Integrating Detailed Petrography, Geochemistry, and Mineralogy to Elucidate Extensive Early Diagenesis in the Eocene Carbonates of Qatar

Western Michigan University ScholarWorks at WMU

Dissertations Graduate College

6-2020

Brooks H. Ryan Western Michigan University, [email protected]

Follow this and additional works at: https://scholarworks.wmich.edu/dissertations

Part of the Geochemistry Commons, and the Geology Commons

Recommended Citation Ryan, Brooks H., "Integrating Detailed Petrography, Geochemistry, and Mineralogy to Elucidate Extensive Early Diagenesis in the Eocene Carbonates of Qatar" (2020). Dissertations. 3562. https://scholarworks.wmich.edu/dissertations/3562

This Dissertation-Open Access is brought to you for free and open access by the Graduate College at ScholarWorks at WMU. It has been accepted for inclusion in Dissertations by an authorized administrator of ScholarWorks at WMU. For more information, please contact [email protected]. INTEGRATING DETAILED PETROGRAPHY, GEOCHEMISTRY, AND MINERALOGY TO ELUCIDATE EXTENSIVE EARLY DIAGENESIS IN THE EOCENE CARBONATES OF QATAR

Brooks H. Ryan

A dissertation submitted to the Graduate College in partial fulfillment of the requirements for the degree of Doctor of Philosophy Geological and Environmental Sciences Western Michigan University June 2020

Doctoral Committee:

Stephen E. Kaczmarek, Ph.D., Chair John M. Rivers, Ph.D. R.V. Krishnamurthy, Ph.D.

INTEGRATING DETAILED PETROGRAPHY, GEOCHEMISTRY, AND MINERALOGY TO ELUCIDATE EXTENSIVE EARLY DIAGENESIS IN THE EOCENE CARBONATES OF QATAR

Brooks H. Ryan, Ph.D.

Western Michigan University, 2020

Eocene carbonates comprising the Umm er Radhuma (UER), Rus, and Dammam Formations, cover 80% of the surface and extend >300 m into the subsurface of Qatar. These rocks record marine sedimentation in shallow sub-tidal to restricted settings. Despite undergoing only shallow burial

(<1000 m), extensive diagenetic alteration has occurred. Given that little work on this topic has been published previously, the objectives of this work are to document diagenesis in the Eocene, and to integrate petrography, mineralogy, and geochemistry to delineate the timing and environments of diagenesis of these Eocene carbonates. Detailed petrographic data show that dolomitization occurred early in the UER, before the formation of diagenetic chert, palygorskite, pyrite, calcite, and gypsum cements. Bulk dolomite δ18O values, coupled with clumped isotope-derived (∆47) temperatures and dolomitizing fluid δ18O values, suggest dolomitization took place in near-normal marine fluids, perhaps during shallow burial. These data challenge the current paradigm of large-scale, top-down hypersaline reflux dolomitization in the UER. Associated depth trends of increasing crystal size and with increasing stoichiometry and cation ordering but variable δ18O further suggest that UER dolomites were extensively recrystallized prior to the formation of all other mineral phases. This finding challenges an analysis of the literature which suggests that extensive dolomite recrystallization only happens late in the diagenetic history and/or with deep burial. Similarly, the overlying Rus Fm. consists of dolomitized peritidal facies interbedded with gypsum, as well as more open marine deposits higher in the formation. However, petrographic data shows dolomite formed prior to gypsum cementation, and bulk δ18O and δ13C data suggest a non-evaporative origin. The intimate association between dolomites and meteoric calcites elsewhere in the section indicate that

Rus dolomites have likely been recrystallized multiple times, possibly in meteoric-related fluids. The key findings of this study are that early diagenesis can be extremely complex, and that current paradigms related to reflux dolomitization and dolomite recrystallization may need to be reevaluated.

ACKNOWLEDGEMENTS

I would like to thank all of my family members, friends, and colleagues who have unconditionally supported me throughout this 4-year journey. Special thanks go to Steve, who has been more than an academic advisor, but also a life advisor, during my time in Kalamazoo. His efforts in shaping my life after the PhD should not go unrecognized. I would also like to thank the criticisms and support of my committee members John and RV. John has essentially been a co- advisor for this dissertation, and I am grateful for his ideas in helping shape much of my interpretations presented herein. I also want to thank John and his family for their hospitality while we stayed with them in Qatar for a week in 2018 and 2019. I also want to acknowledge Dr. David

Budd from the University of Colorado, Boulder for his highly insightful and educational critiques of two chapters of this dissertation that were submitted for publication. His input has been invaluable, and I am very appreciative for his thorough reviews of my work. Special thanks also go to my parents who have always taught me to work hard and do the right thing at all stages of my life. I would like to acknowledge the funding support of ExxonMobil Research Qatar, who not only provided funding for all research aspects of this project but also funded me for 1.5 years as a

Research Assistant. Thanks to Western Michigan University’s Department of Geological and

Environmental Sciences, who funded me the other 2.5 years through a Teaching Assistantship. The

American Association of Petroleum Geologists and the Society for Sedimentary Geology are also acknowledged for funding throughout this dissertation. Lastly, I would like to acknowledge support from my German Shepherd, Dolo, whom I raised during the last half of this dissertation. I named him in honor of the mineral dolomite, which makes up the bulk of this dissertation.

Brooks H. Ryan

ii TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

LIST OF TABLES ...... v

LIST OF FIGURES ...... vi

CHAPTER

I. INTRODUCTION ...... 1

References ...... 6

II. EARLY AND PERVASIVE DOLOMITIZATION BY NEAR-NORMAL MARINE FLUIDS: NEW LESSONS FROM AN EOCENE EVAPORITVE SETTING IN QATAR ...... 10

Abstract ...... 10

Introduction ...... 11

Geological Background ...... 13

Materials and Methods ...... 17

Results ...... 23

Discussion ...... 36

Conclusions ...... 50

Acknowledgements ...... 51

Data Availability Statement ...... 51

References ...... 51

III. EARLY AND EXTENSIVE RECRYSTALLIZATION OF CENOZOIC DOLOMITES DURING SHALLOW BURIAL ...... 62

Abstract ...... 62

Introduction ...... 63

Geological Setting ...... 64

iii Table of Contents—continued

Methods ...... 65

Results ...... 66

Discussion ...... 68

Conclusions ...... 72

Acknowledgements ...... 72

References ...... 72

Supplementary Material ...... 75

IV. DOLOMITE DISSOLUTION: AN ALTERNATIVE DIAGENETIC PATHWAY FOR THE FORMATION OF PALYGORSKITE CLAY ...... 97

Abstract ...... 97

Introduction ...... 98

Geologic Background ...... 100

Materials and Methods ...... 101

Results ...... 103

Discussion ...... 110

Conclusions ...... 127

Acknowledgements ...... 128

References ...... 128

V. MULTI-EPISODIC RECRYSTALLIZATION AND ISOTOPIC RESETTING OF DOLOMITES IN NEAR-SURFACE SETTINGS ...... 137

Abstract ...... 137

Introduction ...... 139

Geologic Background ...... 142

Materials and Methods ...... 144

Results ...... 146

iv Table of Contents—continued

Discussion ...... 157

Implications ...... 171

Conclusions ...... 173

Acknowledgements ...... 174

References ...... 175

v LIST OF TABLES

CHAPTER II

1. Stoichiometry statistics for 4 UER dolomites analyzed by XRD, EMPA, and EDS ...... 32

2. UER Dolomite Stable Isotope Analysis ...... 33

3. Qatar gypsum δ18O analysis ...... 35

CHAPTER III

1. Core 1 clumped isotope analysis ...... 67

Supplementary Material

1. Total Variance Explained ...... 77

2. Component Matrix and Factor Loadings ...... 77

3. Analysis of Natural Recrystallized Dolomites in Literature ...... 78

vi LIST OF FIGURES

CHAPTER II

1. Generalized stratigraphic column of the Eocene sediments of Qatar ...... 14

2. A palaeogeographic map of the eastern half of the Arabian Peninsula ...... 16

3. Mineralogical percentages for the Umm er Radhuma (UER) recovered in cores 1, 2, and 3 ...... 19

4. Plane polarized light (PPL) thin section images of observed dolomite textures ...... 24

5. Thin section images in plane polarized light (PPL) of various allochem features ...... 25

6. Non-carbonate diagenetic features ...... 26

7. Diagenetic relationships between dolomite, palygorskite and gypsum ...... 28

8. Thin section images in plane polarized light (PPL) displaying the relationship between calcite, dolomite, palygorskite and chalcedony ...... 29

9. Thin section images (plane polarized light – PPL) from the unconsolidated section of core 2 ...... 30

10. Core profile of core 1 displaying mineralogy and facies associations (A) with corresponding dolomite stoichiometry (mol% MgCO3) (B), cation ordering (C), bulk dolomite δ18O values (D), bulk dolomite δ13C values (E), and Sr and Na concentrations (F) ...... 30

11. The mineralogy and facies associations of core 2 (A) correlated to dolomite stoichiometry (mol% MgCO3) (B), cation ordering (C), bulk dolomite δ18O values (D), bulk dolomite δ13C values (E), and Sr and Na concentrations (F) ...... 31

12. Depth profile of core 3 exhibiting mineralogy and facies associations (A) corresponding to dolomite stoichiometry (mol% MgCO3) (B), cation ordering (C), bulk dolomite δ18O values (D), bulk dolomite δ13C values (E), and Sr and Na concentrations (F) ...... 31

13. Paragenetic sequence for the Umm er Radhuma (UER) in Qatar based on petrographic cross-cutting relationships and geochemical data ...... 37

CHAPTER III

1. A stratigraphic column of the Paleogene strata of Qatar (modified after Ryan et al., 2020), with the specific study interval (red) ...... 65 vii List of Figures—continued

2. Cross plots of cation ordering and stoichiometry (A) and stable isotopic composition (B) as a function of dolomite texture for all cores ...... 67

CHAPTER IV

1. Map of Qatar and surrounding region ...... 100

2. Stratigraphic column for Well RR-01 ...... 102

3. Photographs showing some of the common sedimentological features observed in the Umm er Radhuma ...... 104

4. Powder X-ray diffraction pattern of a representative sample (110.45 m depth) containing ca 76% dolomite and ca 20% palygorskite from 5° to 55° 2q ...... 106

5. High-resolution, plane polarized light (PPL) thin section images of dolomite (light brown to light grey rhombs) and palygorskite (green haze) from the Umm er Radhuma ...... 108

6. Scanning electron microphotograph (SEM) images of dolomite and palygorskite ...... 109

7. Electron dispersive analysis of the same dolomite crystal as Fig. 6A and C ...... 110

8. Scanning electron microphotograph (SEM) of dolomite and palygorskite from well depth of 89.55 m ...... 111

9. Backscatter SEM photomicrograph showing the locations where energy-dispersive spectroscopy (EDS) spot analyses were collected on the highly polished thin section (78.19 m) ...... 115

10. Evolutionary model of the dolomitization reaction, growth of cloudy-centre clear-rim dolomite fabrics and post-dolomitization dissolution of dolomite with subsequent palygorskite formation ...... 117

11. Stability diagrams of palygorskite as a function of pH, H4SiOH4, and Mg ...... 120

12. Stability diagrams modified from Birsoy (2002) displaying the stability boundaries of various minerals within a seven component MgO-CaO-Al2O3-SiO2-H2O-CO2-HCl system .... 122

CHAPTER V

1. A generalized stratigraphic section of the Paleocene and Eocene sediments of Qatar, as well as a location map of the Arabian Peninsula, Qatar, and the three research cores .... 143

2. Percent mineralogy for cores 1, 2, and 3 for the recovered Rus Fm ...... 144

viii List of Figures—continued

3. Plane polarized light thin section images of various petrographic features of the dolomites in the Traina Mbr ...... 149

4. A plane polarized light thin section image of a dolomitized miliolid packstone with the fabric preserved from the Al Khor Mbr ...... 150

5. Features of the dolomite-calcite contact observed at ~28 m depth in core 1 ...... 150

6. In the Traina Mbr. of core 2, large (> 200 µm) poikilotopic calcite crystals (pink) engulf planar-e dolomite crystals ...... 151

7. Plane polarized light image of the Traina Mbr., exhibiting dolomite crystals (tan) cemented by palygorskite (dark green) ...... 152

8. Plane polarized light image of chalcedony (white) that has partially replaced calcite crystals (pink) (A) and a cross polarized light image of euhedral quartz crystals (grey and black) included in calcite mosaics (pink) (B) in core 2 ...... 152

9. Plots of stoichiometry with depth for cores 1 (A), 2 (B), and 3 (C) ...... 153

10. Cross plot of bulk dolomite δ18O and δ13C values for all Rus samples ...... 155

11. Bulk rock δ18O from the Rus plotted as a function of depth for cores 1 (A), 2 (B), and 3 (C) ...... 156

12. Bulk rock δ13C from the Rus plotted as a function of depth for cores 1 (A), 2 (B), and 3 (C) ...... 157

13. Paragenetic sequence of the relative timing of diagenetic events impacting the Rus in Qatar, based on petrographic cross-cutting relationships ...... 158

ix CHAPTER I

INTRODUCTION

The Greater Ghawar Uplift and Persian Gulf provinces of the Arabian Peninsula are associated with some of the largest hydrocarbon reserves on earth (Pollastro, 2003). The South

Pars-North Field, offshore Qatar, is the largest gas and condensate field in the world, with estimated reserves totaling ~500 trillion cubic feet of gas (TCFG) (Pollastro, 2003; Perotti et al., 2011). The geology of such fields has been well documented, with major reservoirs existing in the Greater

Paleozoic, Jurassic, and Cretaceous Petroleum Systems (Pollastro, 2003). In Qatar, Cenozoic rocks ranging in age from Late Paleocene to Recent overly the Cretaceous System. The Eocene rocks, although not hydrocarbon reservoirs, are of particular importance, as they are valuable groundwater aquifers and provide water resources to the country (Eccleston et al., 1981;

Baalousha, 2016).

The Eocene stratigraphy of Qatar is represented by the Umm er Radhuma, Rus, and

Dammam Formations. Much of what is known about the sedimentology of the Eocene in Qatar comes from early studies of the Arabian Peninsula (Powers et al., 1966), the surficial sediments

(Cavelier, 1970), and the subsurface water resources (Eccleston et al., 1981). A few recent studies, however, have further characterized the sediments, rocks, and stratigraphy of the Qatar Eocene formations, reporting that the rocks are dominated by limestones and dolomites formed in shallow marine to intermittently restricted environments (Abu-Zeid, 1991; Al-Hajari and St. C. Kendall,

1992; Al-Saad, 2003; Al-Saad, 2005; Holail et al., 2005; Rivers et al., 2019). Much focus has also been placed on refining the biostratigraphy and formation boundaries (Smout, 1954; Powers et al. 1966;

Hasson, 1985; Hewaidy, 1994; Boukhary et al. 2011). Most of this knowledge has led to the understanding that Qatar experienced a transgression during the Late Paleocene-Early Eocene, a 1 regression during the Early Eocene, and a subsequent transgression in the Middle Eocene. These events correspond to the deposition of the Umm er Radhuma Fm., Rus Fm., and Dammam Fm., respectively.

Despite the recent efforts to advance the understanding of the Eocene depositional system, little has been done to characterize the diagenetic alterations that have taken place since deposition. Holail and Al-Hajari (1997) and Holail et al. (2005) are the earliest published studies to analyze the diagenetic features in Qatar, but they are limited to the Lower Dammam and Upper

Dammam subformations, respectively. A better understanding of the diagenetic impacts on the

Umm er Radhuma and Rus formations has recently come to light (Rivers et al. 2019), but the general lack of diagenetic understanding in Qatar is puzzling considering that diagenetic alterations of all three Eocene formations have been recorded across the Arabian Peninsula, including Kuwait,

Oman, Saudi Arabia, and the United Arab Emirates (Bou-Rabee and Burke, 1987; Whittle and

Alsharhan, 1994; Whittle et al., 1996; Al-Awadi et al., 1998; El-Saiy and Jordan, 2007; Hersi, 2011;

Khalaf, 2011; Pollitt et al., 2012; Khalaf and Abdullah, 2013; Khalaf et al., 2018).

As part of an in-depth geological study of the Eocene carbonates in Qatar being spearheaded by ExxonMobil Research Qatar in collaboration with Western Michigan University, the purpose of the present study is to bridge the gap between the diagenetic alterations impacting the Eocene successions of Qatar and those documented in the surrounding areas. A diagenetic study of the

Eocene successions of Qatar is of importance for energy exploration, because it is interpreted that the Paleocene-Eocene strata have not been buried further than modern day depths (50-400 m) based on studies focusing on the tectonic and depositional history of Qatar (Van Buchem et al.,

2014; Rivers and Larson, 2018). What this means is that it is unlikely the Eocene rocks have been subjected to deep burial (> 1 km) alterations, and thus the early near-surface diagenetic effects are likely preserved in these rocks. This has sparked the interest of ExxonMobil Research Qatar, as these rocks likely contain information about how early diagenesis impacts the pore system, and can

thus be utilized as an analogue for hydrocarbon exploration and recovery in other localities. This study seeks to address the first part of the problem: characterizing the early diagenetic alterations.

From 2016-2018, 3 wells were drilled and cored in order to better understand the stratigraphy, diagenesis, and hydrology of the Eocene formations. Data was collected from these cores and incorporated into this dissertation to specifically answer the following questions:

(1) What diagenetic alterations are observed in the Eocene rocks of Qatar?

(2) What is the relative order of diagenetic events, specifically with regard to the carbonate

minerals?

(3) What environments and/or processes are responsible for the diagenetic alterations to the

carbonate minerals?

Extensive petrographic methods were utilized to analyze cross cutting relationships and determine the timing of all diagenetic minerals and textures observed. Subsequent mineralogical and geochemical data was integrated to shed light on interpretations relating to diagenetic environments and processes responsible for the alterations. The ultimate outcome is a complete and detailed paragenetic sequence characterizing the Eocene formations, and insights into the diagenetic environments that were present. This study has resulted in the first extensive diagenetic characterization of the Eocene sediments of Qatar, presented in the following chapters and briefly summarized below.

Chapter II presents “Early and pervasive dolomitization by near-normal marine fluids: New lessons from an Eocene evaporative setting Qatar,” which reports on the diagenetic alterations and ultimate paragenetic sequence of the Umm er Radhuma (UER) in Qatar. Previous studies on the

Umm er Radhuma in Saudi Arabia and Kuwait had suggested that dolomitization of the Umm er

Radhuma occurred by refluxing of hypersaline brines, implying that evaporite minerals such as anhydrite and gypsum were deposited and precipitated early, whereas dolomite formed later by the downward flow of dense hypersaline brines (Pollitt et al., 2012; Saller et al. 2014). However,

petrographic relationships presented in Chapter II show that the dolomites in the UER of Qatar formed prior to gypsum cementation. Furthermore, downward trends in stoichiometry (mol%

MgCO3) directly underneath the overlying Rus evaporite bed are inconsistent with that expected from refluxing fluids that show a decreased in Mg/Ca ratios with depth. Lastly, the dolomite geochemical data, including dolomite δ18O, Sr, and Na, are more compatible with a near-normal marine origin (normal seawater with possibly slightly elevated temperatures or salinities), rather than a hypersaline origin. The principle finding from this study was that dolomites overlain and cemented by evaporites do not need to have formed from evaporative fluids. This finding builds upon other recent studies which came to similar conclusions, and challenges the long-held paradigm that dolomites overlain by evaporites formed via reflux dolomitization. This article was first published by Ryan et al. (2020) in the peer-review scientific journal Sedimentology and has been reprinted here in compliance with the copyright license agreement.

Chapter III builds on Chapter II, and utilizes changes in dolomite textures which display increasing crystal size with depth, and correlates these textural changes with changes in dolomite stoichiometry, cation ordering, and δ18O. These data suggest that mimetic dolomites (finest crystal size), which are the least stoichiometric, least ordered, and contain a narrow range of δ18O have been not been significantly recrystallized, whereas the coarser planar-e and nonplanar dolomites, which are more stoichiometric, more well-ordered and variable in δ18O have undergone significant recrystallization. A principal component analysis is consistent with the interpretation that changes in stoichiometry, cation ordering, and δ18O are correlated to dolomite recrystallization.

Furthermore, nonplanar intervals in all cores contain a chert band with silicified planar-e dolomites, and one core contains two such bands—one with planar-e dolomites and one with mimetic dolomites. It is interpreted that these silicified bands capture the early recrystallization of these dolomites. Based on the ∆47-derived temperatures suggesting burial depths < 400 m and

δ18Ow values that are indicative of slightly evaporated Eocene seawater, the main finding in this

chapter is that Cenozoic dolomites can be extensively recrystallized early and in the near-surface realm. This is at odds with a large literature review of recrystallized dolomites that demonstrates that ~90% are post-Cenozoic in age, and ~94% of all studies interpret an intermediate (500-1000 m) to deep (>1000 m) burial origin. This work has been submitted to Geology and is currently under review.

Chapter IV sheds light on the paragenetic sequence after dolomite recrystallization occurred, focusing on the relationship between dolomite and palygorskite in the UER. This chapter presents results showing that planar-e dolomite crystals display cloudy cores and clear rims, in which the cloudy cores are partially to completely dissolved and the clear rims are pristine and intact. Palygorskite is observed coating the outside of dolomite crystals, nucleating on partially dissolved crystals, and has partially to completely filled dissolved dolomite cores. These results suggest that the formation of palygorskite postdates dolomitization and is concurrent with dolomite dissolution. However, the presence of clay minerals within interpreted depositional cycles suggests that palygorskite is an alteration product of a pre-existing clay. This chapter discusses how dolomite dissolution alters porewater chemistry by buffering meteoric-influenced fluids that are slightly acidic and contain H4SiO4, and releasing Mg2+ ions to drive the transformation of smectite to palygorskite. The focus of this chapter is providing direct evidence that dolomite dissolution can drive the formation of palygorskite, which is a hypothesis that has been put forth prior but lacked physical evidence. This study (Ryan et al., 2019) is currently published in Sedimentology.

Chapter V shifts the focus away from the UER and into the overlying Rus Fm. This chapter outlines how dolomites that are interbedded with evaporites likely did not form in evaporative fluids based on dolomite δ18O values, δ13C values, and the observation that dolomite is included in gypsum, suggesting dolomitization occurred prior to deposition of gypsum. This builds upon the principle findings in Chapter II, and further challenges the common paradigm of hypersaline reflux dolomitization. The Rus dolomites also contain a negative δ13C signature that trends towards near-

normal marine values with depth away from an exposure surface containing meteoric calcites. This is in spite of the fact that the dolomites have significantly higher δ18O values, and dolomite crystals are included in calcite crystals, suggesting dolomites formed prior to calcite. The important findings of this chapter are that near-surface diagenesis can be extremely complex, dolomites overlain by bedded evaporites do not need to have formed from reflux dolomitization, and that dolomites are relatively susceptible to multiphase recrystallization, especially in meteoric-influenced fluids. This research is currently under review in the Journal of Sedimentary Research.

References

Abu-Zeid, M.M., 1991, Lithostratigraphy and framework of sedimentation of the sub-surface

Paleogene succession in northern Qatar: Neue Jahrbuch für Geologie und Paleaontologie, v.

4, p. 191-204.

Al-Awadi, E., Mukhopadhyay, A., and Al-Senafy, M.N., 1998, Geology and hydrogeology of the

Dammam Formation in Kuwait: Hydrogeology Journal, v. 6, p. 302-314.

Al-Hajari, S.A., and Kendall, C.G.S.C., 1992, The sedimentology of the Lower Eocene Rus

Formation of Qatar and neighboring regions: J. Univ. Kuwait (Sci.), v. 19, p. 153-172.

Al-Saad, H., 2003, Facies analysis, cyclic sedimentation and paleoenvironment of the Middle

Eocene Rus Formation in Qatar and adjoining areas: Carbonates and Evaporites, v. 18, p. 41-

50.

Al-Saad, H., 2005, Lithostratigraphy of the Middle Eocene Dammam Formation in Qatar, Arabian

Gulf: effects of sea-level fluctuations along a tidal environment: J. Asian Earth Sci., v. 25, p.

781–789.

Baalousha, H.M., 2016, Development of groundwater flow model for the highly parameterized Qatar

aquifers: Model. Earth Syst. Environ., v. 67, 11 p.

Bou-Rabee, F. and Burke, C.D. (1987) Lithostratigraphy and environment of deposition of the

top of the Wafra member, Radhuma formation (Wafra field, Kuwait): Palaeogeography, P

alaeoclimatology, Palaeoecology, v. 59, p. 269-277.

Boukhary, M., Hewaidy, A.G., Luterbacher, H., Bassiouni, M.E., and Al-Hitmi, H., 2011,

Foraminfera and ostracodes of Early Eocene Umm er Radhuma Formation, Dukhan Oil Field,

Qatar: Micropaleontology, v. 57, p. 37-60.

Cavelier, C., 1970, Geologic Description of the Qatar Peninsula, Department of Petroleum

Affairs, Paris, France, 39 pp.

Eccleston, B.L., Pike, J.G., and Harshash, I., 1981, The water resources of Qatar and their

development: Vol. 1. Food and Agricultural Organization (FAO) of the United Nations, Doha,

Technical Report No. 5., 431 pp.

El-Saiy, A.K., and Jordan, B.R., 2007, Diagenetic aspects of tertiary carbonates west of the Northern

Oman Mountains, United Arab Emirates: Journal of Asian Earth Sciences, v. 31, p. 35-43.

Hasson, P.F., 1985, New observations on the biostratigraphy of the Saudi Arabian Umm er Radhuma

Formation (Paleogene) and its correlation with neighboring regions: Micropaleontology, v.

31, p. 335-364.

Hersi, O.S., 2011, Lithologic and diagenetic attributes of the Sharwayn (Maastrichtian) and Umm er

Radhuma (late Paleocene-Eocene) formations and their significance to the K-T

unconformity, Jabal Samhan area, Dhofar, Sultanate of Oman: Arab J Geosci, v. 4, p. 147-160.

Hewaidy, A.G., 1994, Biostratigraphy of the Umm er Radhuma Formation in South-East Qatar,

Arabian Gulf. N. Jb. Geol. Palaont. Abh., v. 193, p. 145-164.

Holail, H., and Al-Hajari, S., 1997, Evidence of an Authigenic Origin for the Palygorskite in a

Middle Eocene Carbonate Sequence from North Qatar. Qatar Univ. Sci. J., v. 17, p. 405-418.

Holail, H.M., Shaaban, M.N., Mansour, A.S., and Rifai, I., 2005, Diagenesis of the Middle

Eocene Upper Dammam subformation, Qatar: Petrographic and isotopic evidence:

Carbonates and Evaporites, v. 20, p. 72-81.

Khalaf, F.I., and Abdullah, F.A., 2013, Petrography and diagenesis of cavity-fill dolocretes, Kuwait:

Geoderma, v. 207-208, p. 58-65.

Khalaf, F.I., Abdullah, F.A., and Gharib, I.M., 2018, Petrography, diagenesis and isotope

geochemistry of dolostones and dolocretes in the Eocene Dammam Formation, Kuwait,

Arabian Gulf: Carbonates and Evaporites, v. 33, p. 87-105.

Pollastro, R.M., 2003, Total Petroleum Systems of the Paleozoic and Jurassic, Greater Ghawar Uplift

and Adjoining Provinces of Central Saudi Arabia and Northern Arabian-Persian Gulf: U.S.

Geological Survey Bulletin, 2202-H, 107 pp.

Powers, R. W., Ramirez, L. F., Redmond, C. D., and Elberg, E. L., 1966, Geology of the

Arabian Peninsula (Sedimentary Geology of Saudi Arabia). Washington DC: U. S. Geological

Survey. Survey, Professional Paper 560D, 147pp.

Perotti, C.R., Carruba, S., Rinaldi, M., Bertozzi, G., Feltre, L., and Rahimi, M., 2011, The Qatar-

South Fars Arch Development (Arabian Platform, Persian Gulf): Insights from Seismic

Interpretation and Analogue Modelling. In: New Frontiers in Tectonic Research – at the

Midst of Plate Convergence (Ed. U. Schattner), p. 325-352.

Pollitt, D.A., Anthonissen, E., Saller, A.H., BouDagher-Fadel, M.K., and Dickson, J.A.D.,

2012, Abrupt early Eocene global climatic change as a control of carbonate facies and

diagenesis: a new record of the Palaeocene-Eocene Thermal Maximum in the Umm er

Radhuma Formation, Saudi Arabia and Kuwait: Terra Nova, v. 24, p. 487-498.

Rivers, J.M., and Larson, K.P., 2018, The Cenozoic kinematics of Qatar: Evidence for high-angle

faulting along the Dukhan ‘anticline’: Marine and Petroleum Geology, v. 92, p. 953-961.

Rivers, J.M., Skeat, S.L., Yousif, R., Liu, C., Stanmore, E., Tai, P. and Al-Marri, S.M., 2019a,

The depositional history of near-surface Qatar aquifer rocks and its impact on matrix flow

and storage properties: Arabian Journal of Geosciences, v. 12,

https://doi.org/10.1007/s12517-019-4498-6.

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., 2019, Dolomite dissolution: An alternative

diagenetic pathway for the formation of palygorskite clay: Sedimentology, v. 66, p. 1803-

1824.

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., 2020, Early and pervasive dolomitization by

near-normal marine fluids: New lessons from an Eocene evaporative setting in Qatar:

Sedimentology (in press), https://doi.org/10.1111/sed.12726.

Saller, A.H., Pollitt, D.A., and Dickson, J.A.D., 2014, Diagenesis and porosity development in

the First Eocene reservoir at the giant Wafra Field, Partitioned Zone, Saudi Arabia and

Kuwait: AAPG Bulletin, v. 98, p. 1185-1212.

Smout, A.H. (1954) Lower Tertiary Foraminifer of the Qatar Peninsula. In: London: British Museum

(Natural History), 90 pp.

Van Buchem, F., Svendsen, N., Hoch, E., Pedersen-Tatalovic, R., and Habib, K., 2014,

Depositional history and petroleum habitat of Qatar. in Marlow, L., Kendall, C.C.G., and Yose,

L.A., eds., Petroleum systems of the Tethyan region: AAPG Memoir, 106, p. 641-678.

Whittle, G.L., and Alsharhan, A.S., 1994, Dolomitization and certification of the Early Eocene Rus

Formation in Abu Dhabi, United Arab Emirates: Sedimentary Geology, v. 92, p. 273-285.

Whittle G.L., Alsharhan, A.S., and El Deeb, W.M.Z., 1996, Facies analysis and early diagenesis of the

Middle-Late Eocene Dammam Formation, Abu Dhabi, United Arab Emirates: Carbonates and

Evaporites, v. 11, p. 32-41.

CHAPTER II

EARLY AND PERVASIVE DOLOMITIZATION BY NEAR-NORMAL MARINE FLUIDS: NEW LESSONS FROM AN EOCENE EVAPORATIVE SETTING IN QATAR

Brooks H. Ryan1, Stephen E. Kaczmarek1, and John M. Rivers2 1Department of Geological and Environmental Sciences, Western Michigan University, Kalamazoo, Michigan 49008, U.S.A. 2 Qatar Center for Coastal Research (QCCR), ExxonMobil Research Qatar, P.O. Box 22500, Qatar Science and Technology Park-Tech, Doha, Qatar

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., 2020, Sedimentology, https://doi.org/10.1111/sed.12726

Abstract

The upper Palaeocene–lower Eocene Umm er Radhuma Formation in the subsurface of Qatar is dominated by subtidal carbonate depositional packages overlain by bedded evaporites. In Saudi

Arabia and Kuwait, peritidal carbonate depositional sequences with intercalated evaporites and carbonates in Umm er Radhuma have been previously interpreted to have been dolomitized via downward reflux of hypersaline brines. Here, textural, mineralogical and geochemical data from three research cores in Qatar are presented which, in contrast, are more consistent with dolomitization by near-normal marine fluids. Petrographic relationships support a paragenetic sequence whereby dolomitization occurred prior to the formation of all other diagenetic mineral phases, including chert, pyrite, palygorskite, gypsum, calcite and chalcedony, which suggests that dolomitization occurred very early, possibly syndepositionally. The dolomites occur as finely crystalline mimetic dolomites, relatively coarse planar-e dolomites, and coarser nonplanar dolomites, all of which are near-stoichiometric (50.3 mol% MgCO3) and well-ordered (0.73). The dolomite stable isotope values (range -2.5‰ to +1‰; mean δ18O = -0.52‰) and trace element

concentrations (Sr = 40–150 ppm and Na = 100–600 ppm) are compatible with dolomitization by near-normal seawater or mesohaline fluids. Comparisons between δ18O values from Umm er

Radhuma dolomite and the overlying Rus Formation gypsum further suggests that dolomitization did not occur in fluids related to Rus evaporites. This study provides an example of early dolomitization of evaporite-related carbonates by near-normal seawater rather than by refluxing hypersaline brines from overlying bedded evaporites. Further, it adds to recent work suggesting that dolomitization by near-normal marine fluids in evaporite-associated settings may be more widespread than previously recognized.

Introduction

Dolomite [CaMg(CO3)2] is a common diagenetic constituent of ancient sedimentary marine successions. Despite its abundance in the rock record, the scarcity of dolomite in modern environments has led to a century-long debate about how, when and where dolomite forms (Van

Tuyl, 1916; Land, 1985; Budd, 1997; Warren, 2000; Machel, 2004; Gregg et al., 2015; Kaczmarek et al., 2017). Geochemical proxy data commonly permit multiple interpretations as to the temperature and chemistry of the dolomitizing fluids and environmental conditions (Machel, 2004), but frequently point to low-temperature, near-surface settings (Machel, 2004; Gregg et al., 2015). To explain platform-scale pervasive dolomitization of limestone, various diagenetic models have been proposed. Some of the most prominent models include dolomitization by normal seawater (Sass &

Katz, 1982; Land, 1985; Carballo et al., 1987; Land, 1991; Manche & Kaczmarek, 2019), hypersaline reflux (Adams & Rhodes, 1960; Deffeyes et al., 1965; Land, 1967; Land, 1985; Warren, 2000; Machel,

2004; Dravis & Wanless, 2018), mixing zone dolomitization (Badiozamani, 1973; Land, 1973; Folk &

Land, 1975; Humphrey & Quinn, 1989) and dolomitization by burial fluids (Mattes & Mountjoy, 1980;

Barnaby & Read, 1992). Although these models differ in their hydrological and geochemical details, most workers agree that widespread dolomitization requires a precursor limestone to provide the

necessary carbonate, a sufficient reservoir of aqueous Mg, and an efficient hydrological pumping mechanism to move the fluids through the limestone (Morrow, 1982; Land, 1985).

Popular among these dolomitization models is the hypersaline reflux model, which is commonly posited as the responsible mechanism for dolomitizing shallow-marine, peritidal carbonates worldwide (Sun, 1995; Warren, 2000; Machel, 2004). In general, this model describes a mechanism whereby hypersaline brines originating from an overlying evaporite pool seep downward and dolomitize underlying carbonate strata (Adams & Rhodes, 1960). From a geochemical perspective, the model is attractive because the precipitation of evaporites, such as gypsum or anhydrite, in warm arid environments lowers the concentration of Ca2+ ions in solution, thus raising the Mg/Ca ratio and causing fluids to become more supersaturated with respect to dolomite (Adams

& Rhodes, 1960; Deffeyes et al., 1965; Land, 1967; McKenzie et al., 1980; Patterson & Kinsman, 1982;

Land, 1985). From the hydrological perspective, evaporation also provides a drive for fluid flow. As evaporative brines become denser than the surrounding pore fluids, they have the potential to seep downward into the underlying limestone (Machel, 2004). Taken together, warm saline fluids with elevated Mg/Ca ratios coupled with hydrological flow have become the basis for a widely accepted explanation for large, platform-scale dolomitization, particularly in rocks associated with overlying evaporites (Adams & Rhodes, 1960; Moore et al., 1988; Ruppel & Cander, 1988; Saller & Henderson,

1998; Cantrell et al., 2004; Fullmer & Lucia, 2010; Dravis & Wanless, 2018; and many others). These interpretations have also been supported by reactive transport modelling studies that suggest reflux of hypersaline brines is a viable mechanism for platform-scale dolomitization (e.g. Jones & Xiao,

2005; Garcia-Fresca & Jones, 2011; Al-Helal et al., 2012; Garcia-Fresca et al., 2012; Xiao et al., 2018).

One example of a peritidal marine carbonate that is pervasively dolomitized is the Umm Err

Radhuma Formation (UER). The UER is a widespread carbonate unit that lies in the subsurface across much of the Arabian Peninsula (Powers et al., 1966; Sharland et al., 2001). Based on mineralogical, textural and geochemical data, it has been hypothesized that the UER was dolomitized by downward

flowing evaporative brines that originated in evaporative lagoons during deposition of the overlying

Rus evaporite beds (Pollitt et al., 2012; Saller et al., 2014). The UER constitutes a major hydrocarbon reservoir in the Wafra Field located between Kuwait and Saudi Arabia (Danielli, 1988; Meddaugh et al., 2007; Barge et al., 2009; Rubin, 2011). The UER has also attracted attention because it contains the Palaeocene–Eocene Thermal Maximum (PETM) as evidenced by a distinct carbon isotope excursion that is correlative across the Arabian Peninsula (Pollitt et al., 2012). Efforts to assess the impact of diagenesis on reservoir properties at Wafra has produced a paragenetic sequence suggesting that gypsum formation pre-dates dolomitization, supporting the contention that dolomitization occurred by refluxing gypsum-saturated fluids (Saller et al., 2014). The present study examines the mechanism of dolomitization in the UER in the subsurface of Qatar using a suite of textural, mineralogical and geochemical data. Specifically, this study seeks to answer the following: what is the mechanism of dolomitization, and how does it differ from the surrounding region? The data presented herein suggest that although UER dolomites are overlain and cemented by evaporites, they were formed early in fluids with a near-normal seawater signature (defined here as normal seawater with possibly elevated temperatures [≤40°C] or salinities [35-50‰], but not to gypsum saturation [140‰]).

Geological Background

The Umm er Radhuma Formation marks the onset of Cenozoic deposition (Fig. 1) during transgression and regression along a north-east-facing ramp (Sharland et al., 2001). The southern margin of the Tethys region was covered by shallow carbonate ramps during much of the late

Palaeocene and early Eocene (Sharland et al., 2001), resulting in subtidal to peritidal carbonate sequences with intercalated clays and minor evaporites that cover the Arabian Peninsula (Powers et al., 1966; Sharland et al., 2001). The UER is composed primarily of metre-scale peritidal carbonate and evaporite sequences toward the eastern borders of Saudi Arabia and Kuwait (Saller et al., 2014),

Geologic Age Group Formation Lithology

Dammam

Middle Fm.

Eocene Rus Fm. Early Hasa

Umm er Radhuma Fm. Paleocene PALEOGENE

Dolomite Limestone Gypsum

Figure 1: Generalized stratigraphic column of the Eocene sediments of Qatar. Ages and lithologies adapted from Cavelier (1970), Al-Saad (2003) and Rivers et al. (2019a). with more subtidal to open marine carbonate deposits in Qatar (Rivers et al., 2019a), where it lies exclusively in the subsurface and is up to 370 m thick (Powers et al., 1966; Cavelier, 1970; Eccleston et al., 1981). The UER overlies the lithologically similar Cretaceous Simsima Formation with a boundary that is biostratigraphically unconformable (Cavelier, 1970; Hewaidy, 1994). The UER represents a large-scale regression, and is unconformably overlain by the lower Eocene Rus

Formation, which is characterized by metre-scale cycles of dolomite, gypsum and clay (Cavelier,

1970; Eccleston et al., 1981; Al-Saad, 2003; Rivers et al., 2019a). Although micropalaeontological studies are inconclusive, the lower UER is interpreted to be upper Palaeocene in age, whereas the uppermost intervals are interpreted to be lower Eocene (Powers et al., 1966; Hasson, 1985; Hewaidy,

1994; Boukhary et al., 2011; Pollitt et al., 2012). Pollitt et al. (2012) dated the uppermost intervals of the UER to be Eocene based on a carbon isotope shift observed across the region, which these authors interpreted to be the PETM interval, as defined globally by Zachos et al. (2001).

Based on the abundance of open marine to semi-restricted inner ramp limestones with associated shales and evaporites, much of the Arabian Shelf has been inferred to be structurally quiescent during Eocene time (Powers et al., 1966). A recent study by Rivers & Larson (2018), however, mapped a series of high-angle normal faults across Qatar that occurred near the end of UER deposition. These faults effectively separated Qatar into a south-western proximal and north-eastern distal basin, relative to the Arabian Plate, at the end of the Palaeocene. Bifurcation of the basin can explain the higher abundance of siliciclastic and evaporitic material in the south compared to the relatively pure carbonate packages in the north (Rivers et al., 2019a).

Previous facies and environmental interpretations

Recent work indicates that the UER in Qatar is primarily composed of subtidal deposits, with a general shallowing-upward trend towards the UER/Rus boundary (Boukhary et al., 2011; Rivers et al., 2019a). In an assessment of the fauna present in the UER in south-western Qatar, including both planktonic and large benthic foraminifera, Boukhary et al. (2011) interpreted the UER to reflect deposition in an open marine to sheltered lagoon environment. Rivers et al. (2019a) analyzed the same research cores utilized in the present study (Fig. 2) and came to similar conclusions with regard to depositional environments. The paragraphs that follow summarize the key observations and interpretations of Rivers et al. (2019).

Figure 2: A palaeogeographic map of the eastern half of the Arabian Peninsula (modified from Ziegler, 2001), along with an inset map of Qatar (modified from Ryan et al., 2019) with locations of cores 1, 2 and 3.

The UER in core 1 is defined as metre-scale fining-upward depositional cycles of dolomitic packstones and wackestones with variable thickness. Cycle bases are characterized by dolomitic packstones representing coarse lag deposits, whereas cycle caps are characterized by centimeter- scale thin-bedded or laminated dolomitic wackestones to mudstones commonly containing palygorskite. These rocks are identified as facies type FA5, defined as bioturbated dolomitic packstones rich in large benthic foraminifera, echinoderms, dasyclad green algae, coral fragments, and both planktonic and small benthic foraminifera. This facies is interpreted to reflect deposition in normal marine waters below normal wave base on a low-energy, mid-ramp setting. In the uppermost

8.5 m of core 1, the UER consists of dolomitic packstones and wackestones, with both straight and crinkly laminations. These upper 8.5 m of the UER in core 1 was designated facies FA3, defined as bioturbated dolomitic mudstones, wackestones and packstones rich in miliolids and rotalids, as well

as ostracods and small bivalves. This facies is interpreted as a protected interior ramp or shallower subtidal lagoon environment than that interpreted for the underlying FA5.

Similar observations to core 1 were made for core 3, the northern-most core, in which the uppermost ca 10 m are composed of FA3, underlain by ca 20 m of FA5, although these interpretations are more ambiguous because they are obscured by coarse dolomite textures. In contrast to core 1, the bottom-most ca 20 m of core 3 are bioturbated dolomitic mud-lean packstones, designated FA4, rich in large bivalves and gastropods, with a variety of small benthic foraminifera and minor planktonic foraminifera and dasyclad green algae. Facies FA4 was interpreted to reflect carbonate sedimentation on an energetic, open shallow subtidal inner ramp.

Due to the obliterative nature of dolomite textures in the UER of core 2, facies associations were not assigned. However, a general description includes both bioturbated and mud-lean dolo- packstones with large benthic foraminifera, dasyclad green algae, echinoderms and molluscs, which are similar to facies FA4 and FA5. The obliterative nature of dolomite textures in all cores also led

Rivers et al. (2019a) to conclude that a highly detailed sequence stratigraphic interpretation of these strata is not possible. The top of the UER in all cores displays an erosive surface, interpreted to be an informal higher order sequence boundary within the Pg20 sequence boundary of Sharland et al.

(2001). One identifiable feature across all three cores is a chert band previously mapped as continuous across Qatar by Eccleston et al. (1981). Two other sequence boundaries were suggested by Rivers et al. (2019), although their precise positions and how they correlate across wells is highly uncertain (see fig. 20 of Rivers et al., 2019a).

Materials and Methods

All petrographic, mineralogical and geochemical data acquired for this study were gathered from three research bore holes drilled in central and northern Qatar (Fig. 2). The analytical methods, sedimentology and general stratigraphy of these cores are described in detail in Rivers et al. (2019a).

Research cores

Although core recovery approached 100% in core 1, approximately 15 m of material extracted from core 2 (90–105 m depth) and core 3 (75–90 m depth) had poor competency due to karstification. Relevant to the current study is that the bottom ca 50–70 m of each core consists of the uppermost part of the Umm er Radhuma Formation (Fig. 3). In total, 391 cylindrical core plugs

(ca 2.5 cm diameter, ca 3.8 cm long) were taken at 0.02 – 2.0 m vertical spacing from Umm er

Radhuma. Sub-samples of rock were taken from those core plugs and ground into powders using an electric rotary drill and ground by hand with a mortar and pestle for ca 5 minutes that were used for

X-ray diffraction, inductively coupled plasma – mass spectrometry (ICP-MS) and stable isotope analysis as described below.

Thin section petrography and scanning electron microscopy

Thin sections were prepared from the ends of 391 core plugs, with each impregnated with blue epoxy in order to analyze pore space, and subsequently stained with Alizarin Red S (ARS) to differentiate dolomite from calcite. Four thin sections from core 1, chosen because they represent an example of each dolomite texture and also span the entirety of the UER, underwent additional polishing to a sub-micron finish for analysis by scanning electron microscopy (SEM) on a JEOL JSM-

IT100 InTouchScope (JEOL Limited, Tokyo, Japan). Analytical parameters include an accelerating voltage of 15 kV, working distance of 10 mm and probe current of 85 eV. Energy dispersive X-ray spectroscopy (EDS) was used in conjunction with the SEM to determine elemental composition at the micrometre-scale. The EDS reports elemental mass%, which was then converted to mol% in order to calculate dolomite stoichiometry. The EDS spectra that contained elements other than Ca, Mg, C or O were excluded, because this would indicate that anything other than pure dolomite was analyzed. At least 76 SEM-EDS points met these criteria in each of the four samples analyzed.

S N Core 2 Depth (m) (-17 mbsl) 60 Core 1 Depth (m) (-20 mbsl) 60

FA3 70 FA4? 70

80 Core 3 Depth (m) 80 (-41 mbsl) 50

90 FA3? 90 Or 60 Karst 100 FA5? 100 70

FA5? 110

110 80 Karst 120 FA5? 0% 100% 120 Mineralogy 90

Calcite 130 100 Dolomite

0% 100% Clastics Mineralogy 140 FA4? 110 Gypsum Chert band

120 0% 100% Mineralogy Figure 3: Mineralogical percentages for the Umm er Radhuma (UER) recovered in cores 1, 2, and 3, based on X-ray diffraction (XRD) generated by Core Laboratories, as well as facies associations (FA3, FA4 and FA5) adapted from Rivers et al. (2019a,b). Clastic components include clays and chalcedony. Facies designations with question marks represent uncertain facies interpretations due to the obliterative nature of dolomite textures. Intervals in red represent karst sections. Note that the UER tops do not lie at the same depth below the surface in all cores, nor are they the same depth in metres below sea level (mbsl).

X-ray diffraction

Mineralogical characterization of core plugs was carried out in two separate ways, as similarly described in Ryan et al. (2019). Bulk mineralogy for all 391 core plugs was quantified by

Core Laboratories using their standard X-ray diffraction (XRD) procedure. This involves dispersing each sample in a dilute sodium hexa-meta phosphate solution that is then centrifugally size fractioned to isolate clay-sized particles (<2–4 µm). The suspensions are vacuum-deposited on silver membrane filters, air dried, attached to aluminum stubs and analyzed both before and after exposure to ethylene glycol for a minimum of 2 hours at 60°C. The XRD data were collected with a Panalytical automated powder diffractometer (Malvern Panalytical Limited, Malvern, UK) equipped with a 40 kV

Cu source, an X’celerator linear detector using Real Time Multiple Strip Technology (RTMS), a Ni-

Filter and a graphite monochromator. A 2θ scan range of 4° to 70° was used at a rate of 4.2°/minute.

For clay samples, a range of 2.5° to 40° at a rate of 6.4°/minute was used. In order to determine individual clay percentages within the clay fraction, integrated peak areas and empirical reference intensity ratio (RIR) factors were utilized.

In order to more fully characterize percent dolomite (relative to calcite), dolomite stoichiometry and the degree of dolomite cation ordering, additional XRD data was collected at

Western Michigan University on 125 of the 391 core plugs. Standard XRD techniques were employed using CuKα radiation with a Bruker D2 Phaser Diffractometer (Bruker, Billerica, MA, USA). Core plug powders were extracted using an electric rotary drill and further ground by hand with a mortar and pestle to homogenize the sample. Samples were mounted on a Boron-doped silicon P-type zero background diffraction plate. All XRD spectra were collected in the 2θ range of 20 to 40° with a step size of 0.01° and a count time of 1.0 s per step. Dolomite percentages were calculated following the method of Royse et al. (1971), which includes taking the ratio of the dolomite d(104) peak to the sum of the calcite d(104) and dolomite d(104) peak. Dolomite stoichiometry (i.e. mol% MgCO3) was calculated using the equation derived by Lumsden (1979) which empirically relates d-spacing of the dolomite

d(104) peak to the mol% CaCO3. Several internal standards were used to shift peak positions. Reeder

& Sheppard (1984) showed that the equation derived by Lumsden (1979) can lead to inaccuracies of up to 3 mol% CaCO3, and thus the stoichiometry data yielded by XRD was checked with both EDS and electron microprobe data. The degree of cation ordering was calculated by taking the ratio of the dolomite d(015) peak to the dolomite d(110) peak (Goldsmith & Graf, 1958).

Electron microprobe analysis

Electron microprobe analysis (EMPA) was carried out at the Eugene Cameron Electron

Microprobe Lab, University of Wisconsin-Madison. The same four polished thin sections utilized for

SEM-EDS analysis were carbon coated with ca 20 nm carbon and analyzed with a Cameca SXFiveFE electron probe micro-analyzer (Cameca, Gennevilliers, France). Analytical parameters include a 15 kV electron beam, 10 nA current, and a 5–10 µm beam (spot) size, and elemental percentage detection limits of 0.021%, 0.014%, 0.040%, 0.024% and 0.023% for Ca, Mg, Fe, Mn and Sr, respectively. Five or more crystals were analyzed on each thin section containing fabric-destructive dolomite, while at least five allochems were analyzed on one mimetically dolomitized sample. More than 95 points were taken on each sample and analyzed for elemental percent of Ca, Mg, Fe, Mn and

Sr utilizing an internal standard for each element. Only points in which the sum elemental mass% was ³98% were utilized, removing points that included silicates. This resulted in ³25 useful points for each sample. Dolomite stoichiometry was determined by converting the average elemental percent of Mg and Ca for each sample to mol% Mg and Ca.

Inductively coupled plasma – mass spectrometry

Bulk rock samples from 25 core plugs across all three cores were analyzed for trace element concentrations via inductively coupled plasma-mass spectrometry (ICP-MS) at SGS Canada Inc.,

Mineral Services. Rock samples were pulverized and passed through a 75 µm sieve before undergoing

a two-acid aqua regia digestion. This digestion includes a combination of HCl and HNO3 at a 3:1 ratio, respectively. This particular acid composition was used because it yields the most accurate depiction of the carbonate geochemistry because it will dissolve carbonate minerals, but does not dissolve silicates or gypsum. Analytical precision for Sr and Na is ±12% and ±43%, respectively.

Gypsum stable oxygen isotopes

One gypsum sample from the UER Formation and two samples from Rus were analyzed for

δ18O at Queen’s University, Kingston. Samples were dissolved in 2N HCl and subsequently re- precipitated as BaSO4 with a saturated solution of BaCl2. Precipitates were then measured for their oxygen isotopic composition using a MAT 253 Stable Isotope Ratio Mass Spectrometer coupled to a

Thermo Scientific TC/EA High Temperature Conversion Elemental Analyzer (Thermo Fisher

Scientific, Waltham, MA, USA). Precision is ±0.5‰. All gypsum δ18O data is reported relative to

VSMOW (Vienna standard mean ocean water).

Dolomite stable isotope analysis

Bulk rock stable isotope data used here were first reported in Rivers et al. (2019a). However, some minor modifications were made. First, samples containing any quantity of calcite (>1%) were removed due to its potential effect on bulk rock δ18O and δ13C. Secondly, samples recovered from unconsolidated intervals due to karstification are not included, such that only lithified host-rock dolomite samples were considered. This resulted in analysis of 191 samples. Data reported here were measured at the Center for Stable Isotope Biogeochemistry (CSIB), University of California, Berkeley using a GV IsoPrime mass spectrometer with Dual-Inlet and MultiCarb systems (GV Instruments

(Micromass) Limited, Manchester, UK). Samples were reacted with H3PO4 at 90°C for 10 minutes.

Replicates of the international standard NBS19 and two laboratory standards (CaCO3 I and II) are analyzed with each run. Analytical precision is approximately ±0.05‰ for δ13C and ±0.07‰ for δ18O.

Values for all isotopic compositions are reported relative to the Vienna Peedee Belemnite (VPDB) standard.

Results

Bulk rock mineralogy

The bulk mineralogy of the upper UER is dominated by dolomite (62 to 100%) in all three cores, although mineralogical differences do occur laterally between and stratigraphically within cores (Fig. 3). Gypsum is only present in the uppermost UER of core 1 (above 93 m) in quantities ranging from 0–21.2%. Thin (<1 m) intervals of calcite mixed with dolomite and palygorskite are only present towards the top of the UER in cores 2 and 3, with the calcite and minor amounts of chalcedony comprising ≤80% of bulk rock. Palygorskite and chert occur throughout all three cores, with their combined abundance typically being <30%. A distinctive centimetre-thick interval rich in chert is observed in all three cores (for example, 127 m in core 1, 88.5 m in core 2 and 71 m in core

3), with two such bands observed in core 1, and these are discussed in more detail below.

Thin section petrography

Dolomite

Mean dolomite crystal size increases from <10 µm at the top of the UER to >200 µm at the lowest part of core 1. The dominant dolomite texture in the uppermost 5 m of the UER in core 1 is mimetic, whereby microcrystalline dolomite has preserved wackestones and packstones dominated by miliolids and other small benthic foraminifera (Fig. 4A). In contrast, the middle (ca 65–110 m) section of the core consists of relatively coarse (>30 µm) planar-e dolomite crystals with a cloudy- core clear-rim (CCCR) fabric (Fig. 4B). The base of the core is dominated by large (>100 µm) planar- s and nonplanar dolomite mosaics (Fig. 4C). Although most of core 1 consists of fabric-destructive dolomite, some echinoderms, green algae and large benthic foraminifera (LBF) are preserved

A B

100 µm 100 µm

C D

100 µm 100 µm 200 µm Figure 4: Plane polarized light (PPL) thin section images of observed dolomite textures. (A) Core 1, 65.85 m. Mimetic dolomite that has preserved miliolids and other small benthic foraminifera. (B) Core 3, 86.0 m. Planar-e dolomite displaying cloudy core clear rim (CCCR) fabrics, in which cores are commonly partially to completely dissolved but rims remain intact. (C) Core 1, 132.76 m. Nonplanar dolomite with little intercrystalline pore space. (D) Core 2, 97.75 m. A bimodal distribution of large (100 to 200 µm) nonplanar dolomite mosaics and relatively less coarse (<100 µm) planar-e CCCR dolomites is exhibited. through dolomitization (Fig. 5A and B). Additionally, many moulds in the shape of LBF are frequently observed (Fig. 5C).

Cores 2 and 3 differ in that the bottom half of the UER is dominated by planar-s to nonplanar dolomites, while the top half is mainly comprised of planar-e sucrosic dolomites. Mean dolomite crystal size ranges from 30–200 µm in both cores. Finer crystals are typically planar-e to nonplanar, while coarser crystals are nonplanar and planar-s. Infrequently, both textures and crystal sizes occur in a single sample (Fig. 4D). Similar to core 1, the fabric of echinoderms and LBFs are preserved through dolomitization (Fig. 5D), and isolated LBF moulds are also present. Dissimilar to core 1,

however, dolomite towards the top of the UER in cores 2 and 3 is non-mimetic, and miliolids and small benthic foraminifera are absent.

A B

100 µm 250 µm

C D

1 mm 200 µm

Figure 5: Thin section images in plane polarized light (PPL) of various allochem features. (A) Core 1, 100.55 m. Numerous echinoderm fragments have undergone fabric preserving replacement by dolomite (red arrows). (B) Core 1, 103.3 m. Large benthic foraminifera (>250 µm) have been replaced by dolomite that preserved the fabric (red arrows). (C) Core 1, 104.06 m. Moulds <1 mm are observed and take the shape of precursor allochems. (D) Red arrow points to a large benthic foraminifer (Nummulites). The dolomitized large benthic foraminifer illustrates texture-preserving non-mimetic dolomite.

Pyrite and chert bands

Although not detected by XRD, small amounts of pyrite are observed petrographically in all cores, where it occurs within the intercrystalline pores of dolomite crystals (Fig. 6A). Within the chert band spanning all wells, planar-e dolomite crystals (>30 µm) are partially to completely replaced in all three cores (Fig. 6B). In core 1, where two chert bands are observed, the top band is characterized by planar-e dolomites replaced by chert, similar to that observed in cores 2 and 3. A half metre lower

(ca 127 m); however, foraminifera tests preserved in the chert band contain thin rims that appear to be carbonate (Fig. 6C and D). Although SEM-EDS analysis of this thin section was unable to characterize the elemental composition of the rims, bulk XRD shows stoichiometric (49.8 mol%

MgCO3) and well-ordered (0.78) dolomite.

A B

200 µm 200 µm

C D

200 µm 200 µm

Figure 6: Non-carbonate diagenetic features. All images taken in plane polarized light (PPL) unless otherwise noted. (A) Core 3, 81.0 m. Dolomite crystals (tan/grey) are cemented and partially replaced by pyrite (opaque, black). (B) Core 1, 126.83 m. Dolomite crystals (grey) encased in chert. Partial replacement indicated by both chert within dolomite crystals (red arrows) and the fact that most dolomite crystals in centre portion of the image ‘float’ within the chert. (C) Core 1, 127.3 m. Silicified Nummulites displaying tests that still contain carbonate material (red arrow). (D) Cross-polarized image of (C).

Clay, gypsum, calcite and non-banded chalcedony

The other minerals observed in the cores – palygorskite, gypsum and calcite – primarily occur as cements, although some gypsum and calcite is also replacive. The relationship between palygorskite and dolomite is similar in all three wells. Palygorskite always occurs as an

intercrystalline or intracrystalline pore-filling cement within the dolomite, as evidenced by dolomite crystals included within palygorskite and palygorskite observed within partially to completely dissolved dolomite cores (Fig. 7A). Gypsum occurs as an intercrystalline cement that surrounds dolomite and palygorskite (Fig. 7B), and within the intrafossil pores of mimetically replaced foraminifera in the uppermost 5 m (Fig. 7C). Frequently, dolomite rhombs float within gypsum crystals, indicating that some gypsum is replacive (Fig. 7D). The calcite-bearing intervals in cores 2 and 3 are comprised of large (>500 µm) polyhedral calcite crystals, which commonly include partially dissolved dolomite crystals (Fig. 8A) that are partially to completely replaced by the surrounding calcite (Fig. 8B). These intervals also commonly exhibit palygorskite as an intercrystalline cement surrounding calcite and dolomite crystals (Fig. 8C). Some calcite crystals (Fig. 8C) exhibit textures similar to Microcodium (sensu Kabanov et al., 2008) and Microcodium (a) (sensu Esteban, 1974). Some of these coarse calcite crystals are also partially replaced by chalcedony (Fig. 8D).

Karst

Many of the features described above are also observed in the clasts comprising the incompetent intervals of cores 2 and 3 that occur below the karsted intervals. Features include planar dolomite, nonplanar dolomite, microcrystalline dolomite and calcite crystals that include dolomite crystals (Fig. 9).

Dolomite mineralogy

Dolomite stoichiometry data are presented in Figs 10B, 11B and 12B, and show an average of

50.3 mol% MgCO3 (±0.45%). In all cores, dolomite stoichiometry correlates with dolomite texture.

In core 1, mimetic and nonplanar dolomites have similar stoichiometry values averaging 49.7% and

50.1%, respectively. Planar-e CCCR dolomites, in contrast, average ca 1% higher Mg at 50.7%. A t- test shows that this difference is statistically significant at the 99% confidence (� = 0.01) with p-

A B

100 µm 100 µm

C D

200 µm 100 µm Figure 7: Diagenetic relationships between dolomite, palygorskite and gypsum. All images in plane polarized light (PPL). (A) Core 3, 74.3 m. Palygorskite (light green haze) is observed as a cement around dolomite crystals (white arrows), as well as inside of partially dissolved dolomite cores (red arrows). (B) Core 1, 73.44 m. Palygorskite (red arrows) and dolomite are encased within gypsum cement (white). (C) Core 1, 66.48 m. Gypsum (white) has occluded much of the pore space between mimetically dolomitized foraminifera, as well as minor amounts of intrafossil pore space. (D) Core 1, 74.95 m. Planar-e, cloudy core clear rim (CCCR) dolomite crystals (grey) float within gypsum (white) indicating that this gypsum is, in part, replacive. values 7.5 x 10 -9 and 2 x 10-10, respectively. Nonplanar dolomites are significantly more Mg-rich than mimetic dolomites at 99% confidence (� = 0.01) with a p-value of 6.5 x 10-5. In cores 2 and 3, Mg- content is highest at the top of the UER and decreases with depth (Figs 11B and 12B). In core 2, stoichiometry ranges from 50.9% at the top of the UER to 50% at the bottom of the core (R2 = 0.86).

Similar to core 1, planar-e dolomites in core 2 are slightly more Mg-rich (50.6%) than nonplanar dolomites (50%). A t-test determined that this difference is significant (� = 0.01; p-value = 1 x 10-7).

Stoichiometry in core 3 ranges from 50.7% at the top to 49.3% at the bottom (R2 = 0.49). Planar-e

A B

200 µm 100 µm 100 µm

C D

500 µm 500 µm 500 µm Figure 8: Thin section images in plane polarized light (PPL) displaying the relationship between calcite, dolomite, palygorskite and chalcedony. (A) Core 3, 62.5 m. Dolomite crystals (grey) containing partially dissolved cores are included within large crystalline calcite crystals (pink). (B) Core 2, 61.05 m. Remnants of dolomite (dark grey/ black) inclusions float within a large crystalline calcite crystal (pink). (C) Core 2, 61.05 m. A large (>500 µm) calcitic Microcodium structure included within palygorskite (light green haze). Note that dark inclusions are dolomite crystals that have been partially to completely replaced by the central crystal of calcite. (D) Core 2, 65.1 m. Calcite crystals (pink) are partially replaced by chalcedony (white). Dark matrix surrounding crystals consists of palygorskite. dolomites are more Mg -rich in core 3 (50.2%) than nonplanar dolomites (49.7%) (t-test; � = 0.01; p- value = 7 x 10-4).

The stoichiometry determined by EMPA (Table 1) was 48.5% (±0.7%) and 48.6% (±0.01%) for the mimetic and nonplanar dolomite, respectively. For the planar-e samples, the Mg% ranged from 49.4 to 49.7%. As a whole, 170 EMPA analyses indicate that planar-e dolomites are significantly more Mg-rich than mimetic or nonplanar dolomites based on a t-test at 99% confidence (� = 0.01) with p-values of 3.4 x 10-8 and 3.2 x 10-8, respectively. Analysis by SEM-EDS yielded similar results.

The mimetic and nonplanar samples averaged 48.5% and 48.6% Mg, respectively. The two planar-e

A B

100 µm 100 µm 100 µm 500 µm Figure 9: Thin section images (plane polarized light – PPL) from the unconsolidated section of core 2. (A) From 118.1 m. Clasts of dark, microcrystalline dolomite (left), calcite (pink) with inclusions of replaced dolomite crystals (centre), and nonplanar dolomite mosaics (right). (B) Large clasts of nonplanar to planar-s dolomite ‘floating’ within a matrix consisting of 10 to 20 µm sized planar-e dolomite crystals with much intercrystalline pore space.

Figure 10: Core profile of core 1 displaying mineralogy and facies associations (A) with corresponding dolomite stoichiometry (mol% MgCO3) (B), cation ordering (C), bulk dolomite δ18O values (D), bulk dolomite δ13C values (E), and Sr and Na concentrations (F). Dolomite textures are represented as different stratigraphic shades. Red arrows point to negative carbon isotope excursions that potentially represent the Palaeocene–Eocene Thermal Maximum.

samples averaged Mg% between 49.5% and 49.7%. A t-test at 99% confidence (� = 0.01) determined that planar-e dolomites were significantly more Mg-rich than mimetic (p-value = 1.1 x 10-10) and nonplanar (p-value = 3.5 x 10-18) dolomites.

Figure 11: The mineralogy and facies associations of core 2 (A) correlated to dolomite stoichiometry (mol% MgCO3) (B), cation ordering (C), bulk dolomite δ18O values (D), bulk dolomite δ13C values (E), and Sr and Na concentrations (F). Dolomite textures are represented as different stratigraphic shades, and the red arrow points to a negative carbon isotope excursion that potentially represents the Palaeocene–Eocene Thermal Maximum.

Figure 12: Depth profile of core 3 exhibiting mineralogy and facies associations (A) corresponding to dolomite stoichiometry (mol% MgCO3) (B), cation ordering (C), bulk dolomite δ18O values (D), bulk dolomite δ13C values (E), and Sr and Na concentrations (F). Dolomite textures are represented as different stratigraphic shades. Red arrow indicates a negative carbon isotope excursion that potentially represents the Palaeocene–Eocene Thermal Maximum.

Table 1. Stoichiometry statistics for 4 UER dolomites analyzed by XRD, EMPA, and EDS Depth Dolomite XRD ±1 � EDS average ±1 � (m) texture Mg% EMPA average Mg% (%) n Mg% (%) n 65.85 Mimetic 49.6 48.5 0.7 25 48.5 0.01 76 78.19 Planar-e 50.7 49.4 0.6 67 49.5 0.01 82 97.26 Planar-e 50.9 49.7 0.8 42 49.7 0.01 82 126.38 Nonplanar 50 48.6 0.01 40 48.6 0.01 90

Dolomite cation ordering also correlates with dolomite texture (Figs 10C, 11C and 12C).

Mimetic dolomites have the lowest average degree of cation ordering (0.57 ± 0.06), while the cation ordering of planar-e (0.77 ± 0.07) and nonplanar (0.71 ± 0.08) dolomites are significantly higher (t- test, � = 0.01; p-values = 6.5 x 10-9 and 7.5 x 10-7, respectively).

Dolomite stable isotope analysis

Dolomite δ18O and δ13C values are reported relative to depth in Figs 10 to 12 and in Table 2.

In general, dolomite δ18O values are between -2.5‰ and +1‰, whereas δ13C varies between -3.5‰ and +2‰. Average dolomite δ18O and δ13C values from 191 samples are -0.5‰ (s = 0.9) and -0.1‰

(s = 1.2), respectively. δ18O decreases with depth in core 1 (R2 = 0.50), whereas it increases with depth in cores 2 (R2 = 0.57) and 3 (R2 = 0.70). The nonplanar dolomites that lie below the karstified intervals in core 2 (>105 m depth) and core 3 (>91 m depth) are ca 1‰ more positive than all dolomites as a whole (average = 0.39‰ and +0.45‰, respectively). δ13C is more variable in all cores and although no apparent trend was observed with depth, multiple isotope excursions are identified.

In core 1, a -2‰ excursion is observed at 122 m, a +3.5‰ excursion at 104 m, and a -3‰ excursion at 96 m. In cores 2 and 3, a -4.5‰ and -3.8‰ excursion is observed at 105 m and 108 m, respectively.

Table 2. UER Dolomite Stable Isotope Analysis

Core 1 Core 2 Core 3

Depth δ13C δ18O Depth δ13C δ18O Depth δ13C δ18O (m) (‰ VPDB) (‰ VPDB) (m) (‰ VPDB) (‰ VPDB) (m) (‰ VPDB) (‰ VPDB)

60.3 -0.88 0.32 63.00 -0.81 -0.46 51.85 -1.72 -1.67 60.3 -1.22 -1.13 63.90 -0.86 -0.82 52.64 -1.17 -1.86 61.29 -1.77 -0.31 65.95 1.37 -0.47 53.30 0.27 -1.28 61.30 -1.69 -0.45 66.50 1.33 -0.28 54.45 0.01 -1.73 61.8 -2.12 0.29 67.50 -0.39 -0.86 54.78 -1.81 -2.28 62.25 -0.35 0.18 69.05 1.44 -0.52 55.53 0.34 -1.46 62.70 -0.41 -0.31 69.50 1.42 -0.48 56.56 -2.03 -2.00 64.55 -1.06 0.19 70.00 1.01 -0.58 57.40 -0.79 -1.88 65.2 -0.52 0.38 70.50 1.40 -0.88 58.15 -2.87 -1.16 66.5 -0.24 0.16 70.50 1.35 -1.02 58.42 -3.47 -1.36 66.77 -0.30 0.13 70.90 0.99 -0.66 59.04 -0.74 -1.33 68.00 -0.69 -0.18 71.50 0.66 -0.64 63.30 -0.85 -1.45 69.00 -0.79 -0.09 72.00 0.36 -0.80 63.85 -0.56 -1.20 69.74 -1.00 -0.41 73.85 -0.24 -0.89 65.69 -1.79 -1.01 69.97 -0.83 -0.34 74.30 -0.54 -1.03 66.40 -1.82 -1.05 70.65 -0.60 -0.04 75.00 -0.98 -1.17 70.54 -0.88 -0.66 71.00 -0.67 -0.76 76.50 -0.47 -1.14 71.54 -0.91 -0.29 72.00 -1.06 -0.78 77.00 -0.53 -1.19 71.94 -0.74 -1.52 73.00 -0.53 -0.82 77.35 -1.17 -1.85 72.35 -0.91 0.68 74.00 0.17 -0.52 78.50 -0.73 -1.10 74.05 -0.35 -0.24 75.00 -0.09 -0.59 82.00 -0.76 -0.63 91.10 0.16 1.17 75.35 0.13 -0.93 83.00 -0.97 -1.39 92.28 0.05 1.00 75.82 0.06 -1.09 83.80 -0.68 -1.07 94.00 -0.88 0.81 76.15 0.02 -0.73 84.45 -1.05 -0.59 97.75 -0.92 0.31 77.00 -0.36 -0.97 86.00 -0.87 -0.50 98.60 -0.19 0.70 78.20 -0.67 -0.85 87.00 -0.93 -0.03 99.70 1.19 0.20 79.00 -0.63 -1.04 105.50 -2.31 0.15 101.40 -1.82 -0.46 80.00 -0.97 -1.24 105.90 -1.61 -0.11 101.85 -0.55 -0.07 81.00 -0.70 -1.05 106.50 -1.38 -0.01 102.25 0.04 0.15 82.00 -1.17 -1.47 107.05 -1.21 0.20 102.82 0.60 0.46 83.00 -0.72 -1.23 107.55 -0.83 -0.06 103.58 -2.45 -0.89 84.00 -0.87 -2.08 108.00 -2.12 -0.76 104.15 -2.05 -0.83 84.25 -0.41 -1.66 108.30 -1.01 0.24 104.48 0.05 0.54 85.00 -0.46 -1.54 109.00 -0.80 0.40 106.38 -1.36 -0.39 86.00 -0.30 -1.11 109.50 -0.10 0.60 106.90 0.02 0.39 87.00 -0.05 -0.87 110.00 0.08 0.50 107.89 -2.40 -0.67

33 Table 2. (continued) Core 1 Core 2 Core 3

Depth δ13C δ18O Depth δ13C δ18O Depth δ13C δ18O (m) (‰ VPDB) (‰ VPDB) (m) (‰ VPDB) (‰ VPDB) (m) (‰ VPDB) (‰ VPDB)

88.00 0.58 0.05 110.45 0.57 0.89 108.35 -2.57 -0.72 88.92 0.33 -0.16 111.00 0.94 -0.85 109.10 -0.72 0.44 90.00 1.30 -0.69 111.50 1.45 0.82 110.05 -1.59 -0.02 91.00 0.29 -0.98 112.00 1.52 0.87 110.45 0.54 0.29 91.20 0.08 -1.49 112.50 1.41 0.29 111.45 1.20 0.99 92.00 0.34 -0.89 113.20 1.72 0.52 112.00 1.27 1.16 93.00 0.30 -1.18 113.50 1.77 0.73 113.50 1.04 0.64 94.00 0.22 -1.52 113.70 1.25 0.47 114.73 1.59 1.25 95.00 0.33 -1.22 113.70 1.67 0.99 115.22 1.05 0.91 96.00 -1.63 -1.62 114.00 1.65 0.69 115.96 1.39 0.81 97.00 -1.20 -1.20 114.20 1.82 0.74 118.10 1.45 0.87 98.00 -0.52 -0.89 114.90 1.82 0.66 118.95 0.80 0.86 99.00 -0.90 -1.02 115.60 1.78 -1.43 120.65 0.33 1.06 100.1 -0.10 -1.29 115.95 1.79 0.78 121.17 1.30 1.39 100.50 0.40 -1.24 116.25 2.06 0.99 121.30 0.97 1.02 102 0.94 -1.41 117.00 1.80 0.48 122.13 1.23 0.97 103.1 1.23 -1.20 117.50 1.56 1.09 104.35 1.55 -0.92 117.80 1.78 0.89 105 1.16 -0.93 118.36 1.99 0.96 106.00 0.54 -1.30 118.85 2.05 0.40 107.00 0.34 -1.29 119.50 2.10 0.51 108.00 0.01 -1.10 120.00 2.22 -0.18 109.00 -0.71 -1.74 120.00 2.04 0.33 109.20 -0.63 -1.74 110.00 -1.38 -1.63 112.00 -1.94 -2.43 114.00 -0.92 -1.57 116.00 -1.07 -1.24 118.00 -0.55 -1.41 120.00 -1.07 -1.56 120.75 -1.61 -1.52 122.00 -1.58 -1.69 125.00 -0.69 -1.32 125.10 -0.52 -1.09 126.90 0.25 -2.01 127.85 0.41 -1.65

34 Table 2. (continued) Core 1 Core 2 Core 3

Depth δ13C δ18O Depth δ13C δ18O Depth δ13C δ18O (m) (‰ VPDB) (‰ VPDB) (m) (‰ VPDB) (‰ VPDB) (m) (‰ VPDB) (‰ VPDB)

128.00 0.60 -1.49 130.5 0.62 -1.37 131.1 0.20 -2.29 132.7 0.14 -1.33 134 0.26 -1.24 Overall 134.4 0.15 -1.33 total: 134.7 -0.18 -1.57 Average -0.14 -0.52 134.70 -0.32 -1.68 Std dev. 1.17 0.91

Dolomite trace element geochemistry and gypsum stable oxygen isotopes

Results of ICP-MS analysis for Sr and Na for cores 1, 2, and 3 are presented in Figs 10F, 11F and 12F, respectively. In general, Sr levels for all dolomites are low, ranging from ca 40 to 150 ppm, averaging 101 ppm. Sodium concentrations are moderate and range from <100 to 600, averaging 342 ppm.

The results of gypsum δ18O analysis are presented in Table 3. The δ18O of a gypsum sample from the UER was measured to be +15.4‰ VSMOW. Gypsum samples from the overlying Rus

Formation have a similar isotopic composition, ranging from +15.7 to +15.9‰ VSMOW. The δ18O values are slightly more positive than expected based on the secular evaporite δ18O curve of Seal et al. (2000), which shows that Cenozoic marine evaporites range from +12 to +14‰ VSMOW.

Table 3. Qatar gypsum δ18O analysis Formation Depth (m) δ18O (‰VSMOW) Rus 50.45 +15.7 Rus 53.24 +15.9 UER 90.4 +15.4

Discussion

Paragenesis interpretation

The interpreted paragenetic sequence presented in Fig. 13 indicates that the UER of Qatar has a multi-component diagenetic history. Given the petrographic relationships observed, dolomitization appears to be the earliest diagenetic process affecting the UER as evidenced by dolomite crystals either being included within or replaced by all other diagenetic mineral phases.

Dolomite is included within and replaced by the chert band (Fig. 6B to D), which indicates that dolomitization occurred prior to the regional silicification event (Eccleston et al., 1981). Moreover, when dolomite crystals are observed in association with palygorskite, calcite and/or gypsum, the dolomite crystals are either cemented together by these mineral phases or observed as inclusions within authigenic calcite or gypsum crystals. The presence of pyrite in the intercrystalline pores also suggests that dolomitization also occurred prior to pyrite formation. Dolomite dissolution obviously followed dolomitization, and the presence of authigenic palygorskite (Fig. 7A; see also Ryan et al.

2019) and calcite (Fig. 8) in the dissolved cores of many dolomite crystals indicates that dolomite dissolution is concurrent with or pre-dates these phases.

The silicification event resulting in the chert band that has been mapped as a continuous horizon across eastern and southern Qatar (Eccleston et al., 1981; Rivers et al., 2019) is interpreted to post-date dissolution of dolomite. It is, however, difficult to discern whether chertification occurred exclusively prior to, or concurrent with, the other diagenetic minerals due to an absence of cross-cutting relationships. Butts & Briggs (2011) have argued that retention of crystal structures during chertification is an indicator of early replacement. The lack of cross-cutting relations could also mean the other phases were not yet present when that band formed. Given that the chert band

Diagenetic Event Relative Order Dolomitization 1

Dolomite dissolution 2

Chertification 3

Pyrite cementation 4

Smectite replaced 5 by palygorskite

De-dolomitization 6 via Microcodium

Replacement of 7 calcite by chalcedony Gypsum cementation 8 8? 9? Karstification

Figure 13: Paragenetic sequence for the Umm er Radhuma (UER) in Qatar based on petrographic cross-cutting relationships and geochemical data. The timing of dolomitization in the uppermost 50 to 60 m of the cores is estimated to be approximately 56 Ma based on carbon isotope excursions interpreted as the Palaeocene–Eocene Thermal Maximum (see text for details). Late stage gypsum cements interpreted as being associated with brines from the overlying Rus Formation (ca 52 Ma) suggest that diagenetic events 1 to 7 were concluded by 52 Ma. Karst features in cores 2 and 3 cross-cut most of the diagenetic features suggesting that karstification occurred very late in the paragenetic sequence. Karst is not distinct in core 1; however, where the late-stage gypsum cement is present, so the timing of karst relative to gypsum cementation is uncertain.

is stratigraphically lower than intervals containing other diagenetic minerals, it is unlikely that chertification post-dates minerals such as palygorskite, calcite or gypsum. If the chert band formed early, it likely formed before the upper 75 m was deposited and thus pre-dates authigenic phases in the upper 75 m.

The fifth diagenetic event was the formation of palygorskite, which was previously interpreted by Ryan et al. (2019) to be a replacement of smectite that in turn formed from Mg liberated by the dissolution of dolomite. This clay occurs both as an intercrystalline cement and as an intracrystalline cement inside of partially to completely dissolved dolomite cores. Such relationships imply that dolomitization occurred before the palygorskite, and that dolomite dissolution was concurrent with palygorskite precipitation.

Given the petrographic relationship between the coarse calcite crystals and the surrounding palygorskite in cores 2 and 3, palygorskite formation likely followed the formation of Microcodium and crystalline calcite (Fig. 8C). However, dolomite crystals are also included within Microcodium, and are partially to completely replaced by calcite, i.e. dedolomitized (Fig. 8). This calcite is also partially replaced by chalcedony sporadically throughout the UER (Fig. 8D). As a result, the paragenetic sequence proposed in Fig. 13 is dolomitization, de-dolomitization via Microcodium concurrent with alteration of smectite to palygorskite, and finally replacement of calcite by chalcedony.

The observation that gypsum contains dolomite and palygorskite inclusions leads to the interpretation that gypsum (as a cement and a replacement of dolomite) is followed by formation of palygorskite, and thus all Microcodium calcite also. Gypsum is the last diagenetic mineral to have formed in the UER in core 1. Although metre-scale gypsum beds are common in the overlying Rus

Formation, there is no evidence of bedded evaporites in the UER in Qatar (Rivers et al., 2019a). This observation, coupled with the petrographic relationships documented above, suggests it is most likely that authigenic gypsum in the UER originated during deposition of the Rus.

Finally, the UER is overprinted by late stage karst, which is observed only in cores 2 and 3.

Rock powder harvested from poor recovery zones from these cores contains relicts of the diagenetic features described above. It is therefore likely that karstification occurred after deposition and diagenetic alteration of the UER. The timing of karstification relative to the authigenic gypsum in core

1 is unknown, however. The only signs of erosion in core 1 occur at the top of the UER at the contact

with the overlying Rus (Rivers et al., 2019a). It may be possible that this erosional contact is time equivalent to the karstification that occurred in cores 2 and 3, which would suggest that karstification happened after the formation of authigenic gypsum. It is unlikely that karstification in cores 2 and 3 originated from the dissolution of evaporites because, as previously noted, no bedded or authigenic gypsum is present in the UER or Rus in cores 2 and 3. However, given that evidence of karstification is absent in core 1, the timing of karstification relative to authigenic gypsum cannot be conclusively determined.

Regional differences in depositional environment and dolomitization mechanism

Saller et al. (2014) and Pollitt et al. (2012) investigated the paragenetic sequence of the UER and interpreted the mechanism responsible for dolomitization in Saudi Arabia and Kuwait. Saller et al. (2014) and Pollitt et al. (2012) described the UER as a burrowed peloidal wackestones and packstones interbedded with laminated mudstone, wackestone and packstone beds with minor mollusc fragments and benthic foraminifera. The authors also noted that the burrowed facies pass gradationally upward to laminated intervals, the tops of which are sharp with distinct erosional truncations. Saller et al. (2014) interpreted the burrowed intervals to reflect deposition in a subtidal to lower intertidal setting, and the laminated intervals as upper intertidal and supratidal environments.

Based on the observation of supratidal facies and bedded evaporites near the top of the UER,

Saller et al. (2014) interpreted UER dolomites to form as a result of hypersaline reflux. Given that the

UER was deposited in an upper intertidal to supratidal marine setting situated stratigraphically below the overlying Rus evaporites, such an interpretation is geologically reasonable. Saller et al.

(2014) suggested that bedded anhydrite formed prior to UER dolomitization in Saudi Arabia and

Kuwait based on vertical crystal growth structures that were observed in association with laminated dolomitized stromatolitic cyanobacterial mats towards the top of the UER (Saller et al., 2014).

However, Saller et al. (2014) also reported that anhydrite and gypsum commonly include dolomite crystals.

The presence of interbedded evaporites was used by Saller et al. (2014) as one of the principle pieces of evidence in support of the hypothesis that downward refluxing brines were responsible for most of the dolomite in the UER. Building on earlier work by Pollitt et al. (2012), Saller et al. (2014) hypothesized that the waters towards the top of the UER were saturated with respect to gypsum, resulting in gypsum precipitation, and thus a drawdown of Ca2+ from the fluids. Those authors argued that this process drove dolomitization by the downward flow of dense evaporative brines with elevated Mg/Ca (see fig. 16 of Saller et al., 2014). This model is consistent with the observation that microcrystalline (mimetic) dolomites occur near the top of the UER, but coarsen downward (Pollitt et al., 2012; Saller et al., 2014). Previous work suggests that fluids with a high degree of supersaturation with respect to dolomite tend to precipitate microcrystalline, fabric-preserving dolomite compared with less supersaturated fluids (Sibley & Gregg, 1987). Based on this reasoning,

Pollitt et al. (2012) and Saller et al. (2014) both reasoned that as dolomite precipitated in the upper

UER, the saturation state of the downward refluxing fluid was reduced, thus causing fewer dolomite nuclei to form and slower crystal growth rates, which produced the coarser fabric-destructive dolomites observed lower in the UER of Saudi Arabia and Kuwait.

In contrast to the observations of Saller et al. (2014), which show predominantly intertidal and supratidal facies towards the top of the UER, the observations in Rivers et al. (2019a) suggest that UER in Qatar is dominated by open-marine, subtidal facies with some partially restricted facies at the top of the UER in core 1. These differences suggest that the depositional environment in Qatar differed from those experienced in Saudi Arabia and Kuwait, possibly due to different tectonic histories. Saller et al. (2014) note that the UER thins over the Wafra anticline and interpreted this to indicate the structure had already formed a topographic high during deposition in the Palaeocene and Eocene. Saller et al. (2014, fig. 16) interpret multiple structural highs, which resulted in semi-

restricted environments that would have promoted the formation of hypersaline brines. The structural history of the UER in Qatar differs in that the major structural high in the centre of Qatar , which separates core 1 from cores 2 and 3, was forming simultaneously with UER deposition, and had not fully disconnected the two basins until the start of deposition of the overlying Rus (Rivers et al., 2019a). This scenario also explains why most of the facies in the three research cores reflect deposition in subtidal environments lower on the carbonate ramp, with only the uppermost 7 m of core 1 reflecting semi-restricted conditions. It follows that differences in depositional environments may have resulted in different modes of dolomitization in Qatar compared with Saudi Arabia and

Kuwait.

Petrographic relationship between dolomite and gypsum

Petrographic relationships between dolomite and gypsum described above, indicate that dolomite formed before precipitation of gypsum. In particular, gypsum is observed both as a poikolitopic cement as well as a replacement of dolomite crystals (Fig. 7), indicating that authigenic gypsum post-dates dolomitization. This in turn suggests that dolomitization was unrelated to gypsum-saturated fluids, because gypsum clearly formed after dolomite. However, this observation alone does not preclude the possibility that gypsum beds were deposited, and the resulting fluids with elevated Mg/Ca ratios subsequently dolomitized the underlying strata. Reactive transport modelling, for instance, predicts evaporite cements associated with reflux dolomitization based on the premise that dolomitization releases excess Ca2+ into the pore-fluids and drives subsequent gypsum precipitation (Jones & Xiao, 2005; Al-Halal et al., 2012). This implies that although dolomitization precedes the authigenic evaporite cementation, an overlying brine pool associated with the deposition of Rus gypsum beds may have been responsible for reflux dolomitization in the

UER of Qatar, similar to Saudi Arabia and Kuwait. However, due to the lack of bedded evaporites in the UER of in core 1, and the complete lack of evaporites in the UER and the overlying Rus in cores 2

and 3, clear petrographic evidence of refluxing hypersaline brines is missing, although this lack of evidence alone does not preclude a hypersaline reflux origin.

Dolomite textures, stoichiometry and cation ordering

Similar to observations reported in Pollitt et al. (2012) and Saller et al. (2014), the dolomites near the top of core 1 in the present study are finely crystalline and mimetic, whereas coarser planar- e dolomites are underlying mimetic dolomites and the coarsest nonplanar dolomites are observed at the bottom. Although mimetic dolomites are not observed in cores 2 and 3, dolomite crystals coarsen from planar-e dolomite at the top to nonplanar dolomites at bottom. As described above, this increase in crystal size could reflect a decrease in the saturation state of the fluid because Mg/Ca ratios decreased downward as dolomitization occurred (Sibley & Gregg, 1987; Jones & Xiao, 2005).

However, an alternative hypothesis is that dolomite textures in the UER may be facies controlled. Mimetic dolomites, for instance, are associated with facies FA3, which is characterized by muddier sediments occasionally rich in small benthic foraminifera (Rivers et al., 2019a). In contrast, the planar-e and nonplanar dolomites are predominantly observed in facies FA4 and FA5, which are characterized by mud-lean packstones rich in a diverse faunal assemblage (Rivers et al., 2019a). It is therefore likely possible that mud content of the precursor facies was the primary control on subsequent dolomite textures. Specifically, the muddier FA3 facies would have a higher reactive surface area for dolomite nucleation, and nucleation may have outpaced growth rates resulting in the mimetic nature of the dolomite (Sibley & Gregg, 1987). In the coarser, grain-rich FA4 and FA5 facies, there is less reactive surface area and thus crystal growth rates may have been faster than nucleation rates, thus resulting in coarse planar and nonplanar textures (Sibley & Gregg, 1987). Following this logic, dolomite textures are explained by variations in the carbonate precursor facies rather than decreasing Mg/Ca ratios as dolomitizing fluids refluxed downward.

The interpretation that dolomite textures were facies controlled is further supported by the relationship between dolomite textures and stoichiometry. The mimetic dolomites in core 1 are near- stoichiometric and relatively well ordered (Fig. 10B and C; Table 1). Below the mimetic interval in core 1, however, the planar-e and nonplanar dolomites are more stoichiometric than the overlying mimetic dolomites (Fig. 10B; Table 1). This compositional trend argues against downward refluxing fluids related to Mg depletion, but rather that the more Mg-rich dolomites lower in the section perhaps formed from fluids with either higher Mg/Ca ratios (Kaczmarek & Sibley, 2007; 2011), higher temperatures (Kaczmarek & Thornton, 2017), higher salinity (Glover & Sippel, 1967; Cohen

& Kaczmarek, 2017) and/or were recrystallized to a greater extent (Sibley, 1990; Mazzullo, 1992;

Malone et al., 1996; Machel, 1997; Kaczmarek & Sibley, 2014).

In contrast with core 1, an overall decrease in dolomite stoichiometry is observed down section in cores 2 and 3. Dolomites near the top of the UER in cores 2 and 3 are the most stoichiometric, and stoichiometry decreases with depth. Evaporites are absent in the UER and overlying Rus in these cores, however, and based on the presence of dominantly FA4 and FA5 faces in these cores, there is no evidence that a highly evaporative environment was present to cause refluxing of evaporative fluids.

Trace element geochemistry

Strontium and sodium concentrations of the dolomites are also inconsistent with dolomitization in highly evaporative fluids. Rather, these data suggest that dolomitization of the UER in Qatar took place in near-normal marine fluids. Near-normal marine fluids, as defined here, include normal marine to slightly evaporated mesohaline fluids that have not reached gypsum saturation.

The Sr content in the UER dolomites ranges between 40 and 150 ppm (averaging 101 ppm) which is similar to or slightly lower than other Cenozoic dolomites interpreted to have formed in near-normal marine seawater (Vahrenkamp & Swart, 1990; Budd, 1997). Based on the study of Vahrenkamp &

Swart (1990), stoichiometric dolomite should contain ca 100 ppm of Sr if precipitated from fluids with typical seawater ratios of Sr/Ca. This suggests that the UER dolomites in Qatar thus precipitated from near-normal marine fluids.

The Na concentrations of UER dolomites, which range from ca 100 to 600 ppm (averaging

342 ppm), are also consistent with those reported for other marine dolomites (Staudt et al., 1993).

Although Na concentrations in dolomite are equivocal (Budd, 1997), some field studies have shown that they can be used to compare genetically related dolomite populations (Sass & Bein, 1988; Staudt et al., 1993). Sass & Bein (1988), for instance, showed that dolomites associated with gypsum have

Na content as high as 2700 ppm, whereas non-evaporitic marine dolomites range between 150–350 ppm. Staudt et al. (1993) measured the Na concentrations of dolomites in which their origins were constrained by petrographic, geochemical and stratigraphic data. Staudt et al. (1993) showed that dolomites formed from near-normal seawater containing Na values between 0 and 500 ppm, whereas dolomites formed in evaporitic environments span a range between 500 and 2000 ppm. The values reported here are therefore more consistent with near-normal marine seawater such as mesohaline fluids and argue against gypsum-saturated brines.

Dolomite and gypsum stable isotopes

When examined collectively, the bulk δ18O values of UER dolomites in all three research cores

(Figs 10D, 11D and 12D) further support the interpretation that dolomitization did not occur via refluxing of evaporative brines. Assuming a temperature range of 25–35°C, which is compatible with sea surface temperature (SST) estimates during the Palaeocene–Eocene (Pearson et al. 2001; Zachos et al., 2006), and integrating an ice-free normal seawater value of -0.98‰ (Zachos et al., 1994), dolomite formed in near-normal marine fluids should have δ18O values between -1.5‰ and +0.75‰ using the equation of Horita et al. (2014). The range of UER dolomite δ18O values presented here (-

2.5 to +1.0‰, averaging -0.52‰) are slightly more negative than the calculated dolomite δ18O values

utilizing near-normal seawater δ18O and temperatures. This suggests that UER dolomites formed from normal marine seawaters under shallow burial conditions at temperatures slightly greater than

35°C (as high as 40°C based on dolomite δ18O of -2.5‰ and seawater δ18O of -0.98‰), were recrystallized, or a combination of both (Land, 1980). Interestingly, the expected dolomite δ18O range calculated here for a near-normal marine origin is similar to the range of δ18O values (-1.0‰ to

+2.5‰) presented by Saller et al. (2014), suggesting that the UER in Saudi Arabia and Kuwait may have also formed from near-normal marine fluids.

Whether or not dolomitization occurred in fluids saturated with respect to gypsum can further be constrained by comparing the gypsum and dolomite δ18O data. Given that the δ18O in dissolved SO4 from typical seawater is ca +8.6‰, and that the fractionation factor for δ18O from dissolved SO4 to gypsum is ca +3.5‰ (Markovic, 2016; Warren, 2016), gypsum precipitating from seawater with δ18O of 0‰ should theoretically have a δ18Ogypsum of +12‰. Bulk gypsum δ18O values in this study range between +15.4‰ and +15.9‰, which implies that gypsum in the Rus Formation precipitated from seawater with δ18O between +3.7‰ and +3.9‰ based on a back-calculation. This argument is complicated, however, because other factors also control the ultimate δ18O of oceanic sulphate, and thus the δ18Ogypsum. The rate of oxygen exchange between sulphate ions and seawater at oceanic temperatures is on the scale of 106 to 109 years (Lloyd, 1968; Chiba & Sakai, 1985), and the residence time of marine sulphate-bound oxygen (ca 500 kyr) far exceeds the ocean mixing time

(1600 yrs), suggesting that changes in δ18Ogypsum is likely not an indicator of changes in seawater δ18O due to evaporation (Jorgensen & Kasten, 2006; Markovic, 2016). Rather, deviations from the theoretical +12‰ indicate fluxes of pyrite weathering, volcanic degassing, evaporite weathering, evaporite burial and sulphate reduction, among others (Markovic, 2016). Moreover, seawater evaporated to the point of gypsum saturation has δ18O as high as +7 to +10‰ (Knauth & Beeunas,

1986; Rivers et al., 2019b), but modern gypsum has a mean δ18O of ca +13‰ (Seal et al., 2000). If dolomites formed in gypsum-saturated fluids with seawater δ18O values similar to modern fluids

precipitating gypsum (+7 to +10‰) between temperatures of 25°C and 35°C, they should have

δ18Odolomite between +6.5‰ and +11.8‰ (Horita, 2014). This range is significantly more positive than those measured in the UER dolomites in Qatar and those from Saudi Arabia and Kuwait described by

Saller et al. (2014), suggesting that the dolomites did not form from fluids associated with gypsum in the Rus Formation. Of course, recrystallization or shallow burial under elevated temperatures could also cause the δ18O values to decrease from these initial values (Land, 1980).

It is also notable that the dolomites closest to overlying evaporite beds both in this study

(core 1; Fig. 10D) and Saller et al. (2014, fig. 13) commonly have δ18O less than +1‰, and frequently less than +0.5‰, with heavier dolomite δ18O values observed deeper in the section. This is contrary to what would be expected, in that dolomites closest to the overlying evaporite beds should record the heaviest δ18O values due to their proximity to the evaporative fluid source (Warren, 2000). It is also notable that the anticipated trend is observed in core 1 in Qatar (Fig. 10D), but the dolomite δ18O only reaches a maximum value of +0.5‰, which again is broadly indicative of near-normal marine fluids.

Dolomitization model and implications

Given that the UER was dolomitized prior to Rus deposition, as argued from the petrographic and geochemical data presented above, it is possible to constrain the timing of dolomitization and to estimate the time and fluid amount required to completely dolomitize the unit. Deposition of the Rus

Formation commenced during the middle Ypresian (Al-Saad, 2005; Rivers et al., 2019a), thus ca 52

Ma provides the youngest age of when UER dolomitization ended. It is difficult to know when UER dolomitization began, however. The UER dolomite has been reported to be as thick as 370 m in the subsurface (Ecceleston et al., 1981), but only the upper most ca 75 m was cored in the research wells here. Thus, although it not possible to extrapolate the findings presented for the uppermost 75 m to the entire formation, an attempt is made to constrain the timing for the 75 m presented in this study.

A 2.0–4.5‰ negative carbon isotope excursion that is correlative across the study area (Figs

10E, 11E and 12E) provides some constraint on the timing of dolomitization in the upper 75 m. A similar excursion was tracked regionally throughout Saudi Arabia and Kuwait and interpreted by

Pollitt et al. (2012) as the PETM, and thus it is interpreted that the excursions presented here may represent the PETM in Qatar. Assuming that to be the case, the position of the PETM provides some constraint on the timing of dolomitization of the uppermost 75 m of the UER. Given that the PETM represents a ca 100 kyr peak in mean global temperatures approximately 55.53 to 56.33 Ma

(Westerhold et al., 2009), ca 56 Ma represents a conservative estimate for the onset of dolomitization of the uppermost 50–60 m (i.e. depth to the top of the excursion in each core). Thickness of the metre- scale fining-upward cycles in the UER vary between 1 m and 3 m (Rivers et al., 2019), meaning 20 to

60 cycles in 4 Ma, or an average duration of ca 60 to 220 kyr for dolomitization of each cycle. Using the area of Qatar (ca 11 570 km2), a thickness of 60 m, 100% dolomitization, 40% initial porosity

(Enos & Sowatsky, 1981), and a Palaeocene–Eocene seawater Mg concentration of ca 0.033 mol/L

(Tyrell & Zeebe, 2004), it would require 5.0 x 1010 L/year of seawater to dolomitize the upper 60 m of the UER (ca 416 km3) over 4 Myr. This annual estimate of seawater volume is consistent with estimates by Montanez & Read (1992), who posited that 5 x 1012 L/year could be fluxed through 1–

4 m peritidal cycles in the Devonian Knox Group.

Two possible mechanisms of fluid flow are considered for dolomitization of the UER by near- normal marine fluids. Because dolomitization is interpreted to have happened relatively early after deposition, it is possible that sediments were dolomitized while they were in direct contact with the overlying marine fluids, i.e. syndepositionally. This style of dolomitization was posited in Manche &

Kaczmarek (2019) to explain variable dolomitization of metre-scale peritidal deposits, in which differences in cyclical vertical patterns were interpreted to reflect temporal changes in environmental conditions. Manche & Kaczmarek (2019) interpreted finely crystalline, stoichiometric dolomites in supratidal and peritidal deposits to reflect shallow conditions with higher fluid Mg/Ca,

salinity and temperature. In contrast, deeper peritidal and subtidal deposits were characterized by coarsely crystalline and less-stoichiometric dolomites, inferred to reflect deeper water depths and lower fluid Mg/Ca, salinity and temperature (Manche & Kaczmarek, 2019). A similar process, whereby sediments are syndepositionally dolomitized, could have been responsible for early dolomitization of the metre-scale cycles observed in the UER, although no metre-scale trends in stoichiometry, stable isotopes or mean crystal size are observed.

Another possible driver could be small-scale reflux of mesohaline fluids through the underlying sediments. Based on the work of Simms (1984) and Jones & Xiao (2005), reflux of mesohaline fluids not to gypsum saturation (salinity 37–48‰) is a viable mechanism for platform- scale dolomitization, because slight increases in salinity can drive fluids through underlying sediment. Based on reactive transport modelling, pervasive dolomitization by mesohaline fluids can occur over distances of up to 7 km and 500 m depth over 1 Myr (Jones & Xiao, 2005). However, given an estimate of 4 Myr for dolomitization of the UER, it would be expected that dolomitization could extend further than 7 km. Additionally, the general shallowing trend of the UER would have caused fluids to move basinward, which would increase the lateral extent of dolomitization. Furthermore, the fining upward sequences observed in the UER are consistent with modelling showing that grainier sediments, when overlain by muddy sediments, provide conduits to move fluids further laterally (Al-Helal et al., 2012). Thus, it is also possible that small-scale reflux of mesohaline fluids could have pervasively dolomitized the UER.

The observed geochemical and textural signatures also are compatible with dolomite recrystallization, which could explain the near-stoichiometric and relatively well-ordered nature of the dolomites (e.g. Malone et al., 1996; Kaczmarek & Sibley, 2014), the slightly depleted dolomite

δ18O relative to calculated normal marine Eocene seawater values (Land, 1980; Mazzullo, 1992) and the depleted Sr concentrations (e.g. Malone et al., 1996) in the UER dolomites. If the UER dolomites were recrystallized, however, petrographic observations still support a model whereby dolomite

crystals formed prior to all other diagenetic mineral phases. Recrystallization would likely obliterate any trends in the initial dolomite stoichiometry that might reflect changes in fluid flow or chemistry.

However, the δ18O, Sr and Na values of the dolomites, if recrystallized, are still suggestive of near- normal marine fluids albeit with slightly elevated temperatures.

Implications

This study provides an example of early, platform-scale dolomitization by near-normal, mesohaline fluids in a setting that would otherwise suggest refluxing of hypersaline evaporative brines was responsible. This is consistent with other studies that have analyzed evaporite-associated dolomites and provided data that suggest near-normal seawater was responsible for dolomitization

(e.g. Newport et al. 2017; Manche & Kaczmarek, 2019). For example, Newport et al. (2017) analyzed the Upper Albian–Lower Turonian Zebbag Formation in southern Tunisia. The Zebbag Formation consists of metre-scale shallowing-upward peritidal and subtidal cycles, with a 1 m thick gypsum horizon in the middle Kerker Member. Although not present in the study by Newport et al. (2017), the Zebbag is also overlain by bedded evaporites north of the authors’ study area (Newport et al.,

2017). Based on facies distributions, dolomite texture variations, slightly positive isotope signatures, slightly elevated Sr concentrations and a near-absence of evaporites, Newport et al. (2017) hypothesized that the Zebbag Formation was pervasively dolomitized by reflux of mesohaline seawater. As briefly described above, Manche & Kaczmarek (2019) analyzed the Cretaceous Upper

Glenn Rose Formation, composed of metre-scale evaporite-associated peritidal dolomites previously interpreted to have formed by refluxing of hypersaline brines. Cyclical variations in dolomite abundance, crystal size, stoichiometry and δ18O were inconsistent with such an interpretation, however, and instead suggested that the data were more consistent with early syndepositional dolomitization by near-normal marine fluids. Combined with these studies, the present study provides yet another example of dolomitization by near-normal fluids in evaporite-capped

49 carbonates. This in turn suggests that evaporative caps do not need to imply reflux dolomitization by hypersaline brines.

Conclusions

Petrographic, mineralogical and geochemical evidence from the Umm er Radhuma (UER), a subtidal carbonate unit capped by evaporites in Qatar, suggests that dolomitization happened early in near- normal marine seawater. In all occurrences, dolomite is either included in or replaced by chert, pyrite, palygorskite, gypsum, calcite and chalcedony, indicating that dolomitization was the earliest diagenetic phase in the UER. Previous studies concluded that UER dolomites in nearby Saudi Arabia and Kuwait formed by large-scale, top-down refluxing of hypersaline brines driven by gypsum precipitation in the overlying unit. The textural, mineralogical, and geochemical data presented from the UER in this study, however, argue against a large-scale, top-down hypersaline reflux model in central Qatar. Mimetic UER dolomites closest to the overlying Rus evaporite beds are less stoichiometric than dolomites deeper in the section, when the opposite trend is expected for dolomitization by downward refluxing brines. Dolomite δ18O values fall within those expected from dolomitization by near-normal marine seawater, or are slightly more negative possibly due to recrystallization or dolomitization in a slightly warmer (≤40°C), shallow burial environment, or both.

Strontium and sodium concentrations are also consistent with dolomitization in marine-like fluids.

The accumulation of evidence, which includes subtidal precursor facies, diagenetically very early dolomite, and dolomite geochemistry and mineralogy, is consistent with early dolomitization by near-normal marine fluids. This study adds to a growing understanding of platform-scale dolomitization of evaporite-associated sequences by near-normal seawater. Further, this study adds to a growing body of research on shallow water carbonates that suggests early dolomitization by marine fluids may be more common than previously accepted, especially in evaporative settings.

50 Acknowledgements

This project was sponsored by ExxonMobil Research Qatar. We thank ExxonMobil management for permission to publish this study. Katie Sauer and Zander Sorenson are acknowledged for their assistance with XRD sample preparation and analysis. John Fournelle assisted in EMPA data collection at the University of Wisconsin, Madison. Sabrina Skeat provided important logistical support. We thank Cameron Manche for comments on an earlier version of this manuscript. We also wish to thank Chief Editor Giovanna Della Porta, Associate Editor Cathy Hollis and reviewer Beatriz

Garcia-Fresca for detailed reviews and constructive comments, as well as reviewer David Budd for his significant technical and editorial contributions, all of which improved the final manuscript.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

References

Adams, J.E. and Rhodes, M.L. (1960) Dolomitization by seepage refluxion. Bulletin of the

American Association of Petroleum Geologists, 44, 1912-1920.

Al-Saad, H. (2003) Facies analysis, cyclic sedimentation and paleoenvironment of the Middle

Eocene Rus Formation in Qatar and adjoining areas. Carbonates and Evaporites, 18, 41-50.

Al-Saad, H. (2005) Lithostratigraphy of the Middle Eocene Dammam Formation in Qatar,

Arabian Gulf: effects of sea-level fluctuations along a tidal environment. Journal of Asian Earth

Sciences, 25, 781-789.

Al-Helal, A.B., Whitaker, F.F. and Xiao, Y. (2012) Reactive transport modeling of brine

reflux: Dolomitization, anhydrite precipitation, and porosity evolution. Journal of

Sedimentary Research, 82, 196-215.

51 Badiozamani, K. (1973) The Dorag dolomitization model—application to the Middle

Ordovician of Wisconsin. Journal of Sedimentary Petrology, 43, 965-984.

Barge, D., Al Yami, F., Uphold, D., Zahedi, A., Deemer, A. and Carreras, P. (2009)

Steamflood piloting the Wafra Field Eocene reservoir in the Partitioned Neutral Zone,

between Saudi Arabia and Kuwait. Presented at the 2009 Society of Petroleum Engineers

Middle East Oil Show, Bahrain, March 15-18, 2009, SPE Paper 120205, 21 p.

Barnaby, R.J. and Read, J.F. (1992) Dolomitization of a carbonate platform during late burial:

Lower to Middle Cambrian Shady Dolomite, Virginia Appalachians. Journal of Sedimentary

Petrology, 62, 1023-1043.

Boukhary, M., Hewaidy, A.G., Luterbacher, H., Bassiouni, M.E., and Al-Hitmi, H. (2011)

Foraminifera and ostracodes of Early Eocene Umm er Radhuma Formation, Dukhan Oil Field,

Qatar. Micropaleontology, 57, 37-60.

Budd, D.A. (1997) Cenozoic dolomites of carbonate islands: their attributes and origin. Earth

Sci. Rev., 42, 1-47.

Butts S.H. and Briggs D.E.G. (2011) Silicification Through Time. In: Taphonomy: Process and

Bias Through Time (Eds. P.A. Allison and D.J. Bottjer), Aims & Scope Topics in Geobiology Book

Series, 32, 411-434.

Cantrell, D., Swart, P. and Hagarty, R. (2004) Genesis and characterization of dolomite,

Arab-D Reservoir, Ghawar field, Saudi Arabia. GeoArabia, 9, 11-36.

Carballo, J.D., Land, L.S. and Miser, D.E. (1987) Holocene dolomitization of supratidal

sediments by active tidal pumping, Sugarloaf Key, Florida. Journal of Sedimentary Petrology,

57, 153-165.

Cavelier, C. (1970) Geologic Description of the Qatar Peninsula. Department of Petroleum

Affairs, Paris, France, 39 p.

Chiba, H. and Sakai, H. (1985) Oxygen isotope exchange rate between dissolved sulfate and

water at hydrothermal temperatures. Geochimica et Cosmochimica Acta, 49, 993-1000.

Cohen, H. and Kaczmarek, S. (2017) Evaluating the effects of fluid NaCl and KCl

concentrations on reaction rate, major cation composition, and cation ordering during high-

temperature dolomitization experiments: Geological Society of America Abstracts with

Programs, v. 49, no. 6, paper 182-46, https://doi.org/10.1130/abs/2017AM-306364.

Danielli, H.M.C. (1988) The Eoecene reservoirs of Wafra Field. In: Giant oil and gas fields: A

core workshop (Eds A.J. Lomando and P.M. Harris), SEPM Core Workshop, 12, 119-154.

Deffeyes, K.S., Lucia, F.J. and Weyl, P.K. (1965). Dolomitization of Recent and Plio–

Pleistocene sediments by marine evaporite waters on Bonaire, Netherlands Antilles. In:,

Dolomitization and limestone diagenesis (Eds. L.C. Pray, R.C. Murray), SEPM Spec. Publ., 13,

71–88.

Dravis, J.J. and Wanless, H.R. (2018) Reflux dolomitization – A Holocene example beneath a

coastal salina, West Caicos Island, Turks and Caicos Islands. Marine and Petroleum Geology,

97, 311-322.

Eccleston, B.L., Pike, J.G. and Harshash, I. (1981) The water resources of Qatar and their

development: Vol. 1. Food and Agricultural Organization (FAO) of the United Nations, Doha.

Technical Report No. 5. 431 p.

Enos, P. and Sawatsky, L.H. (1981) Pore space in Holocene carbonate sediments. Geological

Society of America Bulletin, 92, 961-985.

Esteban, M. (1974) Caliche textures and Microcodium. Boll. Soc. Geol. Ital., 92, 105-125.

Folk, R.L. and Land, L.S. (1975) Mg/Ca ratio and salinity: Two controls over crystallization of

dolomite. AAPG Bulletin, 59, 60-68.

Fullmer, S.M. and Lucia, F.J. (2010) Rate of reflux dolomitization based on hydrodynamic

modeling of a Glen Rose dolostone, Austin, Texas. Gulf Coast Association of Geological Societies

Transactions, 60, 291-300.

Garcia-Fresca, B. and Jones, G.D. (2011) Apparent stratigraphic concordance of reflux

dolomite: New insight from high-frequency cycle scale synsedimentary reactive transport

models. In: Carbonate Geochemistry: Reactions and Processes in Aquifers and Reservoirs (Eds

A. Summers Engel, S. Engel, P. Moore, H. DuChene), Karst Waters Institute Special Publication,

16, 27-30.

Garcia-Fresca, B., Lucia, J.F., Sharp Jr., J.M. and Kerans, C. (2012) Outcrop-constrained

hydrogeological simulations of brine reflux and early dolomitization of the Permian San

Andres Formation. AAPG Bulletin, 96, 1757-1781.

Goldsmith, J.R. and Graf, D.L. (1958) Structural and compositional variations in some natural

dolomites. Journal of Geology, 66, 678-693.

Glover, E.D. and Sippel, R.F. (1967) Synthesis of magnesium calcites. Geochimica et

Cosmochimica Acta, 31, (603-613).

Gregg, J.M., Bish, D.L., Kaczmarek, S.E. and Machel, H.G. (2015) Mineralogy, nucleation

and growth of dolomite in the laboratory and sedimentary environment: A review.

Sedimentology, 62, 1749-1769.

Hasson, P.F. (1985) New observations on the biostratigraphy of the Saudi Arabian Umm er

Radhuma Formation (Paleogene) and its correlation with neighboring regions.

Micropaleontology, 31, 335-364.

Hewaidy, A.G. (1994) Biostratigraphy of the Umm er Radhuma Formation in South-East Qatar,

Arabian Gulf. N. Jb. Geol. Palaont. Abh, 193, 145-164.

Horita, J. (2014) Oxygen and carbon isotope fractionation in the system dolomite–water–CO2

to elevated temperatures. Geochim. Cosmochim. Acta, 129, 111–124.

Humphrey, J.D. and Quinn, T.M. (1989) Coastal mixing zone dolomite, forward modeling and

massive dolomitization of platform-margin carbonates. Journal of Sedimentary Petrology, 59,

438-454.

Meddaugh, W.S., Dull, D., Garber, R.A., Griest, S. and Barge, D. (2007) The Wafra First

Eocene reservoir, Partitioned Neutral Zone (PNZ), Saudi Arabia and Kuwait: Geology,

stratigraphy, and static reservoir modeling: Presented at the 2007 Society of Petroleum

Engineers Middle East Oil Show, Bahrain, March 11-14, 2007, SPE Paper 105087, 11p.

Jones, G.D. and Xiao, Y. (2005) Dolomitization, anhydrite cementation, and porosity evolution

in a reflux system: Insights from reactive transport models. American Association of

Petroleum Geologists Bulletin, 89, 577-601.

Jorgensen, B.B. and Kasten, S. (2006) Sulfur cycling and methane oxidation. Marine

Geochemistry, 271-309.

Kabanov, P., Anadon, P. and Krumbein, W.E. (2008) Microcodium: An extensive review and

a proposed non-rhizogenic biologically induced origin for its formation. Sedimentary Geology,

205, 79-99.

Kaczmarek, S.E. and Sibley, D.F. (2007) A comparison of nanometer-scale growth and

dissolution features on natural and synthetic dolomite crystals: Implications for the origin of

dolomite. Journal of Sedimentary Research, 77, 424-432.

Kaczmarek, S.E. and Sibley, D.F. (2011) On the evolution of dolomite stoichiometry and cation

order during high-temperature synthesis experiments: An alternative model for the

geochemical evolution of natural dolomites. Sedimentary Geology, 240, 30-40.

Kaczmarek, S.E. and Sibley, D.F. (2014) Direct physical evidence of dolomite recrystallization.

Sedimentology, 61, 1862-1882.

Kaczmarek, S.E. and Thornton, B.P. (2017) The effect of temperature on stoichiometry, cation

ordering, and reaction rate in high-temperature dolomitization experiments. Chemical

Geology, 468, 32-41.

Kaczmarek, S.E., Gregg, J.M., Bish, D.L., Machel, H.G. and Fouke, B.W. (2017) Dolomite,

very-high magnesium calcite, and microbes – implications for the microbial model for

dolomitization. In: Characterization and Modeling of Carbonates – Mountjoy Symposium 1 (Eds

A. MacNeil, J. Lonnee, and R. Wood), SEPM Spec. Publ., 109, 7-20.

Knauth, L.P. and Beeunas, M.A. (1986) Isotope geochemistry of fluid inclusions in Permian

halite with implications for the isotope history of ocean water and the origin of saline

formation waters. Geochimica et Cosmochimica Acta, 50, 419-433.

Land, L.S. (1967) Diagenesis of skeletal carbonates. Journal of Sedimentary Petrology, 37, 914-

930.

Land, L.S. (1973) Holocene meteoric dolomitization of Pleistocene limestones, North Jamaica.

Sedimentology, 20, 411-424.

Land, L.S. (1980) The isotopic and trace element geochemistry of dolomite: The state of the art.

In: Concepts and Models of Dolomitization (Eds. D.H. Zenger, J.B. Dunham and R.L. Ethington),

SEPM Spec. Publ., 28, 87-110.

Land, L.S. (1985) The origin of massive dolomite. Journal of Geological Education, 33, 112-

125.

Land, L.S. (1991) Dolomitization of the Hope Gate Formation (north Jamaica) by seawater:

Reassessment of mixing-zone dolomite: In: Stable Isotope Geochemistry: a tribute to Samuel

Epstein (Eds. H.P. Taylore, J.R. O’Neil and I.R. Kaplan), Geochem. Soc. Spec. Publ., 3, 121-133.

Lloyd, R.M. (1968) Oxygen isotope behavior in the sulfate-water system. Journal of

Geophysical Research, 73, 6099-6110.

Lumsden, D.N. (1979) Discrepancy between thin-section and X-ray estimates of dolomite in

limestone. Journal of Sedimentary Petrology, 49, 429-436.

Machel, H.G. (1997) Recrystallization versus neomorphism, and the concept of ‘significant

recrystallization’ in dolomite research. Sedimentary Geology, 113, 161-168.

Machel, H.G. (2004) Concepts and models of dolomitization: a critical reappraisal. In: The

Geometry and Petrogenesis of Dolomite Hydrocarbon Reservoirs (Eds. C.J.R. Braithwaite, G.

Rizzi and G. Darke), Geol. Soc. Lond. Spec. Publ., 235, 7-63.

Malone, M.J., Baker, P.A. and Burns, S.J. (1996) Recrystallization of dolomite: An

experimental study from 50-200°C. Geochimica et Cosmochimica Acta, 60, 2189-2207.

Manche, C.J. and Kaczmarek, S.E. (2019) Evaluating reflux dolomitization using a novel high-

resolution record of dolomite stoichiometry: A case study from the Cretaceous of central

Texas, USA. Geology, 47, 586-590.

Mattes, B.W. and Mountjoy, E.W. (1980) Burial dolomitization of the Upper Devonian Miette

buildup, Jasper National Park, Alberta. In: Concepts and Models of Dolomitization (Eds.

D.H. Zenger, J.B. Dunham and R.L. Ethington), SEPM Spec. Publ., 28, 259-297.

Markovic, S., Paytan, A., Li, H. and Wortmann, U.G. (2016) A revised seawater sulfate

oxygen isotope record for the last 4 Myr. Geochimica et Cosmochimica Acta, 175, 239-251.

Mazzulo, S.J. (1992) Geochemical and neomorphic alteration of dolomite: A review.

Carbonates and Evaporites, 7, 21-37.

McKenzie, J.A., Hsu, K.J. and Schneider, J.F. (1980) Movement of subsurface waters under

the sabkha, Abu Dhabi, UAE, and its relation to evaporative dolomite genesis. In: Concepts and

Models of Dolomitization (Eds. D.H. Zenger, J.B. Dunham and R.L. Ethington), SEPM Spec. Publ.,

28, 11-30.

Moore, C.H., Chowdhury, A. and Chan., L. (1988) Upper Jurrasic Smackover platform

dolomitization, northwestern Gulf of Mexico: a tale of two waters. In: Sedimentology and

Geochemistry of Dolostones (Eds. V.J. Shukla and P.A. Baker), SEPM Special Publication, 43,

175-189.

Morrow, D.W. (1982) Diagenesis 2. Dolomite—Part 2: Dolomitization models and ancient

dolostones. Geoscience Canada, 9, 95-107.

Newport, R., Hollis, C., Bodin, S. and Redfern, J. (2017) Examining the interplay of climate

and low amplitude sea-level change on the distribution and volume of massive

dolomitization: Zebbag Formation, Cretaceous, Southern Tunisia. The Depositional Record, 3,

38-59.

Patterson, R.J. and Kinsman, J.J. (1982) Formation of diagenetic dolomite in coastal sabkha

along Arabian (Persian) Gulf. AAPG Bulletin, 66, 28-43.

Pollitt, D.A., Anthonissen, E., Saller, A.H., BouDagher-Fadel, M.K. and Dickson, J.A.D.

(2012) Abrupt early Eocene global climatic change as a control of carbonate facies and

diagenesis: a new record of the Palaeocene-Eocene Thermal Maximum in the Umm er

Radhuma Formation, Saudi Arabia and Kuwait. Terra Nova, 24, 487-498.

Powers, R.W., Ramirez, L.F., Redmond, C.D. and Elberg, E.L. (1966) Geology of the

Arabian Peninsula (Sedimentary Geology of Saudi Arabia). U.S. Geological Survey,

Washington DC. Survey, Professional Paper, 560D, 147 pp.

Reeder, R.J. and Sheppard, C.E. (1984) Variation of lattice parameters in some sedimentary

dolomites. American Mineralogist, 69, 520-527.

Rivers, J.M., Skeat, S., Yousif, R., Liu, C., Stanmore, E., Tai, P. and Al-Marri, S.M. (2019a)

The depositional history of near-surface Qatar aquifer rocks, and its impact on matrix flow

and storage properties. Arabian Journal of Geosciences, 12, 1-33.

Rivers, J.M., Varghese, L., Yousif, R., Whitaker, F.F., Skeat, S.L. and Al-Shaikh, I. (2019b)

The geochemistry of Qatar coastal waters and its impact on carbonate sediment chemistry

and early marine diagenesis. Journal of Sedimentary Research, 89, 293-309.

Royse, C.F., Jr., Wadell, J.S. and Petersen, L.E. (1971) X-ray determination of calcite-

dolomite: an evaluation. Journal of Sedimentary Petrology, 41, 483-488.

Rubin, E. (2011) Full field modeling of Wafra First Eocene reservoir 56-year production

history: Society of Petroleum Engineers Heavy Oil Conference and Exhibition, Kuwait City,

Kuwait, December 12-14, 2011, SPE Paper 150575, 15 p.

Ruppel, S.C. and Cander, H.S. (1988) Dolomitization of shallow water carbonates by seawater

and seawater-derived brines: San Andres Formation (Guadalupian), west Texas. In:

Sedimentology and Geochemistry of Dolostones (Eds V.J. Shukla and P.A. Baker), SEPM Spec.

Publ., 43, 245-262.

Ryan, B.H., Kaczmarek, S.E. and Rivers, J.M. (2019) Dolomite dissolution: An alternative

Diagenetic pathway for the formation of palygorskite clay. Sedimentology, 66, 1803-1824.

Saller, A.H. and Henderson, N. (1998) Distribution of porosity and permeability in platform

dolomites: Insight from the Permian of West Texas. American Association of Petroleum

Geologists Bulletin, 82, 1528-1550.

Saller, A.H., Pollitt, D.A. and Dickson, J.A.D. (2014) Diagenesis and porosity development in

the First Eocene reservoir at the giant Wafra Field, Partitioned Zone, Saudi Arabia and Kuwait.

AAPG Bulletin, 98, 1185-1212.

Sass, E. and Katz, A. (1982) The origin of platform dolomites: New evidence. American

Journal of Science, 282, 1184-1213.

Sass, E. and Bein, A. (1988) Dolomites and salinity: A comparative geochemical study. In:

Sedimentology and Geochemistry of Dolostones (Eds. V. Shukla and P.A. Baker), SEPM Spec.

Publ., 43, 223-233.

Sharland, P.R., Archer, R., Case, D.M., Davies, R.B., Hall, S.H., Heward, A.P., Horbury,

A.D. and Simmons, M.D. (2001) Arabian plate sequence stratigraphy. GeoArabia Special

Publication no. 2, Gulf Petrolink, Bahrain, 371 p.

Sibley, D.F. (1990) Unstable to stable transformations during dolomitization. The Journal of

Geology, 98, 739-748.

Sibley, D.F. and Gregg, J.M. (1987) Classification of dolomite rock textures. Journal of

Sedimentary Petrology, 57, 967-975.

Simms, M.A. (1984) Dolomitization by groundwater flow systems in carbonate platforms.

Transactions of the Gulf Coast Association of Geological Sciences, 24, 411-420.

Sun, S.Q. (1995) Dolomite reservoirs: Porosity evolution and reservoir characteristics. AAPG

Bulletin, 79, 249-257.

Tyrell, T. and Zeebe, R.E. (2004) History of carbonate ion concentration over the last 100

million years. Geochimica et Cosmochimica Acta, 68, 3521-3530.

Vahrenkamp, V.C. and Swart, P.K. (1990) New distribution coefficient for the incorporation

of strontium into dolomite and its implications for the formation of ancient dolomites.

Geology, 18, 387-391.

Van Tuyl, F.M. (1916) New points on the origin of dolomite. American Journal of Science, 42,

249-260.

Warren, J. (2000) Dolomite: occurrence, evolution and economically important associations.

Earth Sci. Rev., 52, 1-81.

Warren, J. (2016) Evaporites. Springer International Publishing, 1813 p.

Westerhold, T., Röhl, U., McKarren, H.K. and Zachos, J.C. (2009) Latest on the absolute age

of the Paleocene-Eocene Thermal Maximum (PETM): new insights from exact stratigraphic

position of key ash layers +19 and -17. Earth and Planetary Science Letters, 287, 412-419.

Whittaker, S.G. and Mountjoy, E.W. (1996) Diagenesis of an Upper Devonian carbonate-

evaporite sequence: Birdbear Formation, Southern Interior Plains, Canada. Journal of

Sedimentary Research, 5, 965-975.

Zachos, J.C., Sott, L.D. and Lohmann, K.C. (1994) Evolution of early Cenozoic marine

temperatures. Paleoceanography, 9, 353-387.

Zachos, J., Pagani, M., Sloan, L., Thomas, E. and Billups, K. (2001) Trends, rhythms, and

aberrations in global climate 65 Ma to Present. Science, 292, 686-692.

Zachos, J.C., Schouten, S., Bohaty, S., Quattlebaum, T., Slujis, A., Brinkhuis, H., Gibbs,

S.J. and Bralower, T.J. (2006) Extreme warming of mid latitude coastal ocean during the

Paleocene-Eocene Thermal Maximum: Inferences from TEX86 and isotope data. Geology, 34,

737-740.

Ziegler, M.A. (2001) Late Permian to Holoce Paleofacies Evolution of the Arabian Plate and its

Hyrdrocarbon Occurrences. GeoArabia, 6, 445-504.

CHAPTER III

EARLY AND EXTENSIVE RECRYSTALLIZATION OF CENOZOIC DOLOMITES DURING SHALLOW BURIAL

Brooks H. Ryan1, Stephen E. Kaczmarek1, and John M. Rivers2 1Department of Geological and Environmental Sciences, Western Michigan University, Kalamazoo, MI, U.S.A. 2ExxonMobil Research Qatar, Qatar Science and Technology Park, Doha, Qatar

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., in review, Geology.

Abstract

Recrystallization can alter the textural, mineralogical, and geochemical signatures recorded during dolomitization. However, most examples of extensive recrystallization in ancient dolomites are attributed to deep burial. Textural, mineralogical, and geochemical data from the Paleocene-

Eocene Umm er Radhuma Formation (UER) in Qatar show that, in contrast, extensive dolomite recrystallization can also occur early in a near-surface setting. Dolomite stoichiometry and cation ordering correlate with textures that trend with depth from finely crystalline mimetic to coarse planar-e to coarser nonplanar with depth. Shallow mimetic dolomites are the least stoichiometric and least ordered, whereas coarser planar-e and the coarsest nonplanar dolomites observed with depth are more stoichiometric and better-ordered, suggesting recrystallization with shallow burial.

Furthermore, a principal component analysis of dolomite stoichiometry, cation ordering, δ18O, and

δ13C indicates that variations in the data can be explained by two factors. The first factor is interpreted to be dolomite recrystallization, which shows positive loadings of stoichiometry and cation ordering but a negative loading of δ18O. The second factor is interpreted to be the temporal evolution of dissolved inorganic carbon, as δ13C is the only heavily loaded component. Progressive,

early recrystallization during shallow burial is consistent with petrographic observations of two dm- thick chert bands within a nonplanar dolomite interval that preserve less altered mimetic and planar- e dolomites. ∆47-derived temperatures (35°C to 46°C) and δ18Ow values (+0.8 to +2.6‰) are also indicative of shallow burial recrystallization (< 370 m) by slightly evaporated seawater. Collectively, these findings indicate that early diagenesis may reset most, but perhaps not all textural and geochemical attributes, which stands in contrast to the strongly held supposition in the literature that extended geologic time and deep burial are needed for significant dolomite recrystallization.

Introduction

Dolomite is a common replacement of limestone and is a significant host of hydrocarbons and mineral deposits (Warren, 2000), yet debate continues over the “dolomite problem” (Land, 1985;

Gregg et al. 2015). At issue is how massive, platform-scale dolomitization happens (Morrow, 1982).

Conceptual models, which describe the environments, fluids, and hydrological flow mechanisms of dolomitization, are inferred from textural, mineralogical, and geochemical data. It is commonly assumed, however, that these petrological attributes reflect the initial dolomitization event, despite the potential for recrystallization (Mazzullo, 1992).

Modern and geologically young dolomites are generally Ca-rich and poorly ordered compared to their ancient counterparts (Lumsden and Chimahusky, 1980). Laboratory experiments also show the initial dolomite phase to form during dolomitization is poorly ordered and often Ca- rich (Kaczmarek and Sibley, 2014), which is thermodynamically less stable (Navrotsky and

Capobianco, 1987). Thus, there is the potential for dolomite recrystallization.

A wide range of data has been cited as evidence of dolomite recrystallization, including large crystals, a near stoichiometric composition, a high degree of cation ordering, low Sr and Na concentrations, light O isotope ratios, and mottled cathodoluminescence signatures, among others

(Mazullo, 1992). It has been argued that these data may reflect the conditions of recrystallization

rather than the original dolomitizing conditions. This is especially likely in ancient dolomites that have undergone significant burial diagenesis (Kupecz et al., 2003; Machel, 2004), but what about young dolomites that have not suffered deep burial?

To address this question, we investigated the Umm er Radhuma (UER), a pervasively dolomitized unit in Qatar that has likely not been buried deeper than present day depths of 50-400 m (Van Buchem et al., 2014). Based on petrographic data, recent work on the UER suggests that the replacement of limestone by dolomite (i.e. dolomitization) occurred very early in the diagenetic history (52-56 Ma), prior to the formation of chert, pyrite, palygorskite, calcite, and gypsum (Ryan et al., 2020). Here, we present evidence that suggests despite early dolomitization and limited burial depths, early and extensive dolomite recrystallization is responsible for changes in the textural, mineralogical, and geochemical signatures. In addition, such recrystallization of Cenozoic dolomites under shallow burial is unique as shown by an analysis of several studies of recrystallized dolomite, which suggests that recrystallization is mostly an ancient and deep phenomenon.

Geological Setting

During much of the late Paleocene and early Eocene, shallow carbonate ramps covered the southern margin of the Tethys region (Sharland et al., 2001). Transgression and regression along a northeast-facing ramp on the Arabian Peninsula marked the onset of Cenozoic deposition, resulting in sequences of subtidal to peritidal carbonates comprising the Umm er Radhuma Formation (UER).

The UER is as thick as 370 m, observed only in the subsurface, and is continuous across Qatar and most of the eastern Arabian Peninsula (Eccleston et al., 198; Sharland et al., 2001). The unit is characterized by m-thick, fining upward subtidal sequences dominated by dolomite, intercalated siliciclastics, and sporadic replacive evaporite nodules (Rivers et al., 2019).

Methods

Petrographic, mineralogical, and geochemical data were collected from three shallow research cores from Qatar (Figure 1). Detailed methods for thin section petrography, x-ray diffractometry, stable isotopes, and clumped isotopes are presented in the GSA Data Repository1. A principal component analysis (PCA) was carried out using Statistical Product and Service Solutions statistical software in order to determine how many factors control the variability in 71 samples with corresponding data for stoichiometry, cation ordering, δ18O, and δ13C across all three cores. Variables that correlate to one another are combined into factors, with different loadings that relate to the strength of the association between the factors and variables (Joliffe, 2002). Lastly, a literature review of naturally recrystallized dolomites was conducted which includes only journal articles postdating 1992. Geologic age and recrystallization burial depth (shallow ≤ 0.5 km ≤ intermediate ≤

1 km ≤ deep) were recorded to test the assumption that dolomite recrystallization is associated with significant burial.

Depth (m) 60 Core 1 UER Stratigraphic Geologic Age Lithology N Mimetic A D Unit Arabian Peninsula 70

Dammam Formation Middle Core 3 Persian 80 Gulf Eocene Rus Qatar B E Formation 90 Planar-e Early

Core 2 This study 100

Core 1

110 Umm er C KEY Radhuma Dolomite Clastics Formation 120

Paleocene Nonplanar Gypsum Calcite Saudi 2 1 130 Arabia Chert band PALEOGENE

Kilometres 0% 100% 140 Mineralogy Figure 1: A stratigraphic column of the Paleogene strata of Qatar (modified after Ryan et al., 2020), with the specific study interval outlined in red. Locations of cores 1 (+25.23773°, +51.22826°), 2 (+25.38417°, +51.14507°), and 3 (+25.86184°, +51.26197°) shown on map of Qatar (inset map shows location of Qatar on the Arabian Peninsula). Mineralogic profile of the Umm er Radhuma Fm. in core 1 (depth in meters below the land surface), with corresponding thin section images of mimetic (A), planar-e (B), nonplanar (C) dolomite textures, planar-e dolomite crystals in chert cement (D) from the upper chert band (#2) and a silicified mimetically dolomitized (red arrow) foraminifera (E; modified from Ryan et al. 2020) from the lower chert band. Scale bars in A-E represent 200 µm.

1 GSA Data Repository item DR1, Complete Methodology and Statistical Analyses 65

Results

The upper 60-75 m of the UER were analyzed in all cores, and are dominated by dolomite, comprising 62 to 100% of bulk mineralogy (Fig. 1). Dolomite textures vary with depth in all cores

(Figs. 1 and 2), with a general increase in crystal size with depth. The uppermost ~7 m of core 1 contains microcrystalline dolomite (< 10 µm) that has mimetically replaced the precursor limestone

(Fig. 1). The middle 50 m of core 1, and top 20-30 m of cores 2 and 3 are dominated by relatively coarse (30-100 µm) planar-e dolomite exhibiting cloudy cores and clear rims (CCCR). The bottom 20-

30 m of all three cores consists of coarser (100-200 µm) nonplanar dolomite. A correlative dm-thick chert band with planar-e dolomite crystal inclusions is observed in all three cores. Core 1 has two chert bands separated by < 1 meter (Fig. 1). The lower chert band in core 1 includes mimetic dolomite, whereas the upper one includes only planar-e dolomite (Fig. 1).

Dolomite stoichiometry (ave. 50.3 mol% Mg) and cation ordering (ave. 0.73) values are generally high, with statistically different averages as a function of dolomite texture1. Mimetic dolomites are the least stoichiometric (49.7 mol% Mg) and least ordered (0.57), whereas nonplanar dolomites are more stoichiometric (50.1% mol% Mg) and ordered (0.71), and planar-e dolomites are the most stoichiometric (50.7% mol% Mg) and ordered (0.77) (Fig. 2A). Dolomite δ18O and δ13C values, in contrast, are highly variable with respect to dolomite texture (Fig. 2B). Mimetic dolomites average +0.03‰ (range -0.45 to +0.38‰) δ18O and -0.94‰ (range -2.12 to -0.24‰) δ13C (VPDB).

Nonplanar dolomites average +0.03‰ (-2.29 to +1.39‰) and +0.18‰ (-2.57 to +2.22‰), respectively. Planar-e dolomites have the lowest δ18O averaging -1.06‰ (-2.43 to +0.68‰) with δ13C similar to the other textures (ave -0.50‰; -3.47 to +1.55). In all three cores, a ~2.5-4‰ negative carbon isotope excursion is observed in nonplanar dolomites (Ryan et al., 2020). Temperatures derived from ∆47 analysis range from 35.1-46.2°C and calculated fluid δ18Ow ranges +0.84 to +2.64‰

(VSMOW) (Table 1).

Dolomite 1 δ13C 0.8 A B CC160 1 3 Expected dolomite 0.6 Stoichiometry δ18O values for late 2 140 0.9 0.4 Paleocene to early Cation 10)

1 Eocene seawater 120Dolomite Ordering 1 18 0.2 0.8 δ O 100 0 0 0.7 -4 -3 -2 -1 0 1 2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 80 -1 2 Factor -0.2 0.6 Sr (ppm) 60 -0.4 -2 Planar-e Cation Ordering (015: 0.5 40 -0.6

Nonplanar C (‰ VPDB)

3 Nonplanar

1 -3 Mimetic 20 0.4 Mimetic -0.8 49 49.5 50 50.5 51 51.5 -4 Planar-e Mimetic Planar-e Nonplanar Stoichiometry (mol% Mg) 18O (‰ VPDB) 0 -1 Factor 1

Figure 2: Cross plots of cation ordering and stoichiometry (A) and stable isotopic composition (B) as a function of dolomite texture for all cores. Loading values for stoichiometry, ordering, δ18O, and δ13C as a function of factor are shown in (C).

PCA results yield 2 statistically useful factors with eigenvalues > 0.7 and that explain ~80% of the total variance1 (Joliffe, 2002). Factor 1 accounts for 58% of the total variance and is dominated by high positive loadings of stoichiometry (0.774) and cation ordering (0.823) and a high negative loading of δ18O (-0.846) (Fig. 2C). Factor 2 (23% of total variance) is dominated by a high positive loading of δ13C (0.790).

A total of 71 papers were analyzed from the literature for natural, recrystallized dolomites1.

The results show that 90% of these studies involve pre-Cenozoic dolomites, and that 81% of pre-

Cenozoic dolomites interpreted to be recrystallized were recrystallized under self-reported intermediate to deep burial conditions, although many studies reported depth ranges as well. In contrast, only 6% of these 71 studies include Cenozoic dolomites that underwent shallow burial recrystallization. Of all studies that interpreted shallow burial recrystallization, 71% invoked the possible influence of meteoric fluids.

Table 1. Core 1 clumped isotope analysis Depth Dolomite Average ± 1 SE δ18Odol (‰ ± 1 SD Average ± 1 SE δ18Ow (‰ ± 1 SE (m) texture ∆47 (‰) (‰) VPDB) (‰) Temperature (°C) (°C) VSMOW) (‰) 66.48 Mimetic 0.661 0.006 1.08 0.16 35.1 2.10 1.61 0.52 120.67 Nonplanar 0.634 0.007 -0.15 0.10 46.2 2.90 2.64 0.55 128.51 Nonplanar 0.654 0.025 -0.31 0.17 38.8 9.60 0.84 2.10

Discussion

The subtle but significant differences in stoichiometry, cation ordering, and to a lesser extent

δ18O values, as a function of texture, reflect either differences in dolomitization processes associated with each texture, or varying degrees of recrystallization. Ryan et al. (2020) suggested that dolomite textures in the UER are related to precursor limestone facies, in which mimetic dolomites replaced muddier sediments whereas planar-e and nonplanar textures resulted from grainier sediments.

Furthermore, Ryan et al. (2020) argued that the UER dolomites initially formed from Late Paleocene to early Eocene seawater (δ18Ow of -0.98‰ and SST 25-35°C, Zachos et al., 1994; 2006) based on the similarity between dolomite δ18O values theoretically formed under such conditions (-1.5‰ and

+0.75‰) and those of the UER (-2.5‰ to +1‰). However, Ryan et al. (2020) noted that the more negative range of UER dolomite δ18O values, coupled with the near-stoichiometric and well-ordered nature of the dolomites, permits the possibility of recrystallization. The relationship between texture, stoichiometry, ordering, δ18O presented here further supports a recrystallization hypothesis. The mimetic dolomites are the least stoichiometric, least ordered, and have a tight range of δ18O values nearest the upper limit of the theoretical range. In contrast, the planar-e and nonplanar dolomites are more stoichiometric, more ordered, and occupy a wider range with δ18O as negative as -2.5‰, with nonplanar dolomites of cores 2 and 3 commonly ≥ 0‰ (Fig. 2A). The greater δ18O range in the latter two textures implies that those values record more than the conditions during the initial dolomitization and have likely been recrystallized to some degree either in different temperatures or fluids. Following this, the fine crystals comprising mimetic dolomites, along with lower stoichiometry and ordering values, as well as δ18O values closest to expected, reflect the lowest degree of recrystallization (Mazzullo, 1992). The planar-e and nonplanar and dolomites, in contrast, represent more recrystallized dolomite based on larger crystal sizes, being more stoichiometric and well-ordered, and containing a wider δ18O range.

Results of the PCA are also consistent with the interpretation that changes in stoichiometry, cation ordering, and δ18O are related to dolomite recrystallization. Factor 1, which shows positive loadings of stoichiometry and cation ordering but a negative loading of δ18O, is interpreted here to reflect recrystallization. Recrystallization of dolomite in laboratory experiments demonstrates similar trends, in that stoichiometry and cation ordering concurrently increase (Kaczmarek and

Sibley, 2014) whereas dolomite δ18O generally decreases (Malone et al. 1996). Changes in UER dolomite δ18O possibly also reflect changing temperature, fluid salinity, fluid δ18O, or precursor calcite δ18O (Swart, 2015). Of these, only temperature and salinity also affect dolomite stoichiometry, in which a positive correlation is observed with both parameters (Kaczmarek and Thornton, 2017;

Cohen, 2019). However, none of these conditions affects the rate at which cation ordering increases, which has been experimentally shown to only increase simultaneously with stoichiometry through dolomite recrystallization (Malone et al. 1996; Kaczmarek and Sibley, 2014). Thus, the most reasonable interpretation for simultaneous positive loading of stoichiometry and cation ordering and negative loading of δ18O on Factor 1 is dolomite recrystallization. This suggests that the planar-e dolomites, which are the most stoichiometric, most ordered, and contain the lowest δ18O values, are the most recrystallized of all dolomite types. The nonplanar dolomites have likely also been significantly recrystallized based on even larger crystal sizes and high stoichiometry and ordering values, but their relatively higher δ18O values in cores 2 and 3 suggest either somewhat less alteration, initial dolomitization of precursor a calcite with higher δ18O than in core 1, or recrystallizing fluids with higher δ18O.

The observed increase in crystal size and change in dolomite texture with depth in the three cores suggests that recrystallization likely occurred during shallow burial. This is supported by the observation that the two chert bands within a nonplanar dolomite interval in core 1 preserve the mimetic and planar-e dolomite textures that prevail higher in the section (Fig. 1E, F). The observation that dolomite textures are retained through silicification, and that the chert band is correlative across

Qatar (Eccleston et al., 1981), suggests that silicification was associated with widespread but possibly short-lived exposure events that occurred after both mimetic and planar-e dolomites. The two chert bands in core 1, which capture the mimetic and planar-e dolomites, suggest that there were two episodes of exposure. Rivers et al. (2019) reported a δ18O for the lower chert band in core 1 of -6.5‰ and suggested that meteoric fluids were responsible. The difference between the δ18O of the chert and that of the dolomites both within and outside of the chert band are consistent with an exposure event that introduced silica-saturated meteoric fluids prior to further dolomite recrystallization. The replacement of dolomite by chert also clearly shows that dolomite recrystallization occurred early in the diagenetic history. Further evidence of early recrystallization includes planar-e dolomites with partially dissolved Ca-rich cores and pristine stoichiometric rims (Ryan et al., 2019) that are both cross-cut by later mineral phases including chert, palygorskite, gypsum, and calcite (Ryan et al.,

2020).

It is possible that the meteoric fluids associated with silicification were involved with subsequent dolomite recrystallization, as suggested by less recrystallized dolomites being replaced by chert. The ∆47-derived temperatures and δ18Ow values (Table 1), however, are more consistent with slightly evaporative and slightly warmer Eocene seawater (Zachos et al., 1994; 2006). The ∆47- derived temperatures can provide further constraints on how and where shallow the recrystallization occurred. Considering an SST within 25-35°C, and the interpretation that mimetic dolomites are not significantly recrystallized, then the ∆47-derived 35°C for the mimetic sample likely represents surface conditions. Assuming a geothermal gradient of 30°C/km, then the nonplanar samples, which yield crystallization temperatures of 39-46°C, correlate to deepest recrystallization between 130-370 m. Thus, recrystallization of the UER is interpreted to have occurred by modified seawater during shallow-burial, implying temperatures >50°C may not be required for the formation of nonplanar dolomite (Sibley and Gregg, 1987).

Extensive, early dolomite recrystallization of Cenozoic dolomites in a shallow burial realm is significant for two reasons. First, the overwhelming majority of interpreted recrystallized dolomites are post-Cenozoic in age. Secondly, in most of these post-Cenozoic dolomites, recrystallization is interpreted to have occurred under intermediate (>0.5 km) to deep (>1 km) burial, with 72% being exclusively of deep burial origin. These observations alone indicate that the Paleocene-Eocene UER dolomites are a unique example of geologically young dolomites that have recrystallized under shallow burial conditions. Furthermore, the interpretation that recrystallization happened in only modified Eocene seawater is significant, as 71% of recrystallized dolomites of shallow a burial origin were interpreted to be recrystallized in the possible presence of meteoric fluids. Thus, the interpretations presented here, involving shallow burial and slightly modified seawater, implies that deep burial, exceptional diagenetic fluids, and/or extended time are not necessary for extensive dolomite recrystallization. This suggests that extreme caution must be exercised when interpreting the conditions or fluids responsible for dolomitization, as subsequent recrystallization can happen early and in near-surface fluids and conditions. This opens the question for future studies: what exactly causes dolomites to recrystallize?

Equally important to the topic of early dolomite recrystallization is the observation that δ13C loads heavily on a separate factor as a lone variable. Ryan et al. (2020) proposed that the ~2.5-4‰

δ13C excursions correlated across the three wells in the nonplanar interval was likely the Paleocene-

Eocene Thermal Maximum (REFS). This suggests that while other geochemical and mineralogical properties may be reset during dolomite recrystallization, the δ13C may remain unaltered in certain cases, indicating that information pertaining to past climatic conditions may be preserved even through recrystallization.

Conclusions

Collectively, the textural, mineralogical, and geochemical data support the interpretation that the Umm er Radhuma Fm. dolomites have undergone extensive recrystallization with depth in a shallow burial environment. Results from the principal component analysis support this interpretation, as multiple variables affected by dolomite recrystallization load on a single factor.

Petrographic data from two chert layers further constrain the timing of recrystallization to be very early. These findings suggest that dolomite recrystallization is not only an ancient and deep phenomenon, but also that exceptional conditions other than modified seawater and shallow burial are not required for recrystallization. Lastly, despite significant alteration of δ18O, stoichiometry, and cation ordering, the depositional δ13Ccarb signature of the precursor limestone may have been retained in the UER dolomites, implying that some recrystallized dolomites may be viable candidates from which to extract δ13C records.

Acknowledgements

This work was supported by ExxonMobil Research Qatar. We are grateful to Sabrina Skeat for logistical support, and Dr. Sierra Petersen, University of Michigan, for clumped isotope analysis. We appreciate the thorough review from David Budd, and the comments of Chief Editor Gerald Dickens and an anonymous reviewer.

References

Cohen, H.F., 2019, Evaluating the effects of sodium, potassium, magnesium, and calcium

concentrations on dolomite stoichiometry, cation ordering, and reaction rate [Masters

Thesis], Kalamazoo, Western Michigan University, 71 pp.,

https://scholarworks.wmich.edu/masters_theses/5096.

Eccleston, B.L., Pike, J.G., and Harshash, I., 1981, The water resources of Qatar and their

development: Vol. 1. Food and Agricultural Organization (FAO) of the United Nations, Doha,

Technical Report No. 5, 431 p.

Gregg, J.M., Bish, D.L., Kaczmarek, S.E., and Machel, H.G., 2015, Mineralogy, nucleation and

growth of dolomite in the laboratory and sedimentary environment: A review:

Sedimentology, v. 62, p. 1749-1769.

Joliffe, I.T., 2002, Principal Component Analysis, Second Edition: New York, Springer-Verlag

New York, Inc., 487 p.

Kaczmarek, S.E., and Sibley, D.F., 2014, Direct physical evidence of dolomite recrystallization:

Sedimentology, v. 61, p. 1862-1882.

Kaczmarek, S.E., and Thornton, B.P., 2017, The effect of temperature on stoichiometry, cation

ordering, and reaction rate in high-temperature dolomitization experiments: Chemical

Geology, v. 468, p. 32-41.

Kupecz, J. A., I. P. Montañez, and G. Gao, 1993, Recrystallization of dolomite with time, in R.

Rezak and D. L. Lavoie, eds., Carbonate microfabrics: Springer-Verlag, New York, p. 187–194.

Land, L.S., 1985, The origin of massive dolomite: Journal of Geological Education, v. 33, p.

112-125.

Lumsden, D.N., and Chimahusky, J.S., 1980, Relationship between dolomite nonstoichiometry

and carbonate facies parameters, in Zenger, D.H., Dunham, J.B., and Ethington, R.L., eds.,

Concepts and Models of Dolomitization: Society of Economic Paleontologists and

Mineralogists (SEPM) Special Publication 28, p. 123-137.

Machel, H.G., 2004, Concepts and models of dolomitization: A critical reappraisal, in

Braithwaite, C., Rizzi, G., and Darke, G., eds., The Geometry and Petrogenesis of Dolomite

Hydrocarbon Reservoirs: Geological Society (London) Special Publication 235, p. 7-63.

Malone, M.J., Baker, P.A., and Burns, S.J., 1996, Recrystallization of dolomite: An experimental

study from 50-200°C: Geochemica et Cosmochimica Acta, v. 60, p. 2189-2207.

Mazzullo, S.J., 1992, Geochemical and neomorphic alteration of dolomite: A review: Carbonates

and Evaporites, v. 7, p. 21-37.

Morrow, D.W., 1982, Diagenesis 2: Dolomite—Part 2: Dolomitization models and ancient

dolostones: Geoscience Canada, v. 9, p. 95-107.

Navrotsky, A., and Capobianco, C., 1987, Enthalpies of formation of dolomite and magnesian

calcite: American Mineralogist, v. 79, p. 782-787.

Rivers, J.M., Skeat, S., Yousif, R., Liu, C., Stanmore, E., Tai, P., and Al-Marri, S.M., 2019, The

depositional history of near-surface Qatar aquifer rocks, and its impact on matrix flow and

storage properties: Arabian Journal of Geosciences, v. 12, p. 1-33.

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., 2019, Dolomite dissolution: An alternative

pathway for the formation of palygorskite clay: Sedimentology, v. 66, p. 1803-1824.

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., 2020, Early and pervasive dolomitization by

near-normal marine fluids: New lessons from an Eocene evaporative setting in Qatar:

Sedimentology (in press). https://doi.org/10.1111/sed.12726

Sharland, P.R., Archer, R., Case, D.M., Davies, R.B., Hall, S.H., Heward, A.P., Horbury, A.D.,

and Simmons, M.D., 2001, Arabian plate sequence stratigraphy: GeoArabia Special

Publication 2, 371 p.

Sibley, D.F., and Gregg, J.M, 1987, Classification of dolomite rock textures: Journal of

Sedimentary Petrology, v. 57, p. 967-975.

Swart, P.K., 2015, The geochemistry of carbonate diagenesis: The past, present and future:

Sedimentology, v. 62, p. 1233-1304.

Van Buchem, F., Svendsen, N., Hoch, E., Pedersen-Tatalovic, R., and Habib, K., 2014,

Depositional history and petroleum habitat of Qatar. in Marlow, L., Kendall, C.C.G., and Yose,

L.A., eds., Petroleum systems of the Tethyan region: AAPG Memoir, 106, p. 641-678.

Warren, J., 2000, Dolomite: occurrence, evolution and economically important associations:

Earth-Science Reviews, v. 52, p. 1-81.

Zachos, J.C., Stott, L.D., and Lohmann, K.C., 1994, Evolution of early Cenozoic marine

temperatures: Paleoceanography, v. 9, p. 353-387.

Zachos, J.C., Schouten, S., Bohaty, S., Quattlebaum, T., Slujs, A., Brinkhuis, H., Gibbs, S.J., and

Bralower, T.J., 2006, Extreme warming of mid-latitude coastal ocean during the Paleocene-

Eocene Thermal Maximum: Inferences from TEX86 and isotope data: Geology, v. 34, p. 737-

740.

Supplementary Material

Data Repository File DR1: Complete Methodology and Statistical Analyses

Thin section petrography and X-ray diffractometry

Thin sections were prepared from the ends of 391 blue-epoxy impregnated core plugs. Thin sections were subsequently stained with Alizarin Red S (ARS) to differentiate dolomite from calcite.

X-ray diffractometry (XRD) data was collected at Western Michigan University on 125 core plugs across all three cores. An electric rotary drill was used to extract powders which were then homogenized by hand with a mortar and pestle. Powders were analyzed under CuKα radiation with a Bruker D2 Phaser Diffractometer under the 2θ range of 20 to 40° with a step size of 0.01° and a count time of 1.0 s per step. Dolomite mol% MgCO3 was calculated using the equation derived by

Lumsden (1979) and cation ordering was calculated utilizing the method of Goldsmith & Graf, 1958).

Several internal standards were used to shift peak positions.

Dolomite Stable Isotope Analysis

Bulk rock stable isotope data used here (n = 191) were first reported in Rivers et al. (2019a) and then modified by Ryan et al. (2020) to exclude calcite contaminated samples or samples that may not represent the host rock. Stable isotope data were yielded from the Center for Stable Isotope

Biogeochemistry (CSIB), University of California, Berkeley. Samples were reacted with H3PO4 at 90°C for 10 minutes before being analyzed in a GV IsoPrime mass spectrometer with Dual-Inlet and

MultiCarb systems. Analytical precision is ±0.05‰ for δ13C and ±0.07‰ for δ18O. Values for all isotopic compositions are reported relative to the Vienna Peedee Belemnite (VPDB) standard.

Clumped Isotope Analysis

Four dolomite samples from core 1, spanning the entirety of the recovered UER Formation and representing all three dolomite textures (1 mimetic, 1 planar-e, and 2 nonplanar), were measured for their clumped (∆47) isotopic composition at the University of Michigan Stable Isotope

Laboratory. Three replicate analyses from three out of the four samples were utilized, with the fourth sample (planar-e dolomite; depth 89.55 m) only being replicated twice due to contamination of both analyses as indicated by high ∆48 values. The planar-e sample was thus discarded from the clumped isotope results. Measured dolomite ∆47 values are corrected with an acid fractionation factor of

+0.072‰ at 75°C following Petersen et al. (2019), which shows that this acid fractionation factor can be applied to all carbonate mineralogies. Temperatures were calculated using the calibration equation also developed by Petersen et al. (2019), which combines synthetic samples of 6 carbonate mineralogies. In order to calculate the δ18O of the fluid (δ18Ow), the ∆47-derived temperatures and dolomite δ18O values were integrated into the equation of Horita (2014). All δ18Ow data are reported relative to VSMOW.

Dolomite stoichiometry t-tests

Data from Ryan et al. (2020) are briefly reviewed here. Across all cores, mimetic dolomites have the lowest stoichiometry, averaging 49.7% (± 0.24%). This is followed by nonplanar dolomites, which average 50.0% (0.29%), and planar-e dolomites which average 50.6% (0.29%). A t-test at 95% confidence (� = 0.05) shows that planar-e dolomites are significantly more stoichiometric than both nonplanar and mimetic dolomites, with p-values of 7.9x10-9 and 1.1x10-8, respectively. Nonplanar 76

dolomites are significantly more stoichiometric than mimetic dolomites at 95% confidence (� = 0.05) with a p-value of 0.005. T-tests also show that mimetic dolomites have the lowest average degree of cation ordering (0.57 ± 0.06), compared to that of planar-e (0.77 ± 0.07) and nonplanar (0.71 ± 0.08) dolomites (t-test, � = 0.01; p-values = 6.5x10-9 and 7.5x10-7, respectively).

Principal Component Analysis

DR Table 1. Total Variance Explained

Component Initial Eigenvalues

Total % of Variance Cumulative %

1 2.305 57.6 57.6

2 0.916 22.9 80.5

3 0.454 11.3 91.9

4 0.326 8.1 100.0

DR Table 2. Component Matrix and Factor Loadings

Component

1 2 3 4

Stoichiometry 0.774 0.492 -0.128 0.379

Cation Ordering 0.823 0.191 0.493 -0.206

d13C -0.560 0.790 -0.101 -0.228

d18O -0.846 0.113 0.429 0.296

Review of Natural Recrystallized Dolomites

DR Table 3. Analysis of Natural Recrystallized Dolomites in Literature

Authors Title Year Journal Geologic Age Burial (Shallow Dept Temp If

< 0.5 < h (°C) shallow:

Intermediate < 1 (km) possible

km < Deep) meteoric

influence

Newman Fluid-influenced 1994 GSA Bulletin Cambrian Deep 9 < 300

and Mitra deformation and

recrystallization of

dolomite at low

temperatures along a

natural fault zone,

Mountain City

window, Tennessee

Al-Aasm Chemical and 2000 Aquatic Mississippian Shallow Yes

Isotopic Constraints Geochemistry

for Recrystallization

of Sedimentary

Dolomites from the

Western Canada

Sedimentary Basin

Al-Aasm Chemical and 2000 Aquatic Mississippian Deep 4

Isotopic Constraints Geochemistry

for Recrystallization

of Sedimentary

Dolomites from the

Western Canada

Sedimentary Basin

Al-Aasm Stabilization of early- 2000 Sedimentary Mississippian Deep and formed dolomite: a Geology

Packard tale of divergence

from two

Mississippian

dolomites

Reinhold Multiple episodes of 1998 Sedimentary Jurassic Shallow 66-70 Yes

dolomitization and Geology

dolomite

recrystallization

during shallow burial

in Upper Jurassic

shelf carbonates:

eastern Swabian Alb,

southern Germany

Reinhold Multiple episodes of 1998 Sedimentary Jurassic Shallow 47 Yes

dolomitization and Geology

dolomite

recrystallization

during shallow burial

in Upper Jurassic

shelf carbonates:

eastern Swabian Alb,

southern Germany

Malone et Recrystallization of 1994 Sedimentology Miocene Deep > 80 al. dolomite: evidence

from the Monterey

Formation

(Miocene), California

Malone et Hydrothermal 1996 Journal of Miocene Deep > 80- al. dolomitization and Sedimentary 100

recrystallization of Research

dolomite breccias

from the Miocene

Monterey Formation,

Tepusquet Area,

California

Nielsen et Multiple-step 1994 Sedimentology Mississippian Deep ~3 al. recrystallization

within massive

ancient dolomite

units: an example

from the Dinantian of

Belgium

Lonnee and Pervasive 2006 AAPG Bulletin Devonian to Deep > 1 230-

Machel dolomitization with Mississippian 267

subsequent

hydrothermal

alteration in the

Clarke Lake gas field,

Middle Devonian

Slave Point

Formation, British

Columbia, Canada

Machel et Isotopic evidence for 1996 GSA Bulletin Devonian Deep 5.5- al. carbonate 6.5

cementation and

recrystallization, and

for tectonic

expulsion of fluids

into the Western

Canada Sedimentary

Basin

Touir et al. Polyphased 2009 Cretaceous Cretaceous Shallow N/A

dolomitization of a Research

shoal-rimmed

carbonate platform:

example from the

middle Turonian

Bireno dolomites of

central Tunisia

Lewchuck Late Laramide 2000 Canadian Devonian Deep 100 et al. dolomite Journal of Earth

recrystallization of Sciences

the Husky Rainbow

“A” hydrocarbon

Devonian reservoir,

northwestern

Alberta, Canada:

paleomagnetic and

geochemical

evidence

Yoo and Origin and 1998 Sedimentary Ordovician shallow Yes

Lee modification of early Geology

dolomites in cyclic

shallow platform

carbonates,

Yeongheung

Formation (middle

Ordovician), Korea

Kirmaci Dolomitization of the 2008 Sedimentary Cretaceous intermediate 0.5

late Cretaceous– Geology

Paleocene platform

carbonates, Gö lkö y

(Ordu), eastern

Pontides, NE Turkey

Ye and Dolomitization of 1993 Carbonates and Permian shallow Yes

Mazzullo Lower Permian Evaporites

platform facies,

Wichita Formation,

North Platform,

Midland Basin, Texas

Kirmaci et Multistage 2018 Marine and Jurassic- Deep > 2 170- al. dolomitization in Petroleum Cretaceous 210

Late Jurassic–Early Geology

Cretaceous platform

carbonates (Berdiga

Formation), Başoba

Yayla (Trabzon), NE

Turkey: Implications

of the generation of

magmatic arc on

dolomitization

Adam and Petrologic and 2017 Bulletin of Mississippian Deep

Al-Aasm geochemical Canadian

attributes of calcite Petroleum

cementation, Geology

dolomitization and

dolomite

recrystallization: an

example from the

Mississippian

Pekisko Formation,

west-central Alberta

Lukoczki et Early dolomitization 2020 International Triassic Deep 1.3 65 al. and partial burial Journal of Earth

recrystallization: a Sciences

case study of Middle

Triassic peritidal

dolomites in the

Villá ny Hills (SW

Hungary) using

petrography, carbon,

oxygen, strontium

and clumped isotope

data

Durocher Dolomitization and 1997 AAPG Bulletin Mississippian Deep 1 and Al- Neomorphism of

Aasm Mississippian

(Visean) Upper

Debolt Formation,

Blueberry Field,

Northeastern British

Columbia: Geologic,

Petrologic, and

Chemical Evidence

Veillard et Rock-buffered 2019 Geochmica et Miocene Shallow < 0.7 < 40 No al. recrystallization of Cosmochimica

Marion Plateau Acta

dolomites at low

temperature

evidenced by

clumped isotope

thermometry and X-

ray diffraction

analysis

Wierzbicki Burial dolomitization 2006 AAPG Bulletin Jurassic Deep ~4 et al. and dissolution of

Upper Jurassic

Abenaki platform

carbonates, Deep

Panuke reservoir,

Nova Scotia, Canada

Zempolich Geometry of 1997 AAPG Memoir Jurassic Deep and Hardie Dolomite Bodies 69

Within Deep-Water

Resedimented Oolite

of the Middle Jurassic

Vajont Limestone,

Venetian Alps, Italy:

Analogs for

Hydrocarbon

Reservoirs Created

Through Fault-

Related Burial

Dolomitization

Chung and Dolomitization of the 1997 Carbonates and Cambrian Deep 1.1

Land periplatform Evaporites

carbonate slope

deposit, the Machari

Formation (Middle to

Late Cambrian),

Korea

Saller and Partial 2011 AAPG Bulletin Pennsylvania Deep

Dickson dolomitization of a n to Permian

Pennsylvanian

limestone buildup by

hydrothermal fluids

and its effect on

reservoir quality and

performance

Adabi Multistage 2009 Carbonates and Jurassic deep 1-3.5

dolomitization of Evaporites

Upper Jurrassic

Mozduran

Formation, Kopet-

Dagh Basin, N.E. Iran

Srinivasan Determining fluid 1994 Sedimentology Cambrian Deep 2-3 et al. source and possible km

pathways during

burial dolomitization

of Maryville

Limestone

(Cambrian),

Southern

Appalachians, USA

Swennen et Multiple 2012 Sedimentology Cretaceous Deep al. dolomitization

events along the

Pozalagua Fault

(Pozalagua Quarry,

Basque–Cantabrian

Basin, Northern

Spain)

Haeri- Petrologic and 2013 Marine and Ordovician Shallow N/A

Ardakani et geochemical Petroleum al. attributes of Geology

fracture-related

dolomitization in

Ordovician

carbonates and their

spatial distribution

in southwestern

Ontario, Canada

Haeri- Petrologic and 2013 Marine and Ordovician Deep 70-

Ardakani et geochemical Petroleum 100 al. attributes of Geology

fracture-related

dolomitization in

Ordovician

carbonates and their

spatial distribution

in southwestern

Ontario, Canada

Mueller et Testing the 2020 Sedimentology Jurassic Intermediate 0.4- al. preservation 0.8

potential of early

diagenetic dolomites

as geochemical

archives

Rott and Early dolomitization 2013 Journal of Mississippian intermediate to 0.75-

Qing and recrystallization Sedimentary deep 1.4

in shallow marine Research

carbonates,

Mississipian Alida

Beds, Willistoin

Basin (Canada):

Evidence from

petrography and

isotope geochemistry

Saller et al. Facies control on 2001 Bulletin of Devonian Deep

dolomitization and Canadian

porosity in the Petroleum

Devonian Swan Hills Geology

Formation in the

Rosevear area, west-

central Alberta

Lonnee and Dolomitization and 2000 Bulletin of Devonian Intermediate to

Al-Aasm fluid evolution in the Canadian Deep

Middle Devonian Petroleum

Sulphur Point Geology

Formation, Rainbow

South Field, Alberta:

Petrographic and

geochemical

evidence

Cioppa et Correlating 2000 Sedimentary Mississippian Shallow to Yes al. paleomagnetic, Geology Intermediate

geochemical and

petrographic

evidence to date

diagenetic and fluid

flow events in the

Mississippian Turner

Valley Formation,

Moose Field, Alberta,

Canada

Caird et al. Ediacaran 2017 Sedimentary Precambrian Deep

stromatolites and Geology

intertidal

phosphorite of the

Salitre Formation,

Brazil:

Phosphogenesis

during the

Neoproterozoic

Oxygenation Event

Zheng et al. Stratigraphic and 2019 Minerals Permian Deep < 1.5 < 85

Structural Control on

Hydrothermal

Dolomitization in the

Middle Permian

Carbonates,

Southwestern

Sichuan Basin

(China)

Ambers Geochemical and 1996 Carbonates and Mississippian Deep and Petzold petrologic evidence Evaporites

of the origin and

diagenesis of a Late

Mississippian,

supratidial dolostone

Videtich Dolomitization and 1994 Carbonates and Permian Deep 3-5

H2S generation in Evaporites km

the Permian Khuff

Formation, offshore

Dubai, U.A.E.

Adams et Dolomitization by 2019 Sedimentology Triassic Shallow Yes al. hypersaline reflux

into dense

groundwaters as

revealed by vertical

trends in strontium

and oxygen isotopes:

Upper Muschelkalk,

Switzerland

Lukoczki et Multi-phase 2019 Marine and Triassic Deep 1-5 al. dolomitization and Petroleum km

recrystallization of Geology

Middle Triassic

shallow T marine–

peritidal carbonates

from the Mecsek Mts.

(SW Hungary), as

inferred from

petrography, carbon,

oxygen, strontium

and clumped isotope

data

Swart et al. Evidence for high 2016 Sedimentology Jurassic Deep 2.5

temperature and

18O-enriched fluids

in the Arab-D of the

Ghawar Field, Saudi

Arabia

Kupecz and Progressive 1994 IAS Special Pub Ordovician Shallow Yes

Land recrystallization and

stabilization of early-

stage dolomite:

Lower Ordovician

Ellenburger Group,

west Texas

Kupecz and Progressive 1994 IAS Special Pub Ordovician Deep

Land recrystallization and

stabilization of early-

stage dolomite:

Lower Ordovician

Ellenburger Group,

west Texas

Gregg et al. Early diagenetic 1992 Sedimentology Holocene shallow

recrystallization of

Holocene (<3000

years old) peritidal

dolomites, Ambergis

Cay, Belize

Hartig et al. Dolomite in Permian 2011 Journal of Permian Intermediate to

paleosols of the Sedimentary Deep

Bravo Dome CO2 Research

Field, U.S.A.: Permian

reflux followed by

late recrystallization

at elevated

temperature

Montanez Fluid-rock 1992 Journal of Ordovician Deep 100- and Read interaction history Sedimentary 190

during stabilization Research

of early dolomites,

Upper Knox Group

(Lower Ordovician),

U.S. Appilachians

Nicolaides Origin and 1995 Sedimentology Cambrian Deep

modification of

Cambrian dolomites

(Red Heart Dolomite

and Arthur Creek

Formation), Georgina

Basin, central

Australia

Vandeginst Influence of climate 2012 Journal of Jurassic shallow Yes e and John and dolomite Sedimentary

composition on Research

dedolomitization:

Insights from a

multi-proxy study in

the central Oman

mountains

Zhang et al. Formation of saddle 2009 Marine and Cambrian deep 120-

dolomites in Upper Petroleum 200

Cambrian Geology

carbonates, western

Tarim Basin

(northwest China):

Implications for

fault-related fluid

flow

Coniglio et Dolomitization and 2003 Bulletin of Silurian deep < 2 al. recrystallization of Canadian km

middle Silurian reefs Petroleum

and platformal Geology

carbonates of the

Guelph Formation,

Michigan Basin,

southwestern

Ontario

Gregg et al. Porosity evolution of 1993 Sedimentology Cambrian deep

the Cambrian

Bonneterre

Dolomite, south-

easternMissouri, USA

Soreghan et Glacioeustatic 2000 Journal of Pennsylvania deep al. transgressive reflux: Sedimentary n

Stratiform dolomite Research

in Pennsylvanian

bioherms of the

western Orogrande

Basin, New Mexico

Woo and Burial dolomitization 1996 Carbonates and Cambrian deep

Moore and dedolomitization Evaporites

of the Late Cambrain

Wagok Formation,

Yeongweol, Korea

Gaswirth et The role and impact 2007 Sedimentology Eocene shallow Yes al. of freshwater–

seawater mixing

zones in the

maturation of

regional dolomite

bodies within the

proto Floridan

Aquifer, USA

Rahimpour Dolomitization and 2010 Journal of Permo- Deep 1.4

-Bonab et anhydrite Petroleum Triassic al. precipitation in Geology

Permo-Triassic

carbonates at the

South Pars Gasfield,

offshore Iran:

Controls on reservoir

quality

Dong et al. Hydrothermal 2013 Sedimentary Ordovician Deep

alteration of Geology

dolostones in the

Lower Ordovician,

Tarim Basin, NW

China: Multiple

constraints from

petrology, isotope

geochemistry and

fluid inclusion

microthermometry

Mriheel Dolomitization of the 2000 Journal of Eocene Shallow <0.3 <60 Yes and Early Eocene Jirani Petroleum

Anketell Dolomite Formation, Geology

Gabes-Tripoli Basin,

western offshor,

Libya

Guo et al. Early dolomitization 2020 Marine and Ordovician Deep >1

and recrystallization Petroleum

of the Lower-Middle Geology

Ordovician

carbonates in

western Tarim Basin

(NW China)

Özyurt et Geochemical 2018 Canadian Jurassic- Deep 2.2- 95 al. characteristics of Journal of Earth Cretaceous 4.5

Upper Jurassic – Sciences

Lower Cretaceous

platform carbonates

in Hazine Mag ̆ ara,

Gü mü s ̧hane

(northeast Turkey):

implications for

dolomitization and

recrystallization

Geske et al. Impact of diagenesis 2012 Chemical Triassic Shallow 100-

and low grade Geology (interpreted) 350

metamorphosis on then very deep

isotope (δ26Mg,

δ13C, δ18O and

87Sr/86Sr) and

elemental (Ca, Mg,

Mn, Fe and Sr)

signatures of Triassic

sabkha dolomites

Morad et al. Diagenesis of a 1995 Sedimentology Triassic Deep <65

mixed

siliciclastic/evaporiti

c sequence of the

Middle Muschelkalk

(Middle Triassic), the

Catalan Coastal

Range, NE Spain

Meister et Dolomite formation 2013 Sedimentology Triassic Deep ~1-2 40-70 al. in the shallow seas of

the Alpine Triassic

Martin- Fault-controlled and 2015 Marine and Jurassic- intermediate 0.5- 60-80

Martin et stratabound Petroleum Cretaceous 0.75 al. dolostones in the Geology

Late Aptianeearliest

Albian Benassal

Formation (Maestrat

Basin, E Spain):

Petrology and

geochemistry

constrains

Nader et al. Reflux stratabound 2004 Sedimentology Jurassic- deep ~2 <90

dolostone and Cretaceous

hydrothermal

volcanism-associated

dolostone: a two-

stage dolomitization

model (Jurassic,

Lebanon)

Jameson Models of porosity 1994 AAPG Bulletin Pennsylvania Deep 1.3

formation and their n

impact on reservoir

description, Lisburne

Field, Prudhoe Bay,

Alask

Mazzullo Dolomitization of 1994 Carbonates and Permian Deep 1.2-

periplatform Evaporites 1.8

carbonates (Lower

Permian,

Leonardian),

Midland Basin, Texas

Al-Aasm et Dolomitization and 2009 Journal of Oligocene- Shallow Yes al. related fluid Petroleum Miocene

evolution in the Geology

Oligocene-Miocene

Asmari Formation,

Gachsaran Area, SW

Iran: Petrographic

evidence

Du et al. Genesis of Upper 2018 Marine and Cambrian- Deep

Cambrian-Lower Petroleum Ordovician

Ordovician dolomites Geology

in the Tahe Oilfield,

Tarim Basin, NW

China: Several

limitations from

petrology,

geochemistry, and

fluid inclusions

Guo et al. Multiple 2016 Marine and Cambrian- Deep

dolomitization and Petroleum Ordovician

later hydrothermal Geology

alteration on the

Upper Cambrian-

Lower Ordovician

carbonates in the

northern Tarim

Basin, China

Montanez Late Diagenetic 1994 AAPG Bulletin Ordovician Deep 2-5 80-

Dolomitization of km 160

Lower Ordovician,

Upper Knox

Carbonates: A

Record of the

Hydrodynamic

Evolution of the

Southern

Appalachian Basin

References

Goldsmith, J.R., and Graf, D.L, 1958, Structural and compositional variations in some natural

dolomites: Journal of Geology, v. 66, p. 678-693.

Horita, J., 2014, Oxygen and carbon isotope fractionation in the system dolomite-water-CO2 to

elevated temperatures: Geochimica et Cosmochimica Acta, v. 129, p. 111-124.

Lumsden, D.N., 1979, Discrepancy between thin-section and X-ray estimates of dolomite in

limestone: Journal of Sedimentary Petrology, v. 49, p. 429-436.

Petersen, S.V., Defliese, W.F., Saenger, C., Daeron, M., Huntington, K.W., John, C.M., Kelson,

J.R., Bernasconi, S.M., Colman, A.S., Kluge, T., Olack, G.A., Shauer, A.J., Bajnai, D., Bonifacie, M.,

Breitenbach, S.F.M., Fiebig, J., Fernandez, A.B., Henkes, G.A., Hodell, D., Katz, A., Kele, S.,

Lohmann, K.C., Passey, B.H., Peral, M.Y., Petrizzo, D.A., Rosenheim, B.E., Tripati, A., Venturelli,

R., Young, E.D., and Winklestern, I.Z., 2019, Effects of improved 17O correction on

interlaboratory agreement in clumped isotope calibrations, estimates of mineral-specific

offsets, and temperature dependence of acid digestion fractionation: Geochemistry,

Geophysics, Geosystems, v. 20, p. 3495-3519.

Rivers, J.M., Skeat, S., Yousif, R., Liu, C., Stanmore, E., Tai, P., and Al-Marri, S.M., 2019, The

depositional history of near-surface Qatar aquifer rocks, and its impact on matrix flow and

storage properties: Arabian Journal of Geosciences, v. 12, p. 1-33.

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., 2020, Early and pervasive dolomitization by

near-normal marine fluids: New lessons from an Eocene evaporative setting in Qatar:

Sedimentology (in press). https://doi.org/10.1111/sed.12726

CHAPTER IV

DOLOMITE DISSOLUTION: AN ALTERNATIVE DIAGENETIC PATHWAY FOR THE FORMATION OF PALYGORSKITE CLAY

Brooks H. Ryan1, Stephen E. Kaczmarek1, John M. Rivers2 1Department of Geological and Environmental Sciences, Western Michigan University, Kalamazoo, Michigan 49006, U.S.A. 2Qatar Center for Coastal Research (QCCR), ExxonMobil Research Qatar, P.O. Box 22500, Qatar Science and Technology Park-Tech, Doha, Qatar

Ryan, BH., Kaczmarek, S.E., and Rivers, J.M., 2019, Sedimentology, v. 66, p. 1803-1824. https://doi.org/10.1111/sed.12559

Abstract

Palygorskite is a fibrous, magnesium-bearing clay mineral commonly associated with Late

Mesozoic and Early Cenozoic dolomites. The presence of palygorskite is thought to be indicative of warm, alkaline fluids rich in Si, Al, and Mg. Palygorskite has been interpreted to form in peritidal diagenetic environments, either as a replacement of detrital smectite clay during a dissolution- precipitation reaction or solid-state transformation, or as a direct precipitate from solution. Despite a lack of evidence, most diagenetic studies involving these two minerals posit that dolomite and palygorskite form concurrently. Here, petrological evidence is presented from the Umm er Radhuma

Formation (Paleocene-Eocene) in the subsurface of central Qatar for an alternative pathway for palygorskite formation. The Umm er Radhuma is comprised of dolomitized subtidal to peritidal carbonate cycles that are commonly capped by cm-scale beds rich in palygorskite. Thin section, scanning electron microscopy, and elemental analyses demonstrate that palygorskite fibers formed on both the outermost surfaces of dissolved euhedral dolomite crystals and within partially to completely dissolved dolomite crystal cores. These observations suggest that dolomite and palygorskite formed sequentially, and support a model by which the release of Mg2+ ions and the buffering of solution pH during dolomite dissolution promote the formation of palygorskite. This new

diagenetic model explains the co-occurrence of palygorskite and dolomite in the rock record, and provides valuable insight into the specific diagenetic conditions under which these minerals may form.

Introduction

Palygorskite, with an ideal composition of (Mg,Al)2Si4O10(OH)•4(H2O), is a magnesium-rich monoclinic and orthorhombic clay mineral with a characteristic fibrous habit (Weaver & Beck 1977;

Singer, 1979, 1984, 2002; Galan, 1996; Galan & Carretero, 1999; Guggenheim & Krekeler, 2011;

Murray et al., 2011). Palygorskite is observed in a wide variety of sedimentary deposits worldwide

(Isphording, 1973; Callen, 1977; Callen, 1984; Weaver & Beck, 1977; Torres-Ruiz et al., 1994; Akbulut

& Kadir, 2003), with the largest deposits located within the Middle Tertiary sediments of China,

Senegal, Spain, Turkey, Ukraine, and the southeastern United States (Krekeler, 2004; Krekeler et al.,

2004; Garcia-Romero et al., 2007; Murray et al., 2011; Yeniyol, 2012; Yeniyol, 2014; Kadir et al., 2016;

Kadir et al., 2017). The mineral also serves as an important constituent in soils in arid to semi-arid environments (Singer & Norrish, 1974; Singer, 1984; Singer, 2002).

Palygorskite has been reported in environments of marine, lacustrine, and continental origin

(e.g., Weaver, 1984; Botha & Hughes, 1992; Singer, 2002), but there is little agreement about the controls, timing, and mechanism involved with its formation. Multiple genetic interpretations have been offered to explain the presence of palygorskite in the rock record. These include palygorskite as a primary detrital mineral (Singer & Amiel, 1974; Singer, 1979), a diagenetic replacement of detrital smectite (Weaver & Beck, 1977; Chen et al., 2004; Xie et al., 2013), and a direct precipitate from aqueous solutions (Weaver, 1975; Shadfan et al., 1985a; Çağatay, 1990; Botha & Hughes, 1992;

Holail and Al-Hajari, 1997; Xie et al., 2013). Efforts to understand the fundamental controls on palygorskite formation have, however, been limited by an inability to synthesize it in the laboratory at earth surface conditions (Singer, 1979; 2002). Thermodynamic considerations coupled with its

non-uniform distribution in the rock record suggest that specific physicochemical factors promote palygorskite formation, including warm schizohaline fluids with pH between 7.7 and 9, and high

Mg/Al ratios (Singer & Norrish, 1974; Singer, 1979; Birsoy, 2002; Singer, 2002).

Numerous case studies have documented the co-occurrence of dolomite and palygorskite in the rock record (Weaver, 1975; Weaver & Beck, 1977; Shadfan et al., 1985a; Shadfan et al., 1985b;

Çağatay, 1988; Çağatay, 1990; Ingles & Anadon, 1991; Verrecchia & Le-Coustumer, 1996; Holail & Al-

Hajari, 1997; Krekeler et al., 2007; Xie et al., 2013; Draidia et al., 2016; Kadir et al., 2017). Despite a lack of compelling petrological evidence, a number of models to explain the genesis of palygorskite in association with dolomite have been proposed. The presence of dolomite, for example, has been cited as evidence for elevated Mg concentrations in solution (Singer, 1984; Çağatay, 1990; Xie et al.,

2013), the inference being that both dolomite and palygorskite form concurrently under such conditions (Weaver, 1975; Weaver & Beck, 1977). However, large scale co-precipitation of these minerals while in competition for Mg2+ ions is geochemically unfavorable because the thermodynamic stability of both minerals largely depends on magnesium concentrations in solution.

Alternatively, studies have hypothesized that dolomite dissolution could provide the requisite Mg for palygorskite formation (Ingles & Anadon, 1991; Verrecchia & Le-Coustumer, 1996; Holail & Al-Hajari,

1997). Thermodynamic stability diagrams, however, indicate that both minerals are concurrently stable (or unstable) in similar environments (Birsoy, 2002).

As neither mineral has been successfully synthesized in the laboratory at earth-surface temperatures and pressures (see reviews by Singer, 1979; Kaczmarek et al., 2017), understanding the co-occurrence of palygorskite and dolomite in nature is intriguing. The diagenetic relationship between these two minerals may also have important industrial applications because dolomitization generally improves the quality of hydrocarbon reservoirs (Lucia, 2004) whereas precipitation of clays generally has a negative impact (Aksu et al., 2015). In addition, understanding the geochemical constraints on palygorskite formation may be useful when reconstructing paleoenvironments. Thus,

the present study investigates the co-occurrence of palygorskite and dolomite in the Umm er

Radhuma Formation (UER), Qatar. The petrological data presented provide evidence of palygorskite formation as a direct result of, and concurrent with, dolomite dissolution. The data provided here supports previous studies that have speculated that palygorskite can form as a result of dolomite dissolution (Ingles & Anadon, 1991; Verrecchia & Le-Coustumer, 1996; Holail & Al-Hajari, 1997).

Geologic Background

Qatar can be described as an elliptical structural dome that formed in association with the

North-South striking Qatar Arch (Fig. 1; Cavelier, 1970). During the Late Paleocene-Early Eocene, the area was structurally quiescent (Cavelier, 1970). Semi-restricted, inner ramp limestones were deposited across much of the Arabian Shelf (Powers et al., 1966). These limestones and the associated shales and evaporites are known regionally as the Umm er Radhuma Formation, and are as much as

370 m thick in the subsurface of Qatar (Powers et al., 1966; Cavelier, 1970, Eccleston et al., 1981). Al-

Saad (2003) interpreted the Late Paleocene-Early Eocene as being part of a highstand systems tract

Persian Gulf

Qatar

Well RR-01

Saudi Arabia

Kilometres

Figure 1: Map of Qatar and surrounding region. Location of Well RR-01 outlined by red circle.

100

across a low relief basin. Like many of the limestones across much of the Arabian Peninsula, the Umm er Radhuma has undergone extensive dolomitization (Powers et al., 1966). Other than these early studies, little work has been done on the sedimentologic and diagenetic history of the Umm er

Radhuma in Qatar. Thus, the petrographic observations described here provide a first attempt to unravel the diagenetic history of the Umm er Radhuma in Qatar.

Materials and Methods

All data acquired for this study were extracted from the RR-01 well, located in central Qatar

(Fig. 1). The RR-01 well, the first of three research bore holes drilled by ExxonMobil Research Qatar to investigate the shallow subsurface geology of the country, was drilled in 2016. Approximately 135 m of core was recovered, including material from the upper Umm er Radhuma, Rus, and Dammam

Formations (Fig. 2). A total of 146 cylindrical core plugs (~2.5 cm diameter, ~3.8 cm long) were extracted at 0.02 - 2 m vertical spacing throughout the Umm er Radhuma, the interval of interest for this study. Standard, blue epoxy-impregnated petrographic thin sections were made from the core plugs. High-resolution thin section scans were acquired by Core Laboratories (U.A.E.).

Powders taken from each core plug were analyzed using standard powder X-ray diffraction

(XRD) techniques. Mineral percentages for each sample were quantified by Core Laboratories using the following procedure. Each sample was dispersed in a dilute sodium hexa-meta phosphate solution and centrifugally size fractioned to isolate clay-sized (<2-4 µm) particles. Air-dried mounts were measured before and after exposure to ethylene glycol at 60° C for a minimum of 2 hours. The samples were analyzed with a Panalytical automated powder diffractometer equipped with a 40 kV copper source, an X’celerator linear detector using Real Time Multiple Strip Technology (RTMS), a

101

Mineralogy Period Epoch Formation Depth (m) 0

Dammam

Middle

Rus 40

Eocene

Early

Palaeogene Umm er 100 Radhuma

120

Pal. Late

140

Calcite Dolomite Gypsum Clays and Quartz Figure 2: Stratigraphic column for Well RR-01. Percent mineralogy is based on the X-ray diffraction (XRD) procedures of Core Laboratories (United Arab Emi- rates). Palaeocene–Eocene boundary for Umm er Rad- huma Formation was interpreted by comparison of biostratigraphy and d13C values with those of Pollitt et al. (2012). Age boundaries for the Rus and Dam- mam formations are modified from Whittle et al. (1996), Holail et al. (2005) and Al-Saad (2005).

Ni -Filter, and a Graphite monochromator. Bulk samples were scanned over a 2θ range of 4° to 70° at a rate of 4.2°/minute. Clay samples were scanned over a 2θ range of 2.5° to 40° at a rate of

6.4°/minute. Determinations of individual clay percentages were done utilizing integrated peak areas and empirical reference intensity ratio (RIR) factors. Mineralogy for a select subset of core samples was confirmed with standard powder XRD techniques employed at Western Michigan

102

University using a Bruker D2 Phaser Diffractometer with a CuKα anode. Powders were extracted using an electric rotary drill, crushed by hand using an agate mortar and pestle, and mounted on a

Boron-doped silicon P-type zero background diffraction plates. XRD scan conditions include a 2θ range of 5° to 50°, a step size of 0.004°, and a count time of 2.0 s per step.

Scanning electron microscope (SEM) analysis was conducted on samples with both palygorskite and dolomite as determined by XRD and thin section petrography. SEM photomicrographs were collected on a JEOL JSM-IT100 InTouchScope. Rock samples were prepared by breaking small pieces from core plugs with a rock hammer. Samples were cleaned with compressed air to remove particulate matter, mounted on aluminum stubs with electrically conductive carbon tape, and coated with a thin (~20 nm) layer of carbon. All rock samples were analyzed at an accelerating voltage of 20 kV. Working distance (8-12 mm) and probe current (50-70 eV) were varied as needed to achieve optimal imaging results. SEM energy-dispersive spectroscopy

(SEM-EDS) was used in conjunction with XRD to determine elemental composition at the µm-scale for both coated rock samples and uncoated thin sections. One highly polished thin section and two standard (non-polished) thin sections were analyzed using EDS in backscatter mode at 15 kV - 20 kV.

More than 300 points were analyzed on dolomite cores and rims from 18 crystals in a highly polished thin section from 78.19 m, while more than 250 points were taken from 10 crystals in standard thin sections at 86.08 and 89.55 m depth.

Results

Sedimentological description

The Umm er Radhuma is dominated by sedimentary sequences comprised of carbonate and clay-rich intervals. These intervals are defined based on the mineralogy data (Fig. 2) and core observations. Carbonate intervals range from 5 to 25 m thick, and are characterized by vuggy dolomite that displays preferential dissolution of relatively large (0.5 – 2.0 mm) benthic foraminifera

103

(Fig. 3A), and partial mimetic replacement of allochems, such as crinoids (Fig. 3B) and foraminifera

(Fig. 3C). These features suggest that the precursor limestone was deposited in a subtidal to peritidal A E

Crinoids

B Dolomite Foram moulds

Higher clay Burrows content D

5 cm Dolomite

Figure 3: Photographs showing some of the common sedimentological features observed in the Umm er Radhuma. (A) to (D) High-resolution, plane polarized light (PPL) thin section images. (A) Relatively large (ca 1 mm) vugs formed as a result of preferential dissolution of foraminifera. Depth of 122.25 m; scale bar = 1 mm. (B) Partially and mimetically dolomitized echinoderms. Interparticle palygorskite cement (green) fills the space between echinoderms and dolomite matrix. Depth of 100.55 m; scale bar = 100 µm. (C) Mimetically dolomitized large benthic foraminifera surrounded by a dolomite matrix and palygorskite cement. From 96.85 m depth; scale bar = 200 µm. (D) Mimetically dolomitized miliolid foraminifera (grey) within pore-occluding gypsum cement (white). Depth of 66.48 m; scale bar = 100 µm. (E) Core photograph of the transition from a subtidal vuggy dolomite to a burrow- dominated clay-rich cap and back to a subtidal vuggy dolomite. A core plug was acquired at 122.94 m with the thin section image presented in Fig. 5A.

104

setting. The abundance of mimetically replaced miliolid foraminifera increases toward the top of the

Umm er Radhuma (Fig. 3D), which suggests an overall shallowing upward trend. This is further supported by an increased presence of gypsum in the uppermost part of the formation (Fig. 2; 3D).

In contrast to the carbonate intervals, the clay-rich intervals are generally much thinner, ranging from 0.05 – 5.0 m thick. These intervals are characterized as condensed argillaceous carbonate intervals, containing anywhere up to 1 – 26.5% clay minerals (Fig. 2). These clay-rich intervals commonly contain sedimentary structures, such as extensive burrows (Fig. 3E), which indicate a general shallowing from the underlying carbonate interval and are interpreted to represent depositional cycle caps. Multiple cycles, such as the one just described and depicted in

Figure 3E, are identified throughout the Umm er Radhuma, with varying concentrations of clay material within the cycle caps (Fig. 2).

XRD mineralogy

As shown in Figure 2, the mineralogy of the Umm er Radhuma Formation is dominated by dolomite. Also present in significant amounts are gypsum (0-22%), and a variety of silicate minerals.

The abundance of clays and quartz generally constitutes ≤ 26.5% and ≤ 5% of the bulk rock, respectively. In six of the twelve core plugs examined with SEM-EDS, palygorskite makes up > 90% of the total clay fraction in the rock. In the remaining six core plugs, palygorskite constitutes ≥ 60% of the clays, with the remaining fraction consisting mainly of illite or sepiolite. Figure 4 is a representative XRD diffractogram from the Umm er Radhuma showing a mixture of ~76% dolomite and ~20% palygorskite. The most intense XRD peak on the diffractogram is the (104) dolomite reflection at 31.0° 2θ. Dolomite stoichiometry (i.e., mole% MgCO3) was determined using the d- spacing of the (104) reflection based on the empirical equation derived by Lumsden (1979).

Stoichiometry ranges from 49.8 to 51.1 mole% MgCO3 (x = 50.5%; s = 0.44%). Additionally, the (104) dolomite reflection peaks display signs of asymmetry. The presence of the (101), (015), and (021)

105

Dol(104)

Qtz(101) Dol(113)

Pal(110) *Dol(015) Dol(202) Dol(006) Qtz(100) Dol(110) Dol(012) *Dol(101) *Dol(021) Relative Intensity (log scale)

5 10 15 20 25 30 35 40 45 50 Degrees 2 Figure 4: Powder X-ray diffraction pattern of a representative sample (110.45 m depth) containing ca 76% dolomite and ca 20% palygorskite from 5° to 55° 2q. The dolomite (104) peak is the dominant dolomite peak observed, and dolomite ordering reflection peaks [(101), (015) and (021)] are denoted by an asterisk. Palygorskite is identified by the presence of the dominant (110) peak, and quartz is identified by the (100) and (101) peaks. Dol, dolomite; Pal, palygorskite; Qtz, quartz. Bottom line represents background. reflections at 22.0, 35.3, and 43.8° 2θ, respectively, suggest that the dolomite is ordered (Gregg et al.,

2015; Kaczmarek et al., 2017). The degree of cation ordering ranges from 0.36 to 0.95 (x = 0.72; s =

0.14). The principal XRD peak for palygorskite is the (110) reflection identified at 8.5° 2θ, which is consistent with palygorskite exhibiting a d-spacing of 10.4 Å (VanScoyoc et al. 1979; Chisholm, 1992;

Garcia-Romero et al., 2007; Post & Heaney, 2008).

Thin section petrography

Thin section analysis was used to characterize the various mineral phases and establish paragenetic relationships between dolomite and palygorskite (Fig. 5). Emphasis was given to core samples located between 80 and 123 m, where palygorskite is most abundant based on XRD (Fig. 2).

Dolomite in this depth range is characterized by planar-e and planar-s crystals, with diameters ranging 30 - 100 µm (Fig. 5). Clear, inclusion-poor dolomite rims are commonly observed in association with cloudy euhedral dolomite cores (Fig. 5D and 5E). Cloudy dolomite cores lack calcite inclusions, and in many instances are partially to completely dissolved (Fig. 5A-F). Below 123 m,

106

dolomites are characterized by extremely coarse (100-200+ µm) anhedral to planar-s mosaics. Above

80 m, fine (<10 µm), fabric-preserving mimetic dolomites are observed (Fig. 3D). Throughout the

UER, palygorskite is identified in thin section by its light green color. Palygorskite occurs as a pore- occluding, intercrystalline cement hosted between dolomite rhombs (Figure 5). Palygorskite also commonly occurs on the surfaces of partially dissolved dolomite crystals (Figs. 5A and 5F), and within partially to completely dissolved dolomite cores (Figs. 5B, 5D, and 5E).

SEM-EDS analysis

SEM micrographs show the presence of palygorskite fibers anchored to the crystal faces of partially dissolved dolomite rhombs as well as palygorskite clusters nested within partially to completely dissolved dolomite cores (Figs. 6, 7, and 8). As shown in Figure 6, partially corroded planar-e dolomite crystals are surrounded by a matrix of palygorskite fibers. Higher magnification

SEM images show fine palygorskite fibers occupying dissolution voids on partially corroded euhedral dolomite crystal faces (Figs. 6C and 6D). Elemental compositions of these features (Fig. 7) are consistent with dolomite and palygorskite. Dolomites with cloudy core fabrics exhibit a partially dissolved rhombohedral core filled with palygorskite fibers (Figs. 8A and 8B), and less frequently dolomite cores are completely dissolved and filled with palygorskite fibers (Figs. 8C and 8D). The curved shape of palygorskite fibers near their end is likely an artifact resulting from dehydration and electron beam damage (Krekeler & Guggenheim, 2008).

High-resolution EDS spot data show that dolomite rims are generally more stoichiometric than dolomite cores. Data from the highly polished thin section is associated with a significantly lower uncertainty than the data from the standard petrographic thin sections. The polished sample

(78.19 m) yields an average Mg/Ca ratio of 0.88 (s = 0.05) for the cores and an average Mg/Ca ratio of 0.92 (s = 0.04) for the rims (Fig. 9). The standard thin section from 86.06 yields average Mg/Ca

107

ratios of 0.78 (s = 0.18) and 0.85 (s = 0.19) for cores and rims, respectively. The standard thin section from 89.55 m yields average Mg/Ca ratios of 0.77 (s = 0.17) and 0.85 (s = 0.16), respectively.

A B

C D

E F

Figure 5: High-resolution, plane polarized light (PPL) thin section images of dolomite (light brown to light grey rhombs) and palygorskite (green haze) from the Umm er Radhuma. (A) Depth of 122.94 m. Planar-e and planar-s dolomite with intracrystalline (cores) and intercrystalline porosity filled with palygorskite. (B) Depth of 120.67 m. Similar dolomite as shown in (A), but exhibits an inner dissolved core and intact rim. Palygorskite is observed as an intercrystalline cement, as well as intracrystalline cement in the partially dissolved cores. (C) Depth of 110.45 m. Euhedral dolomite with intercrystalline palygorskite cement. (D) Depth of 99.10 m. Cloudy-core clear- rim dolomite crystals displaying partially dissolved cores with palygorskite filling as intracrystalline cement. Palygorskite is also observed as an intercrystalline cement between dolomite crystals. (E) Depth of 89.55 m. Palygorskite is observed as cement between dolomite crystals. Dissolved dolomite cores are common and palygorskite is observed within the dissolved cores. (F) Depth of 79.77 m. Euhedral dolomite crystals displaying partial dissolution of crystal faces, in which palygorskite is observed. Rare cloudy- core clear-rim dolomites with dissolved cores are also observed. Scale bar in all images represents 100 µm.

108

A B

10 µm 10 µm C D

5 µm 5 µm E F

10 µm 10 µm

Figure 6: Scanning electron microphotograph (SEM) images of dolomite and palygorskite from well depth of 110.45 m (A) to (D) and 93.98 m (E) and (F). (A) Euhedral dolomite crystal engulfed in palygorskite. Dissolution features are observed on the dolomite crystal face. (B) Euhedral dolomite crystal surround by palygorskite. Note the major dissolution feature in the top left quadrant of the crystal. (C) Magnified image of top left corner of dolomite crystal in (A). Fibrous palygorskite crystals are observed extending perpendicular from the partially dissolved dolomite face. (D) Magnified image of the dissolution feature observed in the upper left corner of the dolomite crystal face in (C). Note the palygorskite fibres forming in and on the dissolved surface. (E) Extremely corroded dolomite crystal displaying palygorskite fibre nucleations on the crystal surface. Palygorskite is also observed as intercrystalline cement coating the dolomite crystal. (F) Partially dissolved euhedral dolomite crystal with associated palygorskite on the crystal face and within intercrystalline pores.

109

001 002 5 0 .0 30.0

] C

3 4 0 .0 ]

+ 25.0 C C a 3 + E . E 3 0 .0 . 20.0 x1 [ x1 [ s S i M g s t M g C a t 15.0 2 0 .0 un O un S 10.0 S Co Co O A l S i C a 1 0 .0 S C a S 5 .0

0 .0 0 .0 0 .0 0 0 .5 0 1 .0 0 1 .5 0 2 .0 0 2 .5 0 3 .0 0 3 .5 0 4 .0 0 4 .5 0 5 .0 0 0 .0 0 0 .5 0 1 .0 0 1 .5 0 2 .0 0 2 .5 0 3 .0 0 3 .5 0 4 .0 0 4 .5 0 5 .0 0 ke V ke V

003 1 0 0 .0 ]

3 S i + 8 0 .0 O E . 1 x

[ 6 0 .0

s A l t

n C u 4 0 .0 M g S

Co C a 2 0 .0 S K K C a

0 .0 10 µm 0 .0 0 0 .5 0 1 .0 0 1 .5 0 2 .0 0 2 .5 0 3 .0 0 3 .5 0 4 .0 0 4 .5 0 5 .0 0 ke V

001 002 003

5 µm 5 µm 2 µm Figure 7: Electron dispersive analysis of the same dolomite crystal as Fig. 6A and C. Point 001 was taken in the same area as Fig. 6B. The elements identified are indicative of both dolomite (Ca, Mg, C and O) and palygorskite (Mg, Al, Si and O). Point 002 was taken in the centre of the partially dissolved dolomite face. The elements identified are characteristic of dolomite, with relatively minor counts of Si and S. Note the nm-sized palygorskite crystals nucleating on a partially dissolved dolomite face. Point 003 was taken in the surrounding palygorskite matrix. The elements identified are representative of palygorskite, with relatively minor counts of C, S, K and Ca. Note that energy-dispersive spectroscopy graphs may display two peaks for the same element. All grey peaks represent counts of K-alpha X-rays, while the coloured peaks represent counts of K-beta X-rays.

Discussion

Arabian Peninsula dolomite & palygorskite co-occurrence

Numerous regional studies have documented the co-occurrence of palygorskite and dolomite in Eocene carbonates on the Arabian Peninsula (Shadfan et al., 1985a; Shadfan et al., 1985b; Holail &

Al-Hajari, 1997; Çağatay, 1988; Çağatay, 1990; Draidia et al., 2016). Although a number of genetic models have been offered to explain this, there is a general lack of conclusive petrological evidence.

Shadfan et al. (1985a), for example, reported on the co-occurrence of palygorskite and dolomite in the carbonates of the Paleocene-Lower Eocene Umm er Radhuma Formation and the Lower-Middle

Eocene Dammam Formation of eastern Saudi Arabia. Palygorskite fibers were reported to exhibit

110

A B

10 µm 5 µm

C D

5 µm 2 µm

Figure 8: Scanning electron microphotograph (SEM) of dolomite and palygorskite from well depth of 89.55 m. (A) The dolomite crystal is surrounded by palygorskite crystals. Within the dolomite rim is a partially dissolved dolomite rhomb, similar to observations in Fig. 5D and E. (B) Similar observation to that in (A), displaying a cloudy-core clear-rim dolomite crystal with palygorskite fibres coating the outer surface of the rim and nucleating on the partially dissolved core. (C) A similar dolomite crystal to that in (A), but it is densely covered by palygorskite fibres and displays a more soluble rhombic core. (D) Higher magnification of the void left by the core in (C). The inner rhombohedral core of this dolomite crystal is observed to be completely dissolved and filled with palygorskite. aligned orientations in the Umm er Rhaduma, while in the Dammam, palygorskite fibers displayed uniform sizes and shapes (Shadfan et al., 1985). These observations and an absence of other clays were cited by Shadfan et al. (1985a) as evidence for an authigenic, rather than a detrital origin for the palygorskite. The authors suggested that the pH of the formation water could have increased in response to shallow water conditions and higher than normal temperatures during the Early Eocene, which may have caused an increase in photosynthetic uptake of CO2. As a consequence of increased pH, it was argued, the solubility of silica would have also increased, thus providing a source of dissolved silica for palygorskite formation (Shadfan et al., 1985a). The authors did not, however, propose a source of the Mg necessary for palygorskite, nor did they provide an explanation for the

111

co-occurrence of dolomite and palygorskite. They commented only that the presence of dolomite suggested a marine origin for palygorskite (Shadfan et al., 1985a).

Çağatay (1990) investigated carbonates of the Eocene Rus and Dammam Formations in Saudi

Arabia. Palygorskite was the most abundant clay mineral, with illite and smectite also present.

Clusters of palygorskite fibers were described as blankets surrounding diagenetic dolomite crystals and detrital anatase and feldspar. The fibers were also present as intercrystalline, pore-occluding cements within the dolomitized intervals. The fine, delicate nature of the fibers and the fibrous coatings, coupled with preservation of fine laminae in the shales were interpreted by the authors as evidence to preclude a detrital origin for the palygorskite. Instead, Çağatay (1990) interpreted the observed textures as resulting from direct precipitation from Mg-rich solutions, a hypothesis consistent with textural, sedimentological, and fossil evidence suggesting deposition occurred in a semi-restricted, shizohaline, peripheral basin. It was argued by the author that such an environment would have likely experienced an arid to semi-arid climate, reduced salinity, and an elevated pH, all conditions that are ideal for primary palygorskite formation (Çağatay, 1990).

Based on the observation that palygorskite formed cements around dolomite crystals,

Çağatay (1990) concluded that palygorskite must have precipitated after dolomitization. Çağatay

(1990) further contended that the presence of dolomite is evidence that Mg concentrations in the pore fluids were high enough to support subsequent palygorskite precipitation, with Si and Al supplies coming from the dissolution of detrital silicate minerals under alkaline conditions. In his model, Çağatay (1990) postulated that marine-derived pore fluids could provide the necessary Mg to form dolomite and palygorskite in sequence.

A study of Eocene, palygorskite-bearing carbonates of Qatar by Holail & Al-Hajari (1997) showed that the Simsima Member of the Dammam Formation contained dolomitized shallow-marine limestone and 4% clay minerals, which are mainly palygorskite. SEM observations revealed bundles of palygorskite fibers coating dolomite crystals, suggesting that palygorskite postdates

112

dolomitization. Holail & Al Hajari (1997) also carried out geochemical analyses including X-ray fluorescence and stable oxygen isotopes to further understand the nature and timing of the palygorskite relative to the dolomite. Elemental analysis revealed an excess of CaO in the palygorskite, interpreted as Ca-carbonate inclusions within the clay. Oxygen isotope values of the palygorskite averaged +18.7‰ SMOW, while the dolomites range from +1.1 to +4.7‰ SMOW suggesting that the palygorskite coatings formed from different waters than the dolomite crystals

(Holail & Al-Hajari 1997). No discussion on the difference in oxygen fractionation between palygorskite and dolomite was offered, however. The depositional environment of the Simisma during Eocene time is interpreted by Holail & Al-Hajari (1997) as a shallow marine environment with elevated salinity and an arid climate. Such an environment, it was posited, would have provided ideal conditions for direct precipitation of palygorskite. Holail & Al-Hajari (1997) suggested that their textural and geochemical observations support a model in which palygorskite formed after the dolomite, and that the environment was more suitable for primary precipitation of palygorskite, as opposed to formation via alteration of smectite or deposition as a detrital mineral.

The model of Holail & Al-Hajari (1997) differs from that of Çağatay (1990), in that it suggests that dissolving dolomite provides the Mg reactants for the precipitating palygorskite, with the necessary Si and Al derived from the dissolution of phyllosilicates and an influx of meteoric water

(Holail & Al-Hajari 1997). While it is logical that dissolving dolomite could contribute Mg to the pore fluids for palygorskite precipitation, no evidence of dolomite dissolution was provided. Further, a scenario whereby large-scale dissolution of clays provided an aqueous medium of dissolved Si and

Al rather than a model whereby small-scale dissolution-precipitation reactions are occurring on precursor clays is unlikely. Although the authors report an absence of smectite, it is possible that it had been replaced by palygorskite via a dissolution-precipitation reaction, rather than palygorskite forming as a direct precipitate. This would be analogous to completely dolomitized limestones, which can be 100% dolomite and 0% calcite (Sperber et al., 1984)

113

Serial Formation of Dolomite & Palygorskite

The dolomite and palygorskite textures reported here differ from those reported in previous studies. Weaver & Beck (1977), for example, observed thin disks of palygorskite, and suggested that the disks form during a co-precipitation reaction between montmorillonite clay and dolomite. As evidence for their model, Weaver (1975) and Weaver & Beck (1977) cited the SEM observed texture of the dolomites, which they described as external plates and hollow cores. They attributed this texture to either competitive growth between dolomite and palygorskite or resulting from what they cited as “the normal way dolomites form.” Re-examination of their published images shows remarkable similarity with the hollow dolomite textures observed in the present study, which are interpreted here as partially dissolved cores.

The dolomite crystals in Figures 5 and 8 are similar in appearance to cloudy-center clear-rim

(CCCR) dolomites (Sibley, 1982), in which the inner rhombohedral cores are inferred to be less stable than the surrounding clear dolomite rim (Folk & Siedlecka, 1974). Sibley (1982) argued that cloudy cores are typically less stable because they commonly contain a high density of calcite inclusions, and as a result, the cores generally dissolve or get filled by calcite during subsequent stabilization. This model is consistent with our observations of the dolomites in the UER, which show dissolved cores and pristine rims. No evidence of calcite inclusions is observed, however.

An alternative model, in which the stable rims form at the expense of the less stable cores, is consistent with the stoichiometry and cation ordering observations in synthetic dolomite experiments carried out by Kaczmarek & Sibley (2011; 2014). Kaczmarek & Sibley (2014) demonstrated that stoichiometry and cation ordering increase during recrystallization of the initial dolomite phase. Cation ordering refers to the degree of ordering of Ca and Mg in the dolomite crystal lattice (Goldsmith & Graf, 1958). Dolomites with a higher degree of cation ordering have a lower free energy of formation than dolomites with a lower or no degree of cation ordering, and thus are more

114

stable and therefore less soluble (Navrotsky & Capobianco, 1987). The CCCR fabrics observed in

Figures 5 and 8 are consistent with a process whereby the dolomite cores are formed during initial replacement of the limestone precursor, and contain less stable (i.e. poorly-ordered and/or Ca-rich) dolomite, whereas the clear rims are a more stable (well-ordered and/or stoichiometric) product of recrystallization (Kaczmarek & Sibley, 2014). This interpretation is supported by high-resolution

EDS data, which show that dolomite cores generally are less stoichiometric than the dolomite rims

(Fig. 9). These data are also consistent with CCCR fabrics in other dolomite studies (Sibley, 1982;

Jones, 2004; Jones, 2007). Powder XRD patterns of these samples are also characterized by asymmetric peaks for the principle dolomite reflections [(104), (015), and (110)], which further support the hypothesis that a range of dolomite Mg/Ca compositions are present. It is likely that this is a result of a combination of less stoichiometric (Ca-rich) and poorly-ordered dolomite cores, as well as more stoichiometric and well-ordered dolomite rims.

A B 100 90

80 Dolomite Cores Dolomite Rims 70 n = 158 n = 155 mean = 0.88 mean = 0.92

y 60 = 0.05 = 0.04 c n e 50 qu e r 40 F

0 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 1.1 20 µm Mg/Ca Figure 9: (A) Backscatter SEM photomicrograph showing the locations where energy-dispersive spectroscopy (EDS) spot analyses were collected on the highly polished thin section (78.19 m). The contrast of the photomicrograph was adjusted to accentuate crystal boundaries and the epoxy. White dots represent dolomite rim locations while black spots represent core locations where EDS data were acquired. (B) Frequency histogram showing the compiled results of the EDS spot analyses from 313 points. The data show that dolomite cores are generally less stoichiometric than the corresponding rims.

It is possible, however, that the clear dolomite rims formed as overgrowths on the cores during a second dolomitization event. Although this model is appealing in that the ions released during later dissolution of the core could be used entirely for precipitation of palygorskite, it is unlikely as it implies that the clear dolomite rims precipitated directly from solution. Because the dolomite cores must have formed before the rim, there would have been a limited supply of 115

carbonate reactants for later rim formation, assuming near complete dolomitization during the first event (Sperber et al., 1984; Sibley et al., 1987). As a result, the clear rims would have formed as primary precipitants, rather than as a replacement of a carbonate precursor. It is more likely based on the observations presented that the clear rims formed as a replacement of the cloudy cores during recrystallization, rather than as primary precipitant overgrowths. This model is more consistent with experimental findings of Kaczmarek & Sibley (2014), which show that formation of more ordered, and thus more stable, dolomite happens during recrystallization only after the vast majority of the

Ca-carbonate precursor has been replaced by the initial dolomite phase. It is worth noting that while core dissolution is not necessary or ubiquitous for the formation of clear rims (Rivers et al., 2012), it is most consistent with the observations presented here.

Thin section (Fig. 5) and SEM images (Figs. 6 and 8) provide textural evidence that palygorskite fibers formed during dissolution of less stable dolomite phases after dolomitization and dolomite recrystallization. Palygorskite fibers in the UER commonly occur as an irregular coating around dolomite rhombohedra (Fig. 6), as clusters inside dissolved dolomite cores (Fig. 8), and on the crystal faces of partially dissolved euhedral dolomite (Figs. 6C and 6D). Figures 8A and 8B show palygorskite fibers within a partially dissolved dolomite core, suggesting that precipitation of palygorskite was concurrent or post-dated dissolution of the dolomite core. Figure 8C further shows a completely dissolved dolomite core filled with palygorskite fibers, suggesting that palygorskite formed after dolomite dissolution. This model is attractive in that the Mg released during dissolution of the dolomite cores could be used for penecontemporaneous precipitation of palygorskite.

The genetic model presented in Figure 10 illustrates the total diagenetic process from initial dolomitization, to recrystallization, to palygorskite nucleation, and finally palygorskite growth and filling of the dissolved dolomite core. First, a small dolomite core is formed during the initial replacement stage of dolomitization. Second, the clear rim forms during a recrystallization stage concurrent with partial dissolution of the relatively poorly-ordered, Ca-rich dolomite core. Third,

116

Initial Replacement Primary Recrystallization Post-dolomitization Complete Core Dissolution Dissolution

Non-stoichiometric, poorly ordered dolomite Stoichiometric, better ordered dolomite Palygorskite fibres

Figure 10: Evolutionary model of the dolomitization reaction, growth of cloudy-centre clear-rim dolomite fabrics and post-dolomitization dissolution of dolomite with subsequent palygorskite formation. The initial replacement stage contains non-stoichiometric, relatively poorly-ordered dolomite that continuously replaces the calcite reactant until 97% depletion. Once 97% depletion is marked, the dolomitization reaction enters the primary recrystallization stage. During the primary recrystallization stage, non-stoichiometric, relatively poorly-ordered dolomite is dissolved while stoichiometric, relatively well-ordered dolomite is precipitated. It is interpreted in this figure that the clear rim is made up of stoichiometric, relatively well-ordered dolomite resulting from the dissolution of the non-stoichiometric, relatively poorly-ordered dolomite core. After the dolomitization reaction has commenced, post-dolomitization fluids introduced to the system dissolve the less stable dolomite core while the clear rim remains stable. Palygorskite forms subsequently with this post-dolomitization dissolution stage, as partial to complete dissolution of the less stable core occurs. Figure 8A represents the post-dolomitization dissolution stage, while Fig. 8C represents the final product of the model presented here. dolomite dissolution at a time after the dolomitization and recrystallization reaction results in further core dissolution (while the clear rim remains stable) and subsequent palygorskite precipitation. This process continues until the dolomite core is completely dissolved and entirely filled by palygorskite.

The paragenetic model presented is further supported by the petrographic evidence showing that palygorskite formation occurred after the clear rim dolomite growth. Assuming that the clear rims form from the redistributed material from the dissolved cores, it is possible that palygorskite and the dolomite rims could have formed coevally. However, the dolomite rims shown in Figure 8 display no evidence of palygorskite inclusions, as would be expected if the minerals formed concurrently. The palygorskite fibers are instead found only as coatings and as single fibers attached to the outer surface of the dolomite crystals or as clusters within the partially dissolved cores. This supports the contention that the palygorskite precipitated after the dolomite rims were fully formed.

This also suggests that the dolomite cores are still susceptible to chemical reactions with pore fluids

117

after the growth of the clear rims, as put forth by Sibley (1982), and supports the idea that subsequent formation of palygorskite can occur within the dolomite cores from later fluids undersaturated with respect to the less stable dolomite phase.

Geochemical relationship between dolomite & palygorskite

Geochemical arguments also support the model of palygorskite genesis by dolomite dissolution, and highlight one of the major limitations of the coeval formation model. The stability of dolomite is a function of temperature and the [Mg2+]/[Ca2+] ratio of the fluid (Carpenter, 1980;

Machel & Mountjoy, 1986; Kaczmarek & Sibley 2011; Kaczmarek & Thornton 2017; and others).

Dolomitization, the process by which calcite is replaced by dolomite, is represented by the following equation:

2CaCO3 + Mg2+ ßà CaMg(CO3)2 + Ca2+ (1)

The Gibbs Free Energy of Reaction (DGrxn) and the equilibrium constant (K) for this reaction depend on the [Ca2+]/[Mg2+] in solution. According to calculations based on this reaction, a fluid is supersaturated with respect to dolomite at normal earth surface conditions (i.e., 25° C and 1 atm) when the [Mg2+]/[Ca2+] of the aqueous solution is greater than ~1 (Sibley, 1982), although large uncertainties exist in the standard free energy of dolomite (Carpenter, 1980). Therefore, the presence of dolomite as evidence for elevated [Mg2+] in solution to form both dolomite and palygorskite is not necessarily correct. Dolomite stability depends on the [Mg2+] relative to [Ca2+], while the stability of palygorskite formation is independent of the [Ca2+] as discussed below.

Empirical observations suggest that palygorskite formation is dependent on the [Mg2+]/[Al3+] in solution (Millot, 1970; Isphording, 1973; Singer, 1979; Birsoy, 2002), among other parameters such as pH, salinity, and H4SiO4. This is evident in many sedimentary successions, whereby more

118

aluminum-rich clays such as kaolinite and smectite are found closer to the source of detrital material, and the MgO:Al2O3 ratio increases into the basin until Mg-rich clays such as palygorskite and sepiolite are observed (Millot, 1970; Isphording, 1973). In pure smectite, the octahedral Al3+/Mg2+ ratio is on the order of 7 to 9, while in pure palygorskite the octahedral Al3+/Mg2+ ratio is between 1 and 1.5, corresponding to a Mg2+/Al3+ ratio of 0.67 - 1, which is at least 4.8 times greater than that of pure smectite (Singer, 1979). This implies that much more Mg is needed relative to Al in order for palygorskite to form instead of smectite. Thus, field observations of palygorskite coupled with the thermodynamics of dolomitization suggest it is unlikely that these minerals would form coevally in

Mg-rich waters as they compete for Mg2+ ions.

Thermodynamic studies have been done to further understand the SiO2-MgO-Al-H2O system as it pertains to the stability of palygorskite in various solutions. Many studies have formulated stability diagrams with varying parameters such as pH, Mg, Al, and H4SiO4 (Singer and Norrish, 1974;

Weaver and Beck, 1975; Elprince et al., 1979; Birsoy, 2002). Under constant Al activities, palygorskite stability increases with increasing pH, Mg, and H4SiO4. Decreases or increases in one parameter, however, affect how much adjustment in the remaining two is required. For example, at relatively high pH levels (pH > 9), palygorskite is stable at low H4SiO4 if [Mg2+] is high, or at high H4SiO4 if [Mg2+] is low (Singer and Norrish, 1974). At low pH (e.g., pH < 6) palygorskite is stable only if both [Mg2+] and H4SiO4 are elevated (Fig. 11A).

Additional work to characterize the stability boundary between smectite and palygorskite in natural systems has been carried out to understand the impact in variations of pH, Mg, Al, and H4SiO4.

Initial work was done by Weaver & Beck (1977) who analyzed the montmorillonite-palygorskite system at 25° C and constant log[Al(OH)4-] of -5.5. The results demonstrated that, similar to the work of Singer and Norrish (1974), low pH values (pH < 6) require elevated [Mg2+] and H4SiO4 for palygorskite to be favored over montmorillonite (Fig. 11B). Weaver & Beck (1977) calculated that palygorskite can directly precipitate from a solution at a pH ~8 and log[H4SiO4] = -5, with increases

119

A B

-2 -3 PALYGORSKITE -4 11 -5 10 -6 -7 log[Mg2+] PALYGORSKITE SOLUTION 9 -8 pH SOLUTION -9 8 SOLUTION -4 -10 7 -3 -11 MONT. -2 log[Mg2+] 3 -12 4 6 -1 5 -6 -5 -4 -3 -2 6 7 -2.6 log[H SiO 0] pH 8 ] 4 4 9 -4.1 log[Si(OH)4

Figure 11: Stability diagrams of palygorskite as a function of pH, H4SiOH4, and Mg. (A) Three-dimensional stability diagram of palygorskite and an aqueous solution. Diagram is modified from Singer & Norrish (1974). (B) Three- dimensional stability diagram displaying the stability fields of palygorskite, montmorillonite, and an aqueous solution at 25°C and log[Al(OH)4-] = -5.5. Modified from Weaver and Beck (1977). in either parameter further favoring palygorskite stability. In contrast, decreases in log[H 4SiO4] cause both minerals to become unstable, while lower pH values favor montmorillonite (Fig. 11B). In addition to determining the stability of montmorillonite and palygorskite in solution, Weaver & Beck

(1977) also calculated the stability field boundary for the transformation between montmorillonite and palygorskite. This demonstrated that an increase in pH, [Mg2+], or H4SiO4 favors palygorskite over montmorillonite (Eq. 2).

log[Mg2+] + 2pH + 2log[H4SiO40] = 5.75 (2)

The results of Singer & Norrish (1974) & Weaver and Beck (1977) are helpful in understanding the stability of palygorskite in aqueous solutions, as well as the transformation reaction between montmorillonite and palygorskite. The fundamental control Al may have on the system is less well understood, however. Weaver & Beck (1977) suggested that transformation from

120

phyllosilicates other than smectite would require chemical modifications energetically unfavorable at low temperatures, due to higher tetrahedral Al occupancy.

Extensive and more recent work by Birsoy (2002) has shed light on the impact of pH, Mg, Al, and H4SiO4 on the stability of palygorskite and minerals commonly associated with it, with specific emphasis on varying Al concentrations. Key results of Birsoy (2002) are summarized in Figure 12.

They show that increasing Al causes an increase in the stability field of both montmorillonite and palygorskite. In all stability diagrams constructed, the solution is in equilibrium with montmorillonite and palygorskite at log[aMg2+/(aH+)2] values < ~12 and log[aAl3+/(aH+)3] £ 5.5, with the exception of those systems that contain quartz instead of amorphous silica. As Al continually increases, however, the reaction must proceed first through montmorillonite by increasing H4SiO4, and palygorskite stability is only reached through further increasing H4SiO4 or by increasing Mg (Fig.

12; Birsoy, 2002). Thus, in a fluid exhibiting relatively constant H4SiO4 concentrations, palygorskite is more stable than montmorillonite at elevated Mg/Al ratios, supporting the field observations discussed above.

As increases in either pH, Mg, or H4SiO4 favor the stability of palygorskite, how they individually affect the stability of dolomite must be assessed. Because palygorskite and dolomite are both stable at similar pH values close to seawater (~8), it is unlikely that changes in pH alone can cause both undersaturation with respect to dolomite and supersaturation with respect to palygorskite. Despite the fact that dolomite and palygorskite must compete for Mg2+ ions if they form coevally, Figure 12 clearly indicates that both minerals can theoretically form together in similar Mg- rich fluids. The stability diagram in Figure 12, however, shows that it is possible that fluctuations in

Mg could cause variations in stability between dolomite and palygorskite. The dolomite stability boundary lies at log[aMg2+/(aH+)2] = ~12, above which it is unstable, while palygorskite remains stable with increasing Mg as long as log[H4SiO4] values are greater than approximately 3. Thus, it is possible that fluids containing excessively high Mg concentrations could be responsible for causing dolomite

121

21 21 20 chlorite A 20 chlorite B 19 19 ] ] 2 chrysotile 2 ) 18 ) 18 + + chrysotile H H 17 sepiolite 17 sepiolite /(a /(a 2+ 16 2+ 16 brucite brucite palygorskite Mg Mg 15 palygorskite 15 log[a 14 log[a 14 magnesite magnesite 13 solution 13 solution 12 12 dolomite dolomite 11 11 quartz Mg-montmorillonite 10 quartz Mg-montmorillonite 10 9 9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 log[a ] log[a ] H4SiO4 H4SiO4

21 21 20 chlorite C 20 chlorite D 19 19 ] ]

2 18 chrysotile 2 18

) ) chrysotile + + H H 17 sepiolite 17 sepiolite /(a /(a 2+ 16 2+ 16 palygorskite Mg brucite Mg brucite 15 palygorskite 15

log[a 14 log[a 14 magnesite magnesite 13 solution 13 solution 12 12 dolomite dolomite 11 11 amorphous Mg-montmorillonite 10 silica Mg-montmorillonite 10 9 9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 log[a ] log[a ] H4SiO4 H4SiO4

Figure 12: Stability diagrams modified from Birsoy (2002) displaying the stability boundaries of various minerals within a seven component MgO-CaO-Al2O3-SiO2-H2O-CO2-HCl system. Each diagram is constructed at 25°C and 1bar with log[aH2O] =1, log[aCO2] = - 3.5, and log[aCa2+/(aH+)2]=13.06. A) Stability diagram constructed at log[aAl3+/(aH+)3] = 4.5 and suppression of talc, kerolite, and amorphous silica. B) Same diagram as (A), but increased Al at log[[aAl3+/(aH+)3] = 5.5. C) Same diagram as (A), but with suppression of talc, kerolite, and quartz. D) Same diagram as (B), but with suppression of talc, kerolite, and quartz.

to become unstable and palygorskite to form. With regards to H4SiO4 levels, dolomite stability is unaffected, but they must remain elevated in order for palygorskite to form. Therefore, the theoretical stability diagrams indicate that a slightly alkaline fluid that contains elevated levels of Mg and H4SiO4 may be responsible for both undersaturation with respect to dolomite and

122

supersaturation with respect to palygorskite. This causes issues, however, as such a system still falls in the dolomite stability field at log[aMg2+/(aH+)2] values < 12 (Fig. 12).

One possibility is that all three parameters (pH, Mg, H4SiOH4) did not evolve simultaneously, but were sequentially developed until the system was undersaturated with respect to dolomite and subsequently saturated with respect to palygorskite. Meteoric water is a likely candidate responsible for starting a series of reactions and has been interpreted as the source of silica or responsible solution for palygorskite formation in other localities (Singer and Norrish, 1974; Soong, 1992;

Draidia et al., 2016; Kadir et al., 2016; Kadir et al., 2017). Such a fluid could contain elevated levels of

H4SiO4 (Knauth, 1979; Bennet & Siegel, 1987) as well as high pCO2 (James & Choquette, 1990). At the earliest stage, the elevated levels of H4SiO4 alone may not cause palygorskite to precipitate due to low pH levels from the dissolved CO2. Dissolved CO2 would, however, cause the meteoric fluid to become slightly more acidic, and thus corrosive to carbonate minerals (James & Choquette, 1990; Liu et al.,

2010). An obvious candidate for preferential dissolution is the less stable, Ca-rich and poorly-ordered dolomite cores (Figs 5 and 8). As the dolomite dissolves, it would buffer the solution pH by forming

HCO3- ions (Rau & Caldeira, 1999). Assuming that the concentration of H4SiO4 in the meteoric solution is high enough, palygorskite will then begin to precipitate once the pH is elevated. In addition to buffering the solution pH, dolomite dissolution will also release Mg2+ ions into solution, and further push the fluid into the stability field of palygorskite (Fig. 12). Therefore, H4SiO4-rich fluids likely initiate the precipitation of palygorskite, but dolomite dissolution plays a vital role in maintaining a suitable solution pH and increasing the Mg which further favors the formation of palygorskite.

Finally, the process(es) responsible for the intercrystalline palygorskite cement may be similar to the process responsible for the intracrystalline palygorskite within dolomite cores. As

H4SiO4 likely provides the dominant control on palygorskite formation in this study, the formation of palygorskite within the intercrystalline pores may occur as a result of the initial stages of dolomite dissolution, causing the solution to buffer and increasing the Mg content of the fluid. As dolomite

123

continues to dissolve, the dissolved cores would provide conduits for fluid flow within the intracrystalline pores. The saturation state of the fluid with respect to palygorskite would also continue to increase with further dolomite dissolution, increasing the reaction rate. This process may result in the precipitation of intracrystalline palygorskite within the dissolved cores, as observed in

Figure 8.

Precursor conditions for palygorskite formation

There are two principal hypotheses for palygorskite formation in shallow-marine environments: the transformation of smectite, or as a direct precipitate from a schizohaline solution with elevated Mg, Si, Al, and pH between 8 and 9. As described above, the fluids responsible for dolomite dissolution would contain some amount of H4SiO4, and dolomite dissolution would buffer the solution pH in addition to providing Mg to the pore fluids. It is unlikely, however, that these fluids alone would provide the requisite concentrations of Si and Mg for palygorskite formation. Thus, a substantial source Si, Mg, and Al are required, whether they originate from a precursor clay, such as smectite, or a solution with ample amounts of dissolved Si, Mg, and Al. Using this information, both hypotheses can be evaluated to further elucidate the genesis of palygorskite in the Umm er Radhuma of Qatar.

The interpretation that palygorskite formed through diagenetic replacement of smectite is problematic for two reasons. First, there is no petrographic, mineralogical, or geochemical evidence of smectite in the Umm er Radhuma. XRD analysis failed to identify smectite, and SEM-EDS and thin section petrographic observations also yield no evidence of smectite. However, the argument that transformation did not occur because smectite is not present (Isphording, 1973; Singer, 1979; Holail

& Al-Hajari, 1997) is tenuous. As discussed earlier, many dolomites consist of 100% dolomite and

0% calcite, yet it is hypothesized that most dolomite forms by replacement of calcite through the dolomitization reaction (Land, 1985; Machel, 2004). Thus, it is not accurate to claim that smectite

124

was not replaced by palygorskite just because smectite is not observed. Additionally, smectite is abundant in the overlying Rus Fm., associated with palygorskite (Fig. 2). This suggests that smectite may have possibly been present in the Umm er Radhuma but was completely replaced by palygorskite.

Secondly, due to the failure to replicate the reaction in the laboratory, there is a general lack of understanding about the thermodynamics and kinetics of the smectite to palygorskite transformation (Singer, 1979). Although the transformation from palygorskite to smectite has been observed in nature (Krekeler et al., 2005) and in the laboratory (Golden et al., 1985; Golden & Dixon,

1990), the reverse reaction of smectite to palygorskite has been unsuccessful, thus making it difficult to evaluate the precise reaction mechanisms. Despite the absence of experimental success, authigenic palygorskite fibers have been observed on the edges of smectite plates in natural settings (Chen et al., 2004; Yeniyol, 2012; Xie et al., 2013). To explain this, Chen et al. (2004) proposed a model whereby Mg2+ enters smectite and reorganizes the positions of both the octahedral and tetrahedral sheets and reverses the positions of free oxygens as to attain the palygorskite structure.

Despite the lack of smectite in the samples analyzed and failure to synthesize the reaction in laboratory settings, the depositional (Fig. 3E) and diagenetic (Fig. 5) relationships in the Umm er

Radhuma suggest the replacement reaction between smectite and palygorskite is more likely than direct precipitation from solution. The Umm er Radhuma formation was deposited as a peritidal to subtidal limestone (Bou-Rabee & Burke 1987; Pollitt et al., 2012), with numerous clay-rich intervals capping the subtidal sequences in Well RR-01 of the present study (Figs. 2 and 3E). Extensive bioturbation through such intervals, such as that exhibited in Figure 3E, provide evidence that these clay-rich intervals are likely depositional in origin. Based on the paragenetic sequence discussed above, however, palygorskite fibers postdate both dolomitization and subsequent dolomite dissolution. Therefore, it is highly likely that a precursor clay was deposited in the mudflat cycle caps and was later altered diagenetically. It is presumed that smectite was the precursor clay, as other

125

phyllosilicates contain too much tetrahedral Al that would require excessively high temperatures for transformation to palygorskite (Weaver & Beck, 1977). In the absence of direct evidence of a smectite-palygorskite transformation, this mechanism is not certain. It is simply the most reasonable one based on literature focused on the transformation of clay minerals to palygorskite (Weaver &

Beck, 1977; Chen et al., 2004).

Shadfan et al. (1985b) posited that the occurrence of palygorskite in the Arabian Peninsula in Eocene rocks is a direct precipitate from solution. The Arabian Shield represents the highlands on the western half of the Arabian Peninsula, while the Arabian Shelf is composed of marine sediments from cycles of peritidal to closed-basin settings in the east. Shadfan et al. (1985b) suggested that the weathering of the Arabian Shield could provide enough dissolved Si and Al to the basin in the east to precipitate palygorskite directly from solution. In order for this process to occur, however, ample amounts of dissolved Si and Al would need to flow into the marine water without transporting detrital material for a replacement reaction. It would also require elevated temperatures and an evaporative environment (Singer, 2002). Such conditions were present during deposition and early diagenesis of the Umm er Radhuma, as indicated by the temporal proximity of these deposits to the

Paleocene-Eocene Thermal Maximum (PETM) (Zachos et al., 2001; 2005; Pollitt et al., 2012) and the presence of diagenetic gypsum towards the top of the formation (Fig. 2).

Explaining the palygorskite cycle caps via primary precipitation is problematic, however. As the burrows presented in Figure 3E are observed within the palygorskite cycle caps, and the paragenetic sequence suggests that palygorskite formed after dolomite, then the bioturbation must have occurred after extensive diagenetic alteration following the primary precipitation hypothesis.

While this may be possible, it is more likely that the burrows were preserved through post- depositional diagenesis involving the transformation of a precursor clay.

Another issue with direct precipitation is based on the solubility relationship between dissolved Al and pH. In general, the solubility of Al decreases with increasing pH. Aluminum is also

126

relatively insoluble in solutions in the pH range 6 – 8, with increasing solubility as pH levels attain values < 6 or > 8 (Driscoll & Schecher, 1990). As a result, solutions in the 7 – 9 optimal pH range for palygorskite formation likely do not contain ample amounts of dissolved Al. Thus, a process that would transport slightly basic fluids with dissolved Al through the section is not so straightforward and is thermodynamically unfavorable.

Although the source of Si, Mg, and Al cannot be directly identified, evidence from the Umm er

Radhuma suggests that dolomite dissolution can play an important role in the formation of palygorskite. In the absence of elevated Mg or H4SiO4 concentrations in slightly basic solutions, smectite will remain stable and palygorskite will not likely form (Fig. 12). Therefore, any process that increases [Mg2+], H4SiO4, and pH in a solution will promote palygorskite formation (Singer & Norrish,

1974; Weaver & Beck, 1977; Singer, 2002). The combination of meteoric fluids and dolomite dissolution would provide additional H4SiO4, buffer the solution pH, and increase Mg, promoting the formation of palygorskite over smectite (Weaver & Beck, 1977; Birsoy, 2002).

Conclusions

This study provides petrographic and geochemical evidence for an alternative model to explain the co-occurrence of dolomite and palygorskite in the Eocene carbonates of the Umm er

Radhuma Formation. This model is based on the observation that palygorskite post-dates both an initial phase of dolomitization as well as a subsequent recrystallization event whereby the unstable cores are dissolved at the expense of more stable rims. The observations presented preclude a detrital origin, direct precipitation from solution, or concurrent precipitation with dolomite. Rather, the evidence provided here is more consistent with the interpretation that palygorskite forms as a diagenetic product concurrent with dolomite dissolution. Thin section and scanning electron microscope images document authigenic palygorskite fibers growing both on the outer surfaces of partially dissolved dolomite crystals, and filling the voids formed by dissolution of the dolomite cores.

127

Although the precise mechanism by which palygorskite replaces smectite is still unclear, these observations suggest that dissolution of metastable dolomite plays an important role in the formation of palygorskite by buffering solution pH and releasing Mg2+ ions into solution. This evidence provides a new model to explore the relationship between dolomite and palygorskite in the geologic record as well as the geochemical and diagenetic conditions responsible for their co- occurrence.

Acknowledgements

This project was supported by a research grant from ExxonMobil Upstream Research Qatar

(awarded to SEK). We thank Cameron Manche and Hanna Cohen for comments on earlier versions of this manuscript. Cameron Manche assisted in the drafting of Figure 1 and in XRD analysis of Figure

4. The quality of the final manuscript was much improved by detailed reviews from two anonymous reviewers, and substantial comments and suggestions from the Associate Editor, Dr. Nick Tosca. We are also grateful to ExxonMobil for permission to publish the results of this study.

References

Akbulut, A. and Kadir, S. (2003) The geology and origin of sepiolite, palygorskite and saponite

in Neogene lacustrine sediments of the Serinhisar-Acipayam basin, Deinizli, SW Turkey.

Clays and Clay Minerals, 51, 279-292.

Aksu, I., Bazilevskaya, E. and Karpyn, Z.T. (2015) Swelling of clay minerals in

unconsolidated porous media and its impact on permeability. GeoResJ, 7, 1-13.

Al-Saad, H. (2003) Facies analysis, cyclic sedimentation and paleoenvironment of the Middle

Eocene Rus Formaiton in Qatar and adjoining areas. Carbonates and Evaporites, 18, 41-50.

Al-Saad, H. (2005) Lithostratigraphy of the Middle Eocene Dammam Formation in Qatar,

128

Arabian Gulf: effects of sea-level fluctuations along a tidal environment. Journal of Asian

Earth Sciences, 25, 781-789.

Bennet, P. and Siegel, D.I. (1987) Increased solubility of quartz in water due to complexing by

organic compounds. Nature, 326, 684-687.

Birsoy, R. (2002) Formation of sepiolite-palygorskite and related minerals from solution. Clays

and Clay Minerals, 50, 736-745.

Botha, G.A. and Hughes, J.C. (1992) Pedogenic palygorskite and dolomite in a late Neogene

sedimentary succession, northwestern Transvaal, South Africa. Geoderma, 53, 139-154.

Bou-Rabee, F. and Burke, C.D. (1987) Lithostratigraphy and environment of deposition of the

top of the Wafra member, Radhuma formation (Wafra field, Kuwait). Palaeogeography,

Palaeoclimatology, Palaeoecology, 59, 269-277.

Çağatay, M.N. (1988) Clay mineralogy and chemistry of some argillaceous rocks in central and

eastern Saudi Arabia. The Arabian Journal for Science and Engineering, 13, 47-63.

Çağatay, M. N. (1990) Palygorskite in the Eocene Rocks of the Dammam Dome, Saudi

Arabia. Clays and Clay Minerals, 38, 299-307.

Callen, R. A. (1977) Late Cainozoic environments of part of Northeastern South Australia.

Journal of the Geological Society of Australia, 24, 151-169.

Callen, R.A. (1984), Clays of the palygorskite-sepiolite group: Depositional environment, age

and distribution. Principles of Sequence Stratigraphy, 37, 1-37.

Carpenter, A.B. (1980) The chemistry of dolomite formation I: The stability of dolomite. In:

Concepts and Models of Dolomitization (Eds D.H. Zenger, J.B. Dunham, and R.L. Ethington),

SEPM Spec. Publ., 28, 111–121.

Cavelier, C. (1970) Geologic description of the Qatar Peninsula. Department of Petroleum

Affairs. 39 p.

Chen, T., XU, H., Lu, A., Xu, X., Peng, S. and Yue, S. (2004) Direct evidence of

129

transformation from smectite to palygorskite: TEM investigation. Science in China Ser. D

Earth Sciences, 47, 985-994.

Chisholm, J.E. (1992) Powder-diffraction patterns and structural models for palygorskite.

Canadian Mineralogist, 30, 61-73.

Draidia, S., El Ouahabi, M., Daoudi, L., Havenith, H. and Fagel, N. (2016) Occurrences and

genesis of palygorskite/sepiolite and associated minerals in the Barzaman formation,

United Arab Emirates. Clay Minerals, 51, 763-779.

Driscoll, C.T. and Schecher, W.D. (1990) The chemistry of aluminum in the environment.

Environ Geochem Health, 12, 28-49.

Eccleston, B.L., Pike, J.G., and Harshash, I. (1981) The water resources of Qatar and their

development: Vol. 1. Food and Agricultural Organization (FAO) of the United Nations, Doha.

Technical Report No. 5. 431 p.

Folk, R.L. and Siedlecka, A. (1974) The schizohaline environment: its sedimentary and

diagenetic fabrics as exemplified by Late Paleozoic rocks of Bear Island, Svalbard.

Sedimentary Geology, 11, 1-15.

Galan, E. (1996) Properties and Applications of Palygorskite-Sepiolite Clays. Clay Minerals,

31, 443-453.

Galan, E. and Carretero, M.I. (1999) A new approach to compositional limits for sepiolite and

palygorskite. Clays and Clay Minerals, 47, 399-409.

Garcia-Romero, E., Suarez, M., Santaren, J. and Alvarez, A. (2007) Crystallochemical

Characterization of the Palygorskite and Sepiolite from the Allou Kagne Deposit, Senegal.

Clays and Clay Minerals, 6, 606-617.

Golden, D.C., Dixon, J.B., Shadfan, H. and Kippenberger, L.A. (1985) Palygorskite and

sepiolite alteration to smectite under alkaline conditions. Clays and Clay Minerals, 33, 44-50.

Golden, D.C. and Dixon, J.B. (1990) Low-temperature alteration of palygorskite to smectite.

130

Clays and Clay Minerals. 38, 401-408.

Goldsmith, J.R. and Graf, D.I. (1958), Structural and compositional variations in some natural

dolomites. Journal of Geology, 66, 678-693.

Gregg, J.M., Bish, D.L., Kaczmarek, S.E. and Machel, H.G. (2015) Mineralogy, nucleation

and growth of dolomite in the laboratory and sedimentary environment: A review.

Sedimentology, 62, 1749-1769.

Guggenheim, S., and Krekeler, M.P.S. (2011) The structures and microtextures of the

palygorskite-sepiolite group minerals. In: Developments in Clay Science (Eds E. Galan and A.

Singer), 3, 3-32.

Holail, H. and Al-Hajari, S. (1997) Evidence of an Authigenic Origin for the Palygorskite in a

Middle Eocene Carbonate Sequence from North Qatar. Qatar Univ. Sci. J., 17, 405-418.

Ingles, M. and Anadon, P. (1991) Relationship of clay minerals to depositional environment in

the non-marine Eocene pontils group, SE Ebro Basin (Spain). Journal of Sedimentary

Petrology, 61, 926-939.

Isphording, W. C. (1973) Discussion of the Occurrence and Origin of Sedimentary

Palygorskite-Sepiolite Deposits. Clays and Clay Minerals, 21, 391-401.

James, N.P. and Choquette, P.W. (1990) Limestone: the meteoric diagenetic environment. In:

Diagenesis: Geological Association of Canada (Eds. I. McIlreath and D. Morrow), Reprint

Series, 4, 35-74.

Jones, B. (2004) Petrography and significance of zoned dolomite cements from the Cayman

Formation (Miocene) of Cayman Brac, British West Indies. Journal of Sedimentary Research,

74, 95-109.

Jones, B. (2007) Inside-out dolomite. Journal of Sedimentary Research, 77, 539-551.

Kaczmarek, S.E. and Sibley, D.F. (2011) On the evolution of dolomite stoichiometry and cation

131

order during high-temperature synthesis experiments: An alternative model for the

geochemical evolution of natural dolomites. Sedimentary Geology, 240, 30-40.

Kaczmarek, S.E. and Sibley, D.F. (2014) Direct physical evidence of dolomite recrystallization.

Sedimentology, 61, 1862-1882.

Kaczmarek, S.E. and Thornton, B.P. (2017) The effect of temperature on stoichiometry, cation

ordering, and reaction rate in high-temperature dolomitization experiments. Chemical

geology, 468, 32-41.

Kaczmarek, S.E., Gregg, J.M., Bish, D.L., Machel, H.G. and Fouke, B.W. (2017) Dolomite,

very-high magnesium calcite, and microbes – implications for the microbial model for

dolomitization. In: Characterization and Modeling of Carbonates – Mountjoy Symposium 1

(Eds A. MacNeil, J. Lonnee, and R. Wood), SEPM Special Publication, 109.

Kadir, S., Erkoyun, H., Eren, M., Huggett, J. and Onalgil, N. (2016) Mineralogy,

Geochemistry, and Genesis of Sepiolite and Palygorskite in Neogene Lacustrine Sediments,

Eskisehir Province, West Central Anatolia, Turkey. Clays and Clay Minerals, 64, 145-166.

Kadir, S., Eren, M., Irkec, T., Erkoyun, H., Kulah, T., Onalgil, N. and Huggett, J. (2017) An

approach to genesis of sepiolite and palygorskite in lacustrine sediments of the Lower

Pliocene Sakarya and Porsuk Formations in the Sivrihisar and Yunusemre-Bicer Regions

(Eskisehir), Turkey. Clays and Clay Minerals, 65, 310-328.

Knauth, L.P. (1979) A model for the origin of chert in limestone. Geology, 7, 274-277.

Krekeler, M.P.S. (2004) Improved Constraints on Sedimentary Environments of Palygorskite

Deposits of Hawthorne Formation, Southern Georgia from a Detailed Study of a Core. Clays

and Clay Minerals, 52, 253-262.

Krekeler, M.P.S., Hammerly, E., Rakovan, J. and Guggenheim, S. (2005) Microscopy studies

of the palygorskite-to-smectite transformation. Clays and Clay Minerals. 53, 92-99.

Krekeler, M.P.S. and Guggenheim, S. (2008) Defects in microstructure in palygorskite-

132

sepiolite minerals: A transmission electron microscopy (TEM) study. Applied Clay Science,

39, 98-105.

Krekeler, M.P.S., Probst, P., Samsonov, M., Tselepis, C.M., Bates, W., Kearns, L., and

Maynard, J.B. (2007) Investigations of Subsurface-flow Constructed Wetlands and Associated

Geomaterial Resources in the Akumal and Reforma Regions, Quintana Roo, Mexico.

Environmental Geology, 53, 709-726.

Land, L.S. (1985) The origin of massive dolomite. Journal of Geological Education, 33, 112–

125.

Liu, Z., Dreybrodt, W., Wang, H. (2010) A new direction in effective accounting for the

atmospheric CO2 budget: Considering the combined action of carbonate dissolution, the

global water cycle and photosynthetic uptake of DIC by aquatic organisms. Earth-Science

Reviews, 99, 162-172.

Lucia, F.J. (2004) Origin and petrophysics of dolostone pore space. Geological Society, London,

Special Publications, 235, 141-155.

Lumsden, D.N. (1979) Discrepancy between thin-section and X-ray estimates of dolomite in

limestone. Journal of Sedimentary Petrology, 49, 429–436.

Machel, H.G. (2004) Concepts and models of dolomitization: a critical reappraisal. In: The Geometry

and Petrogenesis of Dolomite Hydrocarbon Reservoirs (Eds C.J.R. Braithwaite, G. Rizzi and G.

Darke), Geol. Soc. Lond. Spec. Publ., 235, 7–63.

Machel, H.G. and Mountjoy, E.W. (1986) Chemistry and environments of dolomitization—A

reappraisal. Earth Science Reviews, 23, 175-222.

Millot, G. (1970) Geology of Clays: Springer-Verlag, New York, p. 178-234.

Murray, H. H., Pozo, M. and Galán, E. (2011) An Introduction to Palygorskite and Sepiolite

Deposits—Location, Geology and Uses. In: Developments in Clay Science (Eds E. Galan and A.

Singer), 3, 85-99.

133

Navrotsky, A. and Capobianco, C. (1987) Enthalpies of formation of dolomite and magnesian

calcite. American Mineralogist, 79, 782-787.

Pollitt, D. A., Anthonissen, E., Saller, A. H., BouDagher-Fadel, M. K. and Dickson, J. A. D.

(2012) Abrupt early Eocene global climatic change as a control of carbonate facies and

diagenesis: a new record of the Palaeocene-Eocene Thermal Maximum in the Umm er

Radhuma Formation, Saudi Arabia and Kuwait. Terra Nova, 24, 487-498.

Post, J.E. and Heaney, P.J. (2008) Synchrotron powder X-ray diffraction of the structure and

dehydration behavior of palygorskite. American Mineralogist, 93, 667-675.

Powers, R. W., Ramirez, L. F., Redmond, C. D. and Elberg, E. L. (1966) Geology of the

Arabian Peninsula (Sedimentary Geology of Saudi Arabia). Washington DC: U. S. Geological

Survey. Survey, Professional Paper 560D: 147p.

Rau, G.H. and Caldeira, K. (1999) Enhanced carbonate dissolution: a means of sequestering

waste CO2 as ocean bicarbonate. Energy Conversion & Management, 40, 1803-1813.

Rivers, J.M., Kyser, T.K., and James, N.P. (2012) Salinity reflux and dolomitization of

southern Australian slope sediments: the importance of low carbonate saturation levels.

Sedimentology, 59, 445-465.

Shadfan, H., Mashhady, A.S., Dixon, J.B. and Hussen, A.A. (1985a) Palygorskite from

Tertiary Formations of Eastern Saudi Arabia. Clays and Clay Minerals, 33, 451-57.

Shadfan, H., Hussen, A. A. and Alaily, F. (1985b) NOTE: Occurrence of Palygorskite in

Teritiary Sediments of Western Egypt. Clay Minerals, 20, 405-413.

Sibley, D. F. (1982) The origin of common dolomite fabrics: clues from the Pliocene. Journal of

Sedimentary Petrology, 52, 1087-1100.

Sibley, D.F., Dedoes, R.E., and Bartlett, T.R. (1987) Kinetics of dolomitization. Geology, 15,

1112-1114.

Singer, A. (1979) Palygorskite in Sediments: Detrital, Diagenetic or Neoformed—A Critical

134

Review. Geologische Rundschau, 68, 96–1008.

Singer, A. (1984), Pedogenic palygorskite in the arid environment. Palygorskite-Sepiolite

Occurences, Genesis and Uses, 37, 169-175.

Singer, A. (2002) Palygorskite and Sepiolite - the Enigmatic Clay Minerals. Ber. Dtsch. Ton

Tonmineralgruppe, 9, 203-216.

Singer, A. and Amiel, A.J. (1974) Characteristics of Nubian sandstone derived soils in Israel.

Israel J. Earth Sci., 25, 310-319.

Singer, A. and Norrish, K. (1974) Pedogenic palygorskite occurrences in Australia. American

Mineralogist, 59, 508-517.

Sperber, C. M., Wilkinson, B.H., and Peacor, D. R. (1984) Rock composition, dolomite

stoichiometry, and rock/water reactions in dolomitic carbonate rocks. Journal of Geology,

92, 609-622.

Torres-Ruiz, J., Lòpez-Galindo, A., Gonzàlez-Lòpez, J.M. and Delgado, A. (1994)

Geochemistry of Spanish sepiolite-palygorskite deposits: Genetic considerations based on

trace elements and isotopes. Chemical Geology, 112, 221-245.

VanScoyoc, G.E., Serna, C.J. and Ahlrichs, J.L. (1979) Structural changes in palygorskite

during dehydration and dehydroxylation. American Mineralogist, 64, 215-223.

Verrecchia, E.P. and Le Coustumer, M-N. (1996) Occurrence and genesis of palygorskite and

associated clay minerals in a Pleistocene calcrete complex, Sde Boqer, Negev Desert, Israel.

Clay Minerals, 31, 183-202.

Weaver, C.E. (1975), Construction of limpid dolomite. Geology, 3, 425-428.

Weaver, C.E. and Beck, K.C. (1977) Miocene of the S.E. United States: a model for chemical

sedimentation in a peri-marine environment. Dev. in Sedimentology, 22, Elsevier, 234 p.

Whittle, G.L., Alsharhan, A.S. and El Deeb, W.M.Z. (1996) Facies Analysis and Early

Diagenesis of the Middle-Late Eocene Dammam Formation, Abu Dhabi, United Arab

135

Emirates. Carbonates and Evaporites, 11, 32-41.

Xie, Q., Chen, T., Zhou, H., Xu, X., Xu, H., Ji, J., Lu, H. and Balsam, W. (2013) Mechanism

of palygorskite formation in the Red Clay Formation on the Chinese Loess Plateau,

northwest China. Geoderma, 192, 39-49.

Yeniyol, M. (2012) Geology and Mineralogy of a sepiolite-palygorskite occurrence from SW

Eskisehir (Turkey). Clay Minerals, 47, 93-104.

Yeniyol, M. (2014), Characterization of two forms of sepiolite and related Mg-rich clay minerals

from Yenidogan (Sivrihisar, Turkey). Clay Minerals, 49, 91-108.

Zachos, J., Pagani, M., Sloan, L., Thomas, E. and Billups, K. (2001) Trends, Rhythms, and

Aberrations in Global Climate 65 Ma to Present. Science, 292, 686-692.

Zachos, J. C., Röhl, U., Schellenberg, S.A., Sluijs, A., Hodell, D. A., Kelly, D.C., Thomas,

E., Nicolo, M., Raffi, I., Lourens, L.J., McCarren, H. and Kroon, D. (2005) Rapid Acidification of

the Ocean During the Paleocene-Eocene Thermal Maximum. Science, 308, 1611-1615.

136

CHAPTER V

MULTI EPISODIC RECRYSTALLIZATION AND ISOTOPIC RESETTING OF DOLOMITES IN NEAR- SURFACE SETTINGS

Brooks H. Ryan1, Stephen E. Kaczmarek1, John M. Rivers2

1Department of Geological and Environmental Sciences, Western Michigan University, Kalamazoo, Michigan 49008, U.S.A. 2 Qatar Center for Coastal Research (QCCR), ExxonMobil Research Qatar, P.O. Box 22500, Qatar Science and Technology Park-Tech, Doha, Qatar

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., in review, Journal of Sedimentary Research.

Abstract

The Lower Eocene Rus Formation in Qatar reflects carbonate deposition in a restricted to semi-restricted marine setting on a shallow ramp. Petrographic, mineralogical, and geochemical evidence from three research cores show early diagenesis has extensively altered nearly every petrological attribute of these rocks despite not having been deeply buried. In southern Qatar, the lower Rus (Traina Mbr.) is comprised of fabric-retentive dolomite intervals that preserve wackestone-packstone textures that are interbedded with depositional gypsum beds. In northern

Qatar, the same member is dominantly comprised of fabric destructive, planar-e dolomite and lacks evaporites. In both northern and southern Qatar, the upper Rus (Al Khor Mbr.) is comprised of fabric- retentive dolomite intervals as well as limestone intervals rich with Microcodium textures that display evidence of dedolomitization. Geochemical analysis reveals that the limestones have an average δ18O of -10.73‰ VPDB and δ13C of -7.84‰ VPDB, whereas average dolomite δ18O is significantly higher (-1.06‰ VPBD) but dolomite δ13C values (-3.04‰ VPDB; range -10 to 0‰) overlap with limestone δ13C values. Additionally, dolomite δ13C trends toward normal marine values with depth away from the calcite-dolomite contact in all three cores. These geochemical observations suggest that the limestones were first recrystallized in meteoric fluids resulting in negative δ18O and

137

δ13C in association with Microcodium textures, and subsequently dolomitized in marine fluids resulting in relatively higher dolomite δ18O values with retention of the negative δ13C signature associated with initial meteoric diagenesis. Given petrographic observations that dolomite crystals are commonly included in calcite and partially to completely replaced by calcite in these intervals, however, the paragenetic sequence suggests that dolomite formed prior to calcite in the

Microcodium-bearing intervals. Furthermore, the dolomites are commonly cemented by gypsum in the Traina Mbr. in southern Qatar, suggesting dolomitization may also have occurred prior to, or concurrent with, bedded gypsum formation. In sum, these observations indicate that dolomitization occurred early. Following this, the dolomites were replaced by Microcodium-bearing limestones at and immediately below paleo-exposure surfaces, and at greater depths recrystallized in mixed marine-meteoric fluids, producing a negative δ13C signature that trends toward more positive values away from the limestone-dolomite contact. Lastly, the dolomites underwent another phase of recrystallization in either marine-dominated fluids or possibly a well-mixed aquifer setting, resulting in an invariable δ18O signature but retaining the negative δ13C signature and causing the contrast in dolomite δ18O and δ13C observed. This study thus has implications for how carbonate diagenesis is interpreted on a global and temporal scale, as it suggests early diagenesis of geologically young carbonates can be extremely complex, resulting in multiple stages of mineral replacement and isotopic exchange in meteoric and shallow marine fluids, prior to significant burial. Furthermore, this study posits that dolomitization of a limestone may not necessarily prevent additional early diagenesis. Rather, this study indicates that dolomite is unexpectedly reactive in the near-surface environment, especially in response to high water flux coupled with extreme water chemistry variability.

138

Introduction

Carbonate diagenesis occurs in a diverse array of settings that differ in their biological, chemical, and physical processes (McIlreath and Morrow 1990; Tucker and Bathurst 1990; Tucker and Wright 1990; Tucker 1993; Flugel 2004; Swart 2015). Early diagenesis of shallow-marine carbonate rocks is commonly considered to include all processes occurring immediately after grains are deposited on the seafloor, usually involving cementation, micritization, and boring (Tucker and

Bathurst 1990; Tucker and Wright 1990; Tucker 1993). After cementation and lithification on the seafloor, however, carbonate platforms may be subjected to far different conditions in the meteoric and burial realms. These different environments result in diagenetic features that are both texturally and geochemically discernable from one another, which is useful for studies aiming to document diagenetic alterations.

A common diagenetic alteration impacting marine carbonates is the process of dolomitization. The replacement of massive limestone rock bodies by dolomite is interpreted to be common in the rock record (Machel 2004), although there is much debate on the origins of most natural dolomite (Van Tuyl 1916; Land 1985; Budd 1997; Warren 2000; Machel 2004; Gregg et al.

2015; Kaczmarek et al. 2017). Despite the inability to synthesize dolomite under natural earth surface conditions (Land, 1998), many models have been put forth to explain dolomitization of limestones (Morrow 1982; Land 1985). These include dolomitization by seawater (Sass and Katz

1982; Carballo et al. 1987; Land 1991; Budd 1997; Manche and Kaczmarek 2019), hypersaline seawater (Adams and Rhodes 1960; Deffeyes et al. 1965; Warren 2000; Dravis and Wanless, 2018), mixed freshwater-seawater (Badiozamani 1973; Humphrey and Quinn 1989), and burial fluids

(Mattes and Mountjoy 1980; Barnaby and Read 1992; Ryb and Eiler 2018). Ultimately, all dolomitization models are based on dolomite mineralogy, texture, and geochemistry, with the geochemistry of the dolomites interpreted to reflect the nature of the dolomitizing fluids (Warren

2000; Machel 2004).

139

Stable oxygen (δ18O) and carbon (δ13C) isotopes, for example, have long been used to interpret the paleoenvironmental conditions of carbonate rocks (Urey 1947, Urey et al. 1951, and

O’neil et al. 1969). Meteoric water, for example, is generally negative in δ18O, and carbonate rocks forming in such fluids record these negative δ18O values (Swart 2015). Marine waters, in comparison, have higher δ18O values, commonly ³ 0‰, with more positive values with increasing evaporation

(Swart 2015; Rivers et al. 2019b). In contrast, carbonate d13C values are temperature-independent, but are instead influenced by the d13C of dissolved inorganic carbon in the crystallization fluids, which is largely a function of biological processes because CO2 is fixated during photosynthesis (Lohmann

1987; Swart 2015). Environmental and process interpretations based on the measured isotopic composition of carbonate rocks is, however, a function of primary deposition conditions, and the culmination of all subsequent diagenesis. Thus, stable isotope proxies interpreted without petrological context may yield ambiguous interpretations. For example, d18O values in carbonates formed in meteoric environments can be similar to 18O values in carbonates formed in high temperature fluids (Swart, 2015). Complicating paleoenvironmental interpretations of carbonates further is the process of dolomitization, which is thought to cause a ~+3‰ fraction in d18O from low-

Mg calcite formed under the same conditions at 25°C (Fritz and Smith 1970; Matthews and Katz

1977; Vasconcelos et al. 2005; Horita 2014). Thus, there is much ambiguity in environmental interpretations for calcites and dolomites based solely on d18O. In contrast, d13C measurements of carbonate sediment and rocks are generally utilized as stratigraphic markers of global and temporal changes in the earth carbon cycle (Hayes et al. 1999). However, it has been demonstrated that the depositional d13C signature of carbonate rocks can be altered and reset through later diagenetic processes (Lohmann 1987; Swart and Kennedy 2012; Oehlert and Swart 2014; Swart 2015). This adds complexity to the observation that d13C can vary widely even in one environment (Swart 2015).

Therefore, utilizing only d18O and d13C values of carbonate rocks can lead to ambiguous

140

interpretations, and must be integrated with other petrological datasets, such as mineralogy, petrography, and trace element geochemistry.

The Lower Eocene Rus Fm. is a shallow marine carbonate unit that constitutes a diverse suite of carbonate minerals, textures, and d18O and d13C signatures (Rivers et al. 2019a). Al-Saad (2003) divided the Rus into the lower Traina Mbr. and upper Al Khor member. The Traina Mbr. is dominated by interbedded dolomite and gypsum in southern Qatar but relatively pure dolomites in northern

Qatar. The Al Khor Mbr. across all of Qatar is characterized by layered calcites and dolomites, with sharp and abrupt vertical textural and geochemical changes between them. Notably, whereas the d18O of the calcitic intervals differ from the and dolomitic intervals in all cores, the d13C is relatively similar. Rivers et al. (2019a) reported limestones that had been dolomitized and dolomitized limestones that had been dedolomitized, leading them to hypothesize that recrystallization of the rocks is multi-generational. The present study builds upon previous work in order to answer specific questions about these shallowly buried carbonates. First, how was the Rus dolomitized? Secondly, have Rus dolomites been recrystallized, and if so, under what conditions? And lastly, what is the paragenetic and geochemical relationship between the dolomites and calcites? A detailed paragenetic sequence using petrographic cross-cutting relationships coupled with high-resolution mineralogical data is used to determine the relative timing and crystallization of calcite and dolomite.

Stable isotope data is also used to elucidate the diagenetic environments responsible for each phase.

Ultimately, this study demonstrates that a complex diagenetic history resulting in multiple recrystallization events and isotopic alterations has impacted these relatively young and shallow carbonate rocks, with implications for interpreting evaporite associated dolomites, recrystallized dolomites, and carbonate diagenesis based on isotopic data without petrography.

141

Geologic Background

The Lower Eocene (middle to late Ypresian) Rus Fm. in Qatar is interpreted to have been deposited in a variably restricted shallow marine environmental setting (Cavelier 1970; Eccleston et al. 1981; Al-Hajari and Kendall 1992; Al-Saad 2003; Rivers et al. 2019a). The Rus Formation unconformably overlies the Umm er Radhuma and unconformably underlies the Dammam Fm. (Fig.

1; Cavelier 1970; Rivers et al. 2019a). Predominantly observed in the subsurface, the Rus only outcrops at the surface in north-central Qatar (Cavelier 1970). The Rus is formally divided into the lower Traina Mbr. and upper Al Khor Mbr., which vary geographically (Powers et al. 1966; Cavelier

1970; Eccleston et al. 1981; Al-Hajari and Kendall 1992; Al-Saad 2003; Rivers et al. 2019a). The

Traina is composed of intercalated gypsum and dolomitic limestone in southern Qatar, but predominantly dolomite in northern Qatar (Al-Hajari and Kendall 1992; Al-Saad 2003; Rivers et al.

2019a). In contrast, the Al Khor is primarily composed of dolomitic limestone across most of Qatar

(Al-Saad 2003; Rivers et al. 2019a). The stratigraphic thickness of the Rus also varies between 20 and

110 m thick (Cavelier 1970; Abu-Zeid 1991; Al-Hajari and Kendall 1992; Al-Saad 2003; Rivers and

Larson 2018; Rivers et al. 2019a), an observation attributed to structural highs in central Qatar associated with high-angle, syndepositional normal faults that resulted in more restriction in the south and a more open marine environment in the north (Cavelier 1970; Eccleston et al. 1981; Rivers and Larson 2018).

Much work with regard to facies analysis and stratigraphy has been carried out on the Rus and overlying Dammam (Al-Hajari and Kendall 1992; Al-Saad 2003; Al-Saad 2005). More recent work by Rivers et al. (2019a), however, has further refined the sedimentology and stratigraphy of Qatar with respect to the Rus, synthesized in the following lines. The Rus across three shallow research cores consists predominantly of three depositional lithofacies, with a 4th facies only present in the southern half of Qatar. These facies include protected inner ramp facies, tidal flat facies, subaqueous salina facies, and calcitic paleosol facies. In southern Qatar, the Traina is largely composed of

142

Geologic Age Stratigraphic Lithology Unit Arabian Peninsula Dammam

Middle Fm. Core 3 Persian Gulf

Eocene Rus Qatar Fm. Early

Core 2

Core 1

Umm er Radhuma Fm. Paleocene Saudi Arabia PALEOGENE

Kilometres

Dolomite Limestone Gypsum

Figure 1: A generalized stratigraphic section of the Paleocene and Eocene sediments of Qatar, as well as a location map of the Arabian Peninsula, Qatar, and the three research cores (modified from Ryan et al. 2020).1 interbeds of inner ramp and subaqeous salina facies, with few thin intervals of tidal flat facies. In contrast, the subaqueous salina facies is absent in the Traina in northern Qatar where it is dominated by inner ramp facies, although fabric-destructive dolomite textures generally obscure the precursor fabric. The Al Khor in all three cores is dominated by tidal flat facies and calcitic paleosols rich with clay minerals, with few intervals of inner ramp facies. Two calcitic paleosols consisting of recrystallized Microcodium-bearing limestones were identified and correlated across all three cores and interpreted to reflect major exposure events.

143

Materials and Methods

Rock core was recovered from three research boreholes drilled in central and northern Qatar

(Fig. 1), from which all petrographic, mineralogical, and geochemical data were acquired for this study. Detailed descriptions of the methods, sedimentology, and general stratigraphy of these cores are reported in Rivers et al. (2019a). Core recovery through the Rus was 100% in Cores 1 and 3, but approximately 5 intervals from 50-62 m depth in Core 2 that are < 3 m thick consisted of unconsolidated material due to karsting (Rivers et al. 2019a). In all cores, the Rus Fm. is approximately 45 m thick (Fig. 2). 306 cylindrical side-wall core plugs (~2.5 cm diameter, ~3.8 cm long) were taken at 0.02 - 2 m vertical spacing from the Rus, and sample powders from plugs were used for X-ray diffraction and stable isotope analysis as described below. S N Depth (m) Core 2 15

Depth (m) 20 Core 3 Depth (m) Core 1 10 15 25 15 20

Al-Khor 30

Al-Khor 20 25 Al-Khor 35 25 30 40 30 35 45 35 40 50 Traina Traina 40 45 55 Traina 45 50 60 0% 100% 50 55 Mineralogy 0% 100% Mineralogy

60 0% 100% Calcite Mineralogy Dolomite

Clastics

Gypsum

Figure 2: Percent mineralogy for cores 1, 2, and 3 for the recovered Rus Fm. based on XRD generated by Core Laboratories. Clays and chalcedony comprise the clastic component, and all calcite is low-Mg calcite. Cores are datumed to the boundary between the Traina and Al Khor members.

144

X-ray diffraction mineralogy

All core plug powders were analyzed for mineralogy by X-ray diffraction (XRD) methods. Bulk mineralogy was quantified by Core Laboratories utilizing their standard procedure. Each sample is dispersed in a dilute sodium hexa-meta phosphate solution, then centrifugally size-fractioned in order to isolate clay-sized particles (<2-4 µm). Samples are then air-dried and exposed to ethylene glycol for at least 2 hours at 60° C, being measured both before and after exposure. A Panalytical automated diffractometer equipped with an X’celerator linear detector using Real Time Multiple

Strip Technology (RTMS), a graphite monochromator, Ni-Filter, and 40 kV Cu source is used to analyze the samples. Bulk samples are scanned in the 2θ scan range of 4° to 70° at a rate of

4.2°/minute, whereas clay samples are scanned in a range of 2.5° to 40° at a rate of 6.4°/minute.

Because many of the samples contain a mixed mineralogy of calcite and dolomite, additional

XRD analysis was undertaken at Western Michigan University in order to determine percent dolomite relative to calcite, dolomite stoichiometry, and dolomite cation ordering on 68 of the 306 core plugs. Standard XRD procedures were utilized and samples were analyzed with a Bruker D2

Phaser Diffractometer with CuKα radiation. Powders were extracted from core plugs using an electric rotary drill, ground by hand with mortar and pestle for homogenization, and mounted on a Boron- doped silicon P-type zero background diffraction plate. Mounts were analyzed under a 2θ range of

20 to 40° with a step size of 0.01° and a count time of 1.0 s per step. The method of Royse et al. (1971) was employed in order to calculate dolomite percentages (relative to calcite), which takes the ratio of the dolomite d(104) peak to the sum of the calcite d(104) and dolomite d(104) peak. The mol% MgCO3 of the dolomite (i.e. stoichiometry) was calculated by using the d-spacing of the dolomite d(104) peak

(Lumsden 1979). Because Reeder and Sheppard (1984) demonstrated that the equation of Lumsden

(1979) can lead to inaccuracies of up to 3 mol% CaCO3 in Ca-poor dolomites, several internal standards were used to verify peak positions. Cation ordering of the dolomites was determined by taking the ratio of the dolomite d(015) peak to the dolomite d(110) peak (Goldsmith and Graf 1958).

145

Thin section petrography

All thin sections were prepared from the ends of the 306 core plugs analyzed and imaged in a variety of different ways. Thin sections were made and provided by Core Laboratories for all core plugs from each of the three wells. Thin section samples were impregnated with blue epoxy, cut from the ends of the core plugs, and stained with Alizarin Red S (ARS) to differentiate dolomite from calcite

(Dickson 1965). Analysis and imaging of all thin sections was done on a Zeiss Axioplan Microscope.

Additionally, Core Laboratories provided plane-light, high-resolution digital scans of all thin sections from Core 1.

Stable oxygen and carbon isotopes

A total of 211 bulk rock samples were extracted from the Rus across all 3 cores (n = 51, 98, and 62 for Cores 1, 2 and 3, respectively) and analyzed for δ18O and δ13C. Data reported here were first reported in Rivers et al. (2019a). Data were yielded from the Center for Stable Isotope

Biogeochemistry (CSIB), University of California, Berkeley. Samples were reacted with H3PO4 at 90°C for 10 minutes and analyzed by a GV ISoPrime mass spectrometer with Dual-Inlet and MultiCarb systems. Sample weight ranged from 10 to 100 µg, and international standard NBS19 as well as two lab standards were measured with each run. Analytical precision is approximately ±0.05‰ for δ13C and ±0.07‰ for δ18O. Values for all isotopic compositions are reported relative to the Vienna Peedee

Belemnite (VPDB) standard.

Results

Bulk X-ray diffraction mineralogy

The mineralogy of the Traina and Al Khor members varies spatially and temporally within

Qatar (Fig. 2). In core 1, the Traina is composed of sequences characterized by bedded gypsum and dolomite capped by clay (Fig. 2). Most bedded gypsum intervals are £ 5 m thick and contain > 95%

146

gypsum, with minor amounts of dolomite and clays. Similarly, bedded dolomite intervals (> 80 %) are generally £ 3 m with minor (£ 20%) amounts of gypsum and clay (Fig. 2). The clay caps are typically £ 1 m thick. In cores 2 and 3, the Traina differs from core 1 in that it lacks gypsum and contains predominantly dolomite and calcite, with minor amounts of silica and clay in central Qatar.

In core 2, the lower part of the Traina contains low-Mg calcite (LMC) and dolomite, with minor amounts of silica and clay (Fig. 2). In mixed calcite-dolomite samples, calcite can compose > 88% of the bulk rock, whereas other samples are dominated by dolomite comprising > 98% of the bulk rock.

The average silica content in these bulk samples is 4.2% (range 0-70%). Average bulk clay content in the mixed calcite-dolomite samples is 2.2% (range 0-17%). Dolomite dominates in the upper part of the Traina in core 2, and the entire Traina in core 3, comprising > 75% of the bulk rock in all samples, and averaging > 95%. The remaining fraction consists of silica (1.24%) and clay minerals (3.3%).

The mineralogy of the Al Khor is generally constant across Qatar (Fig. 2). The main difference between the Al Khor and the underlying Traina is that m-scale intervals of LMC occur in all three cores, samples comprised of mixed dolomite-calcite are rare, and there is a greater presence of clays

(Fig. 2). The calcite-dominated intervals (< 90%) of the Al Khor also host clay minerals (ave ~12%) and dolomite (< 2%). The dolomite-dominated intervals, by contrast, host similar amounts of the clays and calcite is minor (< 1%). In cores 2 and 3, the abundance of dolomite in calcite-dominated intervals and fractions of calcite in dolomite-dominated intervals are higher than in core 1. For example, dolomite makes up to 7% of bulk rock in calcite-dominated intervals, and calcite up to 20% in dolomite-dominated intervals. Average clay content is also higher in cores 2 and 3 than in core 1, averaging 18.8% and 22.8% of bulk-rock mineralogy, respectively.

Thin section petrography

Dolomite crystal size and texture vary by location and depth. In core 1, the Traina Member

(29.4-61.3 m) consists of microcrystalline (crystal size £ 10 µm) fabric-preserving dolomite (Fig. 3A).

147

Miliolids and other small benthic foraminifera (SBF) are commonly preserved in core 1 (Fig. 3A). In the dolomite-dominated intervals in core 1, dolomite crystals are included in clay (Fig. 3B) and gypsum (Fig. 3C). In cores 2 and 3, by contrast, the Traina Member is characterized by relatively coarse (crystal size £ 100 µm) planar-e dolomites have cloudy to partially corroded cores and clear rims (Fig. 3D). Although the dolomites in cores 2 and 3 are generally fabric destructive, large benthic foraminifera (LBF) and echinoderms (Figs. 3E, F) are rare. In contrast, dolomite in the Al Khor of all cores is predominantly microcrystalline and generally preserves the precursor limestone fabric (Fig.

4).

The dolomite-calcite boundary is sharp (Fig. 5A), although large (>200 µm) replacive calcite crystals are observed (Fig. 5B). Sparry calcite cements that replace dolomite crystals (Fig. 5C) and are void-filling (Fig. 5D) are also observed in the dolomite underlying the dolomite-calcite contact.

The overlying calcite intervals exhibit a wide range of crystal sizes and morphologies. In the Traina of core 2, the only core where calcite is observed in the Traina, calcite crystals are large (> 100 µm) and constitute poikilotopic cements that engulf planar-e dolomite crystals (Fig. 6A). In the Al Khor of core 2, poikilotopic cements are also dominant (Fig. 6B), whereas Microcodium structures are more prevalent in cores 1 and 3 (Figs. 6C, D). Microcodium structures include the “typical Microcodium” as described by Kabanov et al. (2008) (similar to type (a) of Esteban, 1974) which exhibits corn cob or lamellar colonies (Fig. 6C). Additionally, Microcodium type (b) of Esteban (1974) is also observed, with smaller, subquadrangular prisms (Fig. 6D). Calcite is also present as a cement that both fills and partially replaces dolomite crystal cores (Fig. 6E). In all samples in which calcite and dolomite co- occur, dolomite is included in calcite.

In the Traina of core 1, and lesser extent in core 3, clay is observed as an intercrystalline cement (Fig. 7A) between dolomite crystals. In the Al Khor of all cores, clays are observed as intercrystalline cements in calcite intervals (Fig. 6C, D), and form aggregates around which

Microcodium structures have formed (Fig. 7B).

148

A B

200 µm 100 µm

C D

200 µm 100 µm

E F

200 µm 200 µm 100 µm 100 µm Figure 3: Plane polarized light thin section images of various petrographic features of the dolomites in the Traina Mbr. A) Microcrystalline dolomite that has preserved the precursor limestone fabric, displaying dolomitized small benthic foraminifera (core 1, depth 44.28 m). B) Microcrystalline dolomite (red arrows) cemented by palygorskite (green) (core 1, 30.8 m). C) Microcrystalline dolomite (red arrows) encased in gypsum (white) (core 1, 55.68 m). D) Planar-e dolomite displaying cloudy to partially dissolved cores and clear rims (core 2, 53.5 m). E) Planar-e dolomite with preservation of a large benthic foraminfer (core 3, 46.2 m). F) Planar-e dolomite with partial preservation of an echinoderm (red arrow) (core 3, 48.65 m).

149

100 µm

Figure 4: A plane polarized light thin section image of a dolomitized miliolid packstone with the fabric preserved from the Al Khor Mbr. (core 2, 20.65 m).

A B Calcite

Dolomite

2.5 cm 200 µm

C D

200 µm 200 µm 200 µm200200 µmµm

Figure 5: Features of the dolomite-calcite contact observed at ~28 m depth in core 1. A) Image of a sliced-core hand sample displaying the sharp contact between calcite and dolomite. B) Plane-polarized thin section image of calcite (pink) and dolomite (gray), exhibiting the replacive nature of calcite. C) Cross-polarized photomicrograph showing sparry calcite cement that has included and partially replaced dolomite crystals. D) Cross-polarized image displaying void-filling sparry calcite cement within the dolomite interval directly below the dolomite-calcite contact.

150

A B

200 µm 200 µm

C D

500 µm 500 µm

200 µm 200 µm

Figure 6: In the Traina Mbr. of core 2, large (> 200 µm) poikilotopic calcite crystals (pink) engulf planar-e dolomite crystals (A) (core 2, 55 m). However, in the Al Khor Mbr., numerous calcite features are observed as follows. B) Poikilotopic calcite cements (pink) engulfing microcrystalline dolomite (core 2, 15.5 m). C) Corn cob, lamellar colonies of Microcodium (pink) and clay (dark tan/green) (core 1, 24.12 m). D) Subquadrangular prisms of Microcodium (pink) and clay (green) (core 1, 24.64 m). E) Calcite cement (pink) that has filled or partially replaced the dissolved cores of dolomite crystals (grey) (core 3, 8.64 m). All images are plane polarized light photomicrographs.

151

A B

100 µm 100 µm 500 µm

Figure 7: A) Plane polarized light image of the Traina Mbr., exhibiting dolomite crystals (tan) cemented by palygorskite (dark green) (core 3, 51.33 m). B) Plane polarized light image of the Al Khor displaying calcite crystals (Microcodium; pink) included within, as well as surrounding, clay aggregates (green) (core 1, 24.64 m).

Both chalcedony and quartz are observed in intervals that are dominated by calcite.

Chalcedony is frequently observed as partial to complete replacement of calcite crystals (Fig. 8A).

Quartz crystals are commonly coarse (> 200 µm), euhedral to subhedral, and occur within calcite mosaics (Fig. 8B).

A B

200 µm 500 µm

Figure 8: Plane polarized light image of chalcedony (white) that has partially replaced calcite crystals (pink) (A) and a cross polarized light image of euhedral quartz crystals (grey and black) included in calcite mosaics (pink) (B) in core 2 (56.5 m).

152

Dolomite mineralogy

Rus dolomites are near stoichiometric and relatively well-ordered (Fig. 9). Dolomite stoichiometry ranges 48.5-51.0%, averaging 50.05 ± 0.45% and cation ordering ranges 0.4-0.9 averaging 0.62 ± 0.11. The Rus Formation displays no, or very weak, trends of increasing stoichiometry with depth in all cores. Cation ordering increases with depth in cores 2 and 3 (R2 of

0.18 and 0.75, respectively). No depth trend is observed with cation ordering in Core 1 (R2 = 0).

Core 2

Mol% MgCO3 48.5 49 49.5 50 50.5 51 Core 3 15 Mol% MgCO3 48.5 49 49.5 50 50.5 51 Core 1 20 5 Mol% MgCO 3 25 10 48.5 49 49.5 50 50.5 51 15 15 30 20 20 35 25 25 30 Depth (m) 40 30

35 Depth (m) 45 35 40 50 40

Depth (m) 45 B 50 55 45

55 50 C Al-Khor Cation Ordering 60 Traina 0 0.2 0.4 0.6 0.8 1 15 Cation Ordering 65 A 0 0.2 0.4 0.6 0.8 1 20 5

Cation Ordering 10 0 0.2 0.4 0.6 0.8 1 25 15 15 30 20 20 25 35 25 30 Depth (m) 40 30

35 Depth (m) 45 35 40 50 40

Depth (m) 45 E 50 55 45

55 50 F Al-Khor 60 Traina 65 D Figure 9: Plots of stoichiometry with depth for cores 1 (A), 2 (B), and 3 (C). Also displayed are plots of cation ordering as a function of depth for cores 1 (D), 2 (E), and 3 (F). Grey data points are Al Khor samples, while black data points are Traina samples.

153

Somewhat stronger, although inconsistent, trends can be be discerned in specific intervals in the Traina and Al Khor members, however. For example, in core 1, very weak trends are observed for stoichiometry and cation ordering in both the Traina and Al Khor. The Traina exhibits a very weak increase of stoichiometry with depth (R2 = 0.03), with no trend observed in cation ordering (R2 = 0).

Stoichiometry and cation ordering increase with depth in the Al Khor (R2 = 0.05 and 0.12, respectively). However, a strong increase in stoichiometry with depth is observed in the continuous dolomite interval from ~16-21 m (R2 = 0.79). The Traina in core 2, has decreasing stoichiometry (R2

= 0.48) and cation ordering (R2 = 0.63) with depth, whereas the overlying Al Khor displays increasing stoichiometry (R2 = 0.24) and cation ordering (R2 = 0.14) with depth. In core 3, the opposite trend is observed for the Traina, with stoichiometry and cation ordering increasing with depth (R2 = 0.42 and

0.76, respectively). However, in the Al Khor in core 3, stoichiometry decreases with depth (R2 = 0.44), whereas cation ordering displays a very slight increase with depth (R2 = 0.04). Overall, the Traina dolomites are generally more Mg-rich and well-ordered (50.14% ± 0.49% and 0.67 ± 0.12, respectively) than the Al Khor Member (49.98% ± 0.41% and 0.59 ± 0.08, respectively).

Oxygen and carbon isotopes

The average δ18O and δ13C values differ between calcite, dolomite, and mixed calcite-dolomite samples (Fig. 10). Calcite averages δ18O and δ13C values of -10.73‰ and -7.84‰ (VPDB), respectively, whereas dolomite averages -1.06‰ and -3.04‰ (VPDB), respectively. Samples composed of mixed calcite-dolomite yielded δ18O between -6.4‰ and -3.5‰, and δ13C from -4.3‰ to -3.6‰. In core 1, calcite δ18O values are more positive with depth (R2 = 0.73), whereas δ18O values are more negative with depth in core 3 (R2 = 0.79) (Fig. 11). No trend in δ18O is observed in core 2

(R2 = 0). For dolomite, δ18O weakly increases with depth in cores 1 and 2 (R2 = 0.07 and 0.20, respectively) but decreases with depth in core 3 (R2 = 0.39) (Fig. 11). The calcite interval from 24-29 m exhibits a strong increase in δ13C with depth with R2 = 0.79 (Fig. 12). Similarly, the dolomite (30-

154

62 m) below this calcite interval exhibits increasing δ13C with depth (R2 = 0.44). In core 2, dolomite

δ13C becomes more negative with depth until a major calcite interval at 27-29 m (R2 = 0.33). Under the calcite interval at 27-29 m, dolomite δ13C increases with depth (R2 = 0.55). In core 3, dolomite

δ13C increases with depth (R2 = 0.45) below a ~1 m-thick calcite interval.

Calcite Mixed Dolomite 2

0 -14 -12 -10 -8 -6 -4 -2 0 2

-2

-4

C (‰ VPDB) -6 13

-8

-10

-12 18O (‰ VPDB)

Figure 10: Cross plot of bulk dolomite δ18O and δ13C values for all Rus samples. Blue circles are calcite samples, black squares are mixed calcite-dolomite samples, and purple diamonds are dolomite samples.

155

Core 2 18O (‰ VPDB) -14 -12 -10 -8 -6 -4 -2 0 2 Core 3 18 15 O (‰ VPDB) -14 -12 -10 -8 -6 -4 -2 0 2 B 5 C Core 1 20 18 O (‰ VPDB) 10 -14 -12 -10 -8 -6 -4 -2 0 2 15 A 25 15 20 30 20 25 35

Al-Khor Al-Khor 25 Al-Khor 30 Traina 40 Traina Traina

Depth (m) 30 35

45 Depth (m) 35 40 50

Depth (m) 40 45 55 45 50 60 50 Calcite Mixed Dolomite 55 65 Calcite Dolomite 55 60

Calcite Dolomite 65 Figure 11: Bulk rock δ18O from the Rus plotted as a function of depth for cores 1 (A), 2 (B), and 3 (C). Blue circles are calcite samples, black squares are mixed calcite-dolomite samples, and purple diamonds are dolomite samples.

Figure 12 shows that calcite values have the most negative δ18O and δ13C values. The calcite

δ18O and δ13C values exhibit weak covariance (R2 = 0.18). Dolomite δ18O is more positive, predominantly between -2‰ and +1‰, whereas δ13C values span a much wider range from -10‰ to +1‰. Mixed calcite-dolomite samples exhibit δ18O and δ13C values between those of pure calcites and dolomites and display a strong positive covariance between δ18O and δ13C (R2 = 0.9).

The individual Traina and Al Khor members of the Rus exhibit minor differences in dolomite

δ18O but major differences in dolomite δ13C. Traina dolomites δ18O average -0.86 ± 0.61‰ compared to -1.06 ± 0.77‰ in the Al Khor. In contrast to the Traina dolomite δ13C averaging -1.98 ± 1.92‰, the Al Khor dolomite δ13C values are relatively more negative δ13C averaging -4.62 ± 2.12‰. With

156 the exception of one sample, all calcites occur exclusively within the Traina and contain relatively similar δ18O and δ13C values across all three wells (Figs. 10 and 11).

Core 2 13 Core 3 C (‰ VPDB) 13C (‰ VPDB) -12 -10 -8 -6 -4 -2 0 2 15 -12 -10 -8 -6 -4 -2 0 2 5 B C Core 1 20 10 13C (‰ VPDB) -12 -10 -8 -6 -4 -2 0 2 15 25 A 15

20 30 20

25 35 25 Al-Khor Al-Khor Al-Khor 30 Traina 40 Traina Traina 30 Depth (m) Depth (m) 35 45 35

40 50 40 Depth (m)

45 55 45

50 60 50

Calcite Mixed Dolomite 55 65 Calcite Dolomite 55

Calcite Dolomite 65

Figure 12: Bulk rock δ13C from the Rus plotted as a function of depth for cores 1 (A), 2 (B), and 3 (C). Blue circles are calcite samples, black squares are mixed calcite-dolomite samples, and purple diamonds are dolomite samples.

Discussion

Paragenetic sequence

The proposed paragenetic sequence for the Rus Fm. is presented in Figure 13. Based on cross- cutting relationships, dolomitization is interpreted to be the earliest diagenetic process. In the Traina, dolomite is included either in palygorskite clay or gypsum cement (Fig. 3), indicating that

157

dolomitization occurred prior to precipitation of these minerals. In the Al Khor, dolomite is commonly cemented or partially to completely replaced by calcite (Figs. 5 and 6), indicating dolomitization also occurred prior to calcite crystallization. The observation that dolomite crystals included in these mineral phases are partially corroded (Figs. 5-7) suggests that partial dolomite dissolution preceded or was concurrent with precipitation of clays and calcite. Early Late Diagenetic Event Dolomitization

Dolomite dissolution Gypsum cementation Palygorskite cementation Euhedral quartz De-dolomitization via Microcodium Dolomite recrystallization 1 (?) Dolomite recrystallization 2 (?) Clay cementation Replacement of calcite by chalcedony

Figure 13: Paragenetic sequence of the relative timing of diagenetic events impacting the Rus in Qatar, based on petrographic cross-cutting relationships.

Dolomitization in the Rus is followed by clay and gypsum cementation. In sections where dolomite and clay are present, clay occurs as intercrystalline cement that surrounds dolomite

158

crystals (6A), the inference being that depositional clays were diagenetically altered after the dolomitization (e.g., Ryan et al., 2019). In the Traina in Core 1, gypsum is observed as a pore- occluding cement that surrounds dolomite (Fig. 3C), which suggests that gypsum precipitation postdates dolomitization. Following formation of dolomite, clay, and gypsum is the formation of euhedral quartz crystals included in calcite crystals, suggesting either growth of quartz prior to calcite, or co-precipitation of quartz and calcite.

Following the precipitation of quartz crystals is the formation of Microcodium that commonly includes and replaces dolomite crystals and forms poikilotopic low-Mg calcite cements. Given that dolomite crystals are included within and partially replaced by calcite, calcite must have formed after dolomitization. The process of dedolomitization, i.e. dolomite replaced by calcite, suggests that dolomite dissolution happened concurrent with precipitation of calcite, as observed in many of the calcitized sections (Figs. 5 and 6). Based on the observation that calcite crystals are partially replaced by chalcedony in some intervals (Fig. 8A), chalcedony is interpreted to have formed after the

Microcodium. Calcite crystals in the calcite-dominated intervals are also surrounded by clay cement

(Fig. 6D), suggesting that some clay may also have recrystallized after the calcite was emplaced.

Dolomitization interpretation

Dolomitization is interpreted to have occurred early in the paragenetic sequence, but it is important to note that although dolomite is cemented by gypsum, relatively large (>5 m) beds of depositional gypsum are interbedded with dolomitized tidal flat facies in the Traina of core 1 (Rivers et al. 2019a). Dolomitization by hypersaline fluids related to gypsum has been commonly invoked to explain the formation of massive evaporite-associated dolomite platforms. It is generally understood that such fluids have high Mg/Ca ratios conducive for dolomitization due to the removal of Ca2+ by gypsum precipitation, and are also dense and thus have a downward hydrologic drive (e.g., Adams and Rhodes 1960; Deffeyes et al. 1965; Moore et al. 1988; Ruppel and Cander 1988; Saller and

159

Henderson 1998; Cantrell et al. 2004; Kaczmarek and Sibley 2011; Dravis and Wanless 2018; and many others). This interpretation has also been supported by RTM studies that suggest seepage of hypersaline brines is a viable mechanism for platform-scale dolomitization (Jones and Xiao 2005;

Garcia-Fresca and Jones 2011; Al-Helal et al. 2012; Garcia-Fresca et al. 2012). The observation that dolomite is cemented by gypsum does not preclude the possibility that the deposition of gypsum happened first, raised the Mg/Ca ratio of the fluids which resulted in downward dolomitization, and excess Ca2+ released through dolomitization caused subsequent gypsum cementation. Studies using reactive transport modeling (RTM) have demonstrated that this sequence of events is plausible and can explain why dolomites that are cemented by gypsum are located below depositional gypsum beds

(Jones and Xiao 2005; Al-Halal et al. 2012).

Whether or not dolomitization occurred in evaporative fluids related to depositional gypsum can be tested by comparing the observed dolomite δ18O values to hypothetical dolomite δ18O values that would be expected if dolomite formed from evaporative fluids, however, assuming that dolomite

δ18O is reflective of the primary conditions of dolomitization. Given the observed Traina dolomite

δ18O average -0.86‰, it would be expected that the δ18Osw would be -3.8 to +0.7‰ based on a wide range of natural temperatures (20-40°C) (Horita 2014). However, seawater saturated with respect to gypsum has δ18Osw between +7 to +10‰ (Knauth and Beeunas 1986; Rivers et al. 2019b), which is much higher than the range calculated based on the average Traina dolomite δ18O, although the effects on δ18Osw due to evaporation differ in response to humidity variability (e.g. Budd 1997). Rus dolomites would have formed at temperatures of 75°-98°C (Horita 2014) under these fluid chemistries, which is much warmer than modern sabkha surface water temperatures of 20-35°C

(McKenzie 1981; Rivers et al. 2019b). Furthermore, co-occurring aragonite and early-formed dolomite from evaporitic settings have been shown to have δ18Ocarb values between +4‰ and +8‰

(Bellanca and Neri 1986), considerably higher than the dolomite δ18O values of the evaporite- associated Rus dolomites in the present study.

160

Two interpretations are presented to explain why evaporite-associated Traina dolomites have much lower δ18O values than are expected if formed in fluids evaporated to the point of gypsum saturation. The first interpretation is that the dolomites did not form in gypsum-saturated fluids, and rather formed in near-normal marine seawater relatively early after the deposition of the intertidal sediments but prior to significant evaporation and deposition of gypsum. This interpretation is consistent with the observation that dolomite is commonly cemented by gypsum directly below the gypsum deposits as well as the dolomite δ18O suggesting a non-evaporative origin. If the dolomites formed by near-normal seawater, it is likely that this seawater was mesohaline (slightly evaporative but not to the point of gypsum saturation) and had elevated temperatures slightly warmer than expected for Eocene sea surface temperatures (>35°C; Zachos et al. 1994; 2006) due to the abundance of inner ramp and tidal flat facies and interpreted semi-restricted setting (Rivers et al.

2019a). For example, fluids with δ18Osw of +0.7‰ at 40°C can produce dolomites with the average

Traina dolomite δ18O (-0.86‰). A similar dolomite δ18O is attained as temperatures continue to increase and δ18O is more positive due to evaporation (e.g. 45°C and +1.7‰). If the average δ18O of evaporite-associated dolomites (i.e. Traina in core 1) is utilized (-0.42‰), slightly more evaporative fluids are needed at the same temperature (e.g. δ18Osw +1.2‰ at 40°C). In both cases, however, the calculated δ18Osw reflects mesohaline fluids and not seawater that has evaporated to the point of gypsum saturation (Knauth and Beeunas 1986; Rivers et al. 2019b). Fluid values with δ18Osw ³~+1‰ are consistent with slight evaporation of normal Eocene seawater which is interpreted to have δ18Osw of -0.98‰ (Zachos et al. 1994). Small-scale flushing of mesohaline fluids due to fluid density contrasts has been shown to be a viable mechanism for platform-scale dolomitization in RTM studies

(Jones and Xiao 2005), and thus it may be the responsible mechanism for dolomitization of the Traina dolomites. The petrographic observations indicating dolomite formed prior to gypsum, as well as the dolomite δ18O that suggest a fluid origin uninfluenced by evaporation to the point of gypsum saturation, is also consistent with this interpretation.

161

The second interpretation to explain the relatively low Traina dolomite δ18O values is that the dolomites have been recrystallized, and the present dolomite δ18O values reflect the most recent fluids responsible for stabilization. In this interpretation, it is possible that initial dolomitization did occur in fluids that were evaporated to the point of gypsum saturation. Following the possibility of dolomitization by hypersaline brines is the requirement that the dolomites are recrystallized by fluids with near-normal marine δ18O compositions in order to produce the present dolomite δ18O values. It is common for dolomite δ18O to become lower with progressive recrystallization (Mazzulo

1992; Malone et al. 1996), and it is possible that the dolomites were exposed to more normal-marine dominated fluids as the depositional system became increasingly transgressive during deposition of the Al Khor (Cavelier 1970; Rivers et al. 2019a). Average SST estimates during the Eocene are 25-

35°C (Zachos 1994; 2006), which would require marine fluids with δ18Osw between -2.6‰ and -

0.3‰ to produce the average Traina dolomite δ18O of -0.86‰ (Horita 2014). These δ18Osw estimates are within range of the estimated Eocene normal marine δ18Osw of -0.98‰ (Zachos 1994), with decreases in δ18Osw possibly indicating a slight influence of meteoric fluids. Thus, recrystallization of these dolomites in near-normal marine fluids, possibly with a slight meteoric influence, following any initial dolomitization is a plausible explanation for the observed dolomite δ18O.

A slightly different explanation is required for the dolomites of the Al Khor which have relatively similar dolomite δ18O values as the Traina dolomites, but are not observed in association with gypsum in any of the locales included in our study. The Al Khor dolomites average δ18O of -

1.06‰ which suggests that in comparison to the Traina dolomites, Al Khor dolomites formed either in slightly warmer fluids, fluids with slightly more negative δ18O, were recrystallized to a greater extent, or a combination of all three (Land 1980; Mazzullo 1992). Fluids required to produce a dolomite δ18O of -1.06‰ over the temperature range of 25-35°C would have a δ18Osw between -2.8‰ and -0.5‰, within range but slightly lower than the expected -0.98‰ (Zachos 1994; Horita 2014) and not precluding the possibility of some meteoric influence. If it is interpreted that Al Khor

162

dolomite δ18O values reflect initial dolomitization conditions, then near-normal marine fluids are likely responsible for dolomitization. Although the interpretation of dolomitization by near-normal marine seawater is not uncommon (Land 1991; Vahrenkamp and Swart 1994; Mazzulo 1995; Budd,

1997; Manche and Kaczmarek 2019; Ryan et al. 2020), either hydrologic drivers or kinetic enhancers must facilitate dolomitization (Land 1985). Similar to the model proposed for the Traina dolomites, small-scale reflux of mesohaline fluids is a possible mechanism for dolomitization for the Al Khor dolomites. However, given the lack of gypsum and the interpretation that depositional conditions were transitioning from restricted evaporitic settings to more open-marine settings (Rivers et al.

2019a), small increases in salinity may be unlikely. Another common model for dolomitization by near-normal seawater is geothermal (Kohout) convection (Whitaker and Xiao 2010). However, this model is based on convection of seawater entering an isolated carbonate platform, whereas the

Eocene carbonates of Qatar are interpreted to have been deposited on a protected ramp or lagoon

(Rivers et al. 2019a). Tidal pumping and wind-driven circulation of seawater through sediments is also a viable mechanism of normal seawater dolomitization, although such quantities of dolomite are usually volumetrically small (<5%; Carballo 1987; Mazzullo et al. 1995).

Alternatively, and similar to the Traina dolomites, the Al Khor dolomite δ18O values may not presently reflect the initial dolomitization conditions, but rather the conditions of recrystallization.

Also similar to that proposed for the Traina, the dolomite δ18O values suggest that the recrystallization fluids were either marine fluids with elevated temperatures or marine fluids slightly diluted by meteoric water. The various dolomitizing and recrystallization fluids are assessed below in the context of the petrographic relationships between dolomite and calcite, as well as the geochemical relationships with regards to δ13C.

163

Geochemical relationship between dolomite and calcite

Despite their close stratigraphic and petrographic relationships, the stable isotope results in

Figures 10-12 suggest that the calcites and dolomites in the Rus formed in distinctly different environments. The extremely negative δ18O average (-10.73‰) of the calcites suggests formation either in very hot fluids (Urey 1947; Epstein et al. 1953; O’Neil et al. 1969; Swart 2015) or in meteoric fluids (Allan and Matthews 1982; Lohmann 1988; James and Choquette 1990; Hays and Grossman

1991; Swart 2015), both of which tend to have strongly negative δ18O. Given that the associated

Microcodium features observed in the calcite intervals are generally interpreted to reflect biogenic, non-marine, pedogenic environments (Klappa 1978; Kabanov et al. 2008), it is more likely that negative δ18O values in the Rus calcites are related to meteoric fluids. Previous work has shown that pedogenic environments commonly produce Microcodium with δ18O values as negative as -10‰

(Kabanov et al. 2008). In contrast, the dolomites have δ18O values that are suggestive of either initial dolomitization by normal marine Eocene fluids, near-normal marine Eocene fluids with a slight meteoric influence, or recrystallization in such fluids as discussed above.

A shared δ13C trend in the calcites and dolomites complicate interpretations about the chemistry of the diagenetic fluids. The calcite δ13C values are fairly straightforward to interpret.

Average calcite δ13C of -7.84‰ is also consistent with a biogenic and pedogenic origin for

Microcodium (Kabanov et al., 2008). Microbial-derived soil CO2 commonly exerts a stronger influence on Microcodium δ13C than host rock derived carbon, resulting in more negative δ13C than the host rock it replaced (Kabanov et al. 2008). Furthermore, negative shifts in δ13C are common in carbonates altered by meteoric fluids (Allan and Matthews 1982; James and Choquette 1990).

The negative dolomite δ13C values are more complicated to explain. Although the dolomite

δ18O values are consistent with near-normal marine fluids, highly negative dolomite δ13C values (-

10‰) suggest an alternative mechanism for dolomitization. Furthermore, dolomites directly below the major calcite interval in the Al Khor display similarly negative δ13C values to the calcite, and trend

164

positive with depth away from the calcite in all three cores (Fig. 11). It is possible that dolomitization took place in evaporative fluids in the presence of sulfate reduction and oxidation of organic matter, which can cause the dissolved CO2 in the pore fluids to become depleted, and subsequently result in negative dolomite δ13C values (Swart 2015). However, modern sabkha settings have carbonate sediments with δ13C ranging from -1‰ to +5‰ (Rivers et al. 2019b), implying that highly evaporative fluids may not be the culprit for the negative Rus dolomite δ13C values. Another possibility is that dolomitization occurred in fluids associated with formation of the overlying meteoric calcites. The observation that dolomite δ13C is more positive with depth under the calcite interval is consistent with a model whereby the meteoric fluids responsible for the calcite crystallization contain a decreasing percentage of soil-related CO2 with depth (James and Choquette

1990). Such trends in δ13C with depth is common for mixing-zone systems (Allan and Matthews

1982). This results in dolomitization of the underlying strata coinciding with increasing δ13C as seawater is the endmember fluid at the bottom of the section. Although this scenario can explain the isotope data and trends, the petrographic relationships, which consistently show dolomite crystals included in or replaced by calcite, suggest that initial dolomitization occurred prior to calcitization.

The observation that the average dolomite δ18O is -1.06‰, significantly higher than would be expected if the dolomites formed from similar fluids as the calcites, also argues against meteoric fluids. If the dolomites formed under similar conditions as the calcites, dolomite δ18O would be about

-6.9‰ based on the ~+3.8‰ fractionation from calcite to dolomite (Land 1980). However, if dolomitization occurred prior to the formation of the meteoric calcites, and then fluids related to calcite formation mixed with marine water to recrystallize the underlying dolomites, then alterations in dolomite δ18O and δ13C may result in the observed Rus dolomite δ18O and δ13C values.

165

Mixing-zone dolomite recrystallization model

A mixing-zone recrystallization model posits that the petrographic and geochemical observations can be explained by dolomite recrystallization in a mixed marine-meteoric fluid related to the overlying meteoric calcites. In this model, early dolomitization of marine sediments occurs until the middle of the Al Khor, which is characterized by a m-thick calcite interval representing a major exposure event. This exposure caused dedolomitization of the rocks below and the formation of Microcodium at the exposure surface, resulting in the negative δ18O and δ13C observed in the calcites.

Below the major exposure surface lies the rest of the Al Khor and the Traina dolomites, suggesting that the contact between the calcite and dolomite may reflect differences in the pore-fluid chemistry. Specifically, that the observed δ18O and δ13C values in the overlying calcites reflect meteoric fluids, whereas those in the underlying dolomites reflect recrystallization in a freshwater- marine mixing-zone. This model is also consistent with the petrographic relationships observed, in which the dolomites are present as inclusions within calcite crystals and thus formed prior to calcite, but subsequently recrystallized in fluids related to calcite precipitation. This scenario posits that a range of ratios of meteoric water and marine seawater contain a mixed-fluid δ18O that produces dolomites with the observed dolomite δ18O range.

As a starting point for calculating the fluid mixing ratios, it is reasonable to assume that the end-member freshwater component is similar to the fluids responsible for the Microcodium calcites, and that the end-member seawater component is similar to Eocene seawater. Based on an average

Rus calcite δ18O of -10.7‰, the δ18Ow of the freshwater component would be -8.4‰ at 25°C using the equation of Kim and O’neil (1997). The Eocene seawater component would have a δ18Ow of -

0.98‰ based on Zachos et al. (1994). In order to produce dolomite in the temperature range of 25-

30°C with isotopic compositions similar to the average Rus dolomites (δ18O = -1.06‰), a fluid with

δ18Ow -2.8‰ to -1.6‰ (Horita 2014) is required. Using these values in a mass balance equation

166

where the mixing end product δ18Ow is a result of the volumetrically weighted average of the δ18Ow of the freshwater and seawater components, it is possible to calculate the ratios of the mixed fluids

(Eq. 1; e.g. Rohling 2013). The seawater and freshwater fractions required to produce dolomite recrystallization fluids with δ18Ow between -2.8‰ to -1.6‰ are 75% -92%, and 25% -8%, respectively. However, whether these same ratios could produce dolomites with δ13C as low as -10‰ is unknown. δ13C in carbonate rocks is largely controlled by the δ13C of dissolved inorganic carbon

(DIC), which includes CO2 (aq), CO32-, and HCO3- (Swart 2015). Because the DIC in freshwater fluids is usually present as CO2 (aq) whereas DIC in the marine realm is usually present as HCO3- (Swart

2015), a simple mass balance equation cannot be carried out for dolomite δ13C. This is because HCO3- is 8‰ enriched in δ13C compared to atmospheric CO2 (Vogel et al. 1970), and the fractionation between carbonate material and CO2 is temperature dependent (Romanek et al. 1992).

(X)*(freshwater δ18Ow) + (Y)*(seawater δ18Ow) = δ18Ow (mixed) (1)

Where X is the fraction of freshwater and Y is the fraction of seawater, and X + Y = 1

Results in (X)*(-8.4‰) + (1-X)*(-0.98‰) = (-1.6‰ to -2.8‰) (2)

Some aspects of this model are supported by mineralogical observations. The dolomites of the Rus are stoichiometric (50.05%) and relatively well-ordered (0.62), consistent with the idea that recrystallized dolomites are commonly more stoichiometric and well-ordered dolomite (Carpenter

1980; Navrotsky and Capobianco 1987; Sibley 1990; Malone et al. 1996; Kaczmarek and Sibley 2014).

Secondly, the trends in dolomite crystal size are consistent with dolomite recrystallization in a mixing-zone. The Al Khor in cores 2 and 3 consists of finely crystalline (£ 10 µm), fabric-preserving dolomites, whereas the underlying Traina consists of relatively coarser (30-100 µm) planar-e fabric

167

destructive dolomites. This change in texture and increase in crystal size with depth is consistent with dolomite recrystallization (Mazzulo 1992; Gregg et al. 1992).

Despite the theoretical and empirical arguments against mixing zone dolomitization (Hardie

1987; Machel and Mountjoy 1990; Melim et al. 2004), dolomite recrystallization can occur in mixing zones (Hardie 1987; Machel and Burton 1994), though recrystallized dolomites limited in their spatial extent (Ward and Halley 1985; Humphrey 1988; Cander 1994). Gaswirth et al. (2007) showed that mixing zones formed up to 13.5 km3 of dolomite in the Upper Eocene Ocala Limestone and 2.2 km3 in the early Oligocene Suwannee Limestone of southwest Florida. Based on 22-45 m thickness, an approximate area of Qatar of 11,500 km2, and an average Rus dolomite porosity of 30% (Rivers et al. 2019a), the volume of dolomite in the Rus is considerably higher (between ~180 and 360 km3).

Further criticism of this mixing-zone dolomite recrystallization model lies in the interpretation that seawater fractions between 75% and 92% are necessary to form Rus dolomites with the observed δ18O. Comparing this range to theoretical calculations presented in Hardie (1987), such a fluid would be supersaturated with respect to calcite, ideal dolomite, and non-ideal dolomite, and thus fall outside of the “Dorag Zone” which is characterized by fluids supersaturated with respect to dolomite but undersaturated with respect to calcite. Comparing the saturation states for ideal dolomite and non-ideal dolomite in Hardie (1987), it is suggested that 75-92% seawater is supersaturated with respect to both ideal and non-ideal dolomite, implying such a mixed fluid would provide no thermodynamic drive for recrystallization. Of course, this interpretation rests on many assumptions including that the starting water compositions are similar to those used by Badiozami

(1973), which includes modern Yucatan groundwater and modern average sea water.

A mixing-zone dolomite recrystallization model is also inconsistent with the observed increase in δ13C with depth in dolomites below the Microcodium-bearing exposure surface, but the lack of a trend in δ18O. In a typical mixing zone, δ13C and δ18O covary such that minerals that form in the meteoric end member contain more negative δ13C and δ18O than minerals that form in the marine

168

end member (Allan and Matthews 1982). Whereas the dolomite δ13C follows the trend of increasing

δ13C with depth below the Microcodium-bearing exposure surface, no trend is observed in the δ18O.

This suggests that mixing-zone dolomite recrystallization may not be responsible for the dolomite

δ13C and δ18O trends, and that it is probable that a more complex diagenetic history has impacted the

Rus dolomites. A plausible explanation is that mixing-zone dolomite recrystallization occurred after the formation of Microcodium, resulting in the increase in δ13C away from the exposure surface, as detailed above. However, another, subsequent dolomite recrystallization (R2) in marine-dominated fluids reset the dolomite δ18O to present values whereas dolomite δ13C was retained from the initial mixing zone recrystallization. This multi-recrystallization interpretation would explain both the observed trend in dolomite δ13C, as well as the lack of a trend in dolomite δ18O and near-normal marine ranges of dolomite δ18O.

It may be possible that the dolomites were recrystallized, either initially or during R2, by entirely meteoric-dominated fluids rather than marine-dominated fluids. This model may also provide an explanation for some aspects that are inconsistent with a mixing-zone model for the initial recrystallization, mainly the drastic differences in δ18O and δ13C. The model is predicated on the modern groundwater chemistry of the fluids which currently reside in the Rus and underlying Umm er Radhuma formations, which make up Qatar’s groundwater aquifers (Eccleston et al. 1981).

Eccleston et al. (1981) reported that the δ18O of the modern aquifer fluids range from -3‰ to -1‰, and that the δ13C of the fluids range from -12‰ to -5‰. Other than the general observation that dolomite δ18O is lower in dolomites that have been recrystallized (Mazzulo 1992; Malone et al. 1996), there is currently no published data on the subject of isotope fractionation from unstable dolomite

(non-stoichiometric and/or poorly-ordered) to stable dolomite (stoichiometric and well-ordered).

As such, this model assumes that as dolomite recrystallizes, it incorporates the δ18O and δ13C values of the recrystallizing fluids with little to no change. However, this would also assume that dolomite does not retain the precursor δ13C signature, as is widely assumed (Swart, 2015). Another

169

assumption of this model is that predominantly meteoric waters are saturated with respect to dolomite, causing the dissolution and re-precipitation of a more stable dolomite phase. This assumption may not be unreasonable, as data show that the modern Qatar aquifer waters are marginally supersaturated with respect to dolomite (F. Whitaker, pers. comm.).

Incorporating the aquifer δ18Ow values of -3‰ to -1‰ and a near-surface temperature of

25°C, dolomites precipitating from such fluids should have δ18Odolomite between -1.3‰ to +0.7‰, well within range of the average Rus δ18Odolomite of -1.06‰. In a well-mixed aquifer with increasing rock- water interaction with depth, it would not be unlikely for carbonate δ18O values to remain relatively constant while δ13C values are higher with depth as a result of increased rock-water interaction

(Lohmann, 1987). Such a process could be responsible for the trends observed in Rus dolomite δ18O and δ13C. The extremely negative dolomite δ13C directly under the exposure surface is consistent with high fluxes of vadose CO2 which has recently been considered more important for the dissolution of limestone in caves than the mixing of freshwater and seawater (Gulley et al. 2014), and may have implications for the dissolution of metastable dolomite in similar environments. However, a potential downside to this model it is unlikely that the modern Qatar aquifer values are representative of

Eocene meteoric fluids, as regional speleothem studies have shown that meteoric calcite cements have δ18O values between -12‰ and -4‰ during interglacial periods (Fleitmann et al. 2004). These values are similar to the meteoric calcites in this study and suggest that δ18O of meteoric fluids were significantly lower in this study area in the geological past, and not within range of the modern-day aquifer water δ18O values. Thus, utilizing the modern aquifer δ18O and dolomite saturation indices may not be indicative of processes that affected these rocks in the past.

Complicating all interpretations about the recrystallization of Rus dolomites in meteoric fluids related to the calcite exposure surface are the petrographic observations showing dolomite crystals included in calcite crystals (Figs. 5 and 6), suggesting that dolomites formed prior to the precipitation of calcite. Furthermore, dolomite crystals that are included in calcite crystals do not

170

differ in texture compared to dolomites surrounding the calcite intervals (Figs. 5C; 6A and 6B), suggesting that calcites did not preserve an initial dolomite phase that is uniquely different from underlying dolomites interpreted to be recrystallized. From a petrographic standpoint it is relatively easy to discern that dolomitization and subsequent recrystallization occurred prior to meteoric exposure, but the stable isotope data clearly indicate a vastly more complex diagenetic history including post-exposure recrystallization (Fig. 13).

Implications

Recent stable isotopic studies of Qatar’s modern evaporative marine settings (Rivers et al.

2019b) present a conundrum in the interpretation of the local Cenozoic rock record. Modern waters and sediments from restricted evaporative lagoons show significantly elevated δ18O and δ13C values relative to unevaporated seawater and associated deposits (Rivers et al. 2019b). In spite of this, underlying Eocene-age calcitic and dolomitic rocks associated with bedded gypsum deposits point to formation in waters of either near-normal marine or meteoric affiliation (Ryan et al. 2020; this study). Within this context, the current report demonstrates how near-surface and relatively young carbonate deposits are geochemically reset to the degree that all primary isotopic signatures are likely masked. More specifically, our findings demonstrate that early diagenesis of relatively young, shallowly buried carbonate rocks can be extremely complex and result in multiple stages of crystallization, recrystallization, isotopic inheritance, and diagenetic resetting. Despite the documented resetting, a major finding of this study is that none of the petrographic relationships or petrological characteristics of the Rus dolomites reflect dolomitization by hypersaline fluids, despite the intimate relationship with tidal flat facies and overlying bedded gypsum. Thus, the present study adds to a growing body of literature suggesting that it may be more common than presently accepted that dolomites capped by evaporites may be genetically unrelated to the overlying evaporative fluids

(Newport et al. 2017; Manche and Kaczmarek 2019; Ryan et al. 2020).

171

Another implication of this study is that dolomites are susceptible to extensive, and multi- episodic recrystallization in shallow burial to near-surface conditions. Such extensive recrystallization has not been previously documented in young, shallowly buried dolomites.

Published reports of extensive mineralogical and geochemical resetting are invariably associated with geologically older dolomites that have undergone deep (>1000 m) burial (Gregg and Shelton

1990; Gao and Land 1991; Montañez and Read 1992; Kupecz and Land 1994; Malone et al. 1994;

Machel 2004). In contrast, relatively young dolomites and those that have undergone only shallow burial generally display minor alterations in geochemical parameters (McKenzie 1981; Gregg et al.

1992; Veillard et al. 2019). For example, Veillard et al. (2019) analyzed Miocene dolomites in

Australia and interpreted a shallow burial recrystallization trend based on increases in clumped isotope crystallization temperatures and δ18Ow with depth, coupled with mottled CL signatures, increasing crystal size and slightly increasing stoichiometry with depth. However, such dolomites were generally calcium-rich (commonly <47 mol% MgCO3) and poorly ordered (0.2-0.3) (Veillard et al. 2019) compared to the dolomites in this study (average 50.1% and 0.62, respectively). McKenzie

(1981) interpreted recrystallization of dolomites in Abu Dhabi based on increases in crystal size and cation ordering with depth. However, these dolomites were mainly non-stoichiometric (45-48 mol%

MgCO3), unusual for extensively recrystallized dolomites (Mazzulo 1992; Kaczmarek and Sibley

2014). In another study of Holocene dolomites from Belize, Gregg et al. (1992) interpreted recrystallization of dolomites based on subtle increases in crystal size (0.4 µm at the top to 1.0 µm at the base) and cation ordering (0.7-1.0) within ~30 cm, although these dolomites were also non- stoichiometric (40-46 mol% MgCO3). The present study thus builds upon previous studies and demonstrates that relatively young, shallow burial dolomites can be extensively recrystallized with respect to both mineralogy (dolomite stoichiometry and cation ordering) and geochemistry (δ18O and δ13C) multiple times in their early diagenetic history.

172

Our findings broadly support the conclusions of Swart and Kennedy (2012) and Oehlert and

Swart (2014) who cautioned against using carbonate δ13C as a proxy for global biogeochemical events in Earth history. As shown in the present study, post-depositional diagenesis can exert a strong control on carbonate δ13C. Although, it should not be mistaken that δ13C is always the end result of post-depositional alterations. Recent studies in the Paleozoic strata of the Michigan Basin have shown that depositional δ13C trends can still be traced spatially and temporally through lithological changes between calcite and dolomite (Caruthers et al. 2018). This means that carbonate

δ13C may be altered after deposition due to extensive diagenesis, but that it also depends on other factors. Lastly, this study adds to a large body of literature that demonstrates that carbonate geochemistry alone may lead to inaccurate interpretations about the paragenetic sequence of a rock.

Instead, geochemical proxies must be integrated with petrographic observations in order to assess the complete diagenetic history of carbonate rock units (Schlager and James 1978; Melim et al. 1995;

Frank and Bernet 2000; Melim et al. 2002; Gischler et al. 2013). If the isotopic trends presented in

Figures 11 and 12 were assessed without petrographic context, the interpretation that marine dolomitization post-dated exposure events related to Microcodium would be logical given that the dolomite δ13C is increasingly positive deeper from the exposure surface (i.e. retains meteoric-derived

δ13C), but dolomite δ18O is invariable. The only way to eliminate this as a geologic possibility is through detailed petrographic analysis.

Conclusions

Early and extensive near-surface diagenesis has significantly altered the textures, mineralogy, and isotopic signature of the Eocene Rus Formation of Qatar. Major alterations include early dolomitization, dolomite recrystallization, and exposure dedolomitization by Microcodium.

Petrographic evidence indicates that dolomitization occurred early and prior to calcitization, gypsum precipitation, and clay diagenesis. The stable isotopic compositions of dolomite and calcite, however,

173

suggest a rich and complex history involving multiple recrystallization events tied to exposure.

Dolomites invariably exhibit near-normal marine δ18O values (ave -1.06‰ VPBD) but strongly negative δ13C values (-10 to 0‰; ave = -3.04‰ VPDB), despite being closely associated with bedded evaporites. Calcite intervals, in contrast, are characterized by strongly negative δ18O values (ave -

10.73‰ VPDB) but similar δ13C (ave -7.84‰ VPDB) to the dolomites. The trend in increasing dolomite δ13C with depth away from the calcite-dolomite contact, coupled with similar calcite and dolomite δ13C values, suggest that dolomites were recrystallized in fluids related to formation of

Microcodium, despite evidence showing replacement of dolomite by Microcodium. The model presented here to explain petrographically early dolomites with geochemical characteristics of later calcite phases is as follows. Early dolomitization of the Rus took place either in hypersaline fluids related to gypsum precipitation (Traina) or near-normal marine fluids (Traina and Al Khor).

Following this, major exposure events resulting in meteoric calcite and Microcodium resulted both in dedolomitization as well as subsequent dolomite recrystallization, resetting both the dolomite δ18O and δ13C in a mixed freshwater-seawater system. Following this, the dolomites were possibly recrystallized once more either in near-normal marine fluids or possibly a well-mixed aquifer setting, resulting in the present dolomite δ18O but retaining the initial recrystallization freshwater influenced

δ13C signature or recording the aquifer δ13C signature, with more positive δ13C with increased rock- water interaction. Ultimately, this study sheds light on i) the origin of some evaporite-associated dolomites, ii) the degree to which early formed dolomites can recrystallize under shallow burial and in near-surface fluids, and iii) interpreting geochemical relationships without incorporating petrographic observations.

Acknowledgements

We would like to thank ExxonMobil management for permission to publish this study. We would also like to thank Sabrina Skeat for her logistical support and help with data transfer.

174

References

Abu-Zeid, M.M., 1991, Lithostratigraphy and framework of sedimentation of the sub-surface

Paleogene succession in northern Qatar: Neue Jahrbuch für Geologie und Paleaontologie, v. 4,

p. 191-204.

Adams, J.E., and Rhodes, M.L., 1960, Dolomitization by seepage refluxion: Bulletin of the American

Association of Petroleum Geologists, v. 44, 1912-1920.

Al-Hajari, S.A., and Kendall, C.G.S.C., 1992, The sedimentology of the Lower Eocene Rus

Formation of Qatar and neighboring regions: J. Univ. Kuwait (Sci.), v. 19, p. 153-172.

Al-Helal, A.B., Whitaker, F.F., and Xiao, Y., 2012, Reactive transport modeling of brine reflux:

Dolomitization, anhydrite precipitation, and porosity evolution: Journal of Sedimentary

Research, v. 82, p. 196-215.

Al-Saad, H., 2003, Facies analysis, cyclic sedimentation and paleoenvironment of the Middle

Eocene Rus Formation in Qatar and adjoining areas: Carbonates and Evaporites, v. 18, p. 41-

50.

Al-Saad, H., 2005, Lithostratigraphy of the Middle Eocene Dammam Formation in Qatar, Arabian

Gulf: effects of sea-level fluctuations along a tidal environment: J. Asian Earth Sci., v. 25, p.

781–789.

Allan, J.R., and Matthews, R.K., 1982, Isotope signatures associated with early meteoric

Diagenesis: Sedimentology, v. 29, p. 797-817.

Badiozamani, K., 1973, The Dorag dolomitization model—Application to the Middle Ordovician of

Wisconsin: Journal of Sedimentary Petrology, v. 43, p. 965-984.

Bellanca, A., and Neri, R., 1986, Evaporite carbonate cycles of the Messinian, Sicily: Stable isotopes,

mineralogy, textural features, and environmental implications: Journal of Sedimentary

Petrology, v. 56, p. 614-621.

Budd, D.A., 1997, Cenozoic dolomites of carbonate islands: their attributes and origin: Earth-Science

175

Reviews, v. 41, p. 1-47.

Cander, H.S., 1994, An example of mixing-zone dolomite, middle Eocene Avon Park Formation,

Florida Aquifer System: Journal of Sedimentary Research, v. A64, p. 615-629.

Cantrell, D., Swart, P., and Hagarty, R., 2004, Genesis and characterization of dolomite, Arab-D

Reservoir, Ghawar field, Saudi Arabia: GeoArabia, v. 9, p. 11-36.

Carballo, J.D., Land, L.S., and Miser, D.E., 1987, Holocene dolomitization of supratidal sediments by

active tidal pumping, Sugarloaf Key, Florida: Journal of Sedimentary Petrology, v. 57, p. 153-

165.

Carpenter, A.B., 1980, The chemistry of dolomite formation I: the stability of dolomite, in

Zenger, D.H., Dunham, J.B., and Ethington, R.L., eds., Concepts and Models of Dolomitization:

SEPM Special Publication 28, p. 111-121.

Caruthers, A.H., Gröcke, D.R., Kaczmarek, S.E., Rine, M.J., Kuglitsch, J., and Harrison III, W.B., 2018,

Utility of organic carbon isotope data from the Salina Group halite (Michigan Basin): A new

tool for stratigraphic correlation and paleoclimate proxy resource: GSA Bulletin, v. 130, p.

1782-1790.

Cavelier, C., 1970, Geologic Description of the Qatar Peninsula, Department of Petroleum

Affairs, Paris, France, 39 p.

Deffeyes, K.S., Lucia, F.J., and Weyl, P.K., 1965, Dolomitization of Recent and Plio-Pleistocene

sediments by marine evaporite waters on Bonaire, Netherlands, Antilles, in Pray, L.C., and

Murray, R.C., eds., Dolomitization and limestone diagenesis: SEPM Special Publication 13, p.

71-88.

Dickson, J.A.D., 1965, A modified staining technique for carbonates in thin section: Nature, v. 205,

p. 587-587.

Dravis, J.J., and Wanless, H.R., 2018, Reflux dolomitization – A Holocene example beneath a coastal

176

salina, West Caicos Island, Turks and Caicos Islands: Marine and Petroleum Geology, v. 97, p.

311-322.

Eccleston, B.L., Pike, J.G., and Harshah, I., 1981, The Water and Resources of Qatar and Their

Development: Vol. 1. Food and Agricultural Organization (FAO) of the United Nations, Doha.

Technical Report No. 5, 431 p.

Epstein, S., Buchsbaum, R., Lowenstam, H.A., and Urey, H.C., 1953, Revised carbonate-water

isotopic temperature scale: Geological Society of America Bulletin, v. 64, p. 1315-1326.

Esteban, M., 1974, Caliche textures and Microcodium: Boll. Soc. Geol. Ital., v. 92, p. 105-125.

Fleitmann, D., Matter, A., Pint, J.J., and Al-Shanti, M.A., 2004, The speleothem record of climate

change in Saudi Arabia: Saudi Geological Survey, Jeddah, Kingdom of Saudi Arabia.

Frank, T.D., and Bernet, K., 2000, Isotopic signature of burial diagenesis and primary lithological

contrasts in periplatform carbonates (Miocene, Great Bahama Bank): Sedimentology, v. 47, p.

1119-1134.

Fritz, P., and Smith, D.G.W., 1970, The isotopic composition of secondary dolomites: Geochimica et

Cosmochimica Acta, v. 34, p. 1161-1173.

Gao, G., and Land, L.S., 1991, Early Ordovician Cool Creek dolomite, middle Arbuckle Group, Slick

Hills, SW Oklahoma, U.S.A.: Origin and modification: Journal of Sedimentary petrology, v. 61,

p. 161-173.

Garcia-Fresca, B., and Jones, G.D., 2011, Apparent stratigraphic concordance of reflux dolomite: New

insight from high-frequency cycle scale synsedimentary reactive transport models, in

Summers Engel, A., Engel, S., Moore, P., and DuChene, H., Carbonate Geochemistry: Reactions

and Processes in Aquifers and Reservoirs: Karst Waters Institute Special Publication, v. 16, p.

27-30.

Garcia-Fresca, B., Lucia, F.J., Sharp Jr., J.M., and Kerans, Charles, 2012, Outcrop-constrained

177

hydrogeological simulations of brine reflux and early dolomitization of the Permian San

Andres Formation: AAPG Bulletin, v. 96, p. 1757-1781.

Gaswirth, S.B., Budd, D.A., and Farmer, G.L., 2007, The role and impact of freshwater-seawater

mixing zones in the maturation of regional dolomite bodies within the proto Floridan Aquifer,

USA: Sedimentology, v. 54, p. 1065-1091.

Gischler, E., Thomas, A.L., Droxler, A.W., Webster, J.M., Yokoyama, Y., and Schöne, B.R., 2013,

Microfacies and diagenesis of older Pleistocene (pre-last glacial maximum) reef deposits,

Great Barrier Reef, Australia (IODP Expedition 325): A quantitative approach:

Sedimentology, v. 60, p. 1432-1466.

Goldsmith, J.R., and Graf, D.L., 1958, Structural and compositional variations in some natural

dolomites: Journal of Geology, v. 66, p. 678-693.

Gregg, J.M., and Shelton, K.L., 1990, Dolomitization and dolomite neomorphism in the back reef

facies of the Bonneterre and Davis Formations (Cambrian), southeastern Missouri: Journal of

Sedimentary Petrology, v. 60, p. 549-562.

Gregg, J.M., Howard, S.A., and Mazzullo, S.J., 1992, Early diagenetic recrystallization of

Holocene (<3000 years old) peritidal dolomites, Ambergis Cay, Belize: Sedimentology, v. 39,

p. 143-160.

Gregg, J.M., Bish, D.L., Kaczmarek, S.E., and Machel, H.G., 2015, Mineralogy, nucleation

and growth of dolomite in the laboratory and sedimentary environment: A review:

Sedimentology, v. 62, p. 1749-1769.

Gulley, J., Martin, J., and Moore, P., 2014, Vadose CO2 gas drives dissolution at water tables in

eogenetic karst aquifers more than mixing dissolution: Earth Surface Processes and

Landforms, v. 39, 1833-1846.

Hardie, L.A., 1987, Dolomitization: A critical view of some current views: Journal of Sedimentary

Research, v. 57, p. 166-183.

178

Hays, P.D., and Grossman, E.L., 1991, Oxygen isotopes in meteoric calcite cements as indicators

of continental paleoclimate: Geology, v. 19, p. 441-444.

Hayes, J.M., Strauss, H., and 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, v. 161, p. 103-125.

Horita, J., 2014, Oxygen and carbon isotope fractionation in the system dolomite-water-CO2 to

elevated temperatures: Geochimica et Cosmochimica Acta, v. 129, p. 111-124.

Humphrey, J.D., 1988, Late Pleistocene mixing zone dolomitization, southeastern Barbados, West

Indies: Sedimentology, v. 35, p. 327-348.

Humphrey, J.D., and Quinn, T.M., 1989, Coastal mixing zone dolomite, forward modeling, and

massive dolomitization of platform-margin carbonates: Journal of Sedimentary Petrology, v.

59, p. 438-454.

James, N.P., and Choquette, P.W., 1990, Limestone: the meteoric diagenetic environment, in

McIlreath, I.A., and Morrow, D.W., eds., Diagenesis: Geological Association of Canada, Reprint

Series 4, p. 35-74.

Jones, G.D., and Xiao, Y., 2005, Dolomitization, anhydrite cementation, and porosity evolution in a

reflux system: Insights from reactive transport models: AAPG Bulletin, v. 89, p. 577-601.

Kabanov, P., Anadón, P., and Krumbein, W.E., 2008, Microcodium: An extensive review and a

proposed non-rhizogenic biologically induced origin for its formation: Sedimentary Geology,

v. 205, p. 79-99.

Kaczmarek, S.E., and Sibley, D.F., 2011, On the evolution of dolomite stoichiometry and cation order

during high-temperature synthesis experiments: An alternative model for the geochemical

evolution of natural dolomites: Sedimentary Geology, v. 240, p. 30-40.

Kaczmarek, S.E., and Sibley, D.F., 2014, Direct physical evidence of dolomite recrystallization:

Sedimentology, v. 61, p. 1862-1882.

179

Kaczmarek, S.E., Gregg, J.M., Bish, D.L., Machel, H.G., and Fouke, B.W., 2017, Dolomite,

very-high magnesium calcite, and microbes – implications for the microbial model for

dolomitization, in MacNeil, A., Lonnee, J., and Wood, R., eds., Characterization and Modeling

of Carbonates – Mountjoy Symposium 1: SEPM Special Publication 109, p. 7-20.

Kim, S., and O’Neil, J.R., 1997, Equilibrium and nonequilibrium oxygen isotope effects in synthetic

Carbonates: Geochimica et Cosmochimica Acta, v. 61, p. 3461-3475.

Klappa, C.F., 1978, Biolithogenesis of Microcodium: Elucidation: Sedimentology, v. 25, p. 489-522.

Knauth, L.P., and Beeunas, M.A., 1986, Isotope geochemistry of fluid inclusions in Permian halite

with implications for the isotope history of ocean water and the origin of saline formation

waters: Geochimica et Cosmochimica Acta, v. 50, 419-433.

Kupecz, J.A., and Land, L.S., 1994, Progressive recrystallization of early-stage dolomite: Lower

Ordovician Ellenburger Group, west Texas, in Pusur, B., Tucker, M., and Zenger, D., eds.,

Dolomites: A Volume in Honour of Dolomieu: Spec. Publs Int. Ass. Sediment., v. 21, p. 225-

279.

Land, L.S., 1980, The isotopic and trace element geochemistry of dolomite: The state of the

art, in Zenger, D.H., Dunham, J.B., and Ethington, R.L., eds., Concepts and Models of

Dolomitization: SEPM Special Publication 28, p. 87-110.

Land, L.S., 1985, The origin of massive dolomite: Journal of Geoscience Education, v. 33, p. 112-125.

Land, L.S., 1991, Dolomitization of the Hope Gate Formation (north Jamaica) by seawater:

Reassessment of mixing-zone dolomite, in Taylor, H.P., O’Neil, J.R., and Kaplan, I.R., eds., Stable

Isotope Geochemistry: A Tribute to Samuel Epstein: The Geochemical Society, Special

Publication No. 3, p. 121-133.

Land, L.S., 1998, Failure to precipitate dolomite at 25°C from dilute solution despite 1000-fold

oversaturation after 32 years: Aquatic Geochemistry, v. 4, p. 361-368.

Lohmann, K.C., 1987, Geochemical patterns of meteoric diagenetic systems and their

180

application to the study of paleokarst, in James, N.P., and Choquette, P.W., eds., Paleokarst,

58-80.

Lumsden, D.N., 1979, Discrepancy between thin-section and X-ray estimates of dolomite in

Limestone: Journal of Sedimentary Petrology, v. 49, p. 429-436.

Machel, H.G., 2004, Concepts and models of dolomitization: A critical reappraisal, in Braithwaite, C.,

Rizzi, G., and Darke, G., eds., The Geometry and Petrogenesis of Dolomite Hydrocarbon

Reservoirs: Geological Society (London) Special Publication 235, p. 7-63.

Machel, H.G., and Burton, E.A., 1994, Golden Grove Dolomite: Barbados: Origin from modified

Seawater: Journal of Sedimentary Research, v. A64, p. 741-751.

Machel, H.G., and Mountjoy, E.W., 1990, Coastal mixing zone dolomite, forward modeling, and

massive dolomitization of platform-margin carbonates—Discussion: Journal of Sedimentary

Petrology, v. 60, p. 1008-1012.

Malone, M.J., Baker, P.A., and Burns, S.J., 1994, Recrystallization of dolomite: evidence from the

Monterey Formation (Miocene), California: Sedimentology, v. 41, p. 1223-1239.

Malone, M.J., Baker, P.A., and Burns, S.J., 1996, Recrystallization of dolomite: An experimental

study from 50-200°C: Geochimica et Cosmochimica Acta, v. 60, p. 2189-2207.

Manche, C.J., and Kaczmarek, S.E., 2019, Evaluating reflux dolomitization using a novel high-

resolution record of dolomite stoichiometry: A case study from the Cretaceous of central

Texas, USA: Geology, v. 47, p. 586-590.

Matthews, A., and Katz, A., 1977, Oxygen isotope fractionation during the dolomitization of calcium

Carbonate: Geochimica et Cosmochimica Acta, v. 41, p. 1431-1438.

Mazzullo, S.J., 1992, Geochemical and neomorphic alteration of dolomite: A review: Carbonates and

Evaporites, v. 7, p. 21-37.

Mazzullo, S.J., Bischoff, W.D., and Teal, C.S., 1995, Holocene shallow-subtidal dolomitization by near-

normal seawater, northern Belize: Geology, v. 23, p. 341-344.

181

McIlreath, I.A., and Morrow, D.W., 1990, Diagenesis: Geological Association of Canada,

Reprint Series 4, 338p.

McKenzie, J.A., 1981, Holocene dolomitization of calcium carbonate sediments from the coastal

sabkhas of Abu Dhabi, U.A.E.: A stable isotope study: Journal of Geology, v. 89, p. 185-198.

Melim, L.A., Swart, P.K., and Maliva, R.G., 1995, Meteoric-like fabrics forming in marine waters:

Implications for the use of petrography to identify diagenetic environments: Geology, v. 23,

p. 755-758.

Melim, L.A., Westphal, H., Swart, P.K., Eberli, G.P., and Munnecke, A., 2002, Questioning carbonate

diagenetic paradigms: evidence from the Neogene of the Bahamas: Marine Geology, v. 185, p.

27-53.

Melim, L.A., Swart, P.K., and Eberli, G.P., 2004, Mixing-zone diagenesis in the subsurface of Florida

and The Bahamas: Journal of Sedimentary Research, v. 74, p. 904-913.

Montañez, I.P., and Read, J.F., 1992, Fluid-rock interaction history during stabilization of early

dolomites, Upper Knox Group (Lower Ordovician), U.S. Appalachians: Journal of Sedimentary

Petrology, v. 62, p. 753-778.

Moore, C.H., Chowdhury, A., and Chan, L., 1988, Upper Jurassic Smackover platform dolomitization,

northwestern Gulf of Mexico: a tale of two waters, in Shukla, V.J., and Baker, P.A., eds.,

Sedimentology and Geochemistry of Dolostones: SEPM Special Publication, v. 43, p. 175-189.

Morrow, D.W., 1982, Diagenesis 2. Dolomite—Part 2: dolomitization models and ancient

dolostones: Geological Association of Canada, v. 9, p. 95-107.

Navrotsky, A., and Capobianco, C., 1987, Enthalpies of formation of dolomite and magnesian

Calcite: American Mineralogist, v. 79, p. 782-787.

Oehlert, A.M., and Swart, P.K., 2014, Interpreting carbonate and organic carbon isotope

covariance in the sedimentary record: Nature Communications, v. 5, doi:

10.1038/ncomms5672.

182

O’Neil, J.R., Clayton, R.N., and Mayeda, T.K., 1969, Oxygen isotope fractionation in divalent

metal carbonates: Journal of Chemical Physics, v. 51, p. 5547-5558.

Powers, R.W., Ramirez, L.F., Redmond, C.D., and Elberg, E.L., 1966, Geology of the Arabian

Peninsula (Sedimentary Geology of Saudi Arabia), U.S. Geological Survey, Washington D.C.

Survey, Professional Paper 560D, 147 p.

Reeder, R.J., and Sheppard, C.E., 1984, Variations of lattice parameters in some sedimentary

dolomites: American Mineralogist, v. 69, p. 520-527.

Rivers, J.M., and Larson, K.P., 2018, The Cenozoic kinematics of Qatar: evidence for high-angle

faulting along the Dukhan ‘anticline’: Marine and Petroleum Geology, v. 92, p. 953-961.

Rivers, J.M., Skeat, S.L., Yousif, R., Liu, C., Stanmore, E., Tai, P. and Al-Marri, S.M., 2019a,

The depositional history of near-surface Qatar aquifer rocks and its impact on matrix flow

and storage properties: Arabian Journal of Geosciences, v. 12,

https://doi.org/10.1007/s12517-019-4498-6

Rivers, J.M., Varghese, L., Yousif, R., Whitaker, F.F., Skeat, S.L., and Al-Shaikh, I., 2019b, The

geochemistry of Qatar coastal waters and its impact on carbonate sediment chemistry and

early marine diagenesis: Journal of Sedimentary Research, v. 89, p. 293-309.

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, v. 56, p. 419–430.

Royse, C.F., Jr., Wadell, J.S., and Petersen, L.E., 1971, X-ray determination of calcite-dolomite:

an evaluation: Journal of Sedimentary Petrology, v. 41, p. 483-488.

Rohling E.J., 2013, Oxygen Isotope Composition of Seawater, in Elias, S.A., ed., The Encyclopedia

of Quaternary Science, vol. 2, pp. 915-922. Amsterdam: Elsevier.

Ruppel, S.C., and Cander, H.S., 1988, Dolomitization of shallow water carbonates by seawater and

183

seawater-derived brines: San Andres Formation (Guadalupian), west Texas, in Shukla, V.J.,

and Baker, P.A., eds., Sedimentology and Geochemistry of Dolostones: SEPM Special

Publication, v. 43, p. 245-262.

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., 2019, Dolomite dissolution: An alternative diagenetic

pathway for the formation of palygorskite clay: Sedimentology, v. 66, p. 1803-1824.

Ryan, B.H., Kaczmarek, S.E., and Rivers, J.M., 2020, Early and pervasive dolomitization by near-

normal marine fluids: New lessons from an Eocene evaporative setting in Qatar:

Sedimentology, https://doi.org/10.1111/sed.12726

Ryb, U., and Eiler, J.M., 2018, Oxygen isotope composition of the Phanerozoic ocean and a possible

solution to the dolomite problem: PNAS, v. 115, p. 6602-6607.

Saller, A.H., and Henderson, N., 1998, Distribution of porosity and permeability in platform

dolomites: Insight from the Permian of West Texas: American Association of Petroleum

Geologists Bulletin, v. 82, p. 1528-1550.

Schlager, W., and James, N.P., 1978, Low-magnesian calcite limestones forming at the deep-sea

floor, Tongue of the Ocean, Bahamas: Sedimentology, v. 25, p. 675-702.

Sibley, D.F., 1990, Unstable to stable transformations during dolomitization: The Journal of

Geology, v. 98, p. 739-748.

Swart, P.K., 2015, The geochemistry of carbonate diagenesis: The past, present and future:

Sedimentology, v. 62, p. 1233-1304.

Swart, P.K., and Kennedy, M.J., 2012, Does the global stratigraphic reproducibility of δ13C in

Neoproterozoic carbonates require a marine origin? A Pliocene-Pleistocene comparison:

Geology, v. 40, p. 87-90.

Tucker, M.E., 1993, Carbonate diagenesis and sequence stratigraphy, in Wright, V.P., ed.,

Sedimentology Review 1, p. 51-72.

Tucker, M.E., and Bathurst R.G.C., 1990, Carbonate diagenesis, International Association of

184

Sedimentologists Reprint Series 1, 312 p.

Tucker, M.E., and Wright, V.P., 1990, Carbonate Sedimentology, Blackwell Science Ltd.,

Oxford, 482 p.

Urey, H.C., 1947, The thermodynamic properties of isotopic substances: J. Chem. Soc., v. 99, p. 562-

581.

Urey, H.C., Lowenstam, H.A., Epstein, S., and McKinney, C.R., 1951, Measurement of

paleotemperatures and temperatures of the Upper Cretaceous of England, Denmark, and

the southeastern United States: Geological Society of America Bulletin, v. 62, p. 399-416.

Vasconcelos, C., McKenzie, J.A., Warthmann, R., and Bernasconi, S.M., 2005, Calibration of the δ18O

paleothermometer for dolomite precipitated in microbial cultures and natural environments:

Geology, v. 33, p. 317-320.

Veillard, C.M.A., John, C.M., Krevor, S., and Najorka, J., 2019, Rock-buffered recrystallization of

Marion Plateau dolomites at low temperature evidenced by clumped isotope thermometry

and X-ray diffraction analysis: Geochimica et Cosmochimica Acta, v. 252, p. 190-212.

Vogel, J.C., Grootes, P.M., and Mook, W.G., 1970, Isotopic fractionation between gaseous and

dissolved carbon dioxide: Z. Phys., v. 230, p. 225–238.

Ward, W.C., and Halley, R.B., 1985, Dolomitization in a mixing zone of near-seawater composition,

Late Pleistocene, Northeastern Yucatan Peninsula: Journal of Sedimentary Petrology, v. 55, p.

407-420.

Warren, J., 2000, Dolomite: Occurrence, evolution and economically important associations:

Earth-Science Reviews, v. 52, p. 1-81.

Whitaker, F.F., and Xiao, Y., 2010, Reactive transport modeling of early burial dolomitization of

carbonate platforms by geothermal convection: AAPG Bulletin, v. 94, p. 889-917.

Zachos, J.C., Stott, L.D., and Lohmann, K.C., 1994, Evolution of early Cenozoic marine

temperatures: Paleoceanography, v. 9, p. 353-387.

185

Zachos, J.C., Schouten, S., Bohaty, S., Quattlebaum, T., Slujs, A., Brinkhuis, H., Gibbs, S.J., and

Bralower, T.J., 2006, Extreme warming of mid-latitude coastal ocean during the Paleocene-

Eocene Thermal Maximum: Inferences from TEX86 and isotope data: Geology, v. 34, p. 737-

740.

186