Sr/ Sr and δ Ccarb trends for the Kardla drill core, Estonia. …………………………...... 88

87 86 13 4.7 Upper Ordovician to Lower Silurian Sr/ Sr and δ Ccarb trends for the surface exposures of Vaureal and Ellis Bay Formations on Anticosti Island, Quebec, Canada.…………………………...... 90

87 86 13 4.8 Correlation of Upper Ordovician to Lower Silurian Sr/ Sr and δ Ccarb trends from western Laurentia (Nevada composite) to eastern Laurentia (Anticosti Island composite) to Baltica (Kardla drill core, Estonia……………...92

13 C.1 Cross plot of δ Corg versus total organic carbon (TOC) data from Oklahoma (circles), West Virginia (diamonds), and China (triangles) in Chapter 3…………….…………………………...... 124

13 18 C.2 Cross plot of δ Ccarb versus δ O data from Anticosti Island (circles) and Estonia (diamonds) in Chapter 4 and Appendix B….…………….……………125

xvi 13 C.3 Cross plot of δ Corg versus total organic carbon (TOC) data from Anticosti Island (circles) and Estonia (diamonds) in Chapter 4 and Appendix B….……..126

xvii

CHAPTER 1

INTRODUCTION

Background

The Late Ordovician (~ 444 million years ago) was one of three time periods during the past half billion years in which large continental glaciers formed over Earth’s polar regions (Fig. 1.1) (Crowell, 1999). The effects of glaciation were far-reaching and resulted in one of the largest marine mass extinction events in Earth history during the

Late Ordovician (Sheehan, 2001). The cause of the ice age is uncertain, and a paradoxical association with high atmospheric CO2 levels (pCO2) has previously been inferred from models and proxy evidence that all suggest levels for the early Paleozoic at 14-16 times present atmospheric levels (PAL) (Yapp and Poths, 1992; Berner and Kothavala, 2001).

This inconsistency has been puzzling, because fluctuations in atmospheric CO2 are commonly taken to be the main driver of climate change on geologic timescales (e.g.,

Berner, 1990). Although recent work linking galactic cosmic ray fluxes to climatic

variability (celestial driver for Phanerozoic climate) has challenged this atmospheric CO2 driver assumption (e.g., Shaviv and Veizer, 2003). However, the precise link between pCO2 and ice sheet growth remain poorly known because of uncertainty in global time correlation of strata recording geochemical proxy evidence with those containing

physical records of glaciation (sea level changes, tillites) (Finney et al., 1999; Kump et al., 1999; Brenchley et al., 2003).

Additionally, the cause(s) of the atmospheric CO2 drawdown during the

Ordovician remain uncertain, despite the fact that processes that control atmospheric

CO2levels on geologic timescales (e.g., silicate weathering, CO2 out-gassing from volcanoes and mid-ocean ridges, burial of carbonates) are well known (Berner, 1999).

One potential cause is enhanced organic carbon burial, previously proposed for the

Hirnantian δ13C excursion and associated glaciation (Brenchley et al., 1994). However,

13 during the Hirnantian δ C excursion, proxy evidence for fluctuating pCO2 levels (from

13 13 paired δ Corg and δ Ccarb analyses) indicates elevated pCO2 levels (Kump et al., 1999;

Young et al., in review).

A short-lived glaciation hypothesis (~1 m.y.) was proposed based largely on the correlation of a global sea-level fall with coincident positive shifts in stable isotopes

(δ13C, δ18O) in the latest Ordovician time (Brenchley et al., 1994). However, a more long- lived episode (Middle Ordovician through Early Silurian) of glaciation has been argued based on sedimentary evidence (e.g., Frakes et al., 1992). The research presented here,

tests the hypothesis that a lowering of atmospheric CO2 levels, which can ultimately lead to glaciation, continued from the Middle and Late Ordovician through the Silurian, and can be linked to important tectonic (e.g., mountain building) and oceanographic (e.g., changes in ocean circulation and sites of upwelling) events. New data presented provide a

13 13 87 86 series of continuous records of paired δ Corg and δ Ccarb analyses, along with Sr/ Sr

seawater curves, from both critical intervals that have no published curves and from previously published data from the Middle and Late Ordovician.

The important links between climate, tectonics, and oceanography in the Late

Ordovician are established from the construction of aforementioned stable (δ13C) and radiogenic (87Sr/86Sr) isotopic records. These isotopic records will be constructed from well biostratigraphically (e.g., conodonts, graptolites) constrained Middle and Upper

Ordovician marine carbonates. Analysis of both carbonate and organic carbon isotopes from the same samples provides the possibility of interpreting relative changes in

13 atmospheric pCO2. In addition, paired analysis of δ C also avoids the problem of correlation between carbonate and shale dominated sequences (Kump and Arthur, 1999).

13 13 13 The difference between the isotopic measurement of paired δ Corg and δ Ccarb (Δ C)

has been proposed as a paleo-atmospheric pCO2 proxy (Fig. 1.2) (Popp et al., 1989) because this difference results from isotopic discrimination, the photosynthetic isotope

effect, which decreases with lowering aqueous CO2 concentrations. Recent models have

calculated atmospheric pCO2 thresholds (at ≤ 8 times PAL) necessary to initiate glaciation of Gondwana from an Early Ordovician greenhouse climate (Herrmann et al.,

2003), and these thresholds are lower than previous estimates (Gibbs et al, 1997).

Previous efforts to constrain atmospheric pCO2 levels for the early Paleozoic were relatively coarse in resolution (~10 m.y. intervals) and may have missed an earlier Late

Ordovician drawdown (Kump et al., 1995).

A prolonged period of pCO2 drawdown in the Late Ordovician through Early

Silurian is consistent with the sedimentologic evidence for a long-lived Gondwana

glaciation (Grahn and Caputo, 1992; Pope and Read, 1997, 1998; Caputo, 1998;

Hamoumi, 1999; Pope and Steffan, 2003). Recent work on a global positive δ13C excursion during the Late Ordovician Chatfieldian Stage (~10 m.y. earlier than the

Hirnantian shift) has also indicated that CO2 levels may have dropped and glaciation began as a response to enhanced organic carbon burial during the carbon isotope excursion (e.g. Ainsaar et al., 1999; Saltzman and Young, 2005; Young et al., 2005).

Previous stable isotope studies suggest that variations in atmospheric pCO2 levels, enhanced organic carbon burial, and glaciation continued into the Silurian (e.g.

Samtleben et al., 1996; Bickert et al., 1997; Cramer and Saltzman, 2005).

87Sr/86Sr isotopic ratios of seawater can be used as proxies for paleotectonic events

(e.g., mountain building events, rapid sea-floor spreading rates) because these isotopic ratios are controlled by fluxes from different geologic processes (e.g., continental crust from river input, seafloor hydrothermal alteration) (e.g, Kump and Arthur, 1997).

Fluctuations in the 87Sr/86Sr isotopic ratios of well-preserved Ordovician carbonates will be used as a proxy record of Ordovician seawater change linked to paleotectonic events.

Because strontium has a long residence time in oceans (~4 m.y.) and because oceans mix rapidly (~103 years), the seawater strontium isotope record is globally homogeneous at any given time for the modern oceans (Faure, 1986). This has been documented by identical 87Sr/86Sr ratios for coeval marine carbonates throughout the Phanerozoic (e.g.

Veizer et al., 1989; Burke et al., 1982). There are two major sources for strontium in ocean waters (Fig. 1.3): 1) submarine hydrothermal exchange with mid-ocean ridge basalts (~ 0.703) (Spooner, 1976); and 2) weathering of exposed crustal rocks: (a) old

continental crust, which supplies the oceans with the more radiogenic end-member

(~0.711 or higher) (Faure, 1986), and (b) juvenile volcanic rocks, which supply the less radiogenic end-member (~ 0.704 or lower) (Goldstein and Jacobsen, 1987).

Organization of the dissertation

The remaining dissertation is divided into four chapters. Chapter 2-4 contain

87 86 13 13 Sr/ Sr and/or paired δ Corg and δ Ccarb analyses from three intervals of time spanning the Middle to Late Ordovician boundary (460.5 Ma) to the Ordovician-Silurian boundary

(443 Ma) (Fig. 1.4). This organization is to facilitate future publication, with the intent that Chapters 2, 3, and 4 will be submitted to professional peer-reviewed journals. For example, during the course of completing my dissertation research Chapters 2 and portions of Chapter 4 have been through several rounds of reviews in Geology and

Nature, ultimately not being accepted by either journal. However, these two chapters, in addition to Chapter 3 are now currently in review or in the process of being submitted to the following journals: Palaeogeography, Palaeoclimatology, Palaeoecology; Nature-

Geosciences; Earth and Planetary Science Letters; Geochimica et Cosmochimica Acta;

Chemical Geology. Although there are several coauthors in the submitted manuscripts for publication, only the manuscripts written by the author and published as the senior author are incorporated here in the dissertation.

Figure 1.1. Modeled atmospheric CO2 levels throughout the Phanerozoic (modified from Berner, 1990), and major glacial and/or cool climate intervals of time (Frakes et al.,

1992). Grey dashed line represents modern day atmospheric CO2 levels, and black dashed lines represent the potential error bars for modeled CO2 curve.

Figure 1.2. Generalized curve showing the relationship between Δ13C (calculated as the 13 13 isotopic difference between δ Ccarb and δ Corg) and atmospheric pCO2 levels (modified from Kump and Arthur, 1999). At higher atmospheric CO2 levels photosynthesizing organisms can be more selective of their carbon incorporation, and when 13 13 atmospheric CO2 levels are low they incorporate more C resulting in lower Δ C values. This relationship assumes equilibrium between the surface oceans and atmosphere, and ignores the effects of growth rates, temperature, and diagenesis on Δ13C (see discussion Chapter 3) (from Kump and Arthur, 1999).

7 Riverine Inputs: Sr flux: 3.34 X 1010 mol/yr (Richter et al., 1992)

Island Arc Input: Continental Input: 87Sr/86Sr ≤ 0.704 87Sr/86Sr ≥ 0.711 Volcanic Island Arc

Weathering of

8 continental crust

Hydrothermal Input: 10 Seafloor Sr flux: 0.35 X 10 mol/yr Subduction 87 86 Spreading Sr/ Sr ~0.703 (Veizer, 1989)

Figure 1.3. The Strontium cycle, showing the known fluxes and isotopic compositions to the global oceans (Davis et al., 2003; Richter et al., 1992; Stern, 1982; Veizer, 1989).

Figure 1.4. Middle and Upper Ordovician stratigraphic correlation chart between the diagnostic graptolite, conodont zones, and global and North American Series and Stage Names used in this study (after Webby et al., 2004). Also shown are the major tectonic phases that affected eastern North America, distribution of ash beds in North America (Ettensohn, 1990; Kolata et al., 1996). A relative sea-level curve modified from Ross and Ross (1995).

9 CHAPTER 2

A VOLCANIC-WEATHERING HYPOTHESIS FOR THE LATE ORDOVICIAN ICE AGE

Abstract

Previous work indicates that a lowering of pCO2 to initiate glaciation in the Late

Ordovician (~ 445 million years ago) may have been a response to increased weathering of mafic continental volcanic rocks. Seawater 87Sr/86Sr potentially provides a proxy for volcanic weathering, but a causal connection between a drop in 87Sr/86Sr and Late

Ordovician tectonic events remains poorly understood. Here we document the large and rapid decrease in Ordovician 87Sr/86Sr in a continuous, well-dated stratigraphic sequence from Nevada and link the decline to weathering of juvenile arc volcanics uplifted during the closure of the Iapetus (proto-Atlantic) ocean. Seawater 87Sr/86Sr stabilizes at a stratigraphic level marked by proxy evidence (e.g., ‘cool water’ carbonates, bedded cherts and phosphates, positive δ13C excursion) for global cooling. This is consistent with the notion that weathering of mafic volcanic rocks, recognized as an important contributor to global CO2 consumption today, also played a critical role in the transition to a Late Ordovician icehouse world.

Introduction

A large decrease in Phanerozoic seawater 87Sr/86Sr in the Middle to Late

Ordovician has previously been recognized (e.g., Burke et al., 1982; Qing et al., 1998;

Shields et al., 2003). This Ordovician decrease of ~ 0.001 in 87Sr/86Sr is comparable in magnitude to the well-known late Cenozoic 87Sr/86Sr increase that occurred over the past

~40 m.y. (e.g., Burke et al., 1982). The Ordovician has also been identified with widespread explosive volcanic activity (e.g., Kolata et al., 1996) from arc terranes that were present in low latitude settings along the eastern margins of Laurentia (Fig. 2.1;

Scotese and McKerrow, 1990). Stratigraphic sequences in the Appalachian foreland basin also record lithologic and isotopic evidence for rapid deepening and weathering associated with accretion of arc terranes (Ettensohn, 1990; Anderson and Samson, 1995).

These sequences in the Appalachian foreland basin can be correlated with passive margin sequences from Nevada (Fig. 16 of Harris et al., 1979).

Here we present new Sr isotope measurements from a stratigraphic sequence in central Nevada, where seawater 87Sr/86Sr began to decline during a eustatic fall (Ross and

Ross, 1995). A correlation with tectonic events during closure of the Iapetus ocean is consistent with the hypothesis that intense weathering of arc volcanics provided a large flux of non-radiogenic Sr to the global oceans and supports recent models emphasizing the role of weathering arc volcanic rocks in seawater 87Sr/86Sr fluctuations (e.g., Berner,

2006). Lowering of atmospheric pCO2 is predicted to accompany intense episodes of silicate weathering (see also Kump and Arthur, 1997; Dessert et al., 2003), and we evaluate this linkage in the Late Ordovician.

Background and Geologic Setting

A thick (~750 m) sequence of Middle and Upper Ordovician strata in central

Nevada represent a passive margin setting along the western margin of Laurentia (Ross et al., 1989; Finney et al., 1999) (Fig. 2.1). The Middle Ordovician (Whiterockian) succession is unique in recording a time period rarely preserved over much of Laurentia and many other paleocontinents due to exposure and erosion during a long term, eustatic drop (top of Sauk sequence) (Mussman and Read, 1989). Sections in the Antelope and

Monitor Ranges provide a sequence of key conodont and graptolite taxa (Harris et al.,

1979; Finney et al., 1999) for integration into the latest Ordovician global biostratigraphic framework and time scale (Webby et al., 2004). In addition, the Sr isotope data can be integrated with the published carbon isotope (δ13C) curves from this same sequence

(Finney et al., 1999; Kump et al., 1999; Saltzman and Young, 2005).

Methods and Results

Carbonate samples were cleaned, polished, and unaltered fine-grained (micrite) components were selectively micro-drilled from thin-section billets after petrographic screening. These same samples were previously screened and analyzed for δ13C and δ18O

(Saltzman and Young, 2005). Although brachiopod calcite with high-Sr (ppm) is considered the most reliable recorder of primary seawater 87Sr/86Sr, horizons containing well-preserved brachiopods are uncommon in the sections studied here. Two ways of potentially assessing the degree of diagenetic alteration of our samples include plots of

87Sr/86Sr versus Sr concentrations and δ18O values (e.g., Gao and Land, 1991; Qing et al.,

1998). Plots of 87Sr/86Sr against these variables (Fig. 2.2 and 2.3) show no clear

12 covariance that would be expected if the samples were significantly altered. Furthermore the Sr concentrations in our micritic samples (100-700 ppm Sr) compare with micritic samples previously used to delineate primary seawater 87Sr/86Sr trends (e.g., Gao and

Land, 1991; Montañez et al., 1996). There is also good agreement of our 87Sr/86Sr values from the Antelope Valley Limestone with those previously reported from pristine marine cements in the same formation (Fig. 2.4; Qing et al., 1998).

Procedures similar to Montañez et al. (1996) were used to extract Sr from these same micritic components with ultrapure reagents: aliquots of ~25 mg were first pretreated in 1M ammonium acetate (pH 8) and then leached in 4% acetic acid, before spiking with an 84Sr tracer. Samples were then purified for Sr using a cation exchange resin and a 2N HCl based ion-exchange procedure described by Foland and Allen (1991).

Purified Sr was loaded then with HCl on a Re double-filament configuration and isotopic compositions were measured using dynamic multicollection with a MAT-261A thermal ionization mass spectrometer in the Radiogenic Isotopes Laboratory at The Ohio State

University.

The 87Sr/86Sr values are near ~0.7090 in the Lower and Middle Antelope Valley

Limestone (Fig. 2.5), within the H. sinuosa conodont zone (Harris et al., 1979). Values began to fall through the upper part of the Antelope Valley Limestone from 0.7090 to

0.7086, which lies within the H. holdentata/ “pre-flexuosus” zones, with the upper 5 m in the P. serra zone. Values continued to fall from 0.7085 to 0.7080 in the overlying

Copenhagen Formation, within the P. anserinus through A. tvaerensis zones. Samples from the Hanson Creek Formation, within the A. ordovicicus zone, record values from

0.7081 to 0.7079. We also note that our new data do not record an early gradual decline of 87Sr/86Sr prior to the major drop interval as previously reported (see Fig. 2.4; Shields et al., 2003). This could result from biostratigraphic uncertainties and points to the importance of integrating all previously reported curves into a common chronostratigraphic framework that recognizes gaps in individual sections.

As we will argue in the Discussion, the collision, uplift, and exposure of multiple arc terranes from low latitude settings should have released substantial quantities of nonradiogenic Sr to the global oceans. To determine whether the proposed scenario could have produced the observed trends in 87Sr/86Sr during the Ordovician (Fig. 2.5), a simple reservoir model was constructed (using Visual-Basic and Microsoft Excel). A Taconic flux was calculated by using an area of exposed Ordovician island arcs that collided (2.32

× 106 km2, from paleogeographic reconstructions; Scotese and Mckerrow, 1990), the average Sr concentration of basaltic rocks (465 ppm; e.g., Dessert et al., 2003), and a chemical erosion rate of modern day high-standing plate boundary terranes (~400 t/ km2/ yr; e.g., Lyons et al., 2005). The Sr fluxes, 87Sr/86Sr values, and reservoir size (Richter et al., 1992; Davis et al., 2003) were used in addition to the Taconic flux in three stages

(Table 2.2) to produce the model curve in Figure 2.6. The resulting 87Sr/86Sr curve approximates the timing and magnitude of the observed Late Ordovician trends (Fig. 2.5) and supports previous models suggesting that changes in 87Sr/86Sr riverine flux provide the most consistent explanation for long-term changes in Phanerozoic seawater 87Sr/86Sr

(Fig. 10 of Kump, 1989; Kump and Arthur, 1997).

The observed decrease in oceanic 87Sr/86Sr were reproduced well by addition of the calculated Taconic flux, at 465 Ma, and specifying that this additional source of Sr had a 87Sr/86Sr of 0.7035 (Stern, 1982). Furthermore, because an increase in global silicate weathering rates could drain the atmospheric-oceanic CO2 reservoir in as little as

5 m.y. in the absence of additional sources (volcanism or organic matter oxidation)

(Volk, 1993), we suggest that only ‘weatherability’ increased globally and not overall silicate weathering rates (Kump and Arthur, 1997). Weatherability refers to the idea that any increase in silicate weathering in one region (e.g., due to uplift, or changes in percentage of volcanic versus non-volcanic rock weathered) will be offset by decreased

weathering elsewhere around the world due to CO2 draw down and cooling. In this way,

unsustainable imbalances in CO2 removal rates relative to CO2 inputs are restored (Kump and Arthur, 1997). For example in the case of the Taconics, increased weathering of uplifted mafic terranes in a tropical setting is offset by decreased weathering elsewhere as the planet cools. Therefore, the total weathering flux does not change in our model

(although the rate of consumption of CO2 increases due to the Ca rich nature of mafic terranes, causes a transient lowering of the size of the atmospheric reservoir; see

Discussion below).

Discussion

By documenting the Middle to Late Ordovician steady decrease of 87Sr/86Sr in a continuous stratigraphic section with excellent biostratigraphic control, one can more thoroughly evaluate previous hypotheses for this event. Furthermore, by integrating the

Sr data with previously published δ13C data in this same stratigraphic section (Saltzman

15 and Young, 2005), one can focus on possible implications for carbon cycling and the Late

Ordovician greenhouse-icehouse transition.

Controls on Middle to Late Ordovician seawater 87Sr/86Sr. Seawater 87Sr/86Sr trends (Burke et al., 1982) reflect a balance of fluxes from hydrothermal exchange at mid-ocean ridges, weathering of exposed crustal rocks (riverine input), and diagenesis of marine carbonates. The isotopic value of weathered marine carbonates is assumed to be close to the oceanic value, acting to buffer seawater 87Sr/86Sr (Elderfield and Gieskes,

1982). The riverine flux is itself made up of a large variety of rocks with different

87Sr/86Sr. To a good approximation, these can be considered as mixtures of two end- member components that include old continental crust that is relatively radiogenic

(~0.7116 or higher) and juvenile volcanic rocks (~ 0.7035) (Stern, 1982; Davis et al.,

2003). Increased hydrothermal activity at mid-ocean ridges has been proposed to explain declines in seawater 87Sr/86Sr in the mid-Cretaceous that are smaller in magnitude (~<

0.0003) (Jones and Jenkyns, 2001) than the Ordovician decrease. However, evidence for rising sea level in the Late Ordovician (Ross and Ross, 1995), if used as a proxy for increased seafloor spreading, suggests that an increased hydrothermal flux (sea level rise at base of Copenhagen Member B) postdates initiation of the 87Sr/86Sr shift by several million years (Fig. 2.5). Rather, a relative sea-level drop is apparent during the initial

87Sr/86Sr decline (Copenhagen Member A) and may correlate with the Knox

Unconformity in Eastern Laurentia (e.g., Harris et al., 1979; Kolata et al., 1996; Holland and Patzkowsky, 1996). Here, we explore changes in the percentage of total silicate

16 weathering that is volcanic as the cause of the Ordovician 87Sr/86Sr decrease (Berner,

2006).

Weathering of juvenile volcanic rocks. Juvenile arc terranes created during subduction of oceanic crust (Iapetus ocean) in the Ordovician (Anderson and Samson,

1995; Ettensohn, 1990) would have weathered rapidly in these tectonicly active, tropical to sub-tropical regions to yield a high flux of non-radiogenic Sr (Berner, 2006; Dessert et al., 2003; Lyons et al., 2005). Modern rivers draining high standing island terranes associated with plate boundaries in tropical to sub-tropical latitudes can have physical and chemical erosion rates up to two orders of magnitude greater than those from much larger rivers draining geologically older regions of low topography (e.g., Lyons et al.,

2005). Although island arcs probably existed outboard of the Laurentian continental margin during the Early Ordovician (see pre-Taconic model phase I, Fig. 2.7), uplift of these basaltic terranes occurred during convergence in the Middle Ordovician

(Whiterockian) (Mussman and Read, 1989; Anderson and Samson, 1995). This should have significantly enhanced weathering in this region and as a response 87Sr/86Sr began to drop (see Blountian phase II, Fig. 2.7). High εNd values in sedimentary rocks in the adjacent Sevier foreland basin also suggest high proportions of juvenile crustal material weathered at this time (Anderson and Samson, 1995).

Seawater 87Sr/86Sr values continued to decline through the late Whiterockian and early Mohawkian stages of the Middle-Late Ordovician before leveling off near 0.7080

(Fig. 2.7). Ages of volcanic and metamorphic rocks suggest that accretion of most major

Taconic terranes was complete by the mid-Mohawkian (455 Ma) (e.g., Anderson and

Samson, 1995), coinciding with a reversal of subduction polarity and a new phase of explosive volcanism. Peak of volcanism and ash deposition (K-bentonite beds) in the middle Mohawkian (~454 Ma) have signatures of felsic to calc-alkaline parental magmas produced by partial melting of continental crustal rocks (Kolata et al., 1996). εNd values reported for arc terranes and sediments associated with the Taconic foreland basin also suggest increasing continental influence (overlap with Laurentian crustal εNd values)

(Anderson and Samson, 1995). These felsic magmas increased radiogenic Sr in rivers draining the Taconic Highlands, and as a result, 87Sr/86Sr reached a new stable baseline in the middle Mohawkian (Fig. 2.4).

Implications and Conclusions

Chemical weathering of mafic volcanic terranes have a major impact on the

consumption rates of atmospheric CO2 because of the relative abundance of Ca and Mg- rich silicate minerals compared to non-volcanic silicates (e.g., Berner, 2006). Modern river waters draining basaltic terranes are characterized by high molar ratios of Ca/Na,

HCO3/Na, and Mg/Na (Dessert et al., 2003), and experimental studies show that dissolution and Ca-release rates (and CO2 consumption rates) are roughly two orders of magnitude faster in basaltic compared to granitic rocks (Wolff-Boenisch et al., 2006).

Global cooling in the Late Ordovician based on sedimentary proxy indicators such as chert and phosphate (e.g., Holland and Patzkowsky, 1996; Pope and Steffen; 2003) coincides with the termination of the 87Sr/86Sr drop in the middle Mohawkian (Stage II-

III, Blountian-Taconic boundary interval in Figs 2.5, 2.6). A positive δ13C shift also occurs at ~ 454 Ma (Saltzman and Young, 2005) and may reflect upwelling and enhanced

18 organic carbon burial that caused pCO2 levels to fall even lower, possibly to near thresholds needed to initiate glaciation (Patzkowsky et al., 1997; Herrmann et al., 2004).

Subsequently, a brief global warming episode (Boda event) followed in the Richmondian

(Holland and Patzkowsky, 1996; Fortey and Cocks, 2005), and preceded the widespread

Hirnantian glaciation (e.g., Kump et al., 1999). This unsteady Ordovician climate pattern

shares features with the Late Cenozoic, when CO2 drawdown linked to increased weatherability (related to the Himalayan Orogeny) produced an initial, unstable phase of

Antarctic glaciation (Zachos and Kump, 2005). The expanding ice cover in the Late

Ordovician produced a negative feedback on silicate weathering, which together with positive feedbacks due to enhanced thermohaline circulation and organic carbon burial, led to climatic oscillations that persisted for millions of years (Zachos and Kump, 2005).

Figure 2.1. Middle to Late Ordovician paleogeographic map (modified from Scotese and McKerrow, 1990) showing the study area in central Nevada (star) and the juvenile arc terranes in the Iapetus Ocean.

Figure 2.2. 87Sr/86Sr vs. Sr concentration. Data from Table 2.1.

Figure 2.3. 87Sr/86Sr vs. δ18O (‰ VPDB). The δ18O data are from Saltzman and Young, (2005).

Figure 2.4. Generalized 87Sr/86Sr curve (grey) and lithostratigrahic units from Nevada (Fig. 2.2) replotted on Figure 8 of Shields et al. (2003). Antelope Valley Limestone; CPH = Copenhagen Formation; HC = Hanson Creek Formation. Note that A, B, & C refer to the three members of the Copenhagen Formation.

Monitor & Antelope Ranges Composite, Central Nevada, western U.S. Series & -2 0 2 4 6 .7080 .7082 .7084 .7086 .7088 .7090 Stages Form- ations 443 Hirna- ntian N. extraordinarius / P. pacificus Ricmondian CINCINNATIAN Hanson CreeK Formation ( III ) Taconic Maysv. Eureka Edenian Quartzite Chat- fieldian 454 P.tenuis / Turi- UPPERORDOVICIAN Fm. P. undatus MOHAWK. nian Mem. B & C

460.5 Copenhagen A Knox Unc. ? P.anserinus /P. serra Upper Mem. ( II ) Blountian

100 Darriwilian* WHITEROCKIAN I Middle Member 0 m

AntelopeLower Valley Limestone

MIDDLEORDOVICIAN Mem. 470 -3 -1 1 3 5 7 .7080 .7082 .7084 .7086 .7088 .7090 Wackestone/Packstone 13 interbedds δ C (‰ VPDB) 87 86 Legend Sr/ Sr Massive bedded Quartz sandstone Lime mudstone Juvenile basaltic rocks packstone/grainstone and shale interbedds weathered Important conodont Important graptolite Discon- Peak Evolved Felsic + basaltic zonal boundary zonal boundary formity Ash falls rocks weathered

Figure 2.5. Plots of 87Sr/86Sr, δ13C, and a simplified stratigraphic column for the Ordovician Antelope-Monitor Range composite. Radiometric dates (Webby et al., 2004) italicized in the Stages & Series column (*Global Stage Name). Diagnostic microfossils indicative of key biostratigraphic zones are plotted (for complete biostratigraphy of sequence see Harris et al., 1979; Finney et al., 1999) along with three phases of the Taconic Orogeny based on predominant silicate rock types weathered, and peak K-bentonite ash bed deposition (after Ettensohn, 1990; Kolata et al., 1996). The three phases, I, II (Blountian), III (Taconic) also appear in the model simulation in Figure 2.3 (see Table 2.1).

Figure 2.6. Calculated Middle to Late Ordovician seawater 87Sr/86Sr curve (bold line) plotted along with actual measured values (grey rectangles; shape indicates ±1 m.y. uncertainty in age assignment of data point). The calculated strontium seawater curve was produced using information provided in Table 2.1. The three stages (i.e., I, II, III, dashed lines) correspond to changes in model parameters (Table 2.1) for which there is geologic evidence (see text for discussion) to support such variation of these parameters.

Figure 2.7. Tectonic cross sections through Laurentia in the Ordovician showing different fluxes of Sr to the oceans that produced the large seawater 87Sr/86Sr drop recorded in central Nevada (star) (cross sections modified after Draut et al., 2004; Mussman and Read, 1986; Zagorevski et al., 2006). The three phases, I, II (Blountian), III (Taconic) also appear in the model simulation in Figure 2.3 (see Table 2.1).

Age Sr Meters Section Formation Biostrat. Zone 87Sr/86Sr† Uncertainty§ (Ma)* (ppm) 0 BLM Fence AVLS sinuosa 469.0 0.708940 0.000006 117.8 40.5 BLM Fence AVLS sinuosa 468.0 0.708830 0.000014 159.5 135 BLM Fence AVLS sinuosa 467.0 0.708938 0.000010 230.7 157.5 BLM Fence AVLS holdentata 466.0 0.708958 0.000016 140.9 212.5 BLM Fence AVLS holdentata 465.0 0.709039 0.000015 129.4 257.5 BLM Fence AVLS polonicus 464.5 0.708833 0.000005 236.2 311.5 BLM Fence AVLS friendsvillensis 464.0 0.708843 0.000012 260.4 330 Hill 8308 AVLS serra 463.5 0.708693 0.000015 313.2 335 Hill 8308 AVLS serra 463.0 0.708664 0.000012 200.0 360 Hill 8308 CPH anserinus 461.0 0.708484 0.000009 196.6 379.5 Hill 8308 CPH anserinus 460.5 0.708425 0.000013 292.9 402 Hill 8308 CPH tvaernis 458.0 0.708266 0.000017 219.6 409.5 Hill 8308 CPH tvaernis 457.0 0.708239 0.000010 154.0 423 Hill 8308 CPH tvaernis 455.0 0.708090 0.000007 715.9 445.5 Hill 8308 CPH undatus 454.0 0.708179 0.000013 198.9 466 Hill 8308 CPH tenuis 453.5 0.708094 0.000013 544.8 Copenhagen 588.5 Canyon HC ordovicicus 449.0 0.708126 0.000011 551.9 Copenhagen 665 Canyon HC ordovicicus 446.0 0.708051 0.000014 283.8 Copenhagen 746 Canyon HC ordovicicus 444.0 0.707993 0.000011 581.6

Table 2.1. Data table for 87Sr/86Sr values used in Figures 2.2 through 2.5. AVLS = Antelope Valley Limestone; CPH = Copenhagen Formation; HC = Hanson Creek Formation. *Absolute ages based on conodont zonation reported previously (Harris et al., 1979; Finney et al., 1999). †Laboratory value for the SRM 987 standard is (87Sr/86Sr) = 0.710242 ±0.000010 (one- sigma external reproducibility). Normalized to a value of 0.119400 for 87Sr/86Sr. §Two-sigma mean internal reproducibility, typically based upon 100 measured values.

Stage I Stage II* Stage III* Inputs Flux 87Sr/86Sr Flux 87Sr/86Sr Flux 87Sr/86Sr (1010 (1010 (1010 mol/yr) mol/yr) mol/yr)

Riverine1 3.0 0.71000 2.5 0.71000 2.5 0.70985 Riverine (Taconic flux)2 0.5 0.70350 0.5 0.70840 Hydrothermal 0.5 0.70350 0.5 0.70350 0.5 0.70350 Diagenesis 0.3 0.70840 0.3 0.70840 0.3 0.70840

Duration (Ma) 470-465 465-455 455-443 Note: Sr input parameters (except Riverine2), along with oceanic Sr reservoir size (1.25x 1017 mol) from recent studies (Richter et al., 1992). Starting seawater 87Sr/86Sr was 0.7090 (Shields et al., 2003). *For Stage II & III, riverine input of Sr is broken into two end-member components (Riverine1 & Riverine2). 1Radiogenic end member of riverine Sr. Stage I, only flux of riverine Sr to the oceans. Stage II & III, flux is reduced due to new rapidly weathering sources (Kump & Arthur, 1997) of riverine Sr in Taconic Highlands. Stage III, 87Sr/86Sr value decreases due to Tippecanoe transgression over lowlying radiogenic cratons (e.g., Brookfield & Brett, 1988). 2Nonradiogenic end member of riverine Sr. Stage I, no uplifted source to provide a flux to the oceans. Flux in Stage II & III added due to new source in Taconic Highlands (see text for explanation/ calculation of flux). 87Sr/86Sr in Stage III increases due to new radiogenic sources exposed in Taconics (Anderson & Samson, 1995; Kolata et al., 1996).

Table 2.2. Seawater Sr Model Parameters.

CHAPTER 3

13 13 PAIRED δ Ccarb AND δ Corg RECORDS OF UPPER ORDOVICIAN (SANDBIAN- KATIAN) CARBONATES IN NORTH AMERICA AND CHINA: IMPLICATIONS FOR PALEOCEANOGRAPHIC CHANGE

Abstract

Late Sandbian-Early Katian marine carbonates from two sections in North

13 13 America and one in south China were analyzed for paired δ Ccarb and δ Corg, and reveal

13 13 similar δ Ccarb but varying trends in δ Corg stratigraphy while all sections record a well-

13 13 known positive δ Ccarb shift. These δ Corg records are the first through the Guttenberg

13 δ Ccarb excursion (GICE) in China, Oklahoma, and West Virginia. The Pagoda

13 Formation in south China documents a positive excursion in δ Corg associated with the

13 latter part of the positive δ Ccarb excursion and closely resembles trends from Iowa and

Pennsylvania that have been interpreted to reflect a lowering of atmospheric pCO2 levels.

13 13 The paired δ Ccarb and δ Corg trends from Oklahoma and West Virginia differ significantly from what is observed in China. This likely indicates that local changes in nutrient cycling and phytoplankton growth rates were the dominant control on 12C fractionation in Oklahoma and West Virginia, masking the global signal of lowered atmospheric pCO2 levels observed in China.

Introduction

29 It is widely accepted that the Hirnantian Stage (445-443 Ma) of the Ordovician period was a time of major global environmental changes that culminated in the re- emergence of large continental ice sheets over polar landmasses (Brenchley et al., 1994).

However, the circumstances surrounding the initiation of the Hirnantian climate transition remains controversial, and it has been suggested that the Late Ordovician ice age climate actually began ~10 m. y. prior in the Katian Stage (e.g. Frakes et al., 1992;

Pope and Steffen, 2003; Saltzman and Young, 2005). In addition to lithologic and biologic indicators, an important line of evidence that is consistent with global cooling beginning as early as the Katian is a positive δ13C excursion in marine carbonates

13 (δ Ccarb) (e.g. Ludvigson et al., 1996; Patzkowsky et al., 1997; Ainsaar et al., 1999;

13 Young et al., 2005). This δ Ccarb excursion is commonly interpreted to reflect enhanced

13 burial of organic carbon on a global scale. Furthermore, the positive δ Ccarb excursion in the early Katian is also associated with an even larger positive shift in organic matter

13 isotopes (δ Corg) (Hatch et al., 1987; Ludvigson et al., 1996; Patzkowsky et al., 1997).

13 13 13 This decreased isotopic difference between δ Ccarb and δ Corg (Δ C) has been

interpreted to reflect a fall in pCO2 (and CO2(aq)) below a critical threshold for ice-sheet growth in the Late Ordovician (Patzkowsky et al., 1997; Herrmann et al., 2003, 2004).

13 13 However, unlike the widely studied δ Ccarb excursion in the Katian, the δ Corg

13 13 trend has not been globally documented. Paired analysis of δ Ccarb and δ Corg in many

13 13 different regions is critical because the isotopic difference between δ Ccarb and δ Corg

13 (Δ C) is known to respond to variables other than pCO2, such as phytoplankton growth rates, cell geometries (volume to surface area ratios), and the biological source of the

30 organic matter preserved (Francois et al., 1993; Hinga et al., 1994; Bidgare et al., 1997;

Popp et al., 1998). Furthermore, in the case of the early Katian, the epeiric sea water masses that flooded landmasses have been shown to vary in critical parameters (i.e., nutrient concentrations) that affect growth rates and algal ecology (e.g., Holmden et al.,

1998; Young et al., 2005; Panchuk et al., 2006).

13 13 Here, we present paired analyses of δ Ccarb and δ Corg from several Late

Ordovician (late Sandbian to early Katian Stage) carbonate sections in China and North

America that can be correlated to previously studied sections using biostratigraphy and/or

K-bentonite event stratigraphy (Fig. 3.1). The documented Δ13C trends are evaluated alongside with previously published data (Patzkowsky et al., 1997; Pancost et al., 1999) in the context of fluctuating local paleoceanographic conditions during an episode of global environmental change in the Late Ordovician. Our results reveal evidence of local and likely also global carbon cycle effects on δ13C in different sections, and emphasize

13 13 the importance of continued high-resolution study of paired δ Ccarb and δ Corg trends in widely separated oceanic regions.

Geologic Background

Recent work by the International Subcommission on Ordovician Stratigraphy

(ISOS) has led to the division of the Upper Ordovician Series into three globally classified stages, in ascending order, Sandbian, Katian, and Hirnantian (Bergström et al.,

2000; Chen et al., 2006; Goldman et al., 2007). Katian epeiric sea deposits of North

America record regional extinctions of marine benthos and a distinct change in the style of carbonate deposition (“tropical-type” to “temperate-type”) (e.g., Patzkowsky and

31 Holland, 1993; Holland and Patzkowsky, 1996; Patzkowsky et al., 1997; Pope and Read,

13 1998; Pope and Steffan, 2003). The associated positive δ Ccarb shift, known as the

Guttenberg carbon isotope excursion (GICE), is the most widely documented excursion in the Upper Ordovician outside of the Hirnantian. The GICE has been documented in marine limestones from numerous localities in North America, Europe, and recently in

Asia (Fig. 3.1) (e.g., Ludvigson et al., 1996; Patzkowsky et al., 1997; Ainsaar et al., 1999;

Bergström et al., 2007). As a result of intensive study of the GICE over a wide area of

North America, it has been possible to observe local variations in the magnitude of the excursion that are linked to differences in paleoceanographic setting and nutrient cycling

(e.g., Holmden et al., 1998; Young et al., 2005; Panchuk et al., 2006).

The GICE begins near the Phragmodus undatus/ Plectodina tenuis Midcontinent conodont zonal boundary, and ends near the P. tenuis/ Belodina confluens zonal boundary (Young et al., 2005). The GICE lies within the upper part of the

Amorphognathus tvaerensis North Atlantic conodont zone, and has been demonstrated to fall within the North American Corynoides americanus and Orthograptus ruedemanni graptolite zones (Saltzman et al., 2003). These zones comprise the lower part of the newly defined Diplocanthograptus caudatus Graptolite Biozone, which marks the base of the Katian Global Stage of the Upper Ordovician (Goldman et al., 2007). This positive

13 δ Ccarb excursion has also been shown to begin several meters above widespread altered volcanic ash beds, known as the Millbrig and Kinnekulle K-bentonites in eastern North

America and Baltoscandia, respectively (Ainsaar et al., 1999; Saltzman et al., 2003;

Young et al., 2005).

32 13 The positive shift in δ Corg values (up to +7.5‰), associated with the GICE, has been documented from two North American sections in Iowa and Pennsylvania (Hatch et

13 al., 1987; Ludvigson et al., 1996; Patzkowsky et al., 1997). Our new paired δ Ccarb and

13 δ Corg curves include sections through (1) the Pagoda Formation, near Yichang (Hubei

Province) in southeastern China; (2) the Bromide and Viola Springs Formations near

Fittstown (Pontotoc County), Oklahoma, USA; and (3) the Nealmont and Dolly Ridge

Formations, near Riverton (Pendleton County), West Virginia, USA. Previous work on the biostratigraphy (conodonts and graptolites), sequence and event stratigraphy, and

13 δ Ccarb stratigraphy of these three sections allow us to correlate the three sections to other studied regions (Perry, 1972; Alberstadt, 1973; Sweet, 1983; Finney, 1986; An,

1987; Keith, 1989; Chen et al., 1995; Young et al., 2005; Bergström et al., 2007, in review; Goldman et al., 2007).

Methods

Samples were collected from all three sections at 25 to 50 cm intervals and

13 previously analyzed for δ Ccarb (Young et al., 2005; Bergström et al., 2007, in review).

13 For δ Corg studies here, 31 samples from the Fittstown section, 48 samples from the

Dolly Ridge section, and 21 samples from the Puxihe section were analyzed. Fresh rock surfaces were generated by a water-based diamond-blade saw and the resulting thin- section billets were polished, and placed into an ultrasonic bath containing ultrapure

(deionized, 18 MΩ) water to remove surficial organic contaminants. Micritic (fine- grained) components were microdrilled (~1.5 g of powder) from the thin-section billets.

13 Micritic components have been previously demonstrated to record original δ Ccarb

33 signatures from Upper Ordovician marine seas (Gao et al., 1996; Ainsaar et al., 1999;

Young et al., 2005; Panchuk et al., 2006; Fanton and Holmden, 2007;) and also may

13 contain sufficient and well preserved organic matter for δ Corg study (e.g., Patzkowsky et al., 1997; Pancost et al., 1999; Melchin and Holmden, 2006).

Sample powders were accurately weighed and acidified using ~6N HCl to remove carbonate minerals. Insoluble fractions were then repeatedly rinsed in ultrapure water and dried at 85°C. Remaining residues were weighed and homogenized, and then loaded into tin capsules. Samples were combusted with a Costech Elemental Analyzer and the

13 resulting CO2 gas analyzed for δ C through a Finnigan Delta IV stable isotope ratio mass spectrometer under continuous flow using an open-split CONFLO III interface in the

Stable Isotope Biogeochemistry Laboratory at The Ohio State University. Carbon isotope ratios presented here are reported in per mil notation relative to the Vienna Peedee

Belmenite limestone standard (‰ VPDB). Repeated measurements of the IAEA-CH7 standards were ±0.15‰ for δ13C, and ±1.0% for %C (1 standard deviation). Weight percent of total organic carbon (TOC) in samples is determined by comparison of

+ voltages for the ion beam intensities of masses 44, 45, and 46 CO2 between our samples and known wt. % carbon of the gravimetric standard Acetanilide.

Results

Puxihe, Hubei Province, southeastern China. A complete section of the Pagoda

Formation is exposed in the Puxihe Quarry, north of Yichang City in Hubei Province,

13 China (Fig. 3.2). Values of δ Ccarb begin at +0.7‰ in the lowermost beds of the Pagoda

Formation and abruptly trend to +2.4‰ near the first appearance of the diagnostic

34 13 conodont Hamarodus europaeus (Bergström et al., 2007, in review). δ Corg values begin

13 near -29.5‰ and trend positively towards -28.5‰ (Fig. 3.3). As δ Ccarb values begin to

13 decline at 3.5 m, δ Corg values shift abruptly positive reaching a peak value of -26.7‰ at

4.25 m, above the first appearance of the diagnostic conodont Amorphognathus aff. Am.

13 ventilatus in the middle of the Pagoda Formation. δ Corg values shift back to -28.1‰ at

13 5.5 m and stay near this value while at the same time δ Ccarb values return to base-line values of +1.0‰ in the upper part of the Pagoda Formation.

13 13 13 δ Ccarb values peak prior to δ Corg values. The resultant Δ C increases to 31.4‰ initially, but then declines at 4.0 m returning to values of 28.4‰ in the later part of the

13 δ Ccarb excursion.

Fittstown, Arbuckle Mountains, southern Oklahoma. The Fittstown section was sampled along Highway 99 in Pontotoc County, Oklahoma (Fig. 3.4) beginning

13 approximately 10 m below the Bromide/Viola Springs Formation contact. δ Ccarb values begin between -1.0‰ and 0.0‰ (Fig. 3.5) in the upper Bromide Formation (P. undatus

13 Zone), while δ Corg values start out at -28.4‰ and rise to an average value of -27.0‰ near the contact with the Viola Springs Formation. Above, in the first ~10 m of cherty lime mudstone-wackestone facies of the Viola Springs Formation (upper part of the P.

13 undatus and C. bicornis conodont and graptolite zones, respectively) average δ Ccarb

13 values shift to -0.5‰, and δ Corg values shift heavier to -26.5‰. In the overlying 20 m of

13 13 cherty packstone facies both δ Ccarb and δ Corg values trend heavier, in the P. tenuis

13 Zone. δ Corg reaches a peak of -24.7‰ at 25 m and then rapidly falls to -29.1‰ in the

13 next 10 m of section. δ Ccarb values continue to rise to +1.5‰ in the overlying non-

35 13 cherty packstone/grainstone facies, while δ Corg values decline to values near -29.5‰. In the lime mudstone/wackestone facies near the top of the section (B. confluens Zone)

13 δ Corg values shift even lighter to values averaging -31.5‰.

13 The general trend of δ Corg is to heavier values in the lower Viola Springs

Formation, then to lighter values near the disappearance of bedded/nodular chert facies,

13 δ Corg becomes lightest near the packstone/grainstone to wackestone/mudstone facies

13 13 transition. Peak δ Corg values occur prior to peak δ Ccarb values. The resulting isotopic

13 difference (Δ) remains low prior to the δ Ccarb excursion (minimum of ~25‰) and

13 13 increases through the excursion (maximum of ~33‰), with Δ C lowering after δ Ccarb returns to base-line values.

Dolly Ridge, Central Appalachians, West Virginia. The Nealmont and Dolly

Ridge Formations in Germany Valley were sampled from an outcrop along a farm road in

Pendleton County, West Virginia, near the town of Riverton (Fig. 3.6). Lime mudstones of the lower Nealmont Formation that lie below a 0.4 m thick K-bentonite bed, within the

13 P. undatus Zone, record δ Ccarb values from +1.5 to +1.0‰ and stay between 0.0‰ and

13 +0.9‰ through the crinoidal packstones at the top of the Nealmont (Fig. 3.7). δ Corg values fluctuate between -27.2‰ and -28.0‰ but shift abruptly to -25.4‰ in the uppermost few meters of the Nealmont. Across the Nealmont/Dolly Ridge Formation

13 13 contact δ Corg values decline to -27.8‰, while δ Ccarb values begin to shift positively

13 +1.0‰ in the deeper water interbedded dark lime mudstone/ shale facies. As δ Ccarb values continue to rise in the Dolly Ridge Formation, maximum values of +3.3‰ to

+2.7‰ are reached between 87 and 155 m, within the P. tenuis Zone. In this same

36 13 interval δ Corg values also rise to a peak of -26.3‰. Approaching the gradational Dolly

Ridge/ Reedsville Formation contact and the P. tenuis/ B. confluens zonal boundary

13 13 (Keith, 1989), δ Ccarb and δ Corg values decline again to near 0.0‰ and -28.0‰, respectively.

13 13 During the δ Ccarb excursion of ~+2.5‰ in the Dolly Ridge Formation, δ Corg

13 appears to generally track δ Ccarb values (+2‰ magnitude excursion), with the exception

13 13 of the brief +3‰ δ Corg shift in the uppermost Nealmont Formation. Δ C decreases in the uppermost Nealmont and then increases in the overlying Dolly Ridge Formation, but then Δ13C decreases again in the upper Dolly Ridge Formation.

Discussion

The three late Sandbian to early Katian successions presented here record

13 13 different δ Corg trends (Fig. 3.8), despite all sections recording a global positive δ Ccarb

13 13 excursion (GICE). While trends in δ Corg and the resulting Δ C curve recorded from south China are similar to previously documented trends from Iowa and Pennsylvania

(Pancost et al., 1999; Patzkowsky et al. 1997), trends from West Virginia and Oklahoma

13 are markedly different. Here we first evaluate all factors that may affect δ Ccarb and

13 13 δ Corg values to produce stratigraphic trends in Δ C (e.g., Hayes et al., 1999). We then

13 interpret each individual stratigraphic Δ C trend in the context of global changes in pCO2 predicted by models as well as local water mass characteristics (i.e., ‘aquafacies’) observed in previous studies (e.g., Holmden et al., 1998; Young et al., 2005; Panchuk et al., 2006).

37 13 13 Factors affecting the isotopic difference between δ Ccarb and δ Corg

13 Several processes contribute to the isotopic difference between δ Ccarb and

13 13 δ Corg (Δ C), including: 1) the temperature dependent fractionation between dissolved inorganic carbon (DIC) and CO2(aq) in seawater, and the fractionation between DIC and precipitated carbonate minerals (e.g., calcite versus aragonite), 2) photosynthetic fractionation (εp) associated with carbon fixation by primary producers, and 3) organic matter source variation, secondary biological fractionation (i.e., heterotrophy) and/or diagenesis (e.g., due to thermal maturation) (Hayes et al., 1999; Kienast et al., 2001;

Royer et al., 2001).

The large isotopic enrichment (~+8‰ to +12‰) that takes place when CO2(aq) equilibrates with DIC is temperature dependent (0.12‰/°C) (Goericke and Fry, 1994).

This produces differences between water masses at high versus low latitudes (Rau et al.,

1989) and can also be significant on long geological timescales (Hayes et al., 1999). For example, global cooling from the Oligocene to Pleistocene has been suggested to account for ~1‰ of an 8‰ decline in observed Δ13C values over this time period (Hayes et al.,

1999). In contrast, the isotopic enrichment that occurs when DIC is converted to sedimentary carbonate is relatively small (calcite or aragonite, ~+1‰ and +2.5‰, respectively) by comparison, and not significantly correlated with temperature (Romanek et al., 1992).

As with the majority of carbonates forming today, the Katian sections discussed here were deposited in relatively shallow tropical-subtropical settings, and likely experienced only minor fluctuations in temperatures. For example, tropical seas only

38 varied by 1-2°C during the last glacial-interglacial transition in the late Quaternary

(CLIMAP, 1981). Although sea surface temperatures are expected to drop in the later

part of the GICE due to lowered atmospheric pCO2 levels (e.g., Patzkowsky et al., 1997), only minor temperature changes are expected for these Katian carbonates deposited between 5°S and 20°S paleolatitude (see Fig. 1). This is supported by relatively small changes in sea surface temperatures (~24-26°C) that were recently calculated using conodont δ18O data through the Viola Springs Formation at the Fittstown, Oklahoma section (Rosenau et al., 2007). These temperature fluctuations would only account for

~0.25‰ of the 3 to 5‰ shifts in Δ13C values seen in the various sections described here

(Fig. 3.9). Furthermore, there is no evidence for mineralogic differences during the GICE interval in the sections studied here that would have contributed significantly to the observed shifts in Δ13C.

The carbon isotope fractionation associated with photosynthetic production of organic matter is dependent on many factors, including cell growth rates and [CO2(aq)], that likely varied through the Katian and also between individual sections studied here.

Unlike the inorganic isotope fractionation effects discussed above, εp values can be large ranging from 8 to 18‰ in the modern oceans (e.g., Francois et al., 1993; Bidgare et al.,

1997). Before proceeding with discussion of these primary biological factors (see next discussion section below), we first present our case for why secondary and source effects are unlikely to play a dominant role in the patterns we observe.

The sedimentary total organic carbon (TOC) analyzed from our sections is expected to be largely from marine sources during the Ordovician time period that

39 significantly predates the earliest land plant megafossils in the Silurian (e.g., Steemans and Wellman, 2004). Although microfossils (cryptospores) of purported terrestrial origin

(bryophyte-like plants) have been documented in Middle Ordovician marine rocks

(Strother et al., 1996), the earliest report of cryptospores in eastern North America is from younger Hirnantian strata (Richardson and Ausich, 2007). Additionally this terrestrial source is not expected to significantly vary the δ13C of TOC, as extant

Devonian bryophytes have an average δ13C value of -29.4‰ (Jahren et al., 2003), which is well within the range of compound-specific isotopic values of purported Ordovician photoautotrophic sources (Pancost et al., 1999).

Although contributions from terrestrial sourcs are not expected, heterotrophic organisms likely make up some percent of our TOC. The secondary biological process of respiratory remineralization of organic matter in sediments deposited under oxic

13 conditions has been shown to shift δ Corg values positively by ≤+1.5‰ (Hayes et al.,

1989; Gong and Hollander, 1997; Fischer et al., 1998). Further enrichment of up to +3‰ has been documented in sediments underlying oxygen-depleted waters (Laarkamp and

Raymo, 1995; Fischer et al., 1998). The carbonates in our studied sections were mostly deposited under relatively well-oxygenated conditions in shallow epeiric sea settings, although the lowermost ~10 m of the Viola Springs Formation at the Fittstown section may have been deposited under dysoxic conditions. However, in the absence of data from compound-specific studies needed to demonstrate the importance of secondary biological products in our sections, we assume that offsets between primary and secondary producers in our analyzed organic matter were relatively small (≤+1.5‰) and constant

40 (Hayes et al., 1999). Although this assumption is consistent with a study of Late

Devonian sediments (deposited under poorly oxygenated waters) that revealed nearly

13 13 identical trends in δ C of TOC compared to compound-specific δ Corg analyses derived from primary producers (i.e., short chain n-alkanes, pristine, phytane) (Joachimski et al.,

2002), it may not be here.

In addition to organic matter source and secondary biological effects, diagenetic factors can also affect δ13C of TOC. Mobilization of isotopically light organic

13 compounds via thermal maturation can shift the δ Corg values of sedimentary TOC positively by as much as 2‰ (Clayton, 1991; Hayes et al., 1999). An indicator of thermal history is provided by color alteration of conodonts from our study sections that have been assigned conodont alteration index (CAI) values of 1.5, 4.5, and 2 for Oklahoma,

West Virginia, and south China, respectively. The CAI values indicate that Oklahoma and south China experienced burial temperatures ≤ 100°C, while the West Virginia section is classified as mature to supermature experiencing burial temperatures >100°C

(cf. Epstien et al., 1977; An, 1987; Nowlan and Barnes, 1987; Sweet, 1983; Keith, 1989).

13 Thermal history differences could partly explain offsets in δ Corg values between various sections (see Fig. 3.8), but would be an unlikely explanation for the differing stratigraphic

13 13 trends in δ Corg and Δ C because a uniform thermal effect across an entire outcrop is expected.

13 Photosynthetic fractionation εp and its influence on δ C trends through the GICE

Based on the above discussions, we assume that the dominant control on the overall trends in Δ13C observed in different sections was changes in factors that influence

41 12 the magnitude of primary photosynthetic fractionation of C (εp) including CO2(aq) and growth rate. Laboratory and field studies of modern marine phytoplankton have shown that εp decreases as growth rates increase, as the ratio of cell volume to surface area (V/S) increases, or as CO2(aq) concentrations decrease (see Fig. 3.10; Freeman and Hayes, 1992;

Francois et al., 1993; Laws et al., 1995; Bidgare et al., 1997; Popp et al., 1998; Hayes et al, 1999). We first discuss the interpretation of the previously investigated Pennsylvania

Δ13C trend of Patzkowsky et al. (1997) and use this as a reference section for comparison among our studied sections.

εp in the Pennsylvania reference section. Patzkowsky et al. (1997) show that

13 13 δ Ccarb and δ Corg values initially track one another during the beginning of the GICE.

13 13 13 Peak δ Corg values occur after peak δ Ccarb values, resulting in an abrupt fall in Δ C in the latter part of the GICE (Fig. 3.9). Patzkowsky et al. (1997) note the lack of independent indicators of productivity changes (i.e., chert or phosphate deposits; Pope and Steffan, 2003) and the lack of organic matter source variations (i.e.,

Gleocapsomorpha prisca, an anomalous organic-walled microfossil; Hatch et al., 1987)

13 that could have produced the large changes in δ Corg from the analyzed carbonates of the

13 Salona and Coburn Formations. Although compound-specific studies of δ Corg in Iowa through the GICE interval indicate that compounds from G. prisca produced relatively

13 heavy sedimentary δ Corg of TOC in Iowa relative to Pennsylvania (Pancost et al., 1999), this was apparently a localized bloom (Hatch et al., 1987; Fanton and Holmden, 2007).

Instead the Δ13C trends from Pennsylvania were interpreted to reflect fluctuations in

CO2(aq) concentrations (track A in Fig. 3.10). This interpretation is also consistent with

42 the predicted changes in pCO2 during the period of enhanced organic carbon burial

13 suggested by the positive δ Ccarb excursion (GICE).

13 13 The relatively invariant Δ C trend during rising δ Ccarb in the lower part of the

Pennsylvania section was interpreted to reflect high pCO2 levels that lie above the sensitivity of marine phytoplankton (~12-15× PAL). Carbon dioxide levels continued to

13 fall due to enhanced organic carbon burial and by the later part of the δ Ccarb excursion

13 was in the range of sensitivity of phytoplankton, causing δ Corg to abruptly increase and shift Δ13C to lower values (Patzkowsky et al., 1997). A cooling event associated with the

GICE, consistent with the proposed drop in pCO2, is supported by changes in carbonate deposition style, regional extinction of brachiopods, and modeling studies, all of which suggest climatic cooling associated with the GICE (Holland and Patzkowsky, 1996;

Patzkowsky and Holland, 1993; Pope and Read, 1998; Kump and Arthur, 1999).

South China. The overall trend in Δ13C from the paleocontinent of South China is

13 13 very similar to the Pennsylvania Δ C trend (Fig. 3.8 and 3.9), with peak δ Corg values

13 13 occurring later than peak δ Ccarb values. Observed Δ C trends initially increase (Fig.

13 3.9) in the beginning of the GICE and then abruptly fall in the middle of the δ Ccarb excursion. Like Pennsylvania, there are no independent indicators of nutrient or productivity changes through the interbedded shales and lime mudstones of the Pagoda

Formation at the Puxihe section. Therefore a decrease in the concentration of CO2(aq)

(track A in Fig. 3.10) is a more plausible interpretation, and may provide the first confirmation that the GICE organic matter burial event resulted in a global lowering of

pCO2.

43 Oklahoma. The trend in Δ13C from Oklahoma is entirely opposite to those from

Pennsylvania and South China during the GICE (Fig. 3.9). The Δ13C trend in Oklahoma likely differs from what appears to be the global pattern due to local changes in phytoplankton growth rates (track B in Fig. 3.10). This is consistent with the lithologic evidence of bedded cherts and phosphates in Oklahoma that indicate changes in the strength of upwelling along the southern Laurentian margin (Pope, 2004; Young et al.,

2005). Bedded cherts and phosphate in the lower part of the section indicate upwelling of nutrient rich waters, and the associated high growth rates may explain the unusually

13 13 heavy δ Corg values (and low Δ C). Bedded cherts and phosphate disappear upsection, and lowered growth rates may have driven the Δ13C values to increase significantly above this level (Fig. 3.5).

13 The explanation of productivity-driven changes in δ Corg in Oklahoma is consistent with observations in modern upwelling zones. In the modern Peru Upwelling

3- Zone, the correlation of high nutrient (PO4 ) concentrations and low εp values (heavy

13 δ Corg values) has been linked to high phytoplankton growth rates (Bidigare et al., 1997).

Because of these large swings in phytoplankton growth rates in upwelling zones, Late

13 Cenozoic investigations that have examined the use of δ Corg as a proxy for global

changes in pCO2 have focused on oligotrophic waters (e.g., Pagani et al., 1999).

13 13 West Virginia. The δ Corg and Δ C records from Dolly Ridge, West Virginia are somewhat puzzling in that they do not more closely reflect the trends observed by

Patzkowsky et al. (1997) only ~350 km away in Pennsylvania (Fig. 3.11). The

13 Pennsylvania and West Virginia sections record similar δ Ccarb values, lithofacies, and

44 conodont biofacies reflecting the same water mass (‘aquafacies’; Young et al., 2005).

13 13 However, despite these similarities, the contrasting δ Corg and Δ C records suggest that the water mass in West Virginia was influenced by local changes in nutrient cycling that influenced phytoplankton growth rates.

The West Virginia water mass may represent a zone of mixing between nearby water masses over Laurentia (Southern and Taconic aquafacies), or open ocean Iapetus

13 waters (Holmden et al., 1998) that influenced nutrient concentrations and thus δ Corg.

Specifically, the more protracted fall in Δ13C values began at the peak of the GICE (Fig.

3.7), when compared to Pennsylvania and China (Fig. 3.11), is interpreted to reflect a lowering of growth rates in waters stripped of nutrients in nearby regions. This is consistent with the inferred direction of surface water currents flowing from northeast to southwest over eastern North America in the Upper Ordovician (Wilde, 1991; Kolata et al., 2001). In order to explain the different Δ13C trends between West Virginia and

Pennsylvania, growth rates had to be higher in West Virginia before the GICE peak and then lower afterwards (Fig. 3.11 and 3.12). Alternatively, the unique West Virginia trends could reflect locally higher concentrations of CO2(aq) in seawater from this deeper water setting that was closer to the axis of foreland basin relative to the Salona and Coburn

Formations of Pennsylvania (Diecchio, 1986; Fichter and Diecchio, 1986; Slupik, 1999).

Further detailed sedimentologic, biostratigraphic, and compound specific isotope work is needed to determine the exact cause(s) of the unique Δ13C trend from West Virginia.

Implications and Conclusions

45 13 13 The paired δ Ccarb and δ Corg trends from the Pagoda Formation of south China are similar to previously documented records from Pennsylvania and Iowa (Patzkowsky et al., 1997; Pancost et al., 1999) and likely record a global trend that is linked to changes in atmospheric CO2 levels during the GICE. This represents the first evidence of a global change in Δ13C, consistent with the notion that global cooling began in the early Katian

(e.g., Pope and Steffan, 2003; Saltzman and Young, 2005).

The Δ13C trends from Oklahoma and West Virginia are interpreted to reflect local decreases in marine phytoplankton growth rates that overwhelmed the influence of a

global drop in atmospheric pCO2 through the GICE (Fig. 3.12; Hayes et al., 1999;

Joachimski et al., 2002). In Oklahoma, the inferred drop in phytoplankton growth rates is coincident with lithologic evidence for changes in nutrient input associated with the change from upwelling (eutrophic surface waters) to downwelling (oligotrophic surface waters). In West Virginia, the Δ13C offset compared to the global pattern is less significant than in Oklahoma, and was likely due to advection of adjacent surface water masses that had been depleted in nutrients. This study demonstrates the importance of

13 13 high-resolution paired δ Ccarb and δ Corg from multiple sections worldwide in understanding the paleoceanographic conditions during fluctuations of the global carbon cycle.

Figure 3.1. Katian (Upper Ordovician) paleogeographic reconstruction (modified from Witzke, 1990; Scotese and McKerrow, 1991) showing locations of our study sections in Oklahoma (OK), West Virginia (DR), and southeast China (PX) along with sections 13 13 previously investigated (Iowa, IA, and Pennsylvania, PA) for paired δ Ccarb and δ Corg analyses (Ludvigson et al., 1996; Pancost et al., 1999; Patzkowsky et al., 1997).

Figure 3.2. Geologic Map of south-central China, near Yichang City in Hubei Province. Marked on this map is the Puxihe Quarry sample locality (triangle), northeast of Yichang City.

Figure 3.3. Paired carbon isotope data from the Puxihe Quarry (Hubei Province), China, with diagnostic conodonts and graptolites ranges plotted for the Pagoda and Miaopo Shale Formations (biostratigraphy from An, 1987; Bergström et al., 2007, in review; Chen et al., 1995). Three point running average curves are fitted to the data, note the 13 13 positive δ Corg excursion occurs in the declining limb of the δ Ccarb curve, and coincident abrupt drop in Δ13C values.

49 Chinese Series Puxihe, Hubei Province, China aff.

4 m Am. ventilatus Hamarodus europaeus Amorphognathus Pagoda Formation 50 0 NEICHIANSHANIAN UPPERORDOVICIAN

C. bicornus 0 1 2 3 -30 -29 -28 -27 -26 28 29 30 31 32 Zone δ13 δ13 δ13 Shale C (‰ VPDB) C (‰ VPDB) ∆ C (‰ VPDB) Miaopo carb org (carb-org) Legend Skeletal Lime Mudstone/Skeletal First appearance of Packstone/Grainstone Wackestone with shale interbeds Shale diagnostic graptolie species Lime mudstone/Skeletal Skeletal Packstone/ Lime First appearance of Grainstone with bedded chert Wackestone with bedded chert mudstone diagnostic conodont species Disconformity K-bentonite bed

Figure 3.3.

Figure 3.4. A, Map of south-central North America showing state outlines and Ordovician outcrops in Oklahoma (OK). B, Inset map showing the location of the roadcut along Highway 99, south of Fittstown (grey area, outcrops of the Viola Springs Formation) (after Alberstadt, 1973).

Figure 3.5. Paired carbon isotope data from the Fittstown section, Arbuckle Mountains, Oklahoma. Also plotted are important Midcontinent conodont fauna first occurrences from Sweet (1983) and important graptolite occurrences from Finney (1986) and Goldman et al. (2007). Note that the D. caudatus Graptolie Biozone is inferred to begin 13 within the chert-rich carbonates below the δ Ccarb that are barren of graptolites (Goldman et al., 2007). Three point running average curves are fitted to the data, note the negative 13 δ Corg excursion coinciding with disappearance of the chert-rich carbonate facies and the 13 positive δ Ccarb excursion.

52 N. Am. Stage Fittstown, Arbuckle Mnts., Oklahoma

B. confluens

D. caudatus

20 53 Chatfieldian

P. tenuis

Viola Springs Formation 0 m UPPERORDOVICIAN C. bicornis M5 M4

Fm. {Millbrig? Turinian Bromide {Deicke? 210-1-2 -32 -31 -30 -29 -28 -27 -26-25 25 26 27 28 29 30 31 32 33 δ13 δ13 ∆δ13C Ccarb (‰ VPDB) Corg (‰ VPDB) (carb-org) (‰ VPDB)

Figure 3.5.

long a farm road just east of a just of farm east long road A, Map of eastern North America showing Ordovician outcrops (dark grey) insouthern Ordovician the outcrops America North eastern showing of Map A, section for section Formation (Perry, 1972)Dolly a the Ridge outcropping -

Figure 3.6. Inset location map and B, the Region and of central outlines. showing the state Basin Appalachian type Riverton, of Formations). (grey area, Virginia Ridge and West Reedsville Dolly exposure

Figure 3.7. Paired carbon isotope data from the Dolly Ridge section, central Appalachians, West Virginia. Due to the limited biostratigraphy within the Dolly Ridge Formation (Keith, 1989) a detailed conodont biostratigraphy is not yet possible for this 13 section, however the previous K-bentonite work and δ Ccarb work (Perry, 1972; Young et al., 2005) allow us to tie into nearby sections with more detailed biostratigraphy. Three 13 13 point running average curves are fitted to the data, note that Δ C tracks the δ Ccarb curve, 13 with a conspicuous short-lived positive δ Corg excursion at top of the Nealmont Formation.

55 N. Am. Form- Stage ations Dolly Ridge, West Virginia, central Appalachians Reed- sville

40 m

56 0 D o l l y R i d g e

V-7? Chatfieldian

M5 M4 UPPERORDOVICIAN

Millbrig? Nealmont ian Deicke? Turin-

-28-293210-1 -27 -25-26 2726 28 29 30 δ13 δ13 δ13 Ccarb (‰ VPDB) Corg (‰ VPDB) ∆ C(carb-org) (‰ VPDB)

Figure 3.7.

13 Figure 3.8. Comparison of three-point running average δ Corg curves from our three sections presented here, and the previously studied Pennsylvania section (Patzkowsky et 13 al., 1997). Grey rectangle indicates where δ Ccarb values deviate positively from baseline values and later returns to baseline values (GICE) in all sections. Note that the Reedsville, PA section of Patzkowsky et al. (1997) does not record a return to baseline 13 13 δ Ccarb values, and that the Fittstown δ Corg curve is nearly the inverse of the south China and Pennsylvania curves.

Figure 3.9. Comparison of three-point running average Δ13C curves from Oklahoma, south China, and the previously studied Pennsylvania section (Patzkowsky et al., 1997). Note that the Oklahoma Δ13C curve is entirely opposite the south China and Pennsylvania trends.

Figure 3.10. Generalized relationship between the photosynthetic fractionation effect (εp), atmospheric CO2 levels, and modern marine phytoplankton growth rates (µm/day) assuming a constant volume/surface ratio of 1µm (modified from Hayes et al., 1999; 13 13 13 Joachimski et al., 2002). The large isotopic difference between δ Ccarb and δ Corg (Δ C) is mainly due to εp (e.g., Hayes et al., 1999). Track A (black-dashed arrow) would be indicative of areas that marine primary producers maintained consistent grow rates and εp was dominantly influenced by a lowering of CO2(aq) concentrations in seawater (Pennsylvania, Iowa, south China). Track B (grey-dashed arrow) would be indicative of areas where declining growth rates of primary producers were greatly influencing εp, and 13 can accompany a relative lowering of atmospheric pCO2levels (i.e. Δ C increasing in Oklahoma). See text for discussion of upwelling and downwelling waters on growth rates and εp.

Figure 3.11. Comparison of three-point running average Δ13C curves from West Virginia, south China, and the previously studied Pennsylvania section (Patzkowsky et al., 1997). Note that the south China and Pennsylvania Δ13C curves have very similar trends. West Virginia Δ13C values are initially lower than Pennsylvania and south China (dashed line C), but then later during the GICE Δ13C values become higher (dashed line C’). Refer to Figure 12 and text for further explanation.

Figure 3.12. Generalized relationship between εp, atmospheric CO2 levels, and modern marine phytoplankton growth rates (µm/day) (refer to Fig.10 for references). Tracks A and B are the same as in Figure 10. Track C-C’ (time slices in Fig. 11) demonstrates how a relative lowering of growth rates of primary producers can overprint a lowering of 13 atmospheric pCO2 levels, resulting in steady Δ C values of West Virginia during the GICE.

Conodont N. Am. Metersa δ13C δ13C %TOCb Formation carb org Zone Stage 0.0 0.06 -28.43 0.024 P. undatus Turinian Bromide 5.0 0.02 -26.22 0.071 P. undatus Turinian Bromide 8.0 -0.83 -27.70 0.043 P. undatus Turinian Bromide 11.0 0.81 -26.90 0.092 P. undatus Chatfieldian Viola Springs 13.3 -0.03 -26.44 0.021 P. undatus Chatfieldian Viola Springs 14.5 -0.24 -26.65 0.028 P. undatus Chatfieldian Viola Springs 16.0 -0.73 -26.66 0.036 P. undatus Chatfieldian Viola Springs 17.5 -0.21 -26.87 0.036 P. undatus Chatfieldian Viola Springs 19.3 -1.01 -26.33 0.016 P. undatus Chatfieldian Viola Springs 20.5 0.19 -26.32 0.009 P. undatus Chatfieldian Viola Springs 22.3 -0.40 -26.86 0.010 P. undatus Chatfieldian Viola Springs 30.3 -0.18 -26.01 0.004 P. tenuis Chatfieldian Viola Springs 31.0 -0.21 -26.24 0.010 P. tenuis Chatfieldian Viola Springs 33.0 0.15 -26.55 0.004 P. tenuis Chatfieldian Viola Springs 35.0 0.11 -24.73 0.014 P. tenuis Chatfieldian Viola Springs 39.0 0.32 -25.57 0.009 P. tenuis Chatfieldian Viola Springs 41.0 0.70 -25.95 0.003 P. tenuis Chatfieldian Viola Springs 44.0 0.91 -27.84 0.051 P. tenuis Chatfieldian Viola Springs 45.0 0.87 -29.12 0.091 P. tenuis Chatfieldian Viola Springs 47.0 0.97 -29.52 0.104 P. tenuis Chatfieldian Viola Springs 49.0 1.22 -29.14 0.104 P. tenuis Chatfieldian Viola Springs 51.0 1.38 -29.95 0.122 P. tenuis Chatfieldian Viola Springs 53.0 1.33 -29.93 0.089 P. tenuis Chatfieldian Viola Springs 55.0 1.34 -29.57 0.039 P. tenuis Chatfieldian Viola Springs 59.0 0.93 -28.96 0.098 P. tenuis Chatfieldian Viola Springs 61.0 1.09 -29.93 0.106 P. tenuis Chatfieldian Viola Springs 63.0 1.11 -32.15 0.082 P. tenuis Chatfieldian Viola Springs 65.0 1.04 -31.69 0.032 P. tenuis Chatfieldian Viola Springs 67.0 -0.33 -31.86 0.018 P. tenuis Chatfieldian Viola Springs 71.0 -0.62 -31.53 0.019 B. confluens Chatfieldian Viola Springs 73.0 0.38 -31.71 0.028 B. confluens Chatfieldian Viola Springs aSection begins 10m below the Viola Springs/Bromide Contact. bFractional weight of total organic carbon in each sample

Table 3.1. Stable isotope data from Fittstown, Oklahoma section.

Conodont N. Am. Metersa δ13C δ13C %TOCb Formation carb org Zone Stage 0.0 1.44 -27.20 0.066 P. undatus Turinian Nealmont 2.0 1.49 -26.68 0.053 P. undatus Turinian Nealmont 4.0 1.38 -27.30 0.057 P. undatus Turinian Nealmont 5.0 1.18 -27.21 0.062 P. undatus Chatfieldian Nealmont 5.5 1.19 -27.46 0.047 P. undatus Chatfieldian Nealmont 8.1 0.70 -27.48 0.054 P. undatus Chatfieldian Nealmont 10.1 1.00 -27.44 0.062 P. undatus Chatfieldian Nealmont 12.1 0.96 -27.23 0.052 P. undatus Chatfieldian Nealmont 14.1 0.97 -27.64 0.061 P. undatus Chatfieldian Nealmont 16.1 0.29 -28.85 0.018 P. undatus Chatfieldian Nealmont 18.1 1.07 -27.24 0.056 P. undatus Chatfieldian Nealmont 20.1 0.95 -28.07 0.074 P. undatus Chatfieldian Nealmont 22.1 0.92 -26.90 0.028 P. undatus Chatfieldian Nealmont 28.1 0.36 -27.71 0.032 P. undatus Chatfieldian Nealmont 34.1 0.82 -27.77 0.064 P. undatus Chatfieldian Nealmont 42.6 0.88 -26.74 0.056 P. undatus Chatfieldian Nealmont 44.6 0.44 -26.87 0.076 P. undatus Chatfieldian Nealmont 46.6 0.42 -25.93 0.082 P. undatus Chatfieldian Nealmont 48.0 0.49 -25.39 0.092 P. undatus Chatfieldian Nealmont 49.5 0.57 -26.75 0.022 P. undatus Chatfieldian Dolly Ridge 51.5 1.17 -27.32 0.062 P. tenuis? Chatfieldian Dolly Ridge 54.5 0.98 -27.38 0.047 P. tenuis? Chatfieldian Dolly Ridge 59.5 1.01 -27.85 0.032 P. tenuis? Chatfieldian Dolly Ridge 66.5 1.14 -27.23 0.085 P. tenuis Chatfieldian Dolly Ridge 71.5 2.05 -27.48 0.105 P. tenuis Chatfieldian Dolly Ridge 75.5 2.27 -26.83 0.068 P. tenuis Chatfieldian Dolly Ridge 80.5 2.90 -26.85 0.146 P. tenuis Chatfieldian Dolly Ridge 87.5 2.75 -26.77 0.149 P. tenuis Chatfieldian Dolly Ridge 93.5 3.10 -26.52 0.128 P. tenuis Chatfieldian Dolly Ridge 99.5 3.29 -26.74 0.368 P. tenuis Chatfieldian Dolly Ridge 106.5 3.09 -26.63 0.107 P. tenuis Chatfieldian Dolly Ridge 118.5 2.85 -26.91 0.273 P. tenuis Chatfieldian Dolly Ridge 122.5 3.02 -26.28 0.185 P. tenuis Chatfieldian Dolly Ridge 129.5 3.00 -26.29 0.213 P. tenuis Chatfieldian Dolly Ridge 134.2 2.92 -26.44 0.173 P. tenuis Chatfieldian Dolly Ridge 147.2 2.74 -26.75 0.050 P. tenuis Chatfieldian Dolly Ridge 152.2 2.37 -26.77 0.113 P. tenuis Chatfieldian Dolly Ridge

Continued

Table 3.2. Stable isotope data from the Dolly Ridge, West Virginia section. 63 Table 3.2 continued

158.2 2.17 -26.82 0.195 P. tenuis Chatfieldian Dolly Ridge 161.2 2.32 -27.24 0.051 P. tenuis Chatfieldian Dolly Ridge 163.2 1.39 -26.38 0.133 P. tenuis Chatfieldian Dolly Ridge 171.2 1.44 -27.38 0.333 P. tenuis? Chatfieldian Dolly Ridge 175.2 0.50 -26.66 0.064 P. tenuis? Chatfieldian Dolly Ridge 179.2 0.62 -27.71 0.071 B. confluens? Chatfieldian Dolly Ridge 183.2 -0.37 -28.06 0.101 B. confluens? Chatfieldian Dolly Ridge 187.2 1.29 -27.66 0.263 B. confluens? Chatfieldian Reedsville aBase of section 7m below first reported K-bentonite bed N-1 of Perry (1972) in the Nealmont Formation.

North Atlantic a 13 13 Meters δ Ccarb δ Corg %TOC Conodont Formation Zone 0.00 0.70 -29.11 0.063 Am. tvaerensis Pagoda 0.25 1.06 -29.55 0.070 Am. tvaerensis Pagoda 0.50 1.22 -28.96 0.056 Am. tvaerensis Pagoda 1.00 1.70 -28.50 0.047 Am. tvaerensis Pagoda 1.50 1.39 -29.06 0.047 Am. tvaerensis Pagoda 2.00 2.41 -28.66 0.043 Am. tvaerensis Pagoda 2.50 2.11 -28.47 0.032 Am. tvaerensis Pagoda 3.00 2.19 -28.88 0.034 Am. tvaerensis Pagoda 3.25 2.35 -28.86 0.032 Am. tvaerensis Pagoda 3.50 2.01 -29.37 0.074 Am. tvaerensis Pagoda 3.75 1.88 -27.42 0.029 Am. tvaerensis Pagoda 4.00 1.94 -27.61 0.046 Am. tvaerensis Pagoda 4.25 1.60 -26.75 0.032 Am. tvaerensis Pagoda 4.50 1.74 -27.30 0.037 Am. tvaerensis Pagoda 5.00 1.52 -26.90 0.029 Am. tvaerensis Pagoda 5.50 1.82 -28.11 0.027 Am. tvaerensis Pagoda 5.75 1.62 -28.34 0.051 Am. tvaerensis Pagoda 6.00 1.39 -28.57 0.034 Am. tvaerensis Pagoda 6.50 1.40 -27.95 0.043 Am. tvaerensis Pagoda 7.00 1.20 -27.83 0.035 Am. tvaerensis Pagoda 8.00 1.24 -27.82 0.030 Am. tvaerensis Pagoda 8.50 0.91 -28.43 0.029 Am. tvaerensis Pagoda aMeasured section begins at the Miaopo Shale/ Pagoda Formation contact.

Table 3.3. Stable carbon isotope data from the Puxihe, China section.

65 CHAPTER 4

A GLOBAL CHANGE IN δ 13C OF ORGANIC MATTER DURING THE LATE ORDOVICIAN GLACIATION: IMPLICATIONS FOR ATMOSPHERIC COMPOSITION

Abstract

The Late Ordovician (~ 444 million years ago) was one of three episodes during the past half billion years in which large continental glaciers formed over Earth’s polar regions. The effects of this glaciation were far-reaching and resulted in one of the largest marine mass extinction events in Earth history. The cause of this ice age is uncertain, and a paradoxical association with high atmospheric CO2 levels was previously inferred from models and proxy evidence. However, precise linkages between pCO2 and ice sheet growth remain poorly understood due in part to uncertainties in global time correlation of stratigraphic sections and the relatively coarse time resolution of available geochemical

13 13 proxy data. Reported here are high-resolution paired δ Ccarb and δ Corg analyses from well-dated stratigraphic sequences in Estonia and Anticosti Island, Canada. Additionally, high-resolution 87Sr/86Sr analyses are reported from these two sequences and one from

Nevada, USA. These isotopic records, from two separate paleocontinents (Baltica and

Laurentia), are consistent with the notion that pCO2 levels increased during the Late

Ordovician as ice sheets expanded and reduced rates of silicate weathering. Ultimately,

66 13 the highest pCO2 levels coincide with the peak in δ Ccarb and geologic evidence of deglaciation.

Introduction

The end-Ordovician (Hirnantian) mass extinctions are associated with evidence of glaciation and a large positive carbon isotope excursion in marine carbonates documented worldwide (e.g., Long, 1993; Brenchley et al., 1994; Underwood et al., 1997; Finney et

13 al., 1999; Kump et al., 1999). Hypotheses, that link the positive shift in δ Ccarb with glaciation include enhanced oceanic circulation that increased productivity and burial of organic matter (Brenchley et al., 1994), or increased weathering of carbonates exposed during eustatic drop (Kump et al., 1999). The two models also differ in their predictions

13 about the timing of changes in atmospheric CO2 during glaciation and the δ Ccarb excursion. In the productivity model, pCO2 is driven down by organic matter burial and

13 reaches a low at peak δ Ccarb values. In contrast, the weathering hypothesis proposes that

13 pCO2 and δ Ccarb were controlled primarily by changes in the percentage of silicate

13 versus carbonate weathering, and predicts that the peak in δ Ccarb should coincide with high pCO2.

13 13 Here, we present new paired analyses of δ Ccarb and δ Corg from two

13 13 13 paleocontinents and interpret the resulting Δ C (δ Ccarb - δ Corg) trends as a proxy for relative changes in atmospheric pCO2 levels (e.g., Kump et al., 1999). Additionally reported are 87Sr/86Sr analyses that provide evidence for fluctuations in the continental weathering of silicate rocks (e.g., Kump and Arthur, 1997). These data coupled with the sedimentary and biostratigraphic evidence from other key Hirnantian sections result in a

67 13 revised interpretation of the timing and linkages among the of Hirnantian δ Ccarb excursion, glaciation, and pCO2.

Background and Geologic Setting

Upper Ordovician carbonates from Estonia were deposited in an epicontinental marine platform setting on the paleocontinent of Baltica (e.g., Kaljo et al., 2001). The

Kardla drill core, from southern Estonia, contains a thick Upper Ordovician sequence, deposited in a relatively deeper water shelf setting (i.e., the Livonian tongue) (Kaljo et al.,

2001). Exposed on Anticosti Island, Quebec, Canada is a thick (~530 meters), virtually undisturbed sequence of Upper Ordovician through lower Silurian carbonates and siliciclastics, deposited in a carbonate shelf/ramp setting in a foreland basin on the western side of the Iapetus Ocean (Fig. 4.1) (Long and Copper, 1987; Long, 1993).

Ordovician-Silurian boundary sections in both Estonia and Anticosti Island have been intensively examined, and the biostratigraphy of conodonts, chitinozoans, and brachiopods facilitates global correlation (e.g., McCracken and Nowlan, 1986; Soufine and Achab, 2000; Copper, 2001; Kaljo et al., 2001; Männik, 2001; Brenchley et al.,

2003). Specifically, the presence of the chitinozoans Belonechitina gamachiana and

Spinachitina taugourdeaui (Soufiane and Achab, 2000) in the upper Ellis Bay Formation allow for precise correlation with Hirnantian sections from Estonia (Brenchley et al.,

2003). The thick (~200 m) sequence of Upper Ordovician strata in central Nevada represent a passive margin setting along the western margin of Laurentia and contains biostratigraphically useful microfossils for correlation with the sequences from Estonia and Anticosti Island (Ross et al., 1989; Finney et al., 1999) (Fig. 4.1).

68 Methods and Results

13 13 The δ Ccarb and δ Corg curves presented here (Fig. 4.2, 4.3, and 4.4) are constructed from micritic (fine-grained) components with well-preserved calcite cements and have experienced low thermal alteration (Conodont Alternation Index (CAI) ~1; <50-

80°C (McCracken and Nowlan, 1986; Männik, 2001). Samples were cleaned and polished, and least altered components (following petrographic examination) were

13 13 preferentially micro-drilled (~1 mg for δ Ccarb and ~1 g for δ Corg). Brachiopod calcite is considered to be the most reliable component for stable isotope analyses, however both of these sequences lack sufficient horizons with suitable brachiopods for this high-

13 resolution paired analyses study. Additionally, brachiopod based δ Ccarb curves from both of Anticosti Island and Estonian sequences have been reported (i.e., Brenchley et al.,

13 1994, 2003) and yield trends that compare well with the micritic-based δ Ccarb curves

(Long, 1993; Kaljo et al., 2001).

Carbonate powders for this study were analyzed at University of Saskatchewan and in the Stable Isotope Biogeochemistry Laboratory at The Ohio State University.

13 Carbonate powders analyzed for δ Ccarb were roasted in a vacuum oven at 200°C for 1 hour to remove water and volatile organic contaminants. 10-50 micrograms of carbonate were then reacted at 70°C with 3-5 drops of anhydrous phosphoric acid for 180-300 seconds. Stable isotope values were obtained using a Finnigan Kiel-III carbonate preparation device directly coupled to the dual inlet of a Finnigan Delta IV stable isotope

13 ratio mass spectrometer. Samples analyzed for δ Corg were accurately weighed and acidified using ~6N HCl to remove the carbonate minerals. Insoluble fractions were then

69 rinsed repeatedly in ultrapure water and dried at 85°C. The remaining residue was then weighed and homogenized, before being loaded into tin capsules. Samples were combusted with a Costech Elemental Analyzer and the resulting CO2 gas analyzed for

δ13C through a Finnigan Delta IV stable isotope ratio mass spectrometer under continuous flow using an open-split CONFLO III interface. Stable carbon measurements were made where repeated measurements of the IAEA-CH7 standards were ±0.15‰ for

δ13C (relative to the Vienna Peedee Belmenite limestone standard), and ±1.0% for %C (1 standard deviation). Weight percent of total organic carbon (TOC) in the samples is determined by comparison of voltages for the ion beam intensities of masses 44, 45, and

+ 46 CO2 between samples and known wt. % carbon of the gravimetric standard

Acetanilide.

University.

13 13 δ Ccarb and δ Corg values from the Kardla drill core, Estonia, remain relatively steady in the Halliku Formation (rugata to gamachiana zones) before a positive shift in

70 the overlying Bernati Member (taugourdeaui zone) of the Kuldiga Formation (Fig. 4.2).

13 13 Peak δ Ccarb and δ Corg values (+6.8‰ and -25.4‰ respectively) are recorded in the

Edole Member of the Kuldiga Formation (scabra zone). δ13C values still remain elevated in the oolitic Saldus Formation, with pre-excursion δ13C values returning in the lowermost Õhne Formation.

13 13 δ Ccarb and δ Corg values in the Vaureal Formation, Anticosti Island, Canada, are relatively steady ~0 to +1‰ and -28.8 to -27.5‰, respectively (Fig. 4.3). Several small

13 magnitude shifts were observed in δ Ccarb in the lower to middle Ellis Bay Formation.

13 13 The upper Ellis Bay Formation, records relatively steady δ Ccarb and δ Corg values

(gamachiana zone) before a positive shift in the upper Lousy Cove Member

(taugourdeaui zone) (Fig. 4.4). Overlying Laframboise reefs record a sharp increase in

13 13 δ Ccarb (+3.3 to +4.7‰) and declining δ Corg values (-26.6‰ to -28‰). Values return to pre-excursion baselines in the overlying Becscie Formation on Anticosti Island. Although

13 the excursions in δ Ccarb (Long, 1993; Kump et al., 1999; Brenchley et al., 1994, 2003;

13 Bergström et al., 2006) and δ Corg (Underwood et al., 1997; Wang et al., 1997; Melchin and Holmden, 2006) are both previously known from sections worldwide, our paired

13 analyses from Estonia and Anticosti Island demonstrate both excursions in δ Ccarb and

13 δ Corg from the same sections and samples for the first time.

A relatively small ~0.0003 87Sr/86Sr increase and return (0.7079 to 0.7082) is recorded in the P. pacificus zone of the Hanson Creek Formation (Fig. 4.5), with the return of this seawater 87Sr/86Sr increase also being recorded in Estonia from the uppermost Jonstorp and lower Halliku Formations (Fig. 4.6). The seawater 87Sr/86Sr

71 record from Anticosti Island, Quebec, Canada does not record this ~0.0003 increase.

Instead 87Sr/86Sr values are relatively constant only trending from 0.7078 in the lower

Vaureal to 0.7079 in the upper Vaureal Formation and overlying Ellis Bay Formations

(Fig. 4.7).

Discussion

13 13 Paired analysis of δ Ccarb and δ Corg. Analysis of both carbonate and organic carbon isotopes from the same samples provides the possibility of interpreting relative

changes in atmospheric pCO2 while avoiding correlation problems that can hinder comparisons of the two curves if constructed from separate shale versus carbonate successions (Kump et al., 1999). The difference between the isotopic measurement of

13 13 13 paired δ Corg and δ Ccarb (Δ C) has been utilized as a paleoatmospheric pCO2 proxy

(Kump et al., 1999; Joachimski et al., 2002) because this difference results from isotopic

discrimination, which decreases with lowering aqueous CO2 concentrations. However, it has also been shown that changes in phytoplankton growth rates, cell geometries, and photoautotroph assemblages also affect isotope discrimination (Bidgare et al., 1997; Popp

13 et al., 1998; Francois et al., 1993). Bulk δ Corg measurements may contain different species with different isotopic compositions. However, recent comparison of bulk organic

13 matter δ Corg versus source-specific organic molecules in Devonian rocks demonstrated that isotope shifts from both sources are nearly identical in timing and magnitudes

(Joachimski et al., 2002). A similar study in the Ordovician reported that, despite changes

13 in the relative contribution of an anomalous species affected δ Corg locally, the

72 magnitude of the environmental signal was still present and compared well with bulk

13 δ Corg trends from other regions (Pancost et al., 1999).

Although the physiological properties of Ordovician marine algae are not known

(Pancost et al., 1999), Δ13C trends presented here are used only as an indicator of relative

changes in atmospheric pCO2 and are not intended to provide empirical paleo- pCO2

13 13 13 levels. The difference between δ Ccarb and δ Corg (Δ C) suggests that atmospheric pCO2

13 levels were relatively low immediately prior to the δ Ccarb excursion and reached their

13 highest levels at the peak of the δ Ccarb excursion. This pattern is consistent with the

13 notion that pCO2 reached a low and then began to rise during the positive δ Ccarb excursion (Kump et al., 1999). The global increase in Δ13C (Fig. 4.2 and 4.4) is thus

consistent with increasing pCO2 levels during the Late Ordovician as ice sheets reached their maximum areal extent and rates of silicate weathering were reduced (Kump et al.,

1999). However, we also note that an increase in pO2 may also produce an increase in

Δ13C (e.g., Berner et al., 2000), and ongoing sulfur isotope studies may ultimately resolve whether or not this was a contributing or even dominant influence on our carbon isotopic trends.

13 The timing of the Hirnantian glaciation. The coincidence of the increase in Δ C

13 and deglaciation is also consistent with biostratigraphic and δ Ccarb evidence that links major erosional surfaces in low latitude carbonate platforms from midcontinent North

America and Europe to the timing of tillite deposition on Gondwana (Bergström et al.,

2006). In North Africa (Gondwana), major tillite deposits are separated by Hirnantian

73 sequences of deltaic, estuarine, or fluvial sediments interpreted to reflect an interglacial highstand (e.g., Sutcliffe et al., 2000; Ghienne, 2003).

This interpretation of relative pCO2 changes in the Hirnantian is also consistent with the sedimentary proxy evidence for the timing of cooling and ice sheet growth, and

87Sr/86Sr seawater trends from Nevada and Estonia. On Anticosti Island, coarsening upward cycles through the Lousy Cove Member of the Ellis Bay Formation on Anticosti

Island include lime mudstone and shale successions overlain by thin sandstones and interbedded lime mudstones, previously interpreted as evidence for sea-level fall (Long and Copper, 1987). The abrupt contact with the overlying Laframboise Member is erosional (Long and Copper, 1987; Long, 1993) which is consistent with peak glaciation and regression, followed by an increase in Δ13C. The oncolite and reefal carbonates that

13 record the peak of the δ Ccarb excursion in the Laframboise Member contain relatively deeper-water brachiopod and coral faunas (Copper, 2001) that were likely deposited during an interglacial highstand. This sea level interpretation is similar to shallow water successions in Estonia (Marshall et al., 1997; Kaljo et al., 2001; Brenchley et al., 2003) that are correlated with the Kardla drill core we have examined here.

Similarly, the well-studied section in central Nevada, sedimentary indicators of upwelling associated with cooling (e.g., chert and phosphate; Pope and Steffen, 2003)

13 precede the δ Ccarb excursion and are overlain by skeletal carbonates that reveal a sea level history (Finney et al., 1999; Kump et al., 1999) consistent with that described for

13 13 Anticosti Island and Estonia. Paired analyses of δ Ccarb and δ Corg have previously been

reported from this Nevada Hanson Creek sequence to estimate pCO2 (Kump et al., 1999).

74 13 However, the δ Corg values reported by Kump et al. (1999) are non-trending, unlike

13 δ Corg values from Estonia, Anticosti Island (Fig. 4.2 & 4.4), and other regions that

13 record the δ Corg excursion (Underwood et al., 1997; Wang et al., 1997; Melchin and

13 Holmden, 2006). The absence of the δ Corg excursion in Nevada may reflect the relatively high thermal alteration (CAI ~3; >100°C) (Kump et al., 1999) compared to the relatively pristine sections from Estonia and Anticosti Island (CAI ~1; <50-80°C)

(McCracken and Nowlan, 1986; Männik, 2001).

Seawater 87Sr/86Sr and climate. The short-lived seawater 87Sr/86Sr increase observed in the P. pacificus zone in Nevada (Fig. 4.5) and the uppermost Jonstorp and lower Halliku Formations in Estonia (Fig. 4.6) probably resulted from: 1) a short-lived lowering of sea-floor hydrothermal activity at mid-ocean ridges (Jones and Jenkyns,

2001), or 2) an increase in weathering of radiogenic Sr from silicates in the Taconic island arcs and possibly other regions of Gondwana. However, evidence for rising sea level in the Upper Ordovician to reach a Paleozoic maximum (Ross and Ross, 1995) is consistent with increased seafloor spreading that reduced the global volume of the ocean basins and flooded the continents. Therefore, a change in continental weathering patterns that released more radiogenic Sr to the oceans is a more plausible scenario. A global warming event (Boda event) in the Ashgill or late Katian preceded the Hirnantian glaciation and extinction events (Fortey and Cocks, 2005). Evidence for this warming event include marine faunas that migrated from low to high latitudes settings, and the appearance of a large carbonate platform on the margin of Gondwana in otherwise thick clastic sequences deposited at high latitudes (~75°S). This Boda warming event could

75 have increased weathering of radiogenic Sr from felsic volcanics (Draut et al., 2004;

Kolata et al., 1996; Zagorevski et al., 2006) in the Taconic highlands region of Laurentia.

Seawater 87Sr/86Sr values would have increased in response to this enhanced weathering.

The increase in silicate weathering would eventually begin drawing down atmospheric

CO2 levels causing global cooling and ice-sheet formation. Cooling should lead to a decrease in global silicate weathering rates and a lowering of seawater 87Sr/86Sr values.

The seawater 87Sr/86Sr record from Anticosti Island, Quebec, Canada does not record the ~0.0003 increase seen in Nevada and Estonia. Instead 87Sr/86Sr values increase steadily from 0.7078 in the lower Vaureal to 0.7079 in the upper Vaureal Formation and overlying Ellis Bay Formations (Fig. 4.7). Although it is possible that cryptic diastems have removed the more detailed record of the 87Sr/86Sr rise observed in Nevada and

Estonia, the diagnostic chitinozoan H. normalis (Soufine and Achab, 2000), conodont A. ordovicicus (McKracken and Nowlan, 1986; Sweet, 2000), and δ13C stratigraphy indicate a relatively complete sequence on Anticosti (Fig. 4.8). Due to the condensed nature of the

Anticosti sequence in the Hirnantian, more targeted and higher-resolution studies are needed there to determine if this ~0.0003 shift in 87Sr/86Sr is indeed representative of global Upper Ordovician seawater.

Implications and Conclusions

The Estonian, Anticosti Island, and Nevada sequences are consistent with an initial stage of the Hirnantian glaciation, driven by a long-term drop in pCO2 due to increased silicate weathering during the Boda warming event that preceded the

Hirnantian (~0.0003 increase in seawater 87Sr/86Sr values) (Fortey and Cocks, 2005;

76 Kump et al., 1999; Saltzman and Young, 2005) possibly combined with reduced poleward ocean heat transport (Herrmann et al., 2004). As expanding ice sheets reduced the fraction of continental area available for weathering at the glacial maximum, pCO2 began to rise (Kump et al., 1999) and at a threshold in pCO2 the ice sheets began to melt.

During deglaciation, rising sea-level may have shifted the site of deep water formation from high to low latitudes, which led to low oxygen conditions in the deep ocean that enhanced preservation of organic matter and the recycling of phosphorous (Van

Cappellen and Ingall, 1994; Cramer and Saltzman, 2005). Enhanced burial of organic matter in the deep oceans eventually contributed (together with silicate weathering) to a second draw down in pCO2 levels, and a second episode of glaciation (e.g., Sutcliffe et al.,

2000; Ghienne, 2003; Bergström et al., 2006).

Figure 4.1. Latest Ordovician (Hirnantian) paleogeographic map. Location of the study area (stars) Anticosti Island, Quebec, Kardla drill core, Estonia, and Nevada, USA. Previously studied Hirnantian sections (triangles) from Laurentia, Baltica, Scotland, and China are also shown (Kump et al., 1999; Brenchley et al., 1994; Bergström et al., 2006; Melchin and Holmden, 2006; Underwood et al., 1997; Wang et al., 1997).

13 13 Figure 4.2. Upper Ordovician to Lower Silurian δ Ccarb and δ Corg trends for the Kardla drill core, Estonia. B and E, represent the Bernati and Edole members of the Kuldiga Formation. Diagnostic chitinozoan ranges are from previous studies (Kaljo et al., 2001; 13 Brenchley et al., 2003). Additionally, most of the δ Ccarb data is replotted from Kaljo et 13 13 13 al. (2001). Also plotted are Δ C (calculated as δ Ccarb - δ Corg) trends, used as a proxy for relative changes in atmospheric pCO2, along with an inferred Gondwana (North African) ice volume curve (modified from Sutcliffe et al., 2000; Ghienne, 2003; Bergström et al., 2006). Please refer to the lithologic legend shown here for the subsequent Figures 4.3 through 4.8.

79 Kardla Drill Core, Estonia Ice Volume Fm. + - Íhne Silurian Saldus

Interglacial

10 E S. taugourdeaui Kuldiga 80 B 0 m Glacial C. scabra C. rugata UPPERORDOVICIAN Halliku B. gamachiana

0 1 2 3 4 5 6 7 -30 -29 -28 -27 -26 -25 28 29 30 31 32 33 δ13 δ13Ccarb (‰ VPDB) δ13C (‰ VPDB) ∆ C (‰ VPDB) LEGEND org (carb-org)

Skeletal packstone/ Oolitic or sandy Reefs Corals Oncolites Cross-bedding grainstone grainstone Calcareous shale Argillaceous lime mudstone/ Unconformity Pyrite and lime mudstone skeletal wackestone and shale

Figure 4.2.

13 13 Figure 4.3. A composite Upper Ordovician to Lower Silurian δ Ccarb and δ Corg trends for the Vaureal, Ellis Bay, and lower Becscie Formations on Anticosti Island, Quebec, Canada. Also plotted are Δ13C values calculated as the isotopic difference between 13 13 δ Ccarb and δ Corg measured from the same sample.

81 Western Anticosti Island Composite Section, Quebec, Canada Becscie Fm. hassi Sil. Rhud. HIR. Section Fm. Pt. Laframboise Ellis Bay Schmitt Creek

Mill Bay

Joseph Point 80 82

A. ordovicicus 0 m Richmondian Homard Vaureal Formation UPPER ORDOVICIAN

Tower Oil River, English Head to Anses aux Faises sections

Cono- dont -1 0 1 2 3 4 5 -30 -29 -28 -27 -26 -25 27 28 29 30 31 32 Stage Zone δ13 13 δ13 Ccarb (‰ VPDB) δ Corg (‰ VPDB) ∆ C(carb-org) (‰ VPDB)

Figure 4.3.

13 Figure 4.4. Correlation of Upper Ordovician to Silurian δ Ccarb (filled circles) and 13 δ Corg (open circles) trends from Estonia to a section at Point Laframboise, western 13 13 Anticosti Island, Quebec. Note that the scales for δ Ccarb and δ Corg are below and above the data plots, respectively. Light and medium grey shaded intervals represent the correlation of the diagnostic chitinozoan zones (gamachiana and taugourdeaui zones), which appear in both sections (Soufiane and Achab, 2000; Brenchley et al., 2003). The dark grey interval from the Kardla drill core represents the scabra chitinozoan zone, and this interval is likely missing at an unconformity near the base of the Laframboise Member in the Point Laframboise section (e.g., Long and Copper, 1987; Bergström et al., 2006).

83 Kardla Drill Core, Estonia Point Laframboise, Anticosti Island, δ13Corg (‰ VPDB) Quebec, Canada Fm. -30 -29 -28 -27 -26 -25 -24 -23 δ13C (‰ VPDB) Mem- org ber -30 -29 -28 -27 -26 -25 -24 Íhne Silurian Fox Point Saldus esi Fm. Becscie A. ellisbayensis Lafram- boise P. nodifera

B. gamachiana 10 84 Lousy Cove Kuldiga S. taugourdeaui 10 0 m Ellis Bay Formation Prinsta C. scabra C. rugata -1 0 1 2 3 4 5 27 28 29 30 31 32 δ13Ccarb (‰ VPDB) ∆δ13C (carb-org) (‰ VPDB) Halliku S. taugourdeaui UPPERORDOVICIAN B. gamachiana 0 m

0 1 2 3 4 5 6 7 δ13Ccarb (‰ VPDB)

Figure 4.4.

87 86 13 Figure 4.5. Plots of Sr/ Sr and δ Ccarb data through the Hanson Creek Formation, from the Monitor Range, in central Nevada. Diagnostic conodont (*), graptolite, and chitinozoan (H. normalis) biostratigraphy is also indicated (after Finney et al., 1999; 13 Soufine and Achab, 2000; Sweet, 2000). δ Ccarb data from Kump et al., (1999); Saltzman and Young (2005).

85 86 iue4.5. Figure

UPPERORDOVICIAN Stage Quartzite Eureka Richmondian (late Katian) Hirnantian Grap- Zone Hanson Creek Formation upper tolite ornatus pacificus extraordinarius persculptus oio ag opst Mri' ig oehgnCno Sections), Canyon Copenhagen & Ridge (Martin's Composite Range Monitor 4-2 -4 δ 13 C 0 carb ‰VPDB) (‰ 2 eta eaa etr U.S. western Nevada, Central m 0 50 4

8 6 *A. ordovicicus *G. ensifer

77 .7079 .7078 H. normalis .7080 87 Sr/ 86 .7081 Sr .7082 Warming ao Interglacial Major Event Boda .7083 Gondwana Glaciation Glaciation cooling Major Major

87 86 13 Figure 4.6. Upper Ordovician to Lower Silurian Sr/ Sr and δ Ccarb trends for the Kardla drill core, Estonia. Diagnostic chitinozoan ranges are from previous studies (Kaljo 13 et al., 2001; Brenchley et al., 2003). Additionally, most of the δ Ccarb data is replotted from Kaljo et al. (2001).

87 100m

Jon- δ13storp KH ua l dl i gk a u Ccarb (‰ VPDB)

Saldus

Ohne

Kardla Drill Core, Estonia Stage Fm. Rhud. SIL. S. taugourdeaui C. scabra Porkuni (Hirnantian) 88 B. gamachiama C. rugata Pirgu (late Katian) UPPER ORDOVICIAN

0 1 2 3 4 5 6 7 .7079 .7080 .7081 .7082 87Sr/86Sr

Figure 4.6.

87 86 13 Figure 4.7. Upper Ordovician to Lower Silurian Sr/ Sr and δ Ccarb trends for the surface exposures of Vaureal and Ellis Bay Formations on Anticosti Island, Quebec, Canada. Diagnostic chitinozoan (thick black vertical lines) and conodonts (*) are from previous studies (Achab, 1977; McCracken and Nowlan, 1986; Nowlan and Barnes, 1981; Soufine and Achab, 2000). Note the diamond symbols are replotted 87Sr/86Sr values from Shields et al. (2003).

89 Stage Western Anticosti Island Composite Section, Quebec, Canada Becscie Fm. Sil. Rhud. HIR. P. nodifera Section Fm. A. ellisbayensis Pt. Laframboise Ellis Bay

Schmitt S. taugourdeaui Creek B. gamachiana Mill Bay

Joseph Point (late Katian) *G. ensifer 90 *A. ordovicicus

Homard 80 Vaureal Formation Richmondian H. normalis UPPER ORDOVICIAN

Tower 0 m Oil River, English Head to Anses aux Faises sections

-1 0 1 2 3 4 5 .7078 .7079 .7080 .7081 δ13 87 86 Ccarb (‰ VPDB) Sr/ Sr

Figure 4.7.

87 86 13 Figure 4.8. Correlation of Upper Ordovician to Lower Silurian Sr/ Sr and δ Ccarb trends from western Laurentia (Nevada composite) to eastern Laurentia (Anticosti Island composite) to Baltica (Kardla drill core, Estonia). Dashed lines represent the proposed 13 correlation based upon previous biostratigraphy and δ Ccarb trends from these sections (see text and Figure captions 4.5 to 4.8 for complete references). Note the lack of ~0.0003 increase in 87Sr/86Sr values on Anticosti Island.

91 92

iue4.8. Figure UPPERORDOVICIAN Quartzite Eureka Richmondian H i r n a n t i a n Stage Grap- Zone upper tolite ornatus pacificus extraordinarius persculptus Hanson Creek Formation Mri' ig oehgnCno Sections), Canyon Copenhagen & Ridge (Martin's δ -2 13 *A. ordovicicus C 0 carb eta eaa etr U.S. western Nevada, Central 2 ‰VPDB) (‰ oio ag Composite Range Monitor H. normalis 4 8 6 77 .7079 .7078 87 .7080 Sr/ 86 Sr .7081 .7082 .7083 Bec- scie Vaureal Formation Ellis Bay Fm. Composite Island Anticosti Western eto,Qee,Canada Quebec, Section, -1 ‰VPDB) (‰ 0 δ 13 2 1 C carb 100 m 0

3 H. normalis 5 4 *A. ordovicicus 77 .7079 .7078 87 Sr/ 86 .7080 Sr .7081 Saldus Íhne Kuldiga Form- ations storp Jon- Halliku 0 δ 13 2 1 adaDilCr,Estonia Core, Drill Kardla C carb 3 ‰VPDB) (‰ 5 4 6 7 77 78 .7081 .7080 .7079 87 Sr/ 86 Sr .7082 13 18 13 Meters δ Ccarb δ O δ Corg TOC Formation Member 0 -0.29 -3.81 -29.63 0.09 Ellis Bay Prinsta 2 0.77 -3.70 Ellis Bay Prinsta 4 0.93 -4.26 -27.87 0.10 Ellis Bay Prinsta 6 0.62 -3.71 -28.18 0.09 Ellis Bay Lousy Cove 6.3 0.48 -3.32 -29.16 0.07 Ellis Bay Lousy Cove 6.6 0.57 -3.86 -28.09 0.08 Ellis Bay Lousy Cove 8.1 0.36 -4.12 -27.64 0.06 Ellis Bay Lousy Cove 9 0.87 -3.71 -28.54 0.06 Ellis Bay Lousy Cove 10.8 0.30 -3.68 -27.92 0.07 Ellis Bay Lousy Cove 12 1.62 -2.69 -27.56 0.05 Ellis Bay Lousy Cove 12.6 1.84 -2.83 -27.46 0.08 Ellis Bay Lousy Cove 13.2 2.05 -2.65 -27.48 0.11 Ellis Bay Lousy Cove 13.8 2.19 -3.51 -27.17 0.32 Ellis Bay Lousy Cove 14.1 2.09 -3.29 Ellis Bay Lousy Cove 14.4 2.16 -3.49 -27.56 0.06 Ellis Bay Lousy Cove 14.7 2.10 -3.81 -27.30 0.03 Ellis Bay Lousy Cove 15 2.23 -3.63 -25.32 0.01 Ellis Bay Lousy Cove 15.3 2.54 -4.03 -26.70 0.03 Ellis Bay Lousy Cove 15.4 3.49 -2.12 -26.60 0.06 Ellis Bay La Framboise 15.6 3.63 -1.88 -26.72 0.03 Ellis Bay La Framboise 15.8 3.65 -2.48 -26.86 0.04 Ellis Bay La Framboise 16 3.65 -2.18 -26.66 0.05 Ellis Bay La Framboise 16.2 3.69 -2.64 -26.58 0.03 Ellis Bay La Framboise 16.3 3.66 -2.90 -26.76 0.02 Ellis Bay La Framboise 16.4 4.02 -2.37 Ellis Bay La Framboise 16.7 3.43 -3.49 Ellis Bay La Framboise 16.8 3.34 -3.47 Ellis Bay La Framboise 16.9 3.86 -2.50 -26.77 0.03 Ellis Bay La Framboise 17 3.54 -3.01 Ellis Bay La Framboise 17.1 3.58 -3.36 Ellis Bay La Framboise 17.2 3.87 -3.02 Ellis Bay La Framboise 17.3 3.80 -2.47 Ellis Bay La Framboise 17.4 3.91 -2.49 Ellis Bay La Framboise 17.5 3.47 -2.43 -27.88 0.01 Ellis Bay La Framboise

Continued

Table 4.1. δ13C data from Anticosti Island, Quebec, Canada.

93 Table 4.1 continued

17.6 4.70 -1.92 Ellis Bay La Framboise 17.7 3.33 -3.43 Ellis Bay La Framboise 17.8 3.63 -3.06 Ellis Bay La Framboise 17.9 3.78 -2.05 -26.85 0.03 Ellis Bay La Framboise 18 4.14 -2.06 Ellis Bay La Framboise 18.1 3.63 -2.89 -27.33 0.02 Ellis Bay La Framboise 18.2 1.57 -3.26 Becsie Fox Point 18.3 2.14 -4.05 -27.91 0.03 Becsie Fox Point 18.6 0.76 -3.98 Becsie Fox Point 18.9 0.83 -4.69 Becsie Fox Point 19.2 0.48 -3.88 -28.87 0.01 Becsie Fox Point 19.5 0.93 -3.57 Becsie Fox Point 19.8 0.87 -3.74 Becsie Fox Point 20.1 0.65 -3.27 -28.42 0.04 Becsie Fox Point 20.4 0.13 -3.75 Becsie Fox Point 20.7 0.64 -3.45 Becsie Fox Point 21 0.57 -3.78 Becsie Fox Point 21.3 0.80 -3.55 -29.14 0.04 Becsie Fox Point 21.6 0.64 -3.50 Becsie Fox Point 21.9 0.65 -3.47 Becsie Fox Point 22.2 0.56 -3.63 Becsie Fox Point 22.5 0.51 -3.70 Becsie Fox Point 22.8 0.50 -3.83 Becsie Fox Point 23.1 0.67 -3.52 Becsie Fox Point 23.4 0.15 -3.83 Becsie Fox Point 23.7 0.61 -3.63 Becsie Fox Point 24 0.67 -3.63 Becsie Fox Point 24.3 0.41 -4.08 Becsie Fox Point 24.6 0.38 -3.84 Becsie Fox Point 24.9 0.55 -3.55 Becsie Fox Point 25.2 0.33 -3.85 -28.85 0.06 Becsie Fox Point 25.5 -0.24 -3.75 Becsie Fox Point 25.8 0.17 -3.65 Becsie Fox Point 26.1 0.24 -3.40 Becsie Fox Point 26.4 0.11 -3.88 Becsie Fox Point

13 18 13 Meters δ Ccarb δ O δ Corg TOC Formation Member 155.2 0.89 -4.99 -28.52 0.05 Õhne 157.2 0.92 -4.82 -28.23 0.05 Õhne 158.2 0.82 -4.68 -28.93 0.08 Õhne 159.2 0.78 -5.47 -29.09 0.04 Õhne 160.2 0.70 -4.73 -27.60 0.04 Õhne 161.0 2.10 -4.10 -27.86 0.06 Õhne 161.2 4.60 -4.70 -26.68 0.02 Saldus 162.0 5.30 -5.30 -26.51 0.02 Saldus 163.0 5.80 -4.90 -26.77 0.02 Saldus 164.0 5.60 -5.00 -26.25 0.02 Saldus 165.0 5.60 -5.10 -26.46 0.02 Saldus 166.0 5.40 -4.80 -26.28 0.02 Saldus 167.0 6.00 -4.30 -26.72 0.03 Kuldiga Edole 168.0 6.20 -4.00 -26.37 0.04 Kuldiga Edole 170.0 6.60 -4.00 -26.15 0.03 Kuldiga Edole 171.0 6.70 -3.60 -26.42 0.03 Kuldiga Edole 172.5 6.78 -3.30 -26.28 0.05 Kuldiga Edole 173.5 6.00 -3.80 -26.17 0.01 Kuldiga Edole 175.0 6.20 -3.40 -26.02 0.22 Kuldiga Edole 176.0 3.60 -3.20 -25.43 0.19 Kuldiga Edole 177.0 3.50 -3.00 -26.03 0.15 Kuldiga Edole 178.0 4.90 -4.00 -25.48 0.12 Kuldiga Edole 179.0 5.10 -3.80 -25.79 0.15 Kuldiga Edole 180.0 4.60 -3.90 -26.36 0.11 Kuldiga Edole 181.0 3.88 -4.30 -25.94 0.17 Kuldiga Bernati 181.5 3.49 -4.97 -26.04 0.23 Kuldiga Bernati 182.0 2.30 -2.90 -27.89 0.11 Kuldiga Bernati 183.0 2.19 -4.34 -27.90 0.14 Kuldiga Bernati 184.0 1.90 -3.20 -27.89 0.11 Kuldiga Bernati 185.0 1.50 -3.20 -28.78 0.12 Kuldiga Bernati 185.5 0.79 -5.29 -28.71 0.12 Halliku 186.0 0.90 -3.40 -26.39 0.18 Halliku 187.3 0.76 -4.31 -29.69 0.14 Halliku 188.5 0.78 -4.61 -29.58 0.11 Halliku 190.0 1.10 -3.68 -29.16 0.03 Halliku

Table 4.2. δ13C data from the Kardla drill core, Estonia.

95 CHAPTER 5

CONCLUSIONS

Middle and Upper Ordovician (late Darriwilian through Hirnantian Stages) marine carbonates from central Oklahoma, central Nevada, eastern West Virginia,

13 13 Anticosti Island (Canada), and Estonia have been investigated for δ Ccarb, δ Corg, and/or

87Sr/86Sr chemostratigraphy. The large Ordovician decrease in seawater 87Sr/86Sr (0.7090 to 0.7080) has now been documented from a continuous well biostratigraphically dated sequence in Nevada. The much improved time constraints on this decrease of ~ 0.001 in

87Sr/86Sr (~464 to 455 Ma) allow for the striking correlations with tectonic events that affected eastern Laurentia (Taconic Orogeny) and probably caused this large decline in seawater 87Sr/86Sr. Additionally, it was shown in Nevada that the end of the seawater

87Sr/86Sr decline approximately coincided with the beginning of the Guttenberg Carbon

13 Isotope Excursion (GICE) in the early Katian Stage (~454 Ma). Paired δ Ccarb and

13 13 δ Corg analyses of early Katian marine carbonates reveal similar δ Ccarb stratigraphy

13 (GICE) but varying trends in δ Corg stratigraphy across North America and the paleocontinent of South China. These records do however support a previous hypothesis linking the GICE to a drawdown in atmospheric pCO2 levels and global cooling, when

13 13 the differing paired δ Ccarb and δ Corg are interpreted in the context of decreasing growth rates of marine phytoplankton. A relatively small rise (0.7079 to 0.7082) in 96 seawater 87Sr/86Sr values is also recorded in late Katian (~448 Ma) carbonates from

Nevada and Estonia, which subsequently decline (0.7079) immediately prior to the

13 Hirnantian Stage (latest Ordovician; 445 Ma). Large positive shifts in both δ Ccarb and

13 δ Corg are recorded from Hirnantian carbonates in Estonia and Anticosti Island, Canada, which demonstrate the timing and global nature of this perturbation of the carbon cycle.

The 87Sr/86Sr data support the hypothesis that weathering of uplifted volcanic rocks during plate convergence is linked to global cooling and the transition from a greenhouse climate state to an ice age in the Late Ordovician. This scenario is supported by a model simulation, which demonstrates that a flux of nonradiogenic Sr from uplifted island arc terranes could have caused the large seawater 87Sr/86Sr decline. Additionally the

87Sr/86Sr data indicate that silicate weathering played the major role in this extensive cooling episode in Earth’s history.

The Late Ordovician glacial period has perplexed researchers for years because previous models indicated atmospheric CO2 levels ten times greater than today.

Integration of the strontium isotope data with previous carbon isotope and stratigraphic data, indicates how the organic carbon subcycle of the long-term carbon cycle may have responded to the cooling in a positive feedback loop. Proxy evidence for cooling (e.g., appearance of ‘cool water’ carbonates, phosphate deposits and chert) further supports this conclusion. This study also supports recent work in the modern on the role of highly weatherable basaltic rocks in CO2 consumption in the modern (~30%), which should have major implications for levels throughout Phanerozoic. These interpretations challenge the previous interpretation that the Ordovician strontium isotope decline

96 coincides with enhanced rates of sea-floor spreading and sea-level rise, and is consistent with the notion that 87Sr/86Sr are best used as a marker for changes in the types of rocks being weathered on the continents. This enhanced silicate weathering also set the stage for large continental ice-sheet formation in polar regions, and a biological mass extinction event (second largest in Earth’s history). Climatic parallels are now seen between the Late Ordovician and the Late Cenozoic greenhouse to icehouse transitions, with silicate weathering providing the initiator and negative feedback on changes in atmospheric CO2 levels.

13 13 13 Multiple early Katian δ Ccarb and δ Corg results record a positive δ Ccarb excursion (GICE), while only sections from Pennsylvania, Iowa, and China record a

13 13 13 positive shift in δ Corg that occurs later than δ Ccarb shift. The resultant Δ C trends are interpreted to reflect fluctuations in CO2(aq) concentrations in seawater, that indicate that

the GICE began under elevated pCO2 levels initially. Then pCO2 levels were drawdown

13 in the later part of the δ Ccarb excursion due to enhanced organic carbon burial

(Patzkowsky et al., 1997). The Oklahoma and West Virginia Δ13C trends are interpreted in the context of fluctuations in growth rates of marine phytoplankton, that suggest these rates decline through the GICE in these settings. The known relationships between modern marine phytoplankton growth rates, the photosynthetic isotope effect (εp) , and

atmospheric pCO2 levels indicate that a decline in growth rates can also accompany a drop in atmospheric pCO2 levels (Hayes et al., 1999). These data show that the oligotrophic platform interior or basinal settings of Pennsylvania, Iowa, and China where likely in equilibrium with the Ordovician atmosphere, while the platform margin section

97 of Oklahoma was not (influenced by a nearby zone of upwelling; Pope, 2004). The paired data along with the sedimentological proxy evidence suggest that these upwelled eutrophic waters diminished as the GICE began, possibly signaling a shift in the dominant type of ocean circulation (thermohaline to anti-estuarine) circulation in this low latitude carbonate platform setting. Furthermore, shifting the dominant oceanic circulation modes would have created deep-water formation sites at lower-latitudes, and promoted a stratified deep-ocean that the enhanced preservation of organic carbon

(Cramer and Saltzman, 2005, 2007a,b). This enhanced organic carbon burial led to the

decline of atmospheric pCO2 levels and the subsequent lowering of sea-level and cooling event that follows the GICE.

13 13 The Hirnantian δ Ccarb and δ Corg results of this study demonstrate that during this critical interval of Earth’s history, carbon cycling had direct links to the changing oceanography, climate, and the resulting geological and paleobiological record. These new data and a reexamination of key Hirnantian (~445 Ma) sections indicate that ice sheet advance during the Hirnantian glaciation was followed by deglaciation. This warm interglacial period was coincident with high carbon isotope values. Vigorous

thermohaline circulation, lowered atmospheric CO2 and the sea-level fall occurred in response to an initial ice sheet advance. Sea-level and atmospheric CO2 levels rose, in response to climatic warming (due to ice coverage of silicate rocks), and the oceans became stratified, which enhanced the burial of organic matter in the deep oceans.

Ultimately, these factors resulted in a second episode of cooling and glacial advance.

This is consistent with the two pulses of extinction and major glacial advance recognized

98 during the Hirnantian (e.g., Bergström et al., 2006; Brenchley et al., 1994, 2003;

Ghienne, 2003). A more complete understanding of Earth’s climatic sensitivity to major changes in carbon dioxide levels will help provide a perspective on the modern period of global climatic and biotic change affecting Earth today. The integration of radiogenic and stable carbon isotope records provides a way of linking the silicate weathering and organic carbon burial subcycles of the long-term carbon cycle. In the Middle and Late

Ordovician enhanced silicate weathering of island arc volcanics (producing the large

87 86 Sr/ Sr drop) likely produced a major change in atmospheric CO2 levels, and was followed by a climatic cooling episode. This supports the previous hypothesis invoking

fluctuations in atmospheric CO2 levels as the main driver of climate change on geologic timescales (e.g., Berner, 1990; 2006).

99 REFERENCES

Ainsaar, L., Meidla, T., and Martma, T. 1999. Evidence for a widespread carbon isotopic event associated with late Middle Ordovican sedimentological and faunal changes in Estonia. Geological Magazine 136(1):49-62.

Alberstadt, L.P., 1973. Articulate Brachiopods of the Viola Formation (Ordovician) in the Arbuckle Mountains, Oklahoma. Oklahoma Geological Survey Bulletin 117, 1-87.

An Tai-Xiang, 1987. Early Paleozoic conodonts from south China. Beijing University Press, Beijing. 128 pp.

Anderson, C.B., and Samson, S.D., 1995, Temporal changes in Nd isotopic composition of sedimentary rocks in the Sevier and Taconic foreland basins: Increasing influence of juvenile sources: Geology, v. 23, p. 983-986.

Arthur, M.A., Dean, W.E., Neff, E.D., Hay, B.J., King, J., and Jones, G., 1994, Varve- calibrated records of carbonate and organic carbon accumulation over the last 2000 years in the Black Sea. Global Biogeochemical Cycles 8, 195-217.

Barta, N.C., 2004. Investigation of the correlation of the lower Chatfieldian (Upper Middle Ordovician) strata of New York, Ontario, and Pennsylvania using the Guttenberg carbon isotope excursion (GICE). M. S. Thesis, The Ohio State University, Columbus, OH, 1-276.

Bergström, S.M., Chen Xu, Young, S.A., Schmitz, B., and Saltzman, M.R., 2007. The first record of the Ordovician Guttenburg δ13C excursion (GICE) in Asia: Chemostratigraphy of the Pagoda Limestone and Yanwashan Formation in southeastern China. Abstracts with Programs-Geological Society of America 39 (6), 144.

Bergström, S.M., Chen Xu, Schmitz, B., Young, S.A., Rong Jia-Yu, and Saltzman,M.R., in review, First record of the Ordovician Guttenburg δ13C excursion in Asia: Chemostratigraphy of the Pagoda and Yanwashan Formations in southeastern China. Lethia.

100 Bergström, S.M., Finney, S.C., Chen Xu, Palsson, C., Wang Xiao-Feng, and Grahn, Y., 2000, A proposed global boundary stratotype for the base of the Upper Ordovician Series of the Ordovician System: Episodes 23, p. 102-109.

Bergström, S.M., Saltzman, M.R., and Schmitz, B., 2006, First record of the Hirnantian (Upper Ordovician) δ13C excursion in the North American Midcontinent and its regional implications: Geological Magazine, v. 143, p. 657-678.

Berner, R.A., 1990, Atmospheric carbon dioxide levels over Phanerozoic time: Science, V. 249, p.1382-1386.

Berner, R.A., 1999, A new look at the long-term carbon cycle: GSA Today, v. 9, p. 1-6.

Berner, R.A., 2006, Inclusion of the weathering of volcanic rocks in the GEOCARBSULF model: American Journal of Science, v. 306, p. 295-302.

Berner, R.A., and Kothavala, Z., 2001, Geocarb III: a revised model of atmospheric CO2 over Phanerozoic time: American Journal of Science, v. 301, p. 182-204.

Berner, R.A., Petsch, S.T., Lake, J.A., Beerling, D.J., Popp, B.N., Lane, R.S., Laws, E.A., Westley, M.B., Cassar, N., Woodward, F.I., and Quick, W.P., 2000, Isotope

Fractionation and Atmospheric Oxygen: Implications for Phanerozoic O2 Evolution: Science, v. 287, p. 1630-1633.

Bickert, T., Pätzold, J., Samtleben, C., and Munnecke, A., 1997, Paleoenvironmental changes in the Silurian indicated by stable isotopes in brachiopods from Gotland, Sweden: Geochimica et Cosmochimica Acta, v. 97, p. 2717-2730.

Bidigare, R.R., Fluegge, A., Freeman, K. H., Hanson, K.L., Hayes, J.M., Hollander, D., Jasper, J.P., King, L.L., Laws, E.A., Milder, J., Millero, F.J., Pancost, R., Popp, B.N., Steinberg, P.A., and Wakeham, S.G., 1997, Consistent fractionation of 13C in nature and in the laboratory: growth-rate effects in some haptophyte algae: Global Biogeochemical Cycles, v. 11, p. 279-292.

Brenchley, P.J., Marshall, J.D., Carden, G.A.F., Robertson, D.B.R., Long, D.G.F., Meidla, T., Hints, L., and Anderson, T.F., 1994, Bathymetric and isotopic evidence for a short-lived Late Ordovician glaciation in a greenhouse period: Geology, v. 22, p. 295-298.

Brenchley, P.J., Carden, G.A.F., Hints, L., Kaljo, D., Marshall, J.D., Martma, T., Meidla, T., and Nolvak, J., 2003, High-resolution stable isotope stratigraphy of Upper Ordovician sequences: Constraints on the timing of bioevents and environmental changes associated with mass extinction and glaciation: Geological Society of America Bulletin, v. 115, p. 89-104. 102

Burke, W.H., Denison, R.E., Hetherington, E.A., Koepnick, R.B., Nelson, H.F., and Otto, J.B., 1982, Variation of seawater 87Sr/86Sr throughout Phanerozoic time: Geology, v.10, p. 516-519.

Caputo, M.V., 1998, Ordovician-Silurian glaciatins and global sea-level changes, in Landing E., and Johnson, M.E., eds., Silurian cycles: Linkages of dynamic stratigraphy with atmospheric, oceanic and tectonic changes: New York State Museum Bulletin, v. 491, p. 15-25.

Chen Xu, Rong Jia-Yu, Wang Xiao-Feng, Wang Zhi-Hao, Zhang Yuan-Dong and Zhan Ren-Bin, 1995. Correlation of the Ordovician rocks of China. Chart and Explanatory Notes. International Union of Geological Sciences Publication 31, 1- 104.

Chen Xu, Rong Jia-Yu, Fan Jun-Xuan, Zhan Ren-Bin, Mitchell, C. E., Harper, D. A. T., Melchin, M. J., Peng, Ping’An, Finney, S. C., and Wang Xiao-Feng 2006, The Global Stratotype Section and Point (GSSP) for the base of the Hirnantian Stage (the uppermost of the Ordovician System): Episodes 29, p.183-96.

Clayton, C.J., 1991. Effect of maturity on carbon isotope ratios of oils and condensates. Organic Geochemistry 17, 887-899.

CLIMAP Project Members, 1981. Seasonal reconstruction of the Earth’s surface at the last glacial maximum: Geological Society of America Map & Chart Series, MC-36.

Copper, P., 2001, Reefs during the multiple crises towards the Ordovician-Silurian boundary: Anticosti Island, eastern Canada, and worldwide: Canadian Journal of Earth Sciences, v. 38, p. 153-171. Cramer, B.D., and Saltzman, M.R., 2005, Sequestration of 12C in the deep ocean during the early Wenlock (Silurian) positive carbon isotope excursion: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 219, p. 333-349.

Cramer, B.D., and Saltzman, M.R., 2005, Carbonate deposition in epeiric sea settings during positive carbonisotope excursions: Do Silurian excursions play by different rules? Palaeogeography, Palaeoclimatology, Palaeoecology, v. 219, p. 333-349.

13 13 Cramer, B.D., and Saltzman, M.R., 2007. Early Silurian paired δ Corg and δ Ccarb analyses from the Midcontinent of North America: Implications for paleoceonography and paleoclimate. Palaeogeography, Paleoclimatology, Paleoecology 256, 195-203.

103 Crowell, J.C., 1999, Pre-Mesozoic Ice Ages: Their bearing on understanding the climate system, Geological Society of America Memoir 192, Boulder, CO, 41-44.

Davis, A.C., Bickle, M.J., and Teagle, D.A.H., 2003, Imbalance in the oceanic strontium budget: Earth and Planetary Science Letters, v. 211, p. 173-187.

Dessert, C., Dupré, B. Gaillardet, J., François, L.M., and Allègre, C.J., 2003, Basalt weathering laws and the impact of basalt weathering on the global carbon cycle: Chemical Geology, v. 202, p. 257-273.

Diecchio, R.J., 1986. Taconian clastic sequence and general geology in the vicinity of the Allegheny Front in Pendleton County, West Virginia. In: Neathery, T.L. (Ed.), Southeastern Section of the Geological Society of America Centennial Field 6, 85-90.

Elderfield, H., and Gieskes, J.M., 1982, Sr isotopes in interstitial waters of marine sediments from Deep Sea Drilling Project cores: Nature, v. 300, p. 493-497.

Epstein, A.G., Epstein, J.B., and Harris, L., 1977. Conodont color alteration-an index to organic metamorphism. U.S.G.S. Professional Paper 995, 1–27.

Ettensohn, F.R., 1990, Flexural interpretation of relationships between Ordovician tectonism and stratigraphic sequences, central and southern Appalachians, U.S.A., in Barnes, C.R., and Williams, S.H., eds., Advances in Orodvician Geology: Geological Survey of Canada Paper 90-9, p. 213-224.

Fanton, K.C., and Holmden, C., 2007. Sea-level forcing of carbon isotope excursions in epeiric seas: implications for chemostratigraphy. Canadian Journal of Earth Sciences 44, 807-818.

Faure, G., 1986, Principles of Isotope Geology. Wiley.

Fichter, L.S., Diecchio, R.J., 1986. The Taconic sequence in the northern Shenandoah Valley, Virginia. In: Neathery, T.L. (Ed.), Southeastern Section of the Geological Society of America Centennial Field 6, 73-78.

Finney, S. C., 1986. Graptolite Biofacies and Correlation of Eustatic, Subsidence, and Tectonic Events in the Middle to Upper Ordovician of North America. Palaios 1, 435-461.

Finney, S.C., Berry, W.B.N., Cooper, J.D., Ripperdan, R.L., Sweet, W.C., Jacobson, S. R., Soufiane, A., Achab, A. and Noble, P.J., 1999, Late Ordovician mass extinction: A new perspective from stratigraphic sections in central Nevada: Geology, v. 27, p. 215-218.

104

13 Fischer, G., Müller, P.J., and Wefer, G., 1998, Latitudinal δ Corg variations in sinking matter and sediments from the South Atlantic: effects of anthropogenic CO2 and

implications for paleo-pCO2 reconstructions. Journal of Marine Systems 17, 471– 495.

Fortey, R.A., and Cocks, L.R.M., 2005, Late Ordovician global warming-The Boda event: Geology, v. 33, p. 405-408.

Frakes, L.A., Francis, J.E., and Syktus, J.I., 1992, Climate modes of the Phanerozoic: Cambridge, Cambridge University Press, 194p.

Francois, R., Altabet, M.A., Goericke, R., McCorkle, D.C., Brunet, C., and Poisson, A., 1993, Changes in the δ13C of surface water particulate organic matter across the subtropical convergence in the SW Indian Ocean. Global Biogeochemical Cycles, 7, p. 627-644.

Freeman, K.H., and Hayes, J.M., 1992, Fractionation of carbon isotopes by

phytoplankton and estimates of ancient CO2 levels. Global Biogeochemical Cycles 6, 185-198.

Gao, G., and Land, L.S., 1991, Geochemistry of Cambro-Ordovician Arbuckle limestone, Oklahoma: Implications for diagenetic δ18O alteration and secular δ13C and 87Sr/86Sr variation: Geochimica et Cosmochimica Acta, v. 55, p. 2911-2920.

Gao, G., Dworkin, S. I., Land, L. S., and Elmore, R. D., 1996, Geochemistry of Late Ordovician Viola Limestone, Oklahoma: implications for marine carbonate mineralogy and isotopic compositions. Journal of Geology 104, 359-367.

Ghienne, J., 2003, Late Ordovician sedimentary environments, glacial cycles, and post- glacial transgression in the Taoudeni Basin, West Africa: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 189, p. 117-145.

Gibbs, M.T., Barron, E.J., and Kump, L.R., 1997, An atmospheric pCO2 threshold for glaciation in the Late Ordovician: Geology, v. 25, 447-450.

Goericke, R., Fry, B., 1994. Variations of marine plankton δ13C with latitude, temperature, and dissolved CO2 in the world ocean. Global Biogeochemical Cycles 8, 85-90.

Goldstein, S.J., and Jacobsen, S.B., 1987, The Nd and Sr isotopic systematics of river- water dissolved material: Implications for sources of Nd and Sr in seawater. Chemical Geology 66, 245-272.

Gong, C. and Hollander, D.J., 1997. Differential contribution of bacteria to sedimentary organic matter in oxic and anoxic environments, Santa Monica Basin, California. Organic Geochemistry 26, 545–563.

Grahn, Y.G., and Caputo, M.V., 1992, Early Silurian glaciation in Brazil: Palaeogeography, Palaeoclimatology, Palaoecology, v. 99, 9-15.

Hallock, P., 1988. The role of nutrient availability in bioerosion: consequences to carbonate buildups. Palaeogeography, Palaeoclimatology, Palaeoecology 63, 275- 291.

Hallock, P., and Schlager, W., 1986. Nutrient excess and the demise of coral reefs and carbonate platforms. Palaios 1, 389-398.

Hamoumi, N. 1999, Upper Ordovician glaciation spreading and its sedimentary record in Moroccan North Gondwana margin. In P. Kraft and O. Fatka (eds.) Short Papers of the 8th International Symposium of the Ordovician System (Prague, June 1999). Acta Universitatis Carolinae, Geologica 43 (1/2), 111-114.

Harris, A. G., Bergström, S. M., Ethington, R. L., Ross, R. J., Jr., 1979, Aspects of the Middle and Upper Ordovician conodont biostratigraphy of carbonate facies in Nevada and southeastern California and comparison with some Appalachian Successions: Brigham Young University Geology Studies, v. 26, p. 7-41.

Hatch, J. R., Jacobson, S. R., Witzke, B. J., Risatti, J. B., Anders, D. E., Watney, W. L., Newell, K. D., Vuletich, A. K., 1987. Possible late Middle Ordovican carbon isotope excursion: evidence from Ordovician oils and hydrocarbon source rocks, Mid-Continent and East-Central United States. American Association of Petroleum Geologists Bulletin 71, 342-1354.

Hayes, J.M., Popp, B.N., Takigiku, R., and Johnson, M.W., 1989, An isotopic study of biogeochemical relationships between carbonates and organic carbon in the Greenhorn Formation: Geochimica et Cosmochimica Acta 53, p. 2961-2972.

Hayes, J.M., Strauss, H., Kaufman, A.J., 1999, The abundance of 13C in marine organic matter and isotopic fractionation in the global biogeochemical cycle of carbon during the past 800 Ma. Chemical Geology 161, 103–125.

106 Herrmann, A.D., Patzkowsky, M.E., and Pollard, D., 2003, Obliquity forcing with 8-12 times preindustrial levels of atmospheric pCO2 during the Late Ordovician glaciation: Geology, v. 31, 485-488.

Herrmann, A.D., Haupt, B.J., Patzkowsky, M.E., Seidov, D., and Slingerland, R.L., 2004, Response of Late Ordovician paleoceanography to changes in sea level,

continental drift, and atmospheric pCO2: potential causes for long-term cooling and glaciation: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 210, p. 385-401.

Hinga, K.R., Arthur, M.A., Pilson, M.E.Q., Whitaker, D., 1994, Carbon isotope fractionation by marine phytoplankton in culture: the effects of CO2 concentration, pH, temperature, and species. Global Biogeochemical Cycles 8, 91–102.

Holland, S.M., and Patzkowsky, M.E., 1996, Sequence stratigraphy and long-term paleoceanographic change in the Middle and Upper Ordovician of the eastern United States, in Witzke, B.J. et al., eds., Paleozoic sequence stratigraphy: Views from the North American craton: Geological Society of America Special Paper 306, p.117-129.

Holmden, C., Creaser, R.A., Muehlenbachs, K., Leslie, S.A., Bergström, S. M., 1998, Isotopic evidence for geochemical decoupling between ancient eperic seas and bordering oceans: Implications for secular curves. Geology 26 (6), 567-570.

Jahren, A.H., Porter, S., Kuglitsch, J.J., 2003. Lichen metabolism identified in Early Devonian terrestrial organisms. Geology 31, 99-102.

Joachimski, M.M., Pancost, R.D., Freeman, K.H., Ostertag-Henning, C., and Buggisch, W., 2002, Carbon isotope geochemistry of the Frasnian-Famennian transition. Palaeogeography, Palaeoclimatology, Palaeoecology, v. 181, p. 91-109.

Jones, C.E., and Jenkyns, H.C., 2001, Seawater Strontium Isotopes, oceanic anoxic events, and seafloor hydrothermal activity in the Jurassic and Cretaceous. American Journal of Science, v. 301, p. 112-149.

Kaljo, D., Hints, L., Martma, T., and Nõlvak, J., 2001, Carbon isotope stratigraphy in the latest Ordovician of Estonia: Chemical Geology v. 175, p. 49-59.

Keith, C., 1989. Conodonts and Conodont Biostratigraphy of the Dolly Ridge Formation in Virginia and West Virginia. Unpublished B. S. Thesis, University of North Carolina, Wilmington, NC, 47 p.

Kienast, M., Calvert, S.E., Pelejero, C., and Grimalt, J.O., 2001. A critical review of 107 13 marine sedimentary δ Corg-pCO2 estimates: New palaeorecords from the South 13 China Sea and a revisit of other low-latitude δ Corg-pCO2 records. Global Biogeochemical Cycles 15, 113–127.

Kolata, D. R., Huff, W. D., and Bergström, S. M., 1996. Ordovician K-bentonites of Eastern North America: Geological Society of America Special Paper 313, p. 1- 84.

Kolata, D. R., Huff, W. D., Bergström, S. M., 2001. The Ordovician Sebree trough: an oceanic passage to the Midcontinent United States. Geological Society of America Bulletin 113 (8), 1067-1078.

Kump, L.R., and Arthur, M.A., 1997, Global Chemical Erosion during the Cenozoic: Weatherability balances the budgets, in Ruddiman, W.F., ed., Tectonic Uplift and Climate Change: New York, Plenum Press, p. 399-425.

Kump, L.R., and Arthur, M.A., 1999, Interpreting carbon-isotope excursions: carbonates and organic matter: Chemical Geology, v. 161, 181-198.

Kump, L.R., Gibbs, M.T., Arthur, M.A., Patzkowsky, M.E., and Sheehan, P.E., 1995, Hirnantian glaciation and the carbon cycle. In Cooper, J.D. et al. (eds.), Ordovician Odyssey: Short papers for the seventh international syposium on the Ordovician System. Pacific Section, Society for Sedimentary Geology, Fullerton, CA, 299-302.

Kump, L. R., Arthur, M. A., Patzkowsky, M. E., Gibbs, M. T., Pinkus, D. S., and Sheehan, P. M., 1999. A Weathering hypothesis for glaciation at high atmospheric

pCO2 during the Late Ordovician. Palaeogeography, Palaeoclimatology, Palaeoecology 152, 173-187.

Laarkamp, K.L., and Raymo, M.E., 1995. Carbon isotopic composition of particulate organic material from the interior of the Equatorial Pacific Ocean. ICP V Program and Abstracts, 5th International Conference on Paleoceanography. University of New Brunswick, Fredricton, p. 132.

Laws, E.A., Popp, B.N., Bidigare, R.R., Kennicutt, M.C., and Macko, S.A., 1995. Dependence of phytoplankton carbon isotopic composition on growth rate and [CO2]aq: theoretical considerations and experimental results. Geochimica et Cosmochimica Acta 59, 1131-1138.

Long, D.G.F., 1993, Oxygen and carbon isotopes and event stratigraphy near the Ordovician-Silurian boundary, Anticosti Island, Quebec: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 104, p. 49-59.

108 Long, D.G.F., and Copper, P., 1987, Stratigraphy of the Upper Ordovician upper Vaureal and Ellis Bay formations, eastern Anticosti Island, Quebec: Canadian Journal of Earth Sciences, v. 24, 1807-1820.

Ludvigson, G. A., Jacobson, S. R., Witzke, B. J., González, L. A., 1996. Carbonate component chemostratigraphy and depositional history of the Ordovician Decorah Formation, Upper Mississippi Valley. In: Witzke, B. J., Ludvigson, G. A., Day, J. (Eds.), Paleozoic Sequence Stratigraphy: views from the North American Craton. Geological Society of America Special Paper 306, 67-86.

Lyons, W.B., Carey, A.E., Hicks, D.M., and Nezat, C.A., 2005, Chemical weathering in high-sediment yielding watersheds, New Zealand: Journal of Geophysical Research, v. 110, p. 1-11.

Männik, P., 2001, Distribution of conodonts, in Põldvere, A., ed., Estonia Geological Sections Valga (10) drill core: Geological Survey of Estonia Bulletin 3, p. 10-12.

Marshall, J.D., Brenchley, P.J., Mason, P., Wolf, G.A., Astini, R.A., Hints, L., and Meidla, T., 1997, Global carbon isotopic events associated with mass extinction and glaciation in the late Ordovician: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 132, p. 195-210.

McCracken, A.D., and Nowlan, G.S., 1986, The Gamachian Stage and Fauna 13: New York State Museum Bulletin, 462, p. 71-79.

Melchin, M.J., and Holmden, C., 2006, Carbon Isotope chemostratigraphy in Arctic Canada: Sea-level forcing of carbonate platform weathering and implications for Hirnantian global correlation: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 234, p. 186-200.

Montañez, I.P., Banner, J.L., Osleger, D.A., Borg, L.E., and Bosserman, P.J., 1996, Integrated Sr isotope variations and sea-level history of Middle to Upper Cambrian platform carbonates: Implications for the evolution of Cambrian seawater 87Sr/86Sr: Geology, v. 24, p. 917-920.

Mussman, W.J., and Read, J.F., 1986, Sedimentology and development of a passive- to convergent-margin unconformity: Middle Ordovician Knox unconformity, Virginia Appalachians: Geological Society of America Bulletin, v. 97, p. 282- 295.

Nielsen, A.T., 2004. Ordovician sea level changes: A Baltoscandian perspective. In: Webby, B.D., Paris, F., Droser, M.L., and Percival, I.G. (Eds.), The Great Ordovician biodiversification event: New York, Columbia Press, p. 84–93.

109 Nowlan, G.S., and Barnes, C.R., 1987. Application of conodont colour alteration indices to regional and economic geology. In: Austin, R.L. (Ed.), Conodonts: Investigative Techniques and Applications: Chichester, Ellis Horwood Limited, p. 188–202.

Pagani, M., Arthur, M.A., and Freeman, K.H., 1999. Miocene evolution of atmospheric carbon dioxide. Paleoceanography 14, 273-292.

Panchuk, K.M., Holmden, C.M., and Leslie, S.A., 2006. Local controls on carbon cycling in the Ordovician midcontinent region of North America, with implications for carbon isotope secular curves. Journal of Sedimentary Research 76, 200-211.

Pancost, R.D., Freeman, K.H., and Patzkowsky, M.E., 1999, Organic-matter source variation and the expression of a late Middle Ordovician carbon isotope excursion. Geology, v. 27, p. 1015-1018.

Patzkowsky, M.E., and Holland, S.M., 1993, Biotic response to a Middle Ordovician paleoceanographic event in eastern North America: Geology 21, p. 619-622.

Patzkowsky, M.E., Slupik, L.M., Arthur, M.A., Pancost, R.D., and Freeman, K.H., 1997, Late Middle Ordovician environmental change and extinction: Harbinger of the Late Ordovician or continuation of Cambrian patterns?: Geology, v. 25, p. 911- 914.

Perry, W.J., Jr., 1972. The Trenton Group of Nittany Anticlinorium, eastern West Virginia. West Virginia Geological and Economic Survey Circular 13, 1-28.

Pope, M.C., 2004. Cherty carbonate facies of the Montoya Group, southern New Mexico and western Texas and its regional correlatives: a record of Late Ordovician paleoceanography on southern Laurentia. Palaeogeography, Palaeoclimatology, Palaeoecology 210, 367-384.

Pope, M.C., and Read, J. F. 1997. High Resolution Stratigraphy of the Lexington Limestone (Late Middle Ordovician), Kentucky, USA: A Cool-water Carbonate- Clastic Ramp in a Tectonically Active Foreland Basin. In James, N.P. and Clarke, J. (eds.), Cool-Water Carbonates. SEPM Special Publication 56, 411-429.

Pope, M.C. and Read, J.F. 1998. Ordovician meter-scale cycles: implications for climate and eustatic fluctuations in the central Appalachians during a global greenhouse, non-glacial to glacial transition. Palaeogeography, Palaeoclimatology, Palaoecology 138, 27-42.

110 Pope, M.C., and Steffen, J.B., 2003, Widespread, prolonged late Middle to Late Ordovician upwelling in North America: A proxy record of glaciation?: Geology, v. 31, p. 63-66.

Popp, B.N., Takigku, R., Hayes, J.M., Louda, J.W., and Baker, E.W., 1989, The post- Paleozoic chronology and mechanism of 13C depletion in primary marine organic matter: American Journal of Science, v. 289, 436-454.

Popp, B.N., Laws, E.A., Bidigare, R.R., Dore, J.E., Hanson, K.L., and Wakeham, S.G., 1998, Effect of phytoplankton cell geometry on carbon isotopic fractionation. Geochimica et Cosmochimica Acta, v. 62, p. 69-77.

Qing, H., Barnes, C.R., Buhl, D., and Veizer, J., 1998, The strontium isotopic composition of Ordovician and Silurian brachiopods and conodonts: Relationships to geological events and implications for coeval seawater: Geochimica et Cosmochimica Acta, v. 62, p. 1721-1733.

Rau, G.H., Takahashi, T., Marais, D.J.D., 1989. Latitudinal variations in plankton δ13C: implications for CO2 and productivity in past oceans. Nature 341, 516-518.

Richardson, J.G., and Ausich, W.I., 2007. Late Ordovician-Early Silurian cryptospore occurrences on Anticosti Island, Quebec, Canada. Canadian Journal of Earth Sciences 44, 1-7.

Richter, F.M., Rowley, D.B., and DePaolo, D.J., 1992, Sr isotope evolution of seawater: the role of tectonics: Earth and Planetary Science Letters, v. 109, p. 11-23.

Romanek, C.S., Grossman, E.L., and Morse, J.W., 1992, Carbon isotopic fractionation in synthetic aragonite and calcite: Effects of temperature and precipitation rate: Geochimica et Cosmochimica Acta 56, p. 419-430.

Rosenau, N.A., Leslie, S.A., and Herrmann, A., 2007. Sea surface temperature change from Late Ordovician δ18O data, Bromide and Viola Springs Formations, Oklahoma, USA. Abstracts with Programs-Geological Society of America 39 (6), 15.

Ross, C.A., and Ross, J.P.R., 1995, North American Ordovician depositional sequences and correlations. in Cooper, J. D. et al., eds., Ordovician Odyssey: Fullerton California, Pacific Section SEPM (Society for Sedimentary Geology), p. 309-313.

Ross, R.J., Jr., James, N.P., Hintze, L.F., and Poole, F.G., 1989, Architecture and evolution of a Whiterockian (early Middle Ordovician) carbonate platform, Basin Ranges of western U.S.A., in Crevello, P. et al., eds, Controls on carbonate

111 platform and basin development: SEPM (Society for Sedimentary Geology) Special Publication 44, p. 167-185.

Royer, D.L., Berner, R.A., and Beerling, D.J., 2001. Phanerozoic CO2 change: evaluating geochemical and paleobiological approaches. Earth-Science Reviews 54, 349– 392.

Saltzman, M. R., 2002. Carbon isotope (δ13C) stratigraphy across the Silurian-Devonian transition in North America: evidence for a perturbation of the global carbon cycle. Palaeogeography, Palaeoclimatology, Palaeoecology 187, 83-100.

Saltzman, M. R., 2005. Phosphorus, nitrogen and the redox evolution of the Paleozoic oceans. Geology 33, 573-576.

Saltzman, M. R. Brasier, M. D., Rippperdan, R. L., Ergaliev, G. K., Lohmann, K. C., Robinson, R. A., Chang, W. T., Peng, S., and B. Runnegar., 2000. A global carbon isotope excursion (SPICE) during the Late Cambrian: relation to trilobite extinctions. Palaeogeography, Palaeoclimatology, Palaeoecology 162, 211-223.

Saltzman, M.R., Bergström, S.M., Huff, W.D., and Kolata, D.R., 2003. Conodont and graptolite Biostratigraphy and the Ordovican (early Chatfieldian, Middle Caradocian) δ13C Excursion in North America and Batoscandia: implications for the Interpretations of the relations between the Millbrig and Kinnekulle K- bentonites. In: Albanesi, G.L., Beresi, M.S., and Peralta, S.H. (Eds.), Ordovician from the Andes. Instituto Superior De Correlacion Geologica (Insugeo), Serie Correlacion Geologica 17, 137-142.

Saltzman, M.R., and Young, S.A., 2005, A long-lived glaciation in the Late Ordovician?: Isotopic and bathymetric evidence from western Laurentia: Geology, v. 33, p. 109-112.

Samtleben, C., Munnecke, A., Bickert, T., and Pätzold, J., 1996, The Silurian of Gotland (Sweden): facies interpretation based on stable isotopes in brachiopod shells: Geologische Rundschan v. 85, 278-292.

Scotese, C.R., and McKerrow, W.S., 1990, Revised world maps and introduction, in McKerrow, W.S., and Scotese, C.R., eds., Paleozoic paleogeography and biogeography: Geological Society [London] Memoir 12, p. 1-21.

Scotese, C.R., McKerrow, W.S., 1991. Ordovician plate tectonic reconstructions. In: Barnes, C.R. and Williams, S.H. (Eds.), Advances in Ordovician Geology. Geological Survey of Canada Paper 90-9, pp. 271-282.

112 Shaviv, N.J., Veizer, J., 2003. Celestial driver of Phanerozoic climate? GSA Today 13 (7), 4-10.

Sheehan, P. M., 2001. The Late Ordovician mass extinction. Annual Reviews of Earth and Planetary Science 29, 331-364.

Shields, G.A., Carden, G.A., Veizer, J., Meidla, T., Rong, J., and Li, R., 2003, Sr, C, and O isotope geochemistry of Ordovician brachiopods: A major isotopic event around the Middle-Late Ordovician transition: Geochimica et Cosmochimica Acta, v. 67, p. 2005-2025.

Slupik, L.M., 1999. Sedimentology and stable isotope chemostratigraphy of Late Middle Ordovician carbonates in central Pennsylvania. M. S. Thesis, Pennsylvania State University, State College, PA, 135 pp.

Spooner, E.T.C., 1976, The strontium isotopic composition of seawater, and seawater- oceanic crust interaction: Earth and Planetary Science Letters 31, 167-174.

Soufiane, A., and Achab, A., 2000, Chitinozoan zonation of the Late Ordovician and the Early Silurian of the island of Anticosti, Quebec, Canada: Review of Palaeobotany & Palynology, v. 109, p. 85-111.

Steemans, P. and Wellman, C.H., 2004. Miospores and the emergence of Land Plants. In: Webby, B.D., Paris, F., Droser, M.L., and Percival, I.G. (Eds.), The Great Ordovician Biodiversification event: New York, Columbia Press, p. 361–366.

Stern, R.J., 1982, Strontium isotopes from circum-Pacific intra-oceanic island arcs and marginal basins: Regional variations and implications for magmagenesis: Geological Society of America Bulletin, v. 93, p. 477-486.

Strother, P.K., Al-Hajri, S., and Traverse, A., 1996. New evidence for land plants from the lower Middle Ordovician of Saudia Arabia. Geology 24, 55-59.

Sutcliffe, O.E., Dowdeswell, J.A., Whittington, R.J., Theron, J.N., and Craig, J., 2000, Calibrating the Late Ordovician glaciation and mass extinction by the eccentricity cycles of Earth’s orbit: Geology, v. 28, p. 967-970.

Sweet, W.C., 1983. Conodont Biostratigraphy of Fite Formation and Viola Group. Oklahoma Geological Survey Bulletin 132, p. 23-36.

Underwood, C.J., Crowley, S.F., Marshall, J.D., and Brenchley, P.J., 1997, High- resolution carbon isotope stratigraphy of the basal Silurian Stratotype (Dob’s Linn, Scotland) and its global correlation: Journal of the Geological Society, London, v. 154, p. 709-718. 113

Van Cappellen, P., and Ingall, E.D., 1994, Benthic phosphorous regeneration, net primary production, and ocean anoxia: A model of the coupled marine biogeochemical cycles of carbon and phosphorous. Paleoceanography, v. 9, p. 677-692.

Veizer, J., 1989, Strontium isotopes in seawater through time. Annual Reviews of Earth and Planetary Sciences 17,141-167.

Wang, K., Chatterton, B.D.E., and Wang, Y., 1997, An organic carbon isotope record of Late Ordovician to Early Silurian marine sedimentary rocks, Yangtze Sea, South

China: Implications for CO2 changes during the Hirnantian glaciation: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 132, p. 147-158.

Webby, B.D., Cooper, R.A., Bergström, S.M., and Paris, F., 2004, Stratigraphic framework and time slices, in Webby, B.D. et al., eds., The Great Ordovician Biodiversification Event: New York, Columbia Press, p. 41-47.

Wilde, P., 1991. Oceanography in the Ordovician, In: Barnes, C.R. and Williams, S.H. (Eds.), Advances in Ordovician Geology. Geological Survey of Canada Paper 90- 9, 283-298.

Witzke, B.J., 1987. Models for circulation patterns in epicontinental seas applied to Paleozoic facies of Norrth American craton. Paleoceanography 2, 229-248.

Witzke, B.J., 1990. Palaeoclimatic constraints for Paleozoic paleolatitudes of Laurentia and Euramerica. In: McKerrow, W.S. and Scotese, C.R. (Eds.), Palaeogeography and Biogeography. Geological Society Memoir 12, 57-73.

Wolff-Boenish, D., Gislason, S.R., and Oelkers, E.H., 2006, The effect of crystallinity on dissolution rates and CO2 consumption capacity of silicates: Geochimica et Cosmochimica Acta, v. 70, p. 858-870.

Yapp, C.H., and Poths, H., 1992, Ancient atmospheric CO2 inferred from natural geothites: Nature, v. 161, 165-180.

Young, S.A., Saltzman, M.R., Bergström, S.M., 2005, Upper Ordovician (Mohawkian) carbon isotope (δ13C) stratigraphy in eastern and central North America: regional expression of a perturbation of the global carbon cycle: Palaeogeography, Palaeoclimatology, Palaeoecology 222, 53-76.

Zachos, J.C., and Kump, L.R., 2005, Carbon cycle feedbacks and the initiation of Antarctic glaciation in the earliest Oligocene: Global and Planetary Change, v. 47, p. 51-66.

114

APPENDIX A

87Sr/86Sr analyses from late Katian-Hirnantian (Upper Ordovician) carbonates from the western and eastern Laurentia, and Baltica (Nevada, USA; Anticosti Island,

Quebec, Canada; Kardla, Estonia).

115

Sectio Sr Meters n Formation Biostrat. Zone 87Sr/86Sr Uncertainty (ppm) 3 MR Hanson Creek D. ornatus 0.707921 0.000009 161.8 5 MR Hanson Creek D. ornatus 0.707940 0.000011 219.5 12 MR Hanson Creek D. ornatus 0.708008 0.000007 94.7 31.5 MR Hanson Creek D. ornatus 0.707924 0.000009 153.8 37.5 MR Hanson Creek D. ornatus 0.707928 0.000011 191.3 43.5 MR Hanson Creek D. ornatus 0.707951 0.000011 205.0 48.75 MR Hanson Creek D. ornatus 0.707985 0.000010 285.6 60 MR Hanson Creek D. ornatus 0.708052 0.000010 150.7 63 MR Hanson Creek D. ornatus 0.708010 0.000011 166.7 69 MR Hanson Creek D. ornatus 0.707970 0.000009 172.6 75.5 MR Hanson Creek D. ornatus 0.708079 0.000013 157.2 78 MR Hanson Creek D. ornatus 0.708091 0.000006 167.0 84.5 MR Hanson Creek D. ornatus 0.708167 0.000006 161.7 89.75 MR Hanson Creek D. ornatus 0.708192 0.000011 179.9 90 MR Hanson Creek P. pacificus 0.708194 0.000007 166.7 93 MR Hanson Creek P. pacificus 0.708194 0.000008 144.8 101 MR Hanson Creek P. pacificus 0.708073 0.000014 106.5 107 MR Hanson Creek P. pacificus 0.707930 0.000010 310.9 110 MR Hanson Creek P. pacificus 0.707927 0.000010 322.1 113 MR Hanson Creek P. pacificus 0.707924 0.000008 307.0 116 MR Hanson Creek P. pacificus 0.708031 0.000009 260.2 119 MR Hanson Creek P. pacificus 0.708012 0.000012 340.3 125.5 MR Hanson Creek P. pacificus 0.708018 0.000016 228.5 128.5 CC Hanson Creek N. extraordinarius 0.707878 0.000012 320.3 148.4 CC Hanson Creek N. extraordinarius 0.707977 0.000007 100.8 151.4 CC Hanson Creek N. extraordinarius 0.707974 0.000010 170.1 163.4 CC Hanson Creek N. extraordinarius 0.708014 0.000011 133.5 188.4 CC Hanson Creek N. persculptus 0.708056 0.000017 124.0 190.7 CC Hanson Creek N. persculptus 0.708086 0.000009 113.7 202.3 CC Hanson Creek N. persculptus 0.708235 0.000016 90.6 Monitor Range sections include the Maritn's Ridge (MR) and Copenhagen Canyon (CC) localities.

Table A.1. Strontium Isotope data from central Nevada, USA discussed in Chapter 4.

116

Global Meters Section Formation Stage 87Sr/86Sr Uncertainty Sr (ppm) 145 Kardla core Jonstorp Katian 0.708104 0.000008 187.0 150.2 Kardla core Jonstorp Katian 0.708098 0.000010 132.2 155.2 Kardla core Halliku Katian 0.708134 0.000008 139.5 158.2 Kardla core Halliku Katian 0.708111 0.000006 151.3 160.2 Kardla core Halliku Katian 0.708169 0.000011 110.9 161.2 Kardla core Halliku Katian 0.708033 0.000009 85.8 163 Kardla core Halliku Katian 0.708020 0.000008 86.7 166 Kardla core Halliku Katian 0.708063 0.000008 67.4 168 Kardla core Halliku Katian 0.708031 0.000009 192.6 173.5 Kardla core Halliku Katian 0.708074 0.000009 126.3 175 Kardla core Halliku Katian 0.708078 0.000013 85.4 177 Kardla core Halliku Katian 0.708043 0.000009 155.4 181 Kardla core Halliku Katian 0.707999 0.000008 119.5 185.5 Kardla core Halliku Katian 0.708049 0.000010 86.6 187.3 Kardla core Halliku Katian 0.707994 0.000006 171.0 188.5 Kardla core Halliku Katian 0.707961 0.000009 233.7 190 Kardla core Halliku Katian 0.707959 0.000009 200.6 191.2 Kardla core Halliku Katian 0.707987 0.000009 145.2 195.3 Kardla core Kuldiga Hirnantian 0.708009 0.000012 115.8 197 Kardla core Kuldiga Hirnantian 0.708027 0.000009 120.8 199.3 Kardla core Kuldiga Hirnantian 0.708012 0.000015 141.5 201.1 Kardla core Kuldiga Hirnantian 0.707962 0.000008 180.0 203 Kardla core Kuldiga Hirnantian 0.707969 0.000010 181.3 205.3 Kardla core Kuldiga Hirnantian 0.708025 0.000015 123.0 207.3 Kardla core Saldus Hirnantian 0.707985 0.000011 168.2 209.1 Kardla core Saldus Hirnantian 0.708072 0.000009 41.8 211.2 Kardla core Saldus Hirnantian 0.708151 0.000030 48.3 213.25 Kardla core Õhne Rhudanian 0.708023 0.000010 126.5 215.1 Kardla core Õhne Rhudanian 0.708149 0.000018 55.6 217.5 Kardla core Õhne Rhudanian 0.708066 0.000010 80.9

Table A.2. Strontium Isotope data from Estonia discussed in Chapter 4.

117

Meters Section Formation Biostrat. Zone 87Sr/86Sr Uncertainty Sr (ppm) 0 Oil River Vaureal A. ordovicicus 0.707858 0.000008 668.6 20 Oil River Vaureal A. ordovicicus 0.707829 0.000008 1300.2 32 Oil River Vaureal A. ordovicicus 0.707822 0.000009 1334.0 40 Oil River Vaureal A. ordovicicus 0.707825 0.000008 1145.6 62.5 Oil River Vaureal A. ordovicicus 0.707831 0.000008 933.6 69 Oil River Vaureal A. ordovicicus 0.707816 0.000009 1047.9 80 EH to AAF Vaureal A. ordovicicus 0.707850 0.000007 526.9 90 EH to AAF Vaureal A. ordovicicus 0.707808 0.000007 651.1 100 EH to AAF Vaureal A. ordovicicus 0.707835 0.000008 566.9 114 EH to AAF Vaureal A. ordovicicus 0.707841 0.000008 512.3 120 EH to AAF Vaureal A. ordovicicus 0.707845 0.000008 559.1 138 EH to AAF Vaureal A. ordovicicus 0.707842 0.000008 693.7 149 EH to AAF Vaureal A. ordovicicus 0.707864 0.000009 410.4 162 EH to AAF Vaureal A. ordovicicus 0.707856 0.000009 519.9 179 EH to AAF Vaureal A. ordovicicus 0.707864 0.000011 418.6 185 EH to AAF Vaureal A. ordovicicus 0.707827 0.000006 537.9 200 EH to AAF Vaureal A. ordovicicus 0.707855 0.000012 254.3 210 EH to AAF Vaureal A. ordovicicus 0.707872 0.000009 531.6 218 EH to AAF Vaureal A. ordovicicus 0.707863 0.000006 455.0 234 EH to AAF Vaureal A. ordovicicus 0.707902 0.000012 292.1 242 EH to AAF Vaureal A. ordovicicus 0.707854 0.000006 355.0 260 EH to AAF Vaureal A. ordovicicus 0.707902 0.000011 367.1 268 EH to AAF Vaureal A. ordovicicus 0.707872 0.000009 481.1 280 EH to AAF Vaureal A. ordovicicus 0.707883 0.000008 537.2 294.5 EH to AAF Vaureal A. ordovicicus 0.707880 0.000006 608.3 EH to AAF: tidal flat sections from English Head to Anses aux Fraises

Table A.3. Strontium Isotope data from Anticosti Island, Quebec, Canada discussed in Chapter 4.

118

APPENDIX B

Additional δ13C analyses from the Vaureal Formation, and lower two members of

Ellis Bay Formation on Anticosti Island, Quebec, Canada.

119

13 18 13 Meters δ Ccarb δ O δ Corg TOC Formation Member 0 -0.21 -4.62 -27.77 0.06 Vaureal Tower 4 0.19 -4.71 Vaureal Tower 8 0.13 -4.68 Vaureal Tower 12 0.22 -4.41 Vaureal Tower 20 -0.23 -4.81 -28.45 0.03 Vaureal Tower 24 -0.19 -4.42 Vaureal Tower 28 0.12 -4.13 Vaureal Tower 32 -0.04 -4.41 Vaureal Tower 36 -0.51 -4.70 Vaureal Tower 40 -0.46 -5.04 -28.09 0.53 Vaureal Tower 44 0.01 -4.92 Vaureal Tower 48 -0.14 -5.00 Vaureal Tower 52 -0.49 -4.27 Vaureal Tower 56 -0.51 -5.08 Vaureal Tower 62.5 -0.51 -4.54 -28.57 0.05 Vaureal Tower 69 -0.63 -5.10 Vaureal Homard 80 -0.09 -3.89 -28.45 0.06 Vaureal Homard 82 -0.21 -4.86 Vaureal Homard 84 0.02 -4.21 Vaureal Homard 86 -0.26 -4.30 Vaureal Homard 88 -0.42 -4.50 Vaureal Homard 90 -0.24 -4.36 Vaureal Homard 92 0.02 -4.39 Vaureal Homard 94 -0.04 -4.72 Vaureal Homard 96 -0.34 -4.37 Vaureal Homard 98 -0.09 -4.31 Vaureal Homard 100 -0.06 -4.03 -28.41 0.03 Vaureal Homard 103 -0.21 -3.93 Vaureal Homard 105 -0.21 -4.21 Vaureal Homard 107 0.04 -4.16 Vaureal Homard 109 -0.30 -4.22 Vaureal Homard 111 -0.27 -5.01 Vaureal Homard 114 0.05 -3.82 Vaureal Homard 116 -0.21 -4.72 Vaureal Homard 118 0.10 -4.45 Vaureal Homard 120 0.13 -4.41 -27.30 0.04 Vaureal Homard 122 0.11 -4.27 Vaureal Homard 124 0.02 -4.19 Vaureal Homard 126 0.40 -4.04 Vaureal Homard

Continued

Table B.1. Additional Stable Isotope data from Anticosti Island, Quebec, Canada. 120 Table B.1 continued

128 0.37 -3.92 Vaureal Homard 130 0.55 -4.05 Vaureal Homard 132 0.42 -4.54 Vaureal Homard 134 0.41 -4.71 Vaureal Homard 136 0.39 -4.53 Vaureal Homard 138 0.67 -3.91 -27.61 0.04 Vaureal Homard 140 0.63 -3.96 Vaureal Homard 143 0.39 -3.89 Vaureal Homard 145 0.50 -4.11 Vaureal Homard 147 0.51 -4.29 Vaureal Homard 149 0.27 -4.45 Vaureal Homard 151 0.27 -4.14 Vaureal Homard 153 0.36 -5.07 Vaureal Homard 155 0.52 -4.29 Vaureal Homard 158 0.54 -3.87 Vaureal Homard 160 0.49 -4.10 Vaureal Homard 162 0.07 -4.50 -27.44 0.00 Vaureal Homard 164 0.53 -4.27 Vaureal Homard 166 0.43 -4.08 Vaureal Homard 169 0.19 -4.23 Vaureal Homard 171 0.07 -4.51 Vaureal Homard 173 0.40 -3.76 Vaureal Homard 174 0.38 -3.86 Vaureal Joseph Point 177 0.30 -3.93 Vaureal Joseph Point 179 0.26 -3.93 -28.70 0.03 Vaureal Joseph Point 181 0.27 -4.35 Vaureal Joseph Point 183 0.61 -3.79 Vaureal Joseph Point 185 0.52 -3.52 Vaureal Joseph Point 188 0.28 -3.94 Vaureal Joseph Point 192 0.39 -4.15 Vaureal Joseph Point 196 0.17 -4.14 Vaureal Joseph Point 200 0.34 -3.75 -28.17 0.03 Vaureal Joseph Point 204 0.35 -3.70 Vaureal Joseph Point 208 0.65 -3.57 Vaureal Joseph Point 210 0.80 -4.37 Vaureal Joseph Point 214 0.60 -3.82 Vaureal Joseph Point 218 0.59 -3.58 -28.15 0.05 Vaureal Joseph Point 222 0.42 -3.92 Vaureal Joseph Point 226 0.10 -3.60 Vaureal Joseph Point 230 0.44 -5.11 Vaureal Joseph Point 234 0.41 -3.89 Vaureal Joseph Point 238 0.31 -3.62 Vaureal Joseph Point

Continued 121 Table B.1 continued

242 -0.14 -3.83 -28.48 0.05 Vaureal Joseph Point 246 0.22 -3.56 Vaureal Joseph Point 250 0.01 -4.20 Vaureal Joseph Point 256 0.39 -4.01 Vaureal Joseph Point 260 -0.64 -4.18 -28.41 0.08 Vaureal Mill Bay 268 0.56 -3.54 Vaureal Mill Bay 272 0.47 -3.87 Vaureal Mill Bay 280 0.82 -3.46 -28.21 0.08 Vaureal Mill Bay 284 0.78 -3.46 Vaureal Mill Bay 288 0.08 -3.44 Vaureal Mill Bay 291.1 1.03 -3.72 Vaureal Schmitt Creek 294.5 0.69 -3.58 Vaureal Schmitt Creek 296.5 -0.07 -3.91 Ellis Bay Grindstone 298.5 0.53 -4.30 -27.01 0.05 Ellis Bay Grindstone 300.5 0.85 -3.71 Ellis Bay Grindstone 302.5 0.91 -3.74 Ellis Bay Grindstone 304.5 1.34 -3.32 Ellis Bay Grindstone 306.5 1.69 -3.52 Ellis Bay Grindstone 308.5 1.68 -3.30 Ellis Bay Grindstone 310.5 1.43 -3.60 -27.39 0.06 Ellis Bay Grindstone 312.5 1.24 -3.03 Ellis Bay Grindstone 314.5 1.00 -4.06 Ellis Bay Grindstone 316.5 1.12 -3.96 Ellis Bay Grindstone 318.5 1.25 -3.46 -27.68 0.05 Ellis Bay Velleda 320.5 1.37 -3.70 -28.53 0.06 Ellis Bay Velleda 322.5 1.12 -3.32 Ellis Bay Velleda 324.5 1.41 -4.53 -27.97 0.04 Ellis Bay Velleda 326.5 0.18 -2.67 Ellis Bay Velleda 328.5 1.16 -3.72 -28.58 0.08 Ellis Bay Velleda 330.5 0.59 -3.47 Ellis Bay Velleda 332.5 -0.02 -3.31 -28.67 0.06 Ellis Bay Velleda 334.5 1.36 -3.47 -28.23 0.10 Ellis Bay Velleda 336.5 1.32 -3.42 -28.32 0.07 Ellis Bay Velleda 338.5 1.02 -3.59 -27.73 0.10 Ellis Bay Velleda 340.5 0.69 -3.58 -27.66 0.08 Ellis Bay Velleda

122

APPENDIX C

13 13 18 Cross plots of δ Ccarb, δ Corg, δ O, and total organic carbon (TOC) data from

Chapters 3 and 4.

123

(‰ VPDB) (‰ org C 13 δ

TOC (wt. %)

13 Figure C.1. Cross plot of δ Corg versus total organic carbon (TOC) data from Oklahoma (circles), West Virginia (diamonds), and China (triangles) in Chapter 3.There is no 13 systematic relationship between δ Corg and the wt % TOC as might be expected if differential alteration of organic matter in intervals that are poor versus relatively rich in organic carbon.

124

O (‰ O VPDB) 18 δ

13 δ Ccarb (‰ VPDB)

13 18 Figure C.2. Cross plot of δ Ccarb versus δ O data from Anticosti Island (circles) and 13 18 Estonia (diamonds) in Chapter 4 and Appendix B. δ Ccarb and δ O from the Estonian 13 sequence do not covary in the manner expected to reset primary δ Ccarb during meteoric diagenesis. Although there appears to be some covariance in the Anticosti Island data set, 13 18 both δ Ccarb and δ O data are well within the ranges of previously published values from pristine brachiopods and marine cements analyzed from Anticosti Island (Long, 1993; Brenchley et al., 1994)

125

(‰ VPDB) (‰ org C 13 δ

TOC (wt. %)

13 Figure C.3. Cross plot of δ Corg versus total organic carbon (TOC) data from Anticosti Island (circles) and Estonia (diamonds) in Chapter 4 and Appendix B. There is no 13 systematic relationship between δ Corg and the wt % TOC as might be expected if differential alteration of organic matter in intervals that are poor versus relatively rich in organic carbon.

126