[Thesis Title]

ABSTRACT

Transgressive reef morphology evolution:

A qualitative and quantitative comparison of

Uppermost Pleistocene and Upper Cambrian reefs

in offshore and Central Texas (USA)

by Pankaj Khanna

A better understanding of the morphological evolution of modern coralgal reefs can potentially help to decipher the forcing mechanisms that influence the evolution of ancient carbonate reefs. In this study, a comparison is made between the characteristics and interpreted evolution of Uppermost Pleistocene drowned coralgal reefs stretching along the south Texas shelf edge and Upper Cambrian microbial reefs outcropping in Central Texas. The similarities and differences between the two settings make an interesting comparison, and the description of their morphology provides “models” that can be used to better understand other reef examples. Among the major similarities, both settings represent low-latitude, shallow water settings within the photic zone and near the shelf edge, and are mixed carbonate siliciclastic systems, whereas the nature of the reef builders in each case is a main difference. A primary forcing mechanism for the modern example - sea level fluctuation – is well constrained, as the most recent transgression following the Last Glacial Maximum can be directly tied to the morphological evolution of the

Uppermost Pleistocene coralgal reefs. The initial transgression triggered a clear switch from shallow coastal siliciclastics to neritic coralgal reefs on the south Texas shelf edge banks. A similar switch is also observed at the base of the Upper

Cambrian microbial reefs where the system changes from mixed carbonate-clastic to pure carbonate system. Based on these comparisons, it is demonstrated that an

Upper Cambrian transgression led to the establishment of microbial reefs upon a broad ramp-like setting by providing enough accommodation and sufficient water quality (minimal turbidity, good circulation) for buildup growth.

Acknowledgments

गु셁र्ब्ह्र मा गु셁셍वषर ्ुԃ गु셁र्देवो महेश्वरԃ ।

गु셁रेव परंर्ब्ह्म तस्मै श्रीगुरवे नमԃ ॥

With the two lines in Sanskrit, I wish to express my sincere gratitude to my teachers. The shloka (the two lines in Sanskrit) means that a teacher takes a hand, opens a mind, touches a heart, and shapes the future. I would like to thank my advisor Prof. André Droxler for his continuous support, time, efforts, and guidance before and during my research at Rice University. His insight, advice, and confidence in my abilities have strengthened me not just as a scientist but also as a person. He has been my mentor, family, as well as friend. He introduced me to new concepts of geological science and the culture of the new country in which I arrived five years ago. I experienced and participated in technologically advanced research cruise onboard R/V Falkor with him, as well as collected high resolution data with drones for my PhD projects. My PhD experience could have never been as good as I have experienced at Rice with André. It all started in October of 2011, where I first met

André for breakfast in Mumbai, India, and that breakfast led me to the most wonderful and memorable journey of my life, ‘PhD’. Thank you André!!

I would also like to thank Dr. Paul M. (Mitch) Harris, who also mentored me throughout my PhD. He is a source of infinite knowledge and has been an inspiration and a great friend. Thanks to Dr. Jeff Nittrouer for mentoring me on one of my projects. The discussions with Dr. Nittrouer have always helped me focus in the

v right direction. Dr. Michael Pyrcz mentored me during an internship at Chevron in summer 2014 and is also on my PhD committee. His support and guidance in statistical modelling has helped me conduct and showcase my research numerically.

I learn something new every single time I have had discussion with Dr. Pyrcz. I would also like to thank Dr. Daniel Lehrmann for mentoring me on one of my PhD projects. I have learned a lot while working in the field, as well as discussion during research meetings and consortium meetings with Dr. Lehrmann. Thanks to Dr. Julia

Morgan and Dr. Janet Siefert for serving on my PhD thesis committee.

I would like to thank Dr. Vitor Abreu and Keriann Pederson, for selecting me to TA Sequence Stratigraphy class four years in a row at Rice University. The discussions with Dr. Abreu and Ms. Pederson in the class as well as in the field helped me learn and grow more outside my area of expertise. Thanks to Dr. Jeroen

Kenter and Dr. Beatriz Garcia-Fresca for the scientific discussions and for guiding, supporting me.

Most of my research was carried out on private ranches in Mason, Texas which were closed for the past 40 years. I would like to thank Mr. Don Shepard,

Rosey, Scott Zesch, Gene Zesch, and Patsy Zesch, for allowing us to carry out research on their ranches. While in the field, I became very good friend with the managers of the Shepard Ranch. Mark and PK (Mark Krauss, and Priscilla Krauss) have been like a family for me and I cannot thank them enough for their love and support. Tony, Tonya, Scott, Manny are few friends from many others in Mason, who have through these past years been great friends in Mason and provided their vi continuous support. I would also like to thank Robert and Brent for helping out in collecting the drone imagery in Mason, Texas.

I am grateful to my friends - Aakash, Deepak, Divye, Kiboy , Renuka, Sajjad,

Tarini, Tushar, Shreya Sourabh, and my labmates and friends from Rice, who have continuously supported me throughout my journey at Rice. I would especially like to thank the Earth, Environmental and Planetary Sciences staff for their endless assistance and guidance.

Thanks to many collaborators and co-authors at Rice University, Texas A&M

Corpus Christi, Trinity University, University of Bremen - Germany, University of

South Florida, Center for Tropical Marine Ecology, Bremen- Germany, University of

Genoa- Italy, Schmidt Ocean Institute, University of Maimi, University of Texas at

Austin, Chevron, Statoil, Shell, and Conocophillips. I am grateful for the scientific party and crew of R/V Falkor for mapping south Texas Banks which provided high resolution multibeam data and shallow seismic data for my research. I would also like to thank Camerawings - Robert Youens for collecting the drone imagery for one of my PhD projects.

Without endless support, patience from my wife Priyam, my brother and sister in-law, Suraj (My inspiration for a PhD, his dream- now a reality) and

Priyanka Khanna, my sister and brother in law, Swati Khanna Dahiya and Aditya

Dahiya and my parents, Rajinder and Rajnandani Khanna, none of this would have been possible. I would also like to remember my grandparents Lt. Shri Dhanraj Nagi and Lt. Shri Swaran Nagi, and Madan Lal Khanna and Lt. Kanta Khanna, who always supported me and guided me in my early life. vii

This work is dedicated to my parents Rajinder (Dad) and Rajnandani Khanna (Mom).

It was their efforts and sacrifices in life that put me on this path and enabled me to

achieve the milestone called PhD. viii

Contents

Abstract……………………………………………………………………………………………………………………..ii

Acknowledgments ...... iv

Contents ...... viii

List of Figures ...... xi

List of Tables ...... xv

List of Equations ...... xvi

General Introduction ...... xvii Coralgal Reef Morphology Records Punctuated Sea-Level Rise during Last Deglaciation ...... 1 1.1. Introduction ...... 2 1.2. Methods ...... 8 1.2.1. Radiocarbon Date Calibration ...... 8 1.2.2. Data Collection R/V Falkor ...... 9 1.2.3. Hypsometric Curves – Data Analysis ...... 9 1.2.4. Computing Paleo Terrace Depth ...... 10 1.3. Results ...... 11 1.3.1. True Coralgal Reef Morphologies ...... 11 1.3.2. Ultimate Coralgal Reef Demise ...... 14 1.3.3. Terrace Hypsometric Curves ...... 15 1.3.4. Paleo-Terrace Depths and Greenland Climate Record ...... 19 1.4. Discussion ...... 22 Submerged reef terraces in the Maldivian Archipelago (Indian Ocean): heritage of a punctuated postglacial sea level rise? ...... 24 2.1. Introduction ...... 25 2.2. Geological and ecological setting ...... 29 2.3. Material and Methods ...... 32 2.3.1. SCUBA Diving Transects ...... 32 2.3.2. Multibeam Bathymetry ...... 34 ix

2.3.3. Interpretation of Paleo RSL and uncertainties ...... 36 2.3.4. Global database of coral reef terraces ...... 41 2.4. Results ...... 41 2.4.1. SCUBA Diving Transects: identification of terraces ...... 41 2.4.2. Multibeam Bathymetry ...... 43 2.5. Discussion ...... 49 2.5.1. Depth of Maldivian Reef Terraces ...... 49 2.5.2. Paleo relative sea level ...... 52 2.5.3. A postglacial punctuated sea level rise? ...... 53 2.5.4. Inheritance from the Late Quaternary ...... 56 2.5.5. A global sequence of postglacial submerged terraces? ...... 60 Regional and Local Context of Microbial Buildups, Central Texas – A 3 phase morphological evolution...... 68 3.1. Introduction ...... 70 3.2. Geological Background and Stratigraphic Context ...... 73 3.2.1. Upper Cambrian Regional Context ...... 73 3.2.2. Upper Cambrian Local Context, Central Texas ...... 74 3.3. Data Collection, Biases, and Errors ...... 75 3.3.1. Field Survey ...... 75 3.3.2. Aerial Imagery - Data Acquisition ...... 80 3.3.3. Aerial Imagery - Data Processing ...... 82 3.4. Observation and Results ...... 94 3.4.1. Upper and Lower Point Peak, Mason County, Central Texas ...... 94 3.4.2. Upper Point Peak Microbial Buildups, Mason County, Central Texas ...... 95 3.4.2.1. Three growth phases of the upper Pont Peak microbial buildups ...... 95 3.4.2.2. The Microbial Buildup Substratum ...... 97 3.4.2.3. Phase 1: “Colonizing Phase” ...... 102 3.4.2.4. Phase 2: “Aggrading and Laterally Expanding Phase” ...... 107 3.4.2.5. Phase 3: “Capping Phase” ...... 111 3.4.3. Upper Point Peak Microbial Buildups, Macro, Meso, and Micro Scale ...... 111 3.5. Discussion ...... 114 3.5.1. Paleogeographic Setting ...... 115 x

3.5.1.1. Trade Winds ...... 115 3.5.1.2. Tides ...... 116 3.5.1.3. Storms and Waves ...... 119 3.5.2. Depositional Model ...... 119 Spatial Statistical Analysis on High-Resolution Drone Imagery: Upper Cambrian Microbial Buildups, Central Texas ...... 123 4.1. Introduction ...... 124 4.2. Geological Background and Stratigraphic Context ...... 127 4.2.1. Upper Cambrian Central Texas ...... 127 4.2.2. Upper Point Peak Microbial Unit – Wilberns Formation ...... 128 4.2.3. Three phase evolution of Upper Point Peak microbial unit ...... 129 4.3. Data Collection, Biases, and Errors ...... 132 4.3.1. Data Acquisition ...... 133 4.3.2. Data Processing ...... 133 4.3.3. Mapping 1st Phase ...... 136 4.3.4. Sources of error and Biases ...... 138 4.4. Spatial Statistical Analysis ...... 138 4.4.1. Univariate Analysis ...... 139 4.4.2. Multivariate Analysis ...... 142 4.4.3. Multi-dimensional Scaling (MDS) ...... 145 4.4.4. Grouping Analysis ...... 147 4.4.5. Ripley’s K Analysis ...... 149 4.4.6. Getis Ord Gi Statistics ...... 149 4.5. Discussion ...... 152

References ...... 154

Appendix A - South Texas banks ...... 176 Appendix B – 3D Point Clouds Upper Cambrian Microbial Buildups, Mason, Central Texas ...... 184

List of Figures

Figure 1.1 –Modern and Uppermost Pleistocene Texas Shelf where living and drowned coralgal reefs are located...... 5

Figure 1.2 – High Resolution multibeam bathymetric maps illustrate the morphology of four of the ten imagined drown coralgal reefs...... 7

Figure 1.3 – True coralgal reef morphologies as evidenced by high resolution multibeam bathymetric maps...... 12

Figure 1.4 – 3.5 kHz Chirp uninterpreted and interpreted lines acquired over Blackfish Ridge Bank...... 17

Figure 1.5 – Punctuated sea-level rise events over timescales of decades to century based on coralgal reef terrace levels and their connection with warming intervals in the North Greenland Ice Core Project climate record during last deglaciation...... 20

Figure 2.1 – The Maldivian Archipelago ...... 28

Figure 2.2 – Reef geomorphic zonation, model for paleo sea-level calculation, and sensitivity of the paleo RSL ...... 38

Figure 2.3 – Morphology of the Maldivian reefs ...... 44

Figure 2.4 – Cross-Section profiles of Maldivian Reefs from SCUBA ...... 45

Figure 2.5 – Morphological features identified using Multibeam Bathymetry of Male Island ...... 48

Figure 2.6 – Comparison of terrace depths: SCUBA diving transects and Multibeambathymetry data ...... 54

xii

Figure 2.8 – Comparing Maldivian reef terraces with global reef terrace record ...... 62

Figure 3.1 – Upper Cambrian paleogeographic, paleoenvironment, and lithofacies map of Laurentia modified from Lochman-Balk, 1970...... 72

Figure 3.2 Stratigraphic column and measured sections – Upper Cambrian, Central Texas...... 76

Figure 3.3 Arc Map 10.1 base map imagery showing the field area in Mason County, Central Texas, and locating 12 outcrops which include pavements and cliffs exposing the lower and the upper microbial units. The yellow text and lines represent cliffs, red rectangles and text represent pavements...... 78

Figure 3.4 Geological map from Sliger (1957) is corrected for the Point Peak Member (which was mapped in this study) and is overlaid on top of Arc Map 10.1 base map imagery...... 79

Figure 3.5 James River Pavement Orthophotograph showcasing different size and shapes of microbial buildups...... 84

Figure 3.6 James River Cliff Orthophotograph...... 86

Figure 3.7 – Three types of microbial buildups in Lower microbial unit: ...... 87

Figure 3.8 Orthophotograph of eastern part of Shepard Cliff exposed along Llano River...... 89

Figure 3.9 Orthophotograph of western part of Shepard Cliff exposed along Llano River...... 91

Figure 3.10 Orthophotograph of Zesch Cliff exposed along Llano River...... 93

Figure 3.11 Fallen Block – Mill Creek displaying the base of the microbial buildups...... 99

Figure 3.11 The three-phase growth model displayed by the microbial buildups explaied using Mitch Buildup, Zesch Cliff...... 100

Figure 3.13 First phase morphology and contemporaneous inter reef sediments...... 104

Figure 3.14 Phase 2 and 3 growth of the microbial buildups...... 105 xiii

Figure 3.15 – Oblique view of the James River looking south taken from a drone locating the measured sections and gamma ray log correlation ...... 109

Figure 3.16 Palaogeographic controls on depositional environments...... 113

Figure 3.17 Composite Depositional model for Upper Cambrian microbial buildups with two separate carbonate factories on a mixed system ramp .... 118

Figure 4.1 Stratigraphic column and measured sections of Point Peak Member, Wilberns Formation across Llano uplift...... 126

Figure 4.2 Microbial Buildup three phase evolution model – Mitch Herm- Zesch Cliff is used to demonstrate the three different phases based on morphology...... 131

Figure 4.3 – Digital Outcrop model of James River Pavement displaying cross section through 1st Phase of Upper Point Peak microbial buildups, Wilberns Formation...... 134

Figure 4.4 – Scaling relationships of the Phase 1 microbial buildups...... 135

Figure 4.5 – Small , Medium, and Large scale polygons mapped on James River Pavement...... 137

Figure 4.6 – Small , Medium, and Large scale polygons mapped on James River Pavement color coded on the basis of anisotropy a), and orientation, b). The frequency plots are also plotted for anisotropy and rose diagrams for orientation...... 141

Figure 4.7 Small scale polygons mapped on James River Pavement form clusters and make up the interior of the medium scale. For each cluster of small scale, a rose diagram is plotted ...... 143

Figure 4.8 Correlation matrix for Small, Medium, and Large scale polygons. Correlation matrixes display scatter plots and correlation coefficients...... 144

Figure 4.9 Multi- Dimensional Scaling analysis for Small, Medium, and Large scale polygons ...... 146

Figure 4.10 – Grouping analysis for Small, Medium, and Large scale polygons based on length, orientation, circularity, and compactness...... 148 xiv

Figure 4.11 – Ripley’s K analysis for Small , Medium, and Large scale polygons...... 150

Figure 4.12 – Getis Ord Gi Statistics- hot spot analysis for Small , Medium, and Large scale polygons...... 151

List of Tables

Table 1.1 – Estimated Paleo terrace depths and their age range. (see details in the text and methodologies) ...... 18

Table 2.1 – Depth of terraces in the Maldivian archipelago, calculation of paleo relative sea level and values used to estimate sediment coverage and paleo water depth...... 55

Table 2.2 – Timing of the main decelerations in postglacial sea level rise rates and their duration (red lines in Fig. 2c, upper panel) corresponding to reef terrace levels found in this study, calculated paleo RSL (from Table 2) and known climatic or sea level events...... 58

Table 2.3 – Calculation of departure from eustasy due to tectonics and GIA in the Maldivian archipelago for RSL indicators formed 6 and 21 ka BP...... 66

Table 4.1 – Amall , Medium and Large Scale - P10, P50, and P90 values of nine attributes (width, length, area, perimeter, orientation, anisotropy, circularity, eccentricity, and compactness) are listed...... 140

List of Equations

Equation 2.1 – Paleo RSL ...... 36

Equation 2.2 – Total Uncertainty ...... 37

General Introduction

Comparative sedimentology between modern and ancient sedimentary systems is the key to better understanding these systems. Low-latitude mixed carbonate-siliciclastic settings are one of the most complex sedimentary systems, which are generally characterized by lateral juxtaposition and vertical stacking of neritic carbonates and shelfal siliciclastics (Droxler and Jorry, 2013). The distribution of both carbonate and siliciclastic sediments varies over time spatially along mixed margins with respect to climate and sea level changes. The two major sediment sources in mixed systems include terrigenous derived siliciclastics, muds and sands products of weathering and erosion of the hinterland, and neritic biogenic carbonates. Contrary to transported siliciclastics, neritic carbonate sediments are usually biogenically produced in situ, “born and not transported” (James, 1978).

Hence, interrogating the biologically constructed (physically and chemically modified) morphological evolution of modern carbonate reefs could help decipher the forcing mechanisms for the ancient carbonate reefs, resulting in better understanding ancient mixed systems.

This research focuses on two mixed systems (modern and ancient), the

Uppermost Pleistocene drowned coralgal reefs along the south Texas shelf edge, northwestern Gulf of Mexico (Belopolsky and Droxler, 1999; Droxler and Jorry,

2013) and the Upper Cambrian microbial buildups along the western edge of the

Great American Carbonate Bank (GACB; Bridge et al., 1947; Ahr, 1967, 1971; Miller et al., 2012; Morgan, 2012). The major difference in the two study areas is the

xviii different nature of the reef builders. This study compares uppermost Pleistocene drowned coralgal reefs along the south Texas shelf edge with upper Cambrian microbial reefs in Mason, Texas. However, these marine reef builders, although 500 million years apart, have been influenced by similar parameters including sunlight, latitude, and temperature. Although salinity plays an important role in the growth of modern coralgal reefs, it does not affect modern microbial settings (Reid et al.,

2003). Environmental controls such as sea level, wind, waves, tides, and storms have also been similar in modern and ancient settings, although the influence of these environmental controls can vary depending upon the paleogeographic setting.

In the Cambrian, plants had not yet evolved and the continents were devoid of vegetation. The degree of chemical and physical weathering, therefore, was quiet different on exposed lands and influenced the flux and the nature of the siliciclastics to the Great American Carbonate Bank (GACB). Glauconite occurrence was very common in the Cambrian sediments of Central Texas though restricted to the modern outer shelves (Chafetz, H.S. and Reid, A. 2000). In the Cambrian, oceans became more oxygenated, but most likely not as oxygenated as modern oceans.

Forcing mechanisms such as sea level fluctuations are directly tied to the evolution of modern reefs. This study clearly showcases how punctuated sea level had a direct influence on the morphological evolution of the Uppermost Pleistocene coralgal reefs on the south Texas shelf edge. During the latest sea level transgression following the Last Glacial Maximum (LGM, 20 Kyr), a switch from shallow coastal siliciclastics to neritic carbonates is observed at the bases of south Texas shelf edge drowned banks (Belopolsky and Droxler, 1999; Droxler and Jorry, 2013). A similar xix switch also occurred in the rock record at the base of the Upper Cambrian microbial reefs in Mason County where the system changes from a shallow subtidal to intertidal mixed carbonate-clastic to a subtidal pure carbonate system during an initial sea level transgression. Then these microbial reefs grew in a systematic 3- phase evolution influenced by a sea level transgression providing enough accommodation to allow their vertical growth.

Chapter 1

Coralgal Reef Morphology Records Punctuated Sea-Level Rise during Last Deglaciation

The following chapter is a reproduction of Khanna, P., Droxler, A. W.,

Nittrouer, J.A., Tunnell, J.W. Jr., Shirley, T.C. from Nature Communications October

2017.

Abstract

Coralgal reefs preserve the signatures of sea-level fluctuations over Earth’s history, in particular since the Last Glacial Maximum 20,000 years ago, and are used in this study to indicate that punctuated sea-level rise events are more common than previously observed during the last deglaciation. Recognizing the nature of past sea-level rises (i.e., gradual or stepwise) during deglaciation is critical for informing models that predict future vertical behavior of global oceans. Here we

2 present high resolution bathymetric and seismic sonar data sets of ten morphologically similar drowned reefs that grew during the last deglaciation and spread 120 km apart along the south Texas shelf edge. Herein six commonly observed terrace levels are interpreted to be generated by several punctuated sea- level rise events forcing the reefs to shrink and backstep through time. These systematic and common terraces are interpreted to record punctuated sea-level rise events over timescales of decades to century during the last deglaciation, previously recognized only during the late Holocene.

1.1. Introduction

Coralalgal reef establishment and evolution during the last deglaciation have been well documented through chronological, sedimentological, and paleontological studies, and provide unique data sets upon which past sea-level records have been reconstructed (Fairbanks, 1989; Bard et al., 1990a; Bard et al., 1996; Alley et al.,

1997; Yokohama et al., 2000; Hanebuth et al., 2000; Lambeck et al., 2004; Alley et al.,

2005; Bird et al., 2007; De Deckker and Yokohama, 2009; Bird et al., 2010; Bardand

Delanghe-Sabatier, 2010; Deschamps et al., 2012; Camoin et al., 2012; Lambeck et al., 2014) Most of these records display, since the Last Glacial Maximum (LGM), several major intervals of rapid sea-level rise over timescales of several centuries, referred to as Melt-Water Pulses, in the uppermost Pleistocene.

Since the early 1930’s (Trowbridge, 1930), several deep banks, with crests lying in about 60 mbsl, were known to occur along the south Texas shelf edge (Fig. 3

1.1). The coralgal origin of the banks was first proposed in the mid-1970’s (Rezak and Bright, 1976; Bright and Rezak, 1976) based on five banks from which rock samples were collected by piston coring, dredging, box coring, and Van Veen grab.

The rocks consist mainly of dead corals (Agaricia sp., Madracis sp., Madracis asperula, Madracis brueggemanni, Madracis myriaster, and Paracyathus pulchellus) and coralline algal nodules. Only two samples were dated; a coral sample from the top of Dream Bank at 68 m yielded a radiocarbon age of 10,580 ± 155 years BP

(11,901.5 ± 335.5 calendar years BP), and a coralline algal sample from the base of

Southern Bank produced a radiocarbon age of 18, 900 ± 370 years BP (22,361 ± 428 calendar years BP; Rezak and Bright, 1976; Bright and Rezak, 1976). In late 1990’s, a multi-channel seismic grid on one of the reefs, Southern Bank, indicates the thickness of the bank to be about 40-50 m (Belopolsky and Droxler, 1999). It is also concluded that the drowned banks along the south Texas shelf edge were established on paleo highs associated with antecedent siliciclastic topographies such as either beach barrier islands or beach ridges developed during late LGM or earliest deglaciation (Belopolsky and Droxler, 1999). In absence of detailed chronologic dates and based upon the current water depth range of these bank tops at about 60 mbsl, the demise of these reefs was proposed to have occurred between

~12,250 to 11,500 Cal BP. Recent studies show that during the LGM, the south Texas coastal system consisted of a bay bounded by the Rio Grande and Colorado lowstand shelf edge deltas, isolated from the open ocean by a barrier island complex (Weight et al., 2011; Fig. 1.1B). The coralgal reefs likely established themselves on top of this lowstand coastal system, thrived, and grew vertically in less than ~ 8,000 years by 4 tracking the 40-50 m of sea-level rise during the uppermost Pleistocene (Belopolsky and Droxler, 1999; Weight et al., 2011). Ultimately, the south Texas reefs drowned and, starting at ~ 9 ka, were subsequently partially buried by the Holocene Texas

Mud Blanket (Rezak and Bright, 1976; Bright and Rezak, 1976; Belopolsky and

Droxler, 1999; Weight et al., 2011; Droxler and Jorry, 2013; TMB).

The observed 40-50 m vertical accretion of the coralgal banks in about 8,000 years suggests average rates of sea-level rise of 5-6 meters per millennium

(Fairbanks, 1989; Bard et al., 1990a; Bard et al., 2010; Deschamps et al., 2012). This pace could have occurred only with the occurrence of scleractinian coral species, including Acropora palmata and Acropora cervicornis, which display unusually fast growth rates and create the main coral framework of the Caribbean reefs

(Gladfelter, 1978). Although these species are not currently growing at the latitudes of the northern Gulf of Mexico (GoM), except for a few colonies of A. palmata newly established at the Flower Garden Banks (FGB) (Fig. 1.1A) in the past decade

(Zimmer et al., 2006), it is assumed that these species formed the coral framework of the south Texas shelf edge drowned banks. This assumption is bolstered by the recent discovery that A. palmata and A. cervicornis grew in large numbers at the base of the FGB (Precht et al., 2014) as early as 10,200 cal BP, based on radiocarbon dating. The occurrence of these coral species as early as the earliest part of the

Holocene in the northern GoM strengthens the inference that they most likely form the coral framework of the uppermost Pleistocene south Texas shelf drowned banks. Additionally, modern and presumably deglacial near surface

Figure 1.1 –Modern and Uppermost Pleistocene Texas Shelf where living and drowned coralgal reefs are located.

(A) Drowned banks along the south Texas shelf-edge are shown as red dots. The

Flower Garden Banks, living coralgal reefs 150 km south of Galveston Island, are displayed as green dots. Note that, with one exception, the drowned reefs are currently located between 70 and 80 m isobaths. (B) Uppermost Pleistocene south

Texas shelf coastal systems at 20 and 17 ka20 (located in Fig. 1.1A by red rectangle), illustrate a shallow bay up to 35 m deep, isolated from the open ocean by a barrier island complex, on top of which the south Texas reefs (shown in red dots) were established during the early part of last deglaciation.

circulation patterns in the GoM show that it is and was responsible for carrying biotic communities into the GoM from the Caribbean (Rezak et al., 1983).

It has been established, that carbonate production areas shrink through back-stepping so to remain within the photic zone when responding to sea-level rise; as such coralgal reefs form distinct sets of terraces (Schlager, 2006) as they grow vertically keeping up with sea-level rise. During transgressions, therefore, episodic and rapid sea-level rise events result in set of terraces, preserving the nature of sea-level rise and diagnostic morphological features of reefs struggling to keep-up with rising sea-level (Schlager, 2006; Chappell and Polach, 1991; Blanchon and Shaw, 1995; Clark et al., 2001; Peltier and Fairbanks, 2006). Ultimately, when the area of carbonate production has shrunk to a minimum through systematic backstepping, reefs are unable to grow vertically fast enough to keep-up so as to remain within the photic zone and reefs ultimately drown (Schlager, 2006; Chappell and Polach, 1991; Blanchon and Shaw, 1995; Clark et al., 2001; Peltier and

Fairbanks, 2006; Neumann and Macintyre, 1985). The edifices of drowned reefs sit below the euphotic zone, as the series of drown banks along the south Texas shelf edge, which are no more vertically accreting, although their crests are still covered by live ahermatypic wire corals, sea-fans, mollusks, annelids, bryozoans, and red algae, and are known to be excellent fishing grounds (Nash et al., 2013).

The new data presented here provides an opportunity to quantify well- imaged back-stepping terraces and identify nature of the sea-level rise during last 7

Figure 1.2 – High Resolution multibeam bathymetric maps illustrate the morphology of four of the ten imagined drown coralgal reefs.

(A) Baker Bank, (B) Dream Bank, (C) Blackfish Ridge, and (D) Southern Bank (see

Fig. 1.1A for their geographical locations). Spurs (ridges) and grooves (troughs) are usually identified on the eastern (windward) side of the banks; the red rectangle in

Fig. 1.2A locates well-developed spurs and grooves shown in Fig. 1.3A. Dream Bank in Fig. 1.2B represents distinct atoll morphologies, with well-developed lagoons surrounded by rimmed margins (see Fig. 1.3B). Dream Bank also clearly displays a set of distinct terraces, illustrated in Fig. 1.3D and 1.E. Red lines on Fig. 1.2C locate the three 3.5 kHz seismic lines (1, 2, 3 ) shown in Fig. 1.4. 8 deglaciation leading to the development of common backstepping morphologies.

High resolution multibeam mapping and seismic profiling of ten drowned banks, located along a 120 km-long stretch of the south Texas outer shelf, identify six common terrace levels; these identical morphologies provide new opportunities to understand coralgal reef evolution through backstepping and terrace formation, most likely triggered by decade to century-long punctuated sea-level rise during the middle part of last deglaciation. Existing sea-level records do not have the ability to resolve these smaller amplitude variations. Hence, it is pertinent to investigate geological records that directly document spatiotemporal sea-level changes to determine if decadal to century-scale sea-level rise episodes are common occurrences.

1.2. Methods

1.2.1. Radiocarbon Date Calibration

Calib Rev 7.0.4 was used to calibrate the radiocarbon ages collected in 1970’s

17, 20, 21. Calibration dataset marina13.14c is used with Delta R = -30 ± 9. The new calibrated calendar year ages are 11,901.5 ± 335.5 calendar years BP for the top of

Dream Bank and 22,361 ± 428 calendar years BP for the base of the Southern Bank.

These ages are not incorporated into the age model but are used to only indicate that these reefs grew during last deglaciation. 9

1.2.2. Data Collection R/V Falkor

During a 15-day research cruise in September 2012 onboard R/V Falkor, funded by the Schmidt Ocean Institute, high-resolution multibeam sonar datasets and 3.5 kHz seismic data were acquired over 10 drowned carbonate reefs on south

Texas shelf edge. The research vessel was equipped with state-of-the-art instrumentation, including a Kongsberg EM 710 multibeam echo sounder to collect high resolution bathymetric maps (< 1 m) and a high-resolution seabed mapping

(3.5 kHz seismic) system, Knudsen CHIRP 3260, to image the sub sea-floor sedimentary units. The ancillary components of the multibeam system included:

SeaPath 320 heading, attitude, and positioning sensor, CNAV positioning correction service and Valeport SV profiler. Multibeam data was processed using Caris 7.1 and further imported to Arc G.I.S. to build and investigate the bathymetric maps of these banks. The CHIRP data was analyzed utilizing the Echo Post Survey software.

1.2.3. Hypsometric Curves – Data Analysis

Hypsometric curves are generated for nine of the ten drowned coralgal banks

(Fig. 1.5A). Bathymetry data is clipped into sub-data sets encompassing each individual drowned bank. The total surface area of each bank is divided into one meter-depth intervals and the surface area of each interval is calculated by counting the number of pixels (at one square meter). To create a hypsometric curve for each individual bank, the percentage of each depth interval is determined using their calculated surface area. Each peak in the hypsometric curves represent individual terrace. When the nine hypsometric curves are plotted together, overlapping peaks 10 identify common terrace depth zones. For each common terrace depth zone, a median terrace depth is determined. Depth uncertainties are evaluated as the difference between the median terrace depth values and their maximum or minimum depth range.

1.2.4. Computing Paleo Terrace Depth

In the absence of chronologic dates, the current depths of the terrace zones, ice volume sea-level curve (Lambeck et al., 2014), in addition to GIA (Milne and

Mitrovica, 2008) and subsidence rates (Winker, 1979) are used to estimate paleo terrace depths. First, the terrace depth zones (with depth uncertainties) are compared with an ice-volume sea-level curve (Lambeck et al., 2014) to identify the age range (including uncertainties) for the development of each terrace zone. To estimate the paleo terrace depth, depth change for each terrace due to GIA and subsidence is calculated by multiplying the estimated age range of each terrace depth zone with avg. rate of GIA (0.71 mm per year; Milne and Mitrovica, 2008) and subsidence (0.5 mm per year; Winker, 1979). The total change in depth is calculated by adding GIA and subsidence (uplift is considered positive and subsidence is considered negative). The estimated total depth change is added to the current terrace depth to identify the paleo terrace depth for each terrace. The paleo terrace depths are further compared with an ice-volume sea-level curve (Lambeck et al.,

2014) to estimate the age range for the development of each paleo terrace. 11

1.3. Results

1.3.1. True Coralgal Reef Morphologies

Multibeam bathymetric mapping and 3.5 kHz seismic profiling, acquired in

September, 2012, onboard the R/V Falkor (Fig. 1.2), showcase the detailed morphological architecture of the south Texas shelf edge drowned banks. Spurs and grooves, typical morphological adaptations to high energy inner fore reef conditions

(Gischler, 2010; Duce et al., 2016;Fig. 1.2 and 1.3A), are preferentially observed in the high-resolution bathymetry on the eastern margins of several mapped banks and, therefore, coincident with their windward high-energy sides; on their protected western lee sides, these features are conspicuously absent. In mid-1970, spurs and groves were already observed, aligned perpendicular to the slope of the bank, by submarine operations using DRV Diaphus (Lindquist, 1978). Moreover, the new data presented here provides an opportunity to quantify well-imaged backstepping terraces (Fig.1.3D), defined as flat areas bounded by steep slopes, common in nine of the ten surveyed banks (Fig. 1.2 and 1.3D). These terraces, separated by

1-2 m high cliffs of coralgal reef rock as was previously observed using submersibles

(Lindquist, 1978), are quantitatively analyzed based on multibeam data (Fig. 1.3).

Additionally, Dream Bank displays narrow rimmed margins enclosing shallow lagoons at two different back-stepping terrace levels, typical coralgal atoll morphologies (Fig. 1.3B and 1.3C).

Figure 1.3 – True coralgal reef morphologies as evidenced by high resolution multibeam bathymetric maps.

(A) Detailed bathymetry (contour interval 0.5 m) displays a clear example of spurs and grooves on the south-eastern margin of Baker Bank (see Fig.1.2A for location),

(B) Picture of modern spurs and grooves in front of the Belize Barrier Reef east of

Tobacco Range, (Photo by- Brandon Martin) as an analog for Fig. 1.3A; note the similar scales between both fossil and modern spur and grove sets. (C) Oblique bird eye view of Dream Bank displays clear atoll morphologies, rimmed margins enclosing shallow lagoon, at two different levels. (D) Side view of Dream Bank (VE:

20x) displays a series of terraces, characteristic morphology of coralgal reef 13 backstepping in response to punctuated high rates of sea-level rise. (E) Slope angle map for Dream Bank clearly identifies the well-defined terraces and cliffs (shown in

Fig. 1.3D), where red color (steep slopes) represents terrace faces and green color

(gentle slopes) terrace flats. (F) Plotted depths to crest of the ten drowned banks, eight of which lie within a 3-4 m depth range from 57.5 – 61.8 mbsl. Such a narrow depth range testifies to their contemporaneous demise.

1.3.2. Ultimate Coralgal Reef Demise

The 58-61 mbsl depth range in which the crests of eight of the ten drowned coralgal reefs occur, point to their contemporaneous demise (Fig. 1.3F).

Furthermore, this depth range coincides with stranded paleo-shorelines and subtidal shoal complexes observed in the GoM (~ 58 mbsl; Locker et al., 1996),

Caribbean (~57 mbsl; Carter et al., 1986), and Southwest Pacific (~56 mbsl; Carter et al., 1986). These paleo-shorelines and shoals are interpreted to have been abandoned by an ~ 11.5 ka event of rapid rise in global sea-level, linked to the onset of MWP-1B occurring at the end of the Younger Dryas (Fig. 1.5B; Fairbanks, 1989;

Lambeck et al., 2014). It is hypothesized, therefore, that the final demise of the south

Texas drowned banks was triggered by the MWP-1B, at ~11.5 ka. The coralgal reefs could not keep up (Schlager, 2006; Neumann and Macintyre, 1985) with the rapid rise in sea-level because their carbonate production surface areas had shrunk to a minimum through systematic backstepping, as an overall response to the last deglaciation sea-level transgression.

Stressors other than sea-level rise can negatively affect coralgal community growth, such as fluctuations in water turbidity, temperature, and salinity. However, siliciclastic sediment influx into the south Texas shelf edge was minimal during the uppermost Pleistocene transgression (Weight et al., 2011), when coastlines migrated landward. Initial burial by the TMB was initiated at ~ 9 ka, and thereby post-date reef drowning by ~ 2.5 ka. Temperature and salinity likely did not trigger the widespread collapse of the south Texas banks. During the Younger Dryas, sea 15 surface temperatures dropped only by ~ 1.50C to reach 260C, and sea surface salinity increased from 34 to 36.5 parts per thousand in the northern GoM (Flower et al., 2004). These nominal changes likely did not modify the coralgal reef ecology because during the time period of reef development, sea surface temperatures and salinity are estimated to have fluctuated with an even greater magnitude, between

25 0C and 29 0C, and 34 to 38 parts per thousand, respectively (Flower et al., 2004).

1.3.3. Terrace Hypsometric Curves

Hypsometric curves, generated from eight banks, identify sets of backstepping terraces at uniform water depths, within a range spanning 75 to 60 mbsl

(Fig. 1.5A). Four individual terraces are identified at: 74 ± 1, 70.5 ± 1.5, 66.5 ± 1.5, and 63 ± 1 mbsl. The terraces are separated by 2-4-m high steep face (Fig. 1.5A). As imaged in 3.5 kHz seismic lines, a fifth well-developed common terrace, buried by the TMB, is identified at 82 ± 1 mbsl (Fig. 1.4, 1.5A). Moreover, a sixth terrace was mapped at 94 ± 1.5 mbsl only on Harte Bank - (the deepest bank with an exposed crest and base at 82 and 102 mbsl respectively), discovered during the 2012 research expedition aboard the R/V Falkor cruise. Because both subsidence and glacio-isostatic adjustment (GIA) rates are assumed to be identical along this 120 km of the south Texas shelf, the observed five common terrace depth ranges can be considered coeval. Despite the absence of systematic chronologic dates for each of the terraces, their consistent depth ranges, among several reefs growing over such this long stretch of the south Texas shelf edge, are interpreted to reflect 16 contemporaneous and systematic back-stepping linked to punctuated sea-level rises.

Sea-level changes are dependent on ice sheet growth and decay, tectonics, and sediment overloading of the shelf and vary in different parts of the world, referred to as relative sea-level (RSL). RSL curves incorporate eustatic sea-level

(ESL) fluctuations and it is usually difficult to separate the two (RSL and ESL). The

Northwestern GoM is an ideal location for which RSL drivers and their amplitudes are well constrained and provides the opportunity to examine ESL signals. The two main drivers for RSL change in northern GoM are GIA (0.71 mm per year of uplift since past 21,000 Cal years BP; Milne and Mitrovica, 2008), and subsidence (0.5 mm per year from past 21,000 years; Winker, 1979). Considering a linear rate for GIA and subsidence for the last deglaciation, and the current depth of the terraces on drowned banks, the depth of each terrace is re-calculated and used as indicator of

ESL (Table 1.1). The corrected depths are compared with an ice volume sea-level curve (Lambeck et al., 2014) to calculate the corrected age range with uncertainties, during which each of the six terraces was developed. These calculations are based on two assumptions: the GIA and subsidence rates are linear, and the development of terraces occurred at sea-level. The uncertainties associated with the age model are dependent upon the GIA, subsidence, and relation of paleo water depth to terrace depth. Calculating GIA and subsidence with corrected age model demonstrates that the deviation in the GIA and subsidence are less than three percent. Further, atoll and spur-groove morphologies, clearly observed in the high

Figure 1.4 – 3.5 kHz Chirp uninterpreted and interpreted lines acquired over

Blackfish Ridge Bank.

The three seismic lines are located in Fig. 1.2C. Coralgal reef is colored in blue, the

Texas Mud Blanket (TMB), partially burying the reef, in brown; yellow arrows point to clear onlaps on the reef by the TMB. Three terrace levels, the deepest at 82 mbsl and two shallower terraces at 74-75 and 69-71 mbsl, are identified and are marked by black rectangles.

Inferred age Depth of Inferred age from range from Subsidence Total Depth Paleo Terrace Terrace Terrace Depth GIA (m) Paleo Terrace (m) Change (m) Depth (m) (m) (Cal years BP) Depth (Cal years BP)

59.25 ± 1.75 11,375 ± 275 8.07 ± .20 -5.68 ± .13 2.38 ± .33 61.63 ± 2.08 11,200 - 12,400 (Drowning)

63 ± 1 12,300± 200 8.73 ± .14 -6.15 ± .1 2.58 ± .24 65.58 ± 1.24 12,550 - 12,800

66.5 ± 1.5 12,750± 150 9.05 ± .11 -6.37 ± .07 2.67 ± .18 69.17 ± 1.68 12,900 - 13,150

70.5 ± 1.5 13,150± 150 9.33 ± .11 -6.57 ± .07 2.76 ± .18 73.26 ± 1.68 13,300 - 13,550

74 ± 1 13,475± 75 9.56 ± .05 -6.73 ± .03 2.82 ± .09 76.82 ± 1.09 13,650 - 13,800

82 ± 1 14,025± 25 9.94 ± .02 -7 ± .01 2.94 ± .03 84.94 ± 1.03 14,050 - 14,150

94 ± 1.5 14,400± 100 10.22 ± .07 -7.2 ± .05 3.02 ± .12 97.02 ± 1.62 14,450 - 14,550

Table 1.1 – Estimated Paleo terrace depths and their age range. (see details in the

text and methodologies)

19 resolution bathymetric data sets, indicate that the reefs, when flourishing, were keeping up with sea-level.

1.3.4. Paleo-Terrace Depths and Greenland Climate Record

The corrected depths of the observed six common terrace levels, identified on the south Texas banks, are projected onto a global eustatic sea-level curve (Fig.

1.5B; Lambeck et al., 2014) and their equivalent ages with uncertainties are estimated based on these projections. Then, these ages with their associated uncertainties are projected onto the NGRIP δ18O record (North Greenland Ice Core

Project, 2004; Steffensen et al., 2008). This climate record from Greenland is, to our knowledge, the only existing high resolution upper Pleistocene climate record during which the six terraces were formed along the South Texas Shelf. The comparison of both records is the only possible opportunity to attempt to understand the cause and effect relationship between warm climate intervals, As observed in Fig. 1.5B, out of the six terraces, four terraces correspond to warm interstadials, one to a stadial-interstadial transition, and one only to a cold stadial.

The NGRIP δ18O record represents climate variations in Greenland. Fig.1. 5B further illustrates that the number of occurrence of the terrace depth zones are similar to the number of warm events observed on NGRIP δ18O record. Moreover, the correlation of each terrace to a warm interstadial period, with one exception, is noteworthy. These warm periods, therefore, can further be linked to ice-sheet/ice- stream collapse events, causing rapid sea-level rise events of the orders of few 20 meters per century, which leads to the development of common terrace morphologies on south Texas shelf banks.

(A) Hypsometric curves for nine south Texas shelf drowned banks, based on high resolution multibeam bathymetry data and 3.5 kHz seismic lines, identify the occurrence of a series of terraces common to the banks. Each depth range (mbsl) of the four common shallower terraces are based on the multibeam data sets, green

(63 ± 1m), blue (66.5 ± 1.5m), purple (70.5 ± 1.5m), and orange (74 ± 1m). Two of those terraces (purple and orange), in addition to a deeper one at 82 ± 1m, are also identified on the 3.5 kHz seismic lines (Fig. 1.4). An additional terrace level is identified on the multibeam map of Harte bank (deepest bank): grey (94.5 ± 1.5 m). 21

(B) Corrected terrace depths (Table 1.1) with uncertainties are projected onto the

Ice-Volume equivalent sea-level curve15. The age equivalent of each terrace is projected onto the NGRIP δ18O record41, 42 with associated uncertainties (red vertical bands). The Younger Dryas interval is represented by a blue band. Melt

Water Pulses (MWP) 1A and 1B are represented by vertical light red and yellow bands.

1.4. Discussion

In absence of correct chronologic dates, the formation of these terraces, common to ten coralgal reefs, located along a distance of over 120 km on the south

Texas shelf edge, indicates that during the recent peak deglaciation sea-level did not always rise gradually, but rather was characterized by a series of punctuated and rapid sea-level rise events over decades to one century, previously only recognized during late Holocene (Lidz and Shinn, 1991; Törnqvist et al., 2012). Because climate warming and resulting ice-sheet collapses have been predicted for the future decades and centuries (Bamber and Aspinall, 2013; Deconto and Pollard, 2016), the steady and gradual sea-level rise, observed over the past two centuries may, therefore, not be a complete characterization of how sea-level would rise in the future. Furthermore, there is a scientific need to utilize advanced technologies, including high-resolution bathymetry systems combined with systematic drilling of reefs and accurate dating techniques, this study serves as a guide to future research endeavors that seek to inform sea-level rise rate and amplitude projections.

Researchers that model sensitivity of sea-level fluctuations – past and present – require as much information as possible regarding smaller amplitude events, and the best place to find this information is from the geological record. The documentation of decades to century-scale punctuated sea-level rise events with magnitudes of meters implies that deglaciation and associated sea-level rise is a non-steady process. Rate of sea-level rise has been observed to accelerate since the past two decades (Chen et al., 2017), therefore, these results have significant 23 implications for the community of science researchers that examine sea-level rise past and present, and for how society prepares for coastal flooding and inundation hazards in the coming decades to centuries.

Chapter 2

Submerged reef terraces in the Maldivian Archipelago (Indian Ocean): heritage of a punctuated postglacial sea level rise?

The following chapter is a reproduction of Alessio Rovere, Pankaj Khanna,

Carlo Nike Bianchi, André W. Droxler, Carla Morri, David F. Naar (in review) from

Geomorphology.

Abstract

Sea level changes have shaped the world’s carbonate platform margins and continental shelves, leaving typical geomorphic imprints. One of these are drowned reef terraces. In this paper, we present the results of 112 scuba diving transects across seven different Maldivian atolls and one multibeam survey around Malé

Island, where we observed the occurrence of drowned reef terraces down to 120 m 25 depth. We identified six submerged terraces that we consider as paleo sea level indicators. We compare the terrace depths in the Maldives with a database of submerged reef terraces from other 51 globally distributed sites. Our results show that it is challenging to identify common global levels, as often suggested in literature. We argue that this is due to most published studies not providing information on: i) errors in the estimation and calculation of paleo sea level from the measured depths of terraces; ii) uncertainty in the estimates of vertical tectonic displacement; and iii) quantification of glacial isostatic adjustment effects (GIA).

After testing the sensitivity of our study area to GIA and tectonics, we compare the relative sea level calculated from the drowned reef terraces in the Maldives with rates of sea level rise since the LGM, and we propose that they are consistent with periods of deceleration of the postglacial sea level rise.

2.1. Introduction

Since the early work of Darwin (1842), coral reefs have captured the interest of both ecologists and geologists, who attempted to explain the mechanisms of their formation (Braithwaite et al., 1973) and morphology (Stoddart, 1969a), the relationships between ecological and geomorphological features (Lasagna et al.,

2010a), and the inheritance of landforms from past sea levels (Schlager, 2005;

Montaggioni and Braithwaite, 2009). The Maldives represent the archetype of an atoll reef archipelago (Naseer and Hatcher, 2004) and, as such, are a good example of modern atoll morphology (Purdy and Bertram, 1993; Aubert and Droxler, 1996,

1992; Risk and Sluka, 2000; Belopolsky and Droxler, 2004; Morri et al., 2015). 26

Recent geomorphological studies on the Maldivian reefs were aimed to constrain their Holocene/Uppermost Pleistocene inherited features (Gischler et al.,

2008), the role of environmental factors (such as wave energy or coral growth) in shaping island morphology (Kench et al., 2006, 2009), the development of sub-aerial karst morphologies during intervals of low sea level (Colantoni et al., 2003), and the effects of human impacts on coral reef geomorphology (Brown and Dunne, 1988).

Similar to other continental shelves, both in tropical (Blanchon, 2011) and non-tropical (Rovere et al., 2011; Zecchin et al., 2015) areas, earlier studies reported that the Maldives are characterized by submerged terraces that range from few meters below present sea level to 130 m depth (Bianchi et al., 1997; Anderson,

1998; Colantoni et al., 2003; Fürstenau et al., 2009; Rufin-Soler et al., 2013). Some of these submerged reef terraces, both in the Maldives and in other tropical areas

(Blanchon and Jones, 1995) have been often regarded as formed by either reef catch-up or marine planation during periods of deceleration or pauses of the postglacial sea level rise, followed by meltwater pulses (Green et al., 2014; Liu et al.,

2015) that caused sudden rate increases in the rate of sea level rise.

Several questions arise regarding submerged reef terraces. In the Maldives, are the depths of individual terraces widespread in the archipelago? How does the depth of a terrace correlate with the paleo sea level that shaped it? Is it possible to correlate depths of the reef terraces found in the Maldives with other reef terraces worldwide? And, is there a relationship between reef terrace depths and periods preceding meltwater pulses or other periods of slow deceleration of postglacial sea 27 level rise such as those identified by Lambeck et al. (2014)?

In order to investigate the aforementioned questions, we analyze a database of 112 scuba diving transects in 7 different Maldivian atolls. We integrate our scuba surveys with the results of a high-resolution multibeam survey acquired around

Malé Island upper slope, part of the southeastern rim of North Malé atoll. Firstly, we propose how paleo sea level should be calculated from the measured terrace depths.

Then, we compare the results of our surveys against a database of submerged coral reef terraces reported by different authors worldwide and against eustatic postglacial sea level rise rates (Lambeck et al., 2014). Finally, we compare their depths to estimates of paleo sea levels during former interglacial.

We consider that correlating reef terraces among different regions is not straightforward, but we show that each terrace level identified in the Maldives correlates well to multi-centennial periods when the postglacial sea level rise was decelerating. We suggest that the terrace depths identified in this study need chronologic constraints, and should be regarded as first-priority targets to unravel potential rapid eustatic sea level changes during the postglacial sea level rise.

Figure 2.1 – The Maldivian Archipelago

(a)The black dots represent the location of the 112 SCUBA diving surveys, yellow dots show the location of the transects shown in Fig. 2.4; (b) The North Malé atoll;

Google, Landsat, DigitalGlobe, CNS/Astrium. 29

2.2. Geological and ecological setting

The Maldivian Archipelago consists of a double chain of 22 atolls stretching over more than 800 km from 7°04' N to 0°48' S and centered around 73° E in the eastern equatorial Indian Ocean (Fig. 2.1a). The Maldivian atolls represent the top of one of the largest modern carbonate platforms and constitute the central and largest part of the Chagos-Laccadives ridge (Risk and Sluka, 2000). As the Maldives platform has been far from any terrigenous influence for its 50 Ma-long history, this

2-3 km-thick edifice is composed entirely of carbonate sediments. These modern atolls are part of the latest phase in the evolution of this platform, which initially established itself on top of an early Eocene subsiding volcanic plateau (Aubert and

Droxler, 1992, 1996; Purdy and Bertram, 1993; Belopolsky and Droxler, 2004).

Three distinct intervals have been identified in the Cenozoic stratigraphic evolution of the Maldivian carbonate system, corresponding to the Palaeogene

(Eocene to late Oligocene), the Neogene (early Miocene to early Pliocene), and the late Pliocene-Pleistocene (Aubert and Droxler, 1996). High-frequency eustatic and climatic changes in the last 3 Ma triggered increased karstification of the reef framework, which is speculated to form the morphology inherited by the modern reefs (Purdy and Bertram, 1993; Colantoni et al., 2003; Gischler et al., 2014). The

Holocene history of Maldivian (among other Indo-Pacific) coral reef systems has been summarized by Montaggioni (2005), mostly based on Woodroffe (1992, and references therein). The degree of modern reef development appears to be linked to coral community structure. Communities consisting principally of branching and 30 domal corals underwent substantial accretion and produced well-developed reefs, whereas assemblages comprising of foliaceous and encrusting corals produced only incipient reefs. The highest accretion rates of frameworks recorded (up to 20 mm·a-1) related to tabular and arborescent acroporids.

At the Last Glacial Maximum (LGM), from 23 to about 19 ka BP, reefs only developed along what were to become the fore slopes of present reefs, forming accumulations a few meters thick at vertical rates of up to 1 mm per year. The rapid postglacial rise in sea level, from about 19 to 6.5 ka BP, was accompanied by the settlement of three successive reef generations, within the periods 17.5-14.7, 13.8-

11.5 and 10 ka BP to the Present. From the LGM to the early Holocene, coral settlement has probably declined (Montaggioni, 2005).

Holocene reef growth started as early as 8.5 ka BP (Gischler et al., 2008).

Marginal reefs, dominated by robust branching corals and coralline algae, accreted in the keep-up mode with rates of 15 m·ka-1. Rate of sea level rise slowed significantly from 7-6 ka BP and subsequently gradually rose with rates of 1 m·ka-1.

Lagoon reefs, characterized by domal corals and detrital facies, accreted in the catch-up mode with rates of 0.25 to 1 m·ka-1 (Gischler et al., 2008). Recent estimates on living reefs indicate that the bioconstructional potential of oceanic reefs is higher than that of lagoon reefs; however, both were able to exhibit superstratal bioconstruction in undisturbed conditions (Bianchi et al., 2016). The overall

Holocene reef thickness is generally less than 20 m (Kench et al., 2009; Morri et al.,

2015). Submarine cementation in Holocene reefs is rather weak, presumably as a 31 consequence of high accretion-rates, i.e., short time available for consolidation

(Gischler et al., 2008). Present-day living reefs exhibit similar features (Lasagna et al., 2010a; Morri et al., 2010).

Some studies on the topography of Maldivian reefs are available in literature.

The shallow (<130 m depth) submarine geomorphology of Ari atoll was investigated using multibeam sonars (Fürstenau et al., 2009), with the result of detailing the knowledge of submerged reef terraces that were previously recognized solely on the basis of single beam (Anderson, 1998) or scuba diving (Morri et al., 1995;

Bianchi et al., 1997) surveys in localized areas of the archipelago. In general, studies of the Maldives atoll upper slopes have recognized breaks in slope at recurrent depths in Ari (14 dive surveys, Morri et al., 1995, and 2 multibeam profiles,

Fürstenau et al., 2009) and Felidhoo (13 scuba transects, Bianchi et al., 1997) atolls.

The break in the reef slope that marks the Last Glacial Maximum (LGM) shoreline has been found at ~130 m below present sea level (Anderson, 1998; Fürstenau et al., 2009).

The earliest studies of Maldivian reef ecology date back to the turn of the

20th century (Gardiner, 1901-1905; Agassiz, 1903), but thorough field investigations on coral communities started with the Xarifa expedition in the late

1950s (Wallace and Zahir, 2007). Recent references for the ecology of Maldivian coral reefs are provided by Andréfouët (2012) and by Morri et al. (2015). Research on coral zonation highlighted the dominance of tabular and branching acroporid corals, which are responsible for superstratal growth and rapid accretion, in shallow 32 water (Davies et al., 1971; Lasagna et al., 2010b; Risk et al., 1994; Scheer, 1974,

1972); below 20 m the only significant bioconstruction was due to the azooxanthellate tree coral Tubastraea micranthus (Morri et al., 1995; Bianchi et al.,

1997). Overall, bioconstructional capacity of Maldivian coral reefs has shown high through a century of research. However, recent ecological crises resulting from major bleaching episodes, outbreaks of the corallivorous crown-of-thorn starfish

Acanthaster planci, and other disturbances (Bianchi et al., 2006; Lasagna et al., 2014;

Morri et al., 2010, 2015; Saponari et al., 2015), severely reduced bioconstructional capacity (Bianchi et al., 2016) and drew attention to the risk of platform drowning

(Ciarapica and Passeri, 1993).

2.3. Material and Methods

2.3.1. SCUBA Diving Transects

In our study, we focus on the reef slopes of seven Maldivian atolls (Fig. 2.1a):

Ari, Felidhoo, North and South Malé, Rasdhoo, Suvadiva, and Thoddhoo. Across their margins, we surveyed a total of 112 scuba diving transects. Depth measurements were corrected for tide according to local tidal predictions by the UK Hydrographic

Office (Malé Island) and referred to chart datum. In the central atolls, tidal range is generally between 0.3 to 0.7 m (Fürstenau et al., 2009). The transects were collected in a temporal interval between 1997 and 2013, therefore, it is possible that the terrace morphologies were subject to minor inter-annual variations caused by debris flow and sedimentation. 33

During the SCUBA surveys, we started each transect from the deepest reachable part of the reef and took notes along a transect perpendicular to the shoreline until the shallower part of the reef. In general, our transect ended on either the reef flat (Fig. 2.2a), at 2-3 m depth, or on the upper terrace on the reef front (Fig. 2.2a). Coordinates of each dive site were obtained before the dive from a

Garmin handheld GPS or extracted from nautical charts. We estimate that the horizontal accuracy of the positioning is in the range of few meters (GPS) to tens of meters (nautical charts). Before each dive, the type of reef (inner, that is in the lagoon, or outer, that is facing the ocean, Fig. 2.2a) was annotated, and in each dive we surveyed the relevant topographic and geomorphic features along the transect

(Bianchi et al., 2004; Rovere et al., 2011).

In particular, we measured the depth of the edge of the reef flat or the upper terrace (Fig. 2.2a), the depth of the inner margin of reef terraces, that are present both in the inner and in the outer reefs (Fig. 2.2a), and the depth of the base of coral rubble deposits (Fig. 2.2a), which usually accumulate with a slope angle of 25-35°.

Distances were measured with a 200 m-long metered tape or with personal dive sonar (i.e., a device that allows to measure distances underwater based on the round trip time for an acoustic pulse to reflect off an object), and depths were measured with a diving computer (depth accuracy between 0.4 to 3 m, Rovere et al., 2010;

Azzopardi and Sayer, 2012). Slope and directions were measured using a handheld clinometer (~5° accuracy) and a diving compass (~5° accuracy).

We estimated, through repeated measurements of the same points during 34 scuba surveys, that our depths carry an error of ±1 m. In order to calculate the modal depth of terraces in the Maldivian Archipelago, we describe each data point

(e.g., terrace depth) collected during SCUBA surveys as Gaussian with 1σ=1 m, in order to account for measurement errors. Then, we sum all the individual Gaussians to create a composite probability density function graph (the result is shown in Fig.

2.6b). The same procedure is used for a global dataset of submerged terraces. Peaks in the composite probability density function are interpreted as modal depths where terraces are most likely found.

2.3.2. Multibeam Bathymetry

A high-resolution 300 kHz multibeam bathymetry and backscatter dataset was acquired using a Kongsberg EM 3000 system integrated with an Applanix

Pos/MV navigation and motion system (Wright et al., 2002; Wolfson et al., 2007;

Mallinson et al., 2014). The merged system provides 127 1.5° × 1.5° overlapping beams at a pulse width of 0.15 ms within a 130 degree swath. The differential GPS position accuracy of the bathymetry was ~1.0 meter with a depth resolution of ~1 cm and depth accuracy of 5-10 cm RMS. The bathymetry data were processed using

CARIS HIPS. Various grid cell sizes were made depending on water depth. An underwater pressure sensor was deployed in the southwestern portion of Malé

Island Harbour and used to correct for actual sea level changes due to tide and wind during the multibeam survey (Wolfson et al., 2007). An approximation of a MLLW

(Mean Lower Low Water) sea level chart datum was used by using the lowest sea level recorded over the 8 day period (Wolfson et al., 2007). The survey surrounded 35

Malé Island including the southern edge of the North Malé Atoll (Fig. 2.5a).

The survey covers minimum water depths of about 1 m and reaches water depths beyond the slope break coinciding with the depth of the LGM on the southern margin. The southern margin quickly reached depths of ~150 m, at which the 300 kHz multibeam system was unable to receive return beams due to signal attenuation by the warm saline waters of the area. This attenuation is evident in the data gaps that appear in water depths deeper than ~130 m in the detailed multibeam bathymetry (Fig. 2.4 c,d). Similar, attenuation data gaps were observed in the warm saline waters around American Samoa with the same multibeam system (Wright et al., 2002). The survey reaches up to 50-60 m water depth on the northern, western, and eastern margins (Fig. 2.5 a,b). These margins represent the inner lagoon of North Malé Atoll. In addition, the survey extends about 1 km west and 600 m north of Malé Island and covers the channels separating Malé from

Funadhoo and Hulhule Islands (see Fig. 2.5a for details). To identify geomorphic features from this dataset, slope gradient maps were overlapped with bathymetric maps and the edges of reef terraces were traced as vectors in ArcGIS.

A series of twelve transects on the southern margin were selected at different depths to identify terraces from the shallow-most parts of the area up to the depth of the last glacial maximum (~120 m). Hypsometric curves were plotted to identify the terraces at different depths. Minimum-maximum depth values for each terrace were identified from examining all twelve transects. To define the depth of a terrace and its associated uncertainty, we chose the median value and 36 calculated the difference between the median value and the maximum or minimum depths.

2.3.3. Interpretation of Paleo RSL and uncertainties

We treated the inner margins of reef terraces identified in this study as relative sea level (RSL) indicators (Rovere et al., 2016a; Shennan et al., 2015; Vacchi et al., 2016), from which information on paleo RSL can be extracted, similarly to what was done using other submarine geomorphological indicators surrounding insular volcanoes and continental shelves (Romagnoli et al., 2013; Quartau et al.,

2014; Zecchin et al., 2015). The depth of a RSL indicator (in this case the depth of the inner margin of the reef terrace, Fig. 2.2b) does not necessarily represent the position of the paleo relative sea level that shaped it (Casalbore et al., 2017; Rovere et al., 2016a). In order to assess the paleo RSL, it is necessary to know the relationship of a sea level indicator to the paleo mean sea level at the time of its formation (i.e., the indicative meaning, Engelhart and Horton, 2012; Hijma et al.,

2015).

Therefore, to calculate the paleo RSL, we use the following equation:

Equation 2.1 – Paleo RSL

Where Md is the measured depth of the inner margin of the reef terrace and 37

Pwd is the paleo water depth, i.e., the water depth at which the reef terrace was forming (Fig. 2.2b). To obtain the value of Pwd we refer to the modern Maldivian reefs. Our SCUBA diving data indicate that the modern edges of the reef flat (for inner reefs, Fig. 2a) and the upper terrace (for outer reefs, Fig. 2.2a) in the Maldives are found between 6.1±2.6 m depth (see gray band in Fig. 2.6b). We assume that this is the depth at which a reef terrace would have formed at any time in the past.

Therefore, the Pwd used in Eq. 2.1 is 6.1±2.6 m.

Although this range can be considered as reliable only for the terraces presented in this study, we note that it is consistent with the observations that: i) modern reef flats are rarely observed at depths lower than 3 m (Blanchon, 2011); ii) the break in slope marking the transition between the upper terrace and the mid- shelf break in slope is rarely deeper than 8 m (Rovere et al., 2016a).

The total uncertainty (δRSL) associated with the paleo RSL is obtained by adding in quadratic individual errors according to:

Equation 2.2 – Total Uncertainty

Figure 2.2 – Reef geomorphic zonation, model for paleo sea-level calculation, and sensitivity of the paleo RSL

(a) Terminology related to the reef geomorphic zonation used in this study, and measured depths of relevant features (colored circles). The terminology has been taken from Blanchon, 2011. For field photographs, the reader is referred to Fig. 3 of this study; (b) Calculation of paleo RSL from the measured depth of the inner margin of a reef terrace (Md) and the estimate of paleo water depth (Pwd) from the modern analog, represented by the depth of the modern reef flat; (c) Sensitivity of the paleo

RSL calculation to possible perturbations on the measured depth of the inner margin. 1 – no perturbations considered, as in this study; 2 – the measured depth is 39 lower than the original inner margin depth due to reef building since the formation of the inner margin; 3 – the measured depth is higher than the original inner margin depth (green circle) due to erosion since the formation of the inner margin.

Other than the estimate of Pwd, the paleo RSL calculation, as described by

Eqs. 2.1 and 2.2, could be affected by the fact that the inner margin measured today is higher or lower than the original one due to bioconstruction, deposition of coral rubble at the toe of the slope or bio and mechanical erosion. If the inner margin presently measured has been covered by reef growth or if it has been eroded by planation processes since its formation, the Md would be affected as illustrated in

Fig. 2.2c (which also accounts for uncertainties on reef growth and erosion). In this study, we adopt the scenario 1 illustrated in Fig. 2.2c, where we assume that the measured depth of the inner margin corresponds to the original depth where it was shaped by the paleo sea level.

As a note, we remark that Holocene reef accretion rates from more than 60 sites globally average at ~4-5 m·ka-1 (Hubbard and Dullo, 2016, their table 6.1). In the Maldives, values comprised between less than 1 m·ka-1 (Gischler et al., 2008;

Klostermann et al., 2014) up to 15 m·ka-1 (Gischler et al., 2008, for periods of

Holocene rapid sea level rise) have been reported. Less data are available for the planation rates of marine terraces. Blanchon and Jones (1995) report values of marine planation between 40 and 70 m·ka-1 for reefs in the Caribbean. It is also worth noting that reef accretion and planation are dependent upon both sea level and the depth of the reef (Woodroffe and Webster, 2014). The faster a reef terrace is drowned, the faster it will be subject to high reef accretion or planation rates.

The thickness of coral rubble or sediments deposited at the toe of the cliff should be also subtracted from the measured depth of the inner margin. During our 41 surveys, we always considered as the inner margin the innermost part of the terrace with a low inclination (10-15º), noting in our SCUBA transects whether coral rubble deposits were covering the inner margin (e.g. Fig. 2.3e). While it is still possible that the depth of our measured inner margin is covered by coralline sands and coral rubble, this should only be a thin veneer upon the terraced surface.

2.3.4. Global database of coral reef terraces

The database presented in Fig. 2.8 has been assembled from studies reporting depths of reef terraces. The data was extracted from literature following a simple approach. If the literature study reported depths as ranges, we averaged the depth ranges and calculated their standard deviation (e.g., statements such as ‘a terrace is found between 20 and 25 m’ were inserted in the database as one terrace at

22.5±2.5 m). Uncertainties from literature were kept as reported, and if no uncertainties were reported, no uncertainty is inserted in our database (this includes the majority of cases).

2.4. Results

2.4.1. SCUBA Diving Transects: identification of terraces

For the seven atolls investigated (Fig. 2.1a), the results obtained from our

SCUBA diving transects show that the morphology of Maldivian reefs is characterized by sets of reef terraces at recurrent depths (Fig. 2.3a). The same morphology can be found either on the outer reefs, facing the open ocean or in inner 42 reefs, facing the inner atoll (1 and 3, respectively, in Fig. 2.3b). In the next paragraphs, we describe the different morphological elements that have relevance in terms of past relative sea level (RSL) indications.

The edge of the reef flat/upper terrace. In the Maldives, the shallow-water portion of the reef developing from the reef crest towards the inner lagoon and the outer ocean has the morphology of a flat, shallow-water terrace (Fig. 2.3a,c), which ends in a sub-vertical mid-shelf slope (Fig. 2.3d). In our SCUBA diving profiles, we measured the depth of this terrace both on inner (edge of reef flat) and outer (upper terrace) reefs; its average depth is at 6.1±2.6 m (see gray line in Fig. 2.6b). This part of the reef is the most highly affected by modern constructional and erosional processes.

Terrace T1. The mid-shelf slope is interrupted by a first terraced surface

(T1, Fig 2.3a, Fig. 2.4c,d,e,g) at a depth of 33±8 m. Often, the inner margin of this terrace is covered by accumulations of coral rubble deposits (Fig. 2.3a,d), only partially cemented by corallinaceous algae. The rockfall deposits create a slope of

~25-30°, and their toe averages at ~25 m depth (dashed line in Fig. 2.6b). The outer edge of T1 is often sharp and ends in another abrupt break in slope (Fig. 2.3e).

Terraces T2 and T3. While terrace T1 extends in general for 10-20 meters and represents an interruption of the shelf slope, other two levels of terraces characterize the investigated coral reefs between 45-65 m depth. The shallowest level is represented by T2, at 50±2.5 m (Fig. 2.4b,d), followed by T3 at 58.8±3.8 m

(Fig. 2.4c). These terraces are separated by relatively short but almost vertical cliffs 43 and are often 40-60 m wide. The break in slope characterizing these deeper terraces is, therefore, better marked than that of T1 and is less masked by debris at the toe of the slope; however these terraces are usually covered by a thin veneer of coralline sands.

Terrace T4. The deepest terrace found by a consistent number of scuba transects is T4, at 71.5±6 m (Fig. 2.4 a,e,f,g). This terrace is usually up to 30-50 meters wide (Fig. 2.4f).

2.4.2. Multibeam Bathymetry

Our multibeam data (Fig. 2.5) show that the southern margin of Malé Island is characterized by several morphologically distinct sets of terraces, located above the LGM break in slope. A total of nine terraces, named M1-M9, have been identified.

Terraces M1 – M4. The shallower reef terraces are located at 25±2 m (M1),

29.5±1.5 m (M2), 34.5±1.5 m (M3), and 38±2 m (M4). These were observed on the southern margin of Malé Island (Fig. 2.5c,e). These terraces are closely spaced along most of the southern margin. They can be clearly identified in the multibeam bathymetric dataset on the southwestern and southeastern corner of Malé Island

(Fig. 2.5c,e).

Figure 2.3 – Morphology of the Maldivian reefs

(a) General morphology of inner and outer Maldivian reefs, as derived from SCUBA diving transects. Letters c-g refer to approximate locations of photos shown in panels c-g; (b) Aerial view of part of a Maldivian Atoll. 1-outer reef; 2-reef flat; 3- inner reef; (c) Edge of the modern reef flat at 7 m (Gulhi Kuda Giri, S. Malé); (d) Mid- shelf scarp slope between the edge of the modern reef flat and T1 (Boldhuffaru outer reef, S. Malé); (e) Accumulation of coral rubble on top of T1. Slope is around

25-30°. The base of this coral rubble deposit is at 23 m (Thoddhoo outer reef); (f)

Outer edge of T1 (Bodhofoludhoo, Ari); (g) Outer edge of T3 (Thoddhoo outer reef).

Figure 2.4 – Cross-Section profiles of Maldivian Reefs from SCUBA

(a-g) Cross-profiles obtained from SCUBA diving transects. Locations for each panel are shown in Fig. 1a. The blue bands represent the calculated paleo sea level from

Table 2.1. In the panel f), the box on the upper right represents the planar view of a tract of the reef between 53 and 82m drawn during a deep scuba dive by C.N.

Bianchi. 46

Terrace M5 and M6. These two terraces are found at 46±4 m (M5) and

56±4 m (M6) (Fig. 2.5c,d,f). These terraces are wider than M1-M4, and they find a counterpart in T2 and T3 identified by scuba diving transects (Fig. 2.6b).

Terrace M7 and M8. The depths of these terrace levels are 70.5±5.5 m (M7) and 88.5±3.5 m (M8) (Fig. 2.5d,f,g). M7 correlates well to the terrace T4 identified by SCUBA diving transects, while M8 occurs at a depth where SCUBA diving was not attempted (Fig. 2.6b).

Terrace M9. This is the deepest terrace identified by multibeam bathymetry, it is narrow and bounded by two steep cliffs, clearly visible in the bathymetry data.

M9 lies at a depth of 106.5±3.5 m (Fig. 2.5d,g).

The multibeam bathymetry also shows a distinct break in slope at ~120 m depth (Fig. 2.5c,d,g,).

Figure 2.5 – Morphological features identified using Multibeam Bathymetry of Malé

Island a) Bathymetric map of Male Island derived from multibeam surveys, with locations of b,c,d,e,f, and g panels ; b) Oblique view displaying oval depressions between 50 and 60 m depth, interpreted as sinkholes formed during time of subaerial exposure; c) Slope map illustrating terraces from M1 to M6; d) Slope map illustrating terraces from M5 to M9, separated by steep cliffs. A total of two types of color stretch schemes were used to produce the slope maps: histogram equalize (illustrate better shallow terraces) and percent clip (illustrate better deeper terraces); e), f), and g)

Cross profiles show casing terraces M1-M4, M5-M7, and M8-M9. Due to sediment burial, a continuous slope profile does not show all the terraces and, therefore, three different cross profiles are chosen. These profiles located in areas of strong currents and thus less sediment masking. The base of the reef slope is shown by the arrows in (c) and (d) at ~120 m depth and is also located in g as the LGM shoreline.

2.5. Discussion

2.5.1. Depth of Maldivian Reef Terraces

Our scuba and multibeam surveys confirmed earlier reports that several levels of drowned reef terraces are imprinted in the insular shelves of the Maldives

(Morri et al., 1995; Bianchi et al., 1997; Anderson, 1998; Fürstenau et al., 2009;

Rufin-Soler et al., 2013). While former studies were limited to few atolls, our data span most of the archipelago. In first instance, it is worth noting that the depths of the terraces identified through scuba diving transects (Fig. 2.6a,b) show a good match with those identified in the multibeam bathymetry in Malé Island (Fig. 2.6c).

In general, the terraces M1 to M4 identified in the shallower areas provide further details on the terrace T1 identified in the SCUBA transects. M5, M6, and M7 show a good fit with T2, T3, and T4 respectively.

For the shallowest areas, T1 (M1-M4) is found within a large depth range throughout the archipelago. This terrace, which in most transects has a limited width, serves as base for coral rubble deposits, eroded and transported from the upper part of the reef front towards deeper depths (dashed line in Fig. 2.6b). Given its limited depth, it is possible that the large depth span of T1 (M1-M4) can be explained by differential reef growth or erosion at the wave base, which differs from site to site. The multibeam bathymetry data support this hypothesis, as they show that this feature is substantially more morphologically complex than the other terraces identified in this study (Fig. 2.5 c,e). In further support of the dynamic character of the shallower terrace level, we highlight that, three years after the mass 50 coral mortality of 1998, the large amount of newly generated coral rubble had obliterated many morphologies previously visible on the reef slope of the Maldives

(Morri et al., 2015). Previous studies pointed out to the presence, at the same depth of T1, of several notches and caves (Bianchi et al., 1997; Rufin-Soler et al., 2013), which were tentatively correlated to periods of Holocene sea level standstills.

Therefore, T1 is probably still affected by modern geomorphic processes, and probably underwent morphological changes in the last part of the Holocene (i.e., since 6-7 ka) when the pace of Holocene sea level rise slowed down (Lambeck et al.,

2014). Despite this caveat, we retain T1 in our sea level indicators due to its prominence and its widespread character. Further work, including age constraints, would be needed to understand the mechanisms of formation and morphological evolution of this terrace, as well as its relation to former sea levels.

The depths of terraces T2 (M5), T3 (M6) and T4 (M7) are recurrent in the seven atolls investigated (Fig. 2.6a,b) and correspond well with both the terraces identified in the multibeam surveys (M5, M6 and M7) and those identified in the

Maldives by previous studies (Morri et al., 1995; Bianchi et al., 1997; Rufin-Soler et al., 2013) (Fig. 2.6d). The only difference we highlight with respect to the data of

Bianchi et al. (1997), is that we attribute the large modal value that they identify at

20/25 m to accumulation of coral rubble at the toe of T1 (M1-M4) rather than to a terrace level.

Two other terraces, M8 and M9 are found at 88 and 106 m, respectively. This confirms earlier reports that at least three terraces characterize the deeper 51

Maldivian slopes (below 80 m depth). Colantoni et al. (2003) hypothesized that the

Maldivian atolls are characterized by a terrace at about 85 m, which coincides with the bottom of the Blue Hole in Ari Atoll. This depth is consistent with our M8 terrace. Offshore Ari, two terraces were identified by multibeam surveys at 94 and

97 m (Fürstenau et al., 2009). These depths, despite slightly shallower than M9, might be related to this terrace. Due to limitations in SCUBA diving, depths higher than 70-75 m were reached only in few dives. In South Malé atoll (Fig. 2.1a), one dive identified a terraced surface at 82 m, which, although slightly out of the depth range of T4 (M7), could be related to it. A deeper terrace, at 87 m (Fig. 2.4g) was instead identified in Suvadiva atoll (Fig. 2.1a), and correlates well with terrace M8.

The deepest SCUBA observation was made in Rashdhoo atoll (Fig. 2.1a, 2.3f). Here it was possible to identify a break in slope at 95 m. Due to the very high operative depth, it was impossible to track the continuity of this terrace either laterally or offshore, therefore, this point has not been considered in the terrace analysis shown in Fig. 2.6b. Nevertheless, we note that the depth of this terrace is just slightly lower than M9.

Deeper terraces had been identified in the Maldives by multibeam surveys at

125±3 m (Fürstenau et al., 2009) and by single beam sonar at 130±10 m (Anderson,

1998). We argue that these terraces can be correlated with the LGM surface, which we identified at ~120 m (Fig. 2.5c,d,e) in the multibeam surveys.

Other relevant features that characterize the sea bottom of the Malé islands are related to karst morphologies that have been drowned by sea level rise during 52 the last deglaciation. Such karst morphologies are mostly observed in the channel between Malé and Hulhule Islands (Fig. 2.5b) where Holocene sediments cannot accumulate due to strong tidal currents in this narrow channel. The depth of these irregular depressions is 50-60 m, and is similar to other water-filled karst features, such as the blue hole (Colantoni et al., 2003), observed in the Maldivian archipelago

(Fig. 2.5b). According to global eustatic sea level curves (Grant et al., 2014) the channel between Malé and Hulhule Islands has been exposed to subaerial agents several times at least during the last 150 ka. Therefore, it is likely that the karst morphology observed today has been developed during this time frame.

2.5.2. Paleo relative sea level

On the basis of the reef terraces identified by SCUBA dives and multibeam dataset upto 120 m of depth, we support the hypothesis that sea level played a major role in shaping the Maldivian atoll margin upper slope (Bianchi et al., 1997;

Fürstenau et al., 2009). We consider the edge of the modern upper terrace and the edge of the reef flat (respectively in outer and inner reefs) as modern analogs of the submerged reef terraces found in this study (Fig. 2.3a). The only difference between these is that, while we suppose that the submerged reef terraces were formed in very short periods of time and are, therefore, limited in width, the modern analogs could develop to their modern morphology during the last ~6 ka, in conditions of quasi-stable sea levels.

The reef terraces were, therefore, formed near sea level by the interplay between coral reef growth and marine erosion (Blanchon and Jones, 1995). They 53 can be considered as analogs to submerged terraces that form in temperate areas in periods of stable or slowly rising sea level, following processes that can be generalized under the ‘cliff overstep’ transgressive model (Zecchin et al., 2011,

2015).

Therefore, each level of terrace (Fig. 2.6b,c) can be treated as a paleo RSL indicator (Shennan et al., 2015), from which the elevation of paleo RSL - that is, RSL still uncorrected for post-depositional vertical movements due to glacial isostatic adjustment and tectonics (Rovere et al., 2016b) - can be calculated taking into account the modern analog (see Methods section for details). In Table 2.1, we report the calculation of the paleo RSL depth and associated uncertainty from Eq. 2.1 and

Eq. 2.2 (see Materials and Methods section) for the terraces T1-T4 and M1-M9.

2.5.3. A postglacial punctuated sea level rise?

The majority of RSL indicators available since the LGM for both tropical and temperate areas are concentrated in the last 6-7 ka BP (Lambeck et al., 2014; Khan et al., 2015) (Fig. 2.7a,b). For periods older than 11.7 ka BP, corals obtained from cores are most often used to reconstruct paleo RSL histories (Woodroffe and

Webster, 2014). In Fig. 2.7c, we compare the paleo RSL depths identified in the

Maldives with a global ESL curve obtained from dated RSL indicators, corrected for

GIA (Lambeck et al., 2014).

Despite the caveats discussed in the paragraphs above (e.g., the necessity to account for GIA and tectonics before comparing the Maldivian RSL data to global 54

Figure 2.6 – Comparison of terrace depths: SCUBA diving transects and

Multibeambathymetry data

(a) Depth of edge of reef flat/upper terrace (gray circles), inner margin of reef terraces (black circles) and base of coral rubble deposits (white circles) for each

SCUBA diving transect; (b) Relative frequency of terraces plotted by depth, with modal depths of most recurrent terrace levels and associated ranges (black line).

The dashed line indicates rockfalls identified at the toe of the reef slope, the grayline indicates the depth of the reef flat/upper terrace edge; (c) Terraces M1-M9 identified through multibeam bathymetry dataset, Malé Island; (d) Depth of terrace levels identified in the Maldives by previous studies.

Measured Measured Name Name depth RSL ±δRSL depth RSL ±δRSL (Multibeam (SCUBA) (Md) (m) (Md) (m) ) ±δMd (m) ±δMd (m)

M1 25±2 18.9±3.3

M2 29.5±1.5 23.4±3 T1 33±8 26.9±8.4 M3 34.5±1.5 28.4±3

M4 38±2 31.9±3.3

T2 50±2.5 43.9±3.6 M5 46±4 39.9±4.8

T3 58.8±3.8 52.7±4.6 M6 56±4 49.9±4.8

T4 71.5±6 65.4±6.5 M7 70.5±5.5 64.4±6.1

M8 88.5±3.5 82.4±4.4

M9 106.5±3.5 100.4±4.4

Table 2.1 – Depth of terraces in the Maldivian archipelago, calculation of paleo relative sea level and values used to estimate sediment coverage and paleo water depth.

56 eustatic curves, which are accounted for with large uncertainties in our paleo RSL calculations), we remark that the paleo RSLs calculated from reef terraces in the

Maldives (Table 2.1) match particularly well with periods of deceleration of the last deglacial global sea level rise (Lambeck et al., 2014) (red lines in Fig. 2.7c, upper panel). Most of these periods can be related to climatic events (Table 2.2).

The hypothesis that the reef terraces observed in this study were created during short periods of decelerations since the LGM is further supported by the observation that our terraces have widths most often included between 10 and 60 m. At rates of marine planation (Blanchon and Jones, 1995) of 40-70 m·ka-1, reef terraces such as the ones presented in this study would form in ~0.15-1.2 ka. This matches the duration of the periods of deceleration marked in red in Fig. 2.7c (~0.2-

1 ka, Table 2.2). However, higher resolution sea-level fluctuations occurring in a decade to century time scale could lead to the formation of terraces during acceleration events of sea-level rise causing the reefs to backstep under stress, as observed on south Texas shelf coralgal reefs where 6 common terraces developed within 2 Ka over 10 reefs separated by 120 kms (Khanna et al., 2017).

2.5.4. Inheritance from the Late Quaternary

In the paragraph above, we make the hypothesis that the submerged coral reef terraces we measured in the Maldives have been created during the postglacial sea level rise following the last glacial maximum. No radiometric age constraints for reef formations older than ~8 ka (Kench et al., 2009) are available in the Maldives, therefore the sea level vs. age correlation shown in Fig. 2.7c should be considered 57

Figure 2.7 – a) RSL and age of RSL indicators from different literature sources compiled by Lambeck et al., 2014 and Khan et al., 2015; b) Frequency of RSL indicators versus their age from the same sources; c) Paleo RSL elevation estimates from the Maldives (gray, pink, light pink, green, and orange as shown in Table 2.1, uncorrected for GIA and tectonics: see discussion in the text) compared with global eustatic sea level changes (Lambeck et al., 2014) (black line, with gray band for uncertainties) and with dated corals in the Maldives (Gischler et al., 2008; Kench et al., 2009; Lambeck et al., 2014; Woodroffe, 2005). The upper panel represents the rate of ESL change obtained from coral reef data by Lambeck and coworkers

(Lambeck et al., 2014). The deceleration periods observerd are colored in red line which match well with the terrace levels observed. 58

Timing Duration Terrace RSL ±δRSL Climatic/sea level events (ka) (ka) level (m)

18.9±3.3

23.4±3 T1 (M1- 8.8-9 0.2 The period between 11.4 and 8.2 ka is generally M4) characterized by high sea level rise rates (Lambeck et 28.4±3 al., 2014). These are though punctuated by sudden decelerations. It has been proposed that a third meltwater pulse (MWP-1C) happened ~8 ka BP, 31.9±3.3 causing drowning of reef terraces formed before this period (Blanchon, 2011). 9.6-10 0.3 T2 (M5) 39.9±4.8

10.5-11 0.4 T3 (M6) 49.9±4.8

Period of deceleration in the sea level rise, followed by MWP-1B (Bard et al., 2010) (starting at ~11.3 ka BP). This period coincides with the timing of the 11.7- 1.0 T4 (M7) 64.4±6.1 Younger Dryas stadial of the Northern Emisphere 12.7 (Lambeck et al., 2014). This depth corresponds well with a shoreline in South Africa dated ~11.7±0.9 ka BP(Pretorius et al., 2016).

Period of deceleration in the sea level rise, though 13.6-13.9 0.3 M8 82.4±4.4 within a period with generally high rising rates.

Period of near constant sea level following the H1 event, followed by a period of rapid sea level rise at 15-15.3 0.3 M9 100.4±4.4 the onset of the Bølling Allerød warm period, coinciding with the MWP-1A (Liu et al., 2015).

Table 2.2 – Timing of the main decelerations in postglacial sea level rise rates and

their duration (red lines in Fig. 2.2c, upper panel) corresponding to reef terrace

levels found in this study, calculated paleo RSL (from Table 2.2) and known climatic

or sea level events. 59 with caution. Previous works in the Maldives also attribute submerged reef terraces to postglacial and Holocene reef growth on the Pleistocene foundation (Bianchi et al., 1997; Fürstenau et al., 2009; Rufin-Soler et al., 2013).

Similar submerged reef terraces are also common features along the world’s continental shelves, carbonate platform and atoll margins, and are often correlated to meltwater pulses during the last deglaciation (Blanchon and Shaw, 2009). Several published studies carried out in tropical and subtropical areas described in fact levels of submerged reef terraces similar to the ones we report in this study for the

Maldives (Fig. 2.8a-e). Most of these studies, either through correlation with dated features such as relic reef buildups (Blanchon et al., 2002) or through chronostratigraphy and correlation with other sites (Green et al., 2014) supported the hypothesis that reef terraces found at different sites at roughly the same elevations were formed following pauses or decelerations of the deglacial sea level rise (e.g., meltwater pulses; Woodroffe et al., 1983; Wagle et al., 1994; Blanchon and

Jones, 1995; Green et al., 2014).

An alternative hypothesis is that some or all of the submerged reef terraces described in this study were instead shaped by past sea level highstands peaking below modern sea level. These might include MIS 5a and 5c as well as substages of

MIS 7 and MIS 3 (Siddall et al., 2007). Under this assumption, the postglacial sea level rise covered Pleistocene terraces with a thin veneer of reef growth during the

Holocene. This alternative hypothesis can be tested using high-resolution seismic data coupled with coring and U-series ages. A similar approach was recently 60 employed to shed light on the evolution of the Maldivian atolls since the Miocene

(Betzler et al., 2013). If all the terraces described in this study were formed during the Pleistocene, though, the question remains whether each terrace should be attributed to a single highstand, or if the terraces represent different sea level events within a single highstand.

2.5.5. A global sequence of postglacial submerged terraces?

As described in the previous section, several studies cross-correlated levels of submerged reef terraces from different locations to periods of sea level deceleration or standstill in the postglacial sea level rise. Additionally, the observations of high frequency terraces from south Texas shelf banks indicate formation of terraces during sea-level rise events (Khanna et al., 2017). In order to identify the presence of global patterns, a better understanding of each terrace within a geological setting is required, so as to compare terraces formed by similar processes. It is evident that comparing terrace depths from Maldives with reef terraces globally will showcase only a few similarities among several differences with global reef terrace record (Fig. 2.8). From our results we propose that there are four additional reasons for this mismatch.

i) Coral reef accretion. When submerged by sea level rise, reef features such as reef flats and reef crests provide substrates for later coral growth and sediment deposition. The rates of accretion may vary according to both sea level rise rates and other environmental factors, such as wave energy and availability of light

(Woodroffe and Webster, 2014). Therefore, the measured submerged elevations of 61

Figure 2.8 – a) Sites where submerged reef terraces have been reported*; b-e)

Detailed view of, respectively: the Caribbean region, the Red Sea region, the Indian region, the Great Barrier Reef region; f) Comparison between the reported depths of reef terraces and the levels found in the Maldives. The background of panels a-e represents the departure from eustasy at 6ka due to GIA, as calculated by Milne and

Mitrovica, 2008. Contours represent the same at 21 ka.

63 reefs measured today, are affected by different rates of post-drowning coral growth.

This effect is greater the longer the reef remains in shallow waters. The terrace T1 can be taken as an example of this process: because it is located in relatively shallow water, reef growth and erosion processes are still active, and they have been likely active for the past 6 ka. This possibly caused the inner margin to be located at different elevation according to local conditions, with the result of increasing the depth range of this terrace.

ii) Comparing RSL indicators vs. comparing paleo RSL. From the calculations shown in Eq. 2.1 and Eq. 2.2, and from the data presented in Table 2.1, it is evident that the elements to be compared among different regions are not the terrace depths per se (as shown in Fig. 2.8f), but the calculated paleo RSL with associated uncertainties. This would require, for each site, considerations of the paleo water depth associated with each reef terrace, as calculated in this study for the Maldives. Two reef terraces found today at different depths in different places may have formed in different conditions of paleo water depth and represent the same paleo RSL. Guidelines for reporting paleo RSL indicators (Düsterhus et al.,

2016; Hijma et al., 2015) maintain that both measurement and interpretation uncertainties must be reported. While we did so for our dataset, several data we extracted from the literature are reported without depth uncertainties or error bars, which means that there is no control on the uncertainty of many of the points in Fig.

2.8. 64

iii) Tectonic displacement. In our paleo RSL calculations, we assumed that no post-depositional vertical displacement due to tectonic activity affected the observed RSL indicators. In the case of areas where tectonic activity is low, at least since the Pleistocene, this assumption might be right, but in other areas tectonic activity may have caused significant uplift or subsidence of the observed RSL indicators. Also, tectonic or coseismic uplift may have the effect of a sustained relative sea level rise, increasing the number of reef backstepped morphologies (i.e., terraced surfaces).

iv) Departures from eustasy caused by GIA. Even when tectonic factors are accounted for, or where tectonics is negligible, perturbations to the Earth’s gravity field and solid surface associated with GIA cause the observed RSL to depart from eustasy (Milne and Mitrovica, 2008). Depending on the geographic location of the sites to be compared, the departures from eustasy may be significant. An example of how this effect can be significant when comparing shorelines from different locations can be derived from Fig. 2.8b-e. Due to GIA (Milne and Mitrovica,

2008), 6 ka ago, relative sea level was 7.9 m below the eustatic sea level in Bermuda

(site 19 in Fig. 2. 8b), 2.5 m above it in Sharm-el Arab (site 26 in Fig. 2.8c), 0.15 m above it in the Laccadives (site 37 in Fig. 2.8d) and 0.8 m below it in Osprey Reef,

Quensland (site 44, Fig. 2.8e). These values should be subtracted from the paleo RSL estimate derived from a reef terrace dating 6 ka at these locations. Note that GIA predictions also carry an uncertainty that can be large according to the location of the site with respect to the former ice sheets. GIA predictions at sites located near the forebulge of the former ice sheets may be characterized by uncertainties in the 65 range of tens of meters, while predictions for sites in the far field usually have lower uncertainties (Milne and Mitrovica, 2008).

As shown in the previous sections, the depth of a reef terrace can be used to calculate the depth of the paleo RSL that formed it. Paleo RSL is the sum of eustatic sea level (ESL), GIA (Milne and Mitrovica, 2008) and tectonic uplift or subsidence, as described in points ii) and iii) above (Rovere et al., 2016b). Both GIA and tectonic corrections are time-dependent. In absence of dating constraints for our dataset, we test the sensitivity of our RSL data to GIA and tectonic uplift using modeling results for 6 ka and 21 ka (Milne and Mitrovica, 2008) and long-term tectonic rates.

The GIA-related departure from eustasy has a latitudinal dependence in the

Maldivian archipelago. According to published models (Milne and Mitrovica, 2008), a shoreline deposited 6 ka in the Maldives should have been displaced by the GIA by

1.5 m in Rashdoo (North) and by 2 m in Suvadiva (South). Shorelines deposited 21 ka ago should be instead found 5.7 and 3.2 m (in Rashdoo and Suvadiva, respectively) below the eustatic value of ~130 m (Milne and Mitrovica, 2008).

From a tectonic perspective, the Maldives have undergone long-term subsidence. Published subsidence rates vary between 0.035 mm·a-1 to 0.15 mm·a-1 since the onset of the Last Interglacial (Gischler et al., 2008). With the caveat that assuming linearity from long-term tectonic histories can be misleading, we translate these rates into total vertical displacement for shorelines dating 6 ka and 21 ka BP.

We obtain that subsidence could have displaced a 6 ka BP shoreline in the Maldives downwards by 0.2-0.9 m, and a 21 ka BP shoreline by 0.7-3.2 m (Table 2.3). 66

Departure from GIA Age (ka TECTONICS eustasy due to correction ESL (m) RSL (m) BP) (m) tectonics and (m) GIA (m)

From ESL in Suva Rashdoo Min Max Milne and Min Max diva Mitrovica109

6 ka 1.5 2.0 -0.2 -0.9 0 0.6 0.9 0.6 / 0.9

21 ka -5.7 -3.2 -0.7 -3.2 -130 -138.9 -136.3 -8.9 / -6.3

Table 2.3 – Calculation of departure from eustasy due to tectonics and GIA in the

Maldivian archipelago for RSL indicators formed 6 and 21 ka BP.

The results briefly discussed above and shown in Table 2.3 allow to assess that overall, in the Maldivian atolls, the departures from eustasy due to GIA and tectonics can be as small as a few decimeters (for a RSL indicator dating 6 ka BP) and as large as ~9 m (for a RSL indicator dating 21 ka BP). These results indicate that caution should be employed when comparing the RSL estimates obtained from the Maldives (or any other site where departures from eustasy are significant) with global eustatic curves and rates (Lambeck et al., 2014). Nevertheless, we highlight that the large (multimetric) error bars that we obtained in our paleo RSL calculations (Table 2.1) might take into account, to some extent, the departures from eustasy caused by GIA and tectonics.

Chapter 3

Regional and Local Context of Microbial Buildups, Central Texas – A 3-phase morphological evolution

The following chapter is a reproduction of Khanna, P., Droxler, A. W., Hopson,

H.H., Kubik, B., Proctor, J., Singh, P., Trotta, R.P., Zhou, Y., Lehrmann, D., and Harris,

P.M. (in prep.).

Abstract

Although Upper Cambrian microbial buildups were recognized in the Point

Peak Member of the Wilberns Formation in Central Texas nearly 70 years ago, only a few studies focused specifically on the buildups themselves. This study develops the regional and local context of the microbial buildups, describes their growth stages, and details the interrelations with surrounding sediments. Fifteen measured sections in the Llano area (2500 sq. km.) demonstrate the occurrence of two

69 microbial units – the lower and upper Point Peak Member units. This detailed study focuses in a 25 sq. km. field area (along the Llano and James Rivers, and Mill Creek in south Mason County) where the full thickness of the Point Peak Member is exposed.

Field studies, in addition of drone imagery, are used to analyze the lower Point Peak unit (bioherms and biostromes (<1m thick, associated with heterolithic and skeletal/ooid grainstone beds), and upper microbial build up unit (up to 14 m thick, associated with skeletal and ooid grainstone beds intercalated with mixed siliciclastic carbonate silt beds). The Upper microbial unit display a three phase growth model, evolving from an initial ‘colonization’ phase on a transgressive lag, through a ‘vertical aggradation and lateral expansion’ phase, and ultimately into a

‘capping’ phase. The ultimate demise of the microbial buildups in the upper Point

Peak unit is interpreted by an increase of water turbidity caused by an enhanced fine siliciclastic flux. Analyses of the microbial buildup colonization phase 1, on a plan view outcrop on the banks of the James River, offers unique opportunities in scaling their growth at three quantifiable scales: large, medium, and small. The lower and upper Point Peak microbial units are used to develop a composite depositional model for the area where the lower Point Peak microbial unit represents inboard shallow subtidal and intertidal facies, whereas the upper Point

Peak microbial unit represents outboard subtidal facies. 70

3.1. Introduction

Microbialites are “organo-sedimentary deposits which accrete as a result of a benthic microbial community trapping and binding detrital sediment and/or forming the locus of mineral precipitation” (Burne and Moore, 1987). Microbes have been building reefs in shallow marine environments since the Archean, around 3.5

Ga (Riding, 2006; Woo and Chough, 2010; Riding 2011). They diversified and reached their peak in the late Proterozoic, after which metazoans (Riding, 2006b) generally became the prominent reef builders. Microbialite flourished, however, after mass extinction events when metazoans declined (Schubert and Bottjer, 1992;

Riding, 2006b). The Upper Cambrian, with few metazoans (Boucot, 1990; Kiessling,

2009), was a time of extensive development of microbial reefs on Laurentia, Baltica,

Siberia, and Gondwana (Lochman-Balk, 1970a; Lee et al., 2015).

Shallow seas associated with Laurentia were present at tropical latitudes (Fig. 3.1) and, therefore, provided an optimum location for the development of microbial reefs

(Lochman-Balk, 1970a; Van Der Voo, 1988; Golonka et al., 1995; Lee et al., 2015).

Measured sections, regional correlation, and biostratigraphy show that four dominant regressions and three transgressions synchronously occurred across

Laurentia, based on Lochman-Balk (1970a). Microbial buildups whose growth was triggered by the youngest of the three transgressions, corresponding to the Point

Peak Member of the Wilberns Formation in Central Texas, are observed in several outcrops within a 2500 km2 portion of Mason County, Texas. Some of those outcrops are the focus of this study (Fig. 3.2-3.4). These buildups are spectacularly 71 exposed in cross-section (cliffs) and plan view (pavements) outcrops and provide an opportunity to understand in detail their initial establishment, their 3D architecture, and their morphological evolution within a regional depositional context (Fig. 3.5-

3.10).

Microbialite morphologies provide a wealth of information about the biological and physical factors influencing their initiation, overall development, and ultimate demise, in addition to the paleo-environmental conditions in which they grew (Andreas and Reid, 2006). The morphology of the microbial reefs can be interrogated at macroscopic scale such as bioherms or mats, at mesoscopic scale such as stromatolitic columns or thrombolytic growth, and microscopic scale such as Girvanella, Epiphyton, etc. (Lee et al., 2015). The microbial buildup growth at the macroscopic scale, a product of accommodation space, hydrodynamics, and sedimentation patterns, is the focus of this paper with a main objective to define the regional and local growth context of uniquely exposed microbial buildups observed in a series of world class outcrops in newly accessible private ranches, to understand their morphological evolution and to develop a regional depositional model.

Figure 3.1 – Upper Cambrian paleogeographic, paleoenvironment, and lithofacies map of Laurentia modified from Lochman-Balk (1970a). The Laurentia continent consisted of North America, northwest Ireland, Scotland, Greenland, and the

Chukotka peninsula. North America is rotated 90o counterclockwise from its current orientation and the study area in modern–day Texas was just south of the equator.

3.2. Geological Background and Stratigraphic Context

3.2.1. Upper Cambrian Regional Context

Two major orogenies are recognized in the Proterozoic (Pisarevsky et al.,

2008): the Genvillian orogeny around 1100 Ma responsible for the formation of the supercontinent Rhodinia (Hoffman, 1991) and the Cadomian/Pan African orogeny around 570 Ma responsible for the supercontinent Pannotia (Golonka 2003,

Golonka, 2006a,b). Pannotia later divided into four major paleocontinents -

Laurentia, Baltica, Siberia, and Gondwana, with the Laurentian continent including

North America, northwest Ireland, Scotland, Greenland, and the Chukotka peninsula

(Golonka 2000, 2002; Ford & Golonka 2003; Golonka et al., 2003; Golonka et al.,

2007). Laurentia drifted northward during the early and middle Cambrian and rotated counter clock-wise, arriving at low latitudes near the equator (McCausland et al., 2007; Golonka, 2007). Shallow seas covered most of these paleocontinents, giving rise to suitable locations for microbial buildups to establish and flourish.

Much of the Laurentia craton was covered by marine sedimentary basins

(Ford & Golonka 2003, Golonka et al., 2003; Schulz et al., 2008). Limestones and dolomites were widely deposited, commonly referred to as the ‘GACB’ – Great

American Carbonate Bank (Ginsburg, 1982; Trettin & Balkwill 1979, Morgan, 2012).

Sloss et al. (1949) named the latest Pre-Cambrian to Early Ordovician strata of

Laurentia the ‘Sauk sequence’. Palmer (1981b) further subdivided the Sauk sequence, based on trilobite faunal discontinuities, into Sauk I, II, and III, deposited from latest Precambrian-Lower Cambrian, Upper Middle Cambrian to Late middle 74

Cambrian, Late Middle Cambrian to early Ordovician time periods, respectively

(Read and Repetski, 2012; Morgan, 2012). Each Sauk sub-sequence has been further divided into smaller sequences (ranging up to ten) on the basis of detailed biostratigraphic and lithostratigraphic analysis (Morgan, 2012 and the references therein).

3.2.2. Upper Cambrian Local Context, Central Texas

The Llano Uplift in Central Texas is a structural dome of Precambrian igneous (1,030 ± 30 million years) and metamorphic (1,120 ± 25 million years) rocks exposed over an area extending approximately 100 km east/west and 70 km north/south (Garrison et al., 1979). The Paleozoic strata in the Llano region, termed the Moore Hollow Group (Bridge et al., 1947; Morgan 2012; Miller et al., 2012), consist of the Riley and Wilberns Formations. In turn, the Riley Formation consists of the Hickory Sandstone, Cap Mountain Limestone, and Lion Mountain sandstone members. The overlying Wilberns Formation has four members – the Welge sandstone, Morgan Creek Limestone, Point Peak, and San Saba (Paigne, 1911; Bridge et al., 1947). The Moore Hollow Group is overlain by the Ordovician Ellenburger

Formation (Cloud and Barnes, 1948). The unconformable contact between the Lion

Sandstone and Welge Sandstone members is defined as the boundary between the

Sauk II and Sauk III sequences (Miller et al., 2012).

The microbial buildups examined in this study lie within the Point Peak

Member of Wilberns Formation exposed in the southern part of Mason County

(Sliger, 1957; Ahr, 1967, 71; Spincer, 1997). Outcrops are observed along the Llano 75 and James Rivers and Mill Creek in newly accessible private ranches. Previous mapping of the Point Peak Member within the Llano uplift area by Bridge et al.

(1947) determined two trends that are key to the current study: 1) the presence of two lower and upper microbial units, separated on the basis of a key bed - the

Plectotrophia zone; and 2) the absence of microbial buildups from the northeastern portion of the study area (Fig. 3.2). The brachiopod species Plectotrophia bridgei and

Billingsella - both epifaunal suspension feeders, define the Plectotrophia zone. Parke

(1953) and Sliger (1957) were the first to make geological maps of this area. Later,

Ahr (1967, 1971) published a detailed petrologic and sedimentologic study of the

Point Peak and San Saba members. Spincer (1997) studied the paleoecology of the

Upper Cambrian buildups with a focus on the Llano region, and Johns et al. (2007) conducted studies on sponges found associated with the buildups. Morgan (2012) defined the sequence stratigraphy for the GACB and Miller et al. (2012) further examined the biostratigraphy of Cambrian and Lower Ordovician strata in the region.

3.3. Data Collection, Biases, and Errors

3.3.1. Field Survey

As aforementioned, the main outcrops of Point Peak microbial buildups occur either along cliffs or on pavements along the Llano and James Rivers, and Mill

Creeks. During several field seasons in 2013, 2014, and 2015, the extent of the microbial unit in the 25 sq. km. of the field area, was mapped by Llano and James 76

Figure 3.2 Stratigraphic column and measured sections. A) Stratigraphic column for

Cambrian strata of the Llano Uplift (modified from Barnes and Bell (1977) and Kyle and Mcbride (2014)). Legend: Blue color – Limestone, Purple- Dolomite, Yellow-

Sandstones, Green – Shales, Granites – Red; B) 22 sections measured through the

Upper Cambrian strata across the Llano Uplift (modified from Miller et al., 2012).

The Point Peak Member of Wilberns Formation, which contains the microbial interval focused on in this study, has been measured in 15 of these sections; C) 15

Sections of the Point Peak Member, modified from Bridge, Barnes, and Cloud (1947), and Ahr (1967). The base of Point Peak Member is marked with a red line and the top with green. Orange dashed line highlights the Plectotrophia zone which has been identified from most of the sections and clearly distinguishes between the two 77 microbial units. This study is focused on the Upper Microbial unit from Point Peak

Member in section 3.

Figure 3.3 Arc Map 10.1 base map imagery showing the field area in Mason County,

Central Texas, and locating 12 outcrops which include pavements and cliffs exposing the lower and the upper microbial units. The yellow text and lines represent cliffs, red rectangles and text represent pavements. 79

Figure 3.4 Geological map from Sliger (1957) is corrected for the Point Peak

Member (which was mapped in this study) and is overlaid on top of Arc Map 10.1 base map imagery.

Rivers, and Mill Creek as well as their gullies. Encountered microbial buildups outcropping along these transects were systematically photographed with a Nikon

COOLPIX camera and located with the integrated GPS unit. These photographs were subsequently imported to Arc G.I.S. to mark the location of the encountered microbial buildup with an error of 5 to 10 m. Moreover, in the summer of 2015, the top and base of the microbial unit encountered along gully transects were mapped with a hand held GPS, (Garmin Montana 650t, with an accuracy of ±5). Sliger’s

(1957) geological map of the field area was complemented by additional detailed observations on the microbial reef occurrences made during this study (Fig. 3.4).

3.3.2. Aerial Imagery - Data Acquisition

Outcrops along the Llano River include three cliffs, the Shepard Cliff (600 m;

Fig. 3.8-3.9), the Zesch Cliff (400 m; Fig. 3.10), and the Goat’s Graveyard Cliff (300 m), exposing the Upper Point Peak, and two pavements, the Shepard Pavement

(500x100 m) exposing Lower Point Peak microbial facies and the Zesch Pavement

(300 x 200 m) exposing part of the Upper Point Peak microbial facies along the

Llano River (Fig. 3.3). James River localities include two outcrops, a composite of pavement and cliff referred to as the James River Pavement (600x200 m; Fig. 3.5) and a cliff outcrop named the Faulted Buildups. A complete Point Peak section, including its lower and upper units, outcrops along a lower Mill Creek cliff (Lauren’s

Cliff). Further up Mill Creek, four other outcrops, including the Bunkhouse, Upper

Mill Creek, Fallen Block, and DroxRock outcrops expose the Upper Point Peak. A fifth outcrop, a Lower Point Peak microbial pavement, is juxtaposed to the Upper Point 81

Peek DroxRock outcrop, along Schmidt Fault. This study will particularly focus on the Upper Point Peak microbial buildups.

Camerawings, an aerial photography company, was hired to conduct the survey with a drone (quadcopter) equipped with a Sony NEX-7, 24.3 MP camera, the best resolution camera available at the time of the data acquisition, a gimble, (tool which keeps the camera horizontal during flight for pictures acquisition of the pavements) and a GPS, to locate the geographical and elevation position at which the photographs were taken.

An aerial survey was conducted in February 2014 to collect digital photographs A total of eleven outcrops were selected for drone survey; five outcrops along Llano River (three cliffs and two pavements), four outcrops along

Mill Creek (cliffs), and two outcrops along James River (one pavement and an adjacent cliff, and one cliff outcrop). Markers were placed in the field on the pavement outcrops before the drone survey, and the precise locations for several markers were collected using DGPS (Differential GPS). A Trimble Total Station unit was used to determine the location coordinates of all the markers in a local coordinate system with an accuracy of 10 cm. DGPS accuracy of marker location, with respect to each other, is less than what could be achieved from Trimble Total

Station, therefore both DGPS and Total Station were used.

For pavement outcrops, the drone was assigned to fly 40 m above the ground, and the flight path was automated. For cliff outcrops, the drone was flown manually at a distance ranging from 20-40 meters in front of the outcrop face. 82

During the data collection for both pavements and cliffs, a 90% overlap was kept.

The total number of photographs for each outcrop varied according to the size of the outcrop. The largest image dataset was collected for James River pavement where a total of 540 photographs were collected.

3.3.3. Aerial Imagery - Data Processing

The digital photographs and camera positions, saved in EXIF file format

(Exchangeable Image File Format), were imported to Agisoft 1.0 for building virtual outcrop models. The Align Photograph tool was used to refine the camera position for each photo and to build an initial sparse 3D point cloud model in a random 3D coordinate system. Further, build mesh tool is selected to create a polygonal mesh that uses the 3D point cloud to make edges, vertices, and faces. To geo-reference the location of the model, ground control points (GCP’s) were integrated with the polygonal mesh. The geo-referenced model is used to build a dense point cloud to further build a dense high resolution mesh. The last step includes building a texture of the 3D mesh model which is the final product – a virtual outcrop model. The virtual outcrop model is exported in 3 different file formats, TIFF file format

(exports an Orthophotograph of the outcrop), DEM (Digital Elevation Model in xyz format), and KMZ (google earth readable file format). 83

Figure 3.5 (A) James River Pavement (600x200 m) Orthophotograph built using 560 aerial images utilizing photogrammetry workflow in Agisoft 1.0; (B) H3 buildups;

(D) are bird’s eye zoomed in view of microbial buildups showcasing different size and shapes of microbial buildups.

Figure 3.6 James River Cliff orthophotograph. Panel (A) and (C) represents an uninterpreted and panel (B) and (D) represent their interpreted versionThe base of the microbial unit is interpreted with an orange line and the top of the microbial unit is interpreted with a blue line. The individual circular to semi-circular – oval shapes represent microbial buildups. The inter reef grainstones are colored with a translucent orange colour to emphasize the relative continuity of the grainstone beds.

Figure 3.7 – Three types of microbial buildups in Lower microbial unit: A) First

Microbial type - Vertical cross section of Lauren Cliff displaying mixed carbonate- clastic sediments. The red rectangle represents a microbial bed, referred as “Dan

Strom”; B) A close up of the Dan Strom microbial bed in Fig. A displays individual microbial heads; C) Second Microbial Type – Bird’s eye view - Digital outcrop built using 408 photographs. Red rectangle displays the individual microbial heads. This bed is called as “Bread Loaf Reefs”; D) Close up of one perfectly preserved reef among the Bread Loafs; E) Third microbial type – vertical cross section showcasing the conical growth of several microbial heads within one of several biostrome beds;

F) A close up of third microbial bed showing typical conical morphology of some microbial heads.

Figure 3.8 Orthophotograph of Shepard Cliff eastern side exposed along Llano River.

The upper panel represents an uninterpreted and lower panel represents an interpreted version. The base of the microbial unit is interpreted with an orange line and the top of the microbial unit is interpreted with a blue line. The individual circular to semi-circular – oval shapes represent microbial buildups. The inter reef grainstones are colored with a translucent orange colour to strengthen the relative continuity of the grainstone beds.

Figure 3.9 Orthophotograph of the Shepard Cliff western side exposed along Llano

River. The upper panel represents an uninterpreted and lower panel represents an interpreted version. The base of the microbial unit is interpreted with an orange line and the top of the microbial unit is interpreted with a blue line. The individual circular to semi-circular – oval shapes represent microbial buildups. The inter reef grainstones are colored with a translucent orange colour to demonstrate the relative continuity of the grainstone beds. 92

Figure 3.10 Orthophotograph of Zesch Cliff exposed along Llano River. The upper panel represents an uninterpreted and lower panel represents an interpreted version. The base of the microbial unit is interpreted with an orange line and the top of the microbial unit is interpreted with a blue line. The individual circular to semi- circular – oval shapes represent microbial buildups. The inter reef grainstones are colored with a translucent orange colour to demonstrate the relative continuity of the grainstone beds. More grainstones are observed within the inter-reef sediments on Zesch Cliff than Shepard cliff.

3.4. Observation and Results

3.4.1. Upper and Lower Point Peak, Mason County, Central Texas

The Point Peak Member of the Wilberns Formation can be divided by the

Plectrotrophia zone into two separate units (Fig. 3.2). The outcrops exposed along

Llano and James Rivers, and Mill Creek expose the complete thickness of the Point

Peak Member, which represents vertical stacking of proximal (inboard) microbial facies (Lower Point Peak; Fig. 3.7) underlying distal (outboard) microbial facies

(Upper Point Peak; Fig. 3.5 and 3.6). The vertical stacking of the Upper Point Peak distal microbial buildups on top of the older Lower Point Peak proximal mostly biostromal in nature implies a major transgression shifting the microbial facies belt towards the exposed land.

Lower Point Peak outcrops in this locality consist of smaller microbial biostromes and bioherms (<1 m in height; Fig. 3.7), associated with heterolithic facies, glauconitic siltstones, and detrital sands (skeletal and interclastic carbonate).

Downdip on the ramp ( to the SE) thicker microbial buildups (upto 30 m thick) have been identified (Bridge et al., 1947) indicating a wide extent of the microbial facies belt on the platform, up to 100 km wide.

Upper Point Peak outcrops above the Plectrotrophia zone in the study area have spectacular exposures of the large microbial buildup unit (up to 14 m thick) and provide a unique opportunity to quantify the distribution and heterogeneity of microbial buildups within this unit. These buildups are surrounded with thick skeletal and oolitic grainstone (intercalated with thin mixed carbonate siliciclastic 95 silty beds) indicating subtidal environments (with possible little fluctuations in base level bringing siliciclastics within the system).

3.4.2. Upper Point Peak Microbial Buildups, Mason County, Central Texas

Systematic and comparative analyses of the multiple microbial buildups and their inter reef sediments exposed along the different cliffs allow identifying their morphology and facies vertical and lateral similarities and heterogeneities among them within the microbial unit. Moreover pavements provide unique opportunities to observe morphology and facies lateral variability among, within, and between several buildups (Fig. 3.5-3.10). A systematic three-phase growth evolution of the microbial buildup was first observed during the field survey and was confirmed later by drone imagery combined with the qualitative and quantitative analysis of the buildup morphology using photogrammetry data in addition to direct field observations. Each individual buildup was given a name to simplify their overall description and analyses (Fig. 3.7-3.10).

3.4.2.1. Three growth phases of the upper Pont Peak microbial buildups

The Shepard and Zesch cliffs along the Llano River vary from 30-50 meters in height and 400-500 meters in length (Fig. 3.7-3.10) and also expose the strata underlying and overlying the microbial buildup interval and, therefore provide a full record of the depositional environments that existed prior to the initial establishment, growth, and final demise of the microbial buildups. 96

Strata underlying the upper Point Peak microbial reef unit are easily accessible for direct observation and sampling on a cliff, referred to as the Fallen

Blocks on Mill Creek (Fig. 3.11). These glauconite-rich greenish strata consist of dominantly thinly bedded siltstones and sandstones of mixed carbonate/siliciclastic composition (Sliger, 1957) that contain symmetrical cross-bedded mega ripples formed by tidal currents. Fining upward beds of irregular thicknesses are interpreted as tidal channels filled at their bases by flat pebble conglomerates interpreted as storm deposits. Several carbonate grainstone beds with abundant ooids and bioclastic grains are interpreted as a high energy shallow subtidal environment (Kelleher et al., 2016). These strata, therefore, based upon the aforementioned observations and sampling are interpreted to represent a marine shallow subtidal/intertidal depositional coastal environment.

A massive 1-2 meter thick skeletal grainstone bed forming a clear overhang along the Shepard and Zesch cliffs abruptly overlies and dramatically contrasts with the underlying strata. This bed is interpreted to represent a relatively deeper subtidal environment and is referred to as “the switch”, marking the base of the microbial buildup unit (Fig. 3.11). This clear transition from mixed carbonates- siliciclastic to pure carbonate deposition is interpreted to be transgressive in nature.

The Upper Point Peak in the study area consists of many individual large microbial buildups spaced apart unevenly and separated from each other by originally sub- horizontal mixed carbonate siliciclastic inter-reef strata, deformed by preferential compaction in particular in the vicinity of the massive buildups. The strata overlying the large microbial buildups consist of sub parallel layered limestone beds, 97 deformed by preferential compaction of the underlying inter reef sediments relative to the massive microbial buildups, most likely already cemented during their growth.

One of the microbial buildups, referred to as ‘Mitch Herm’, exposed along the eastern side of Zesch Cliff on Llano River, clearly illustrates the general evolution of the buildups, from their establishment, distinct three-phase growth, and ultimate demise (Fig. 3.12). ‘Mitch Herm’, as most of the other microbial buildups in the

Upper Point Peak, sharply contrasts with the underlying mixed carbonate- siliciclastic silty and sandy, thin, mostly recessive beds, interpreted to record very shallow subtidal and intertidal marine environments. The sharp transition, referred to the “Switch” corresponds to the flat base of ‘Mitch Herm’ as the other buildups, and the contemporaneous initial accumulation of a 1-2 m thick inter-reef resistant grainstone bed. This bed usually dips towards each buildup on both sides, thins, and in many instances disappears towards their flat center (Fig. 3.7-3.12).

3.4.2.2. The Microbial Buildup Substratum

The flat center of each buildup is not accessible for direct observation along the Shepard and Zesch cliffs. However, within the Fallen Block outcrop along Mill

Creek, a fully developed 8 m-thick buildup, tilted 90 degrees and lying on its flank, offers in 3D its flat base for observation and sampling (Fig. 3.11). The buildup detached itself from the nearby cliff where a small depression in the first inter reef bed, still matches the shape of the rotated buildup edge. The center of the buildup 98 base consists of a 10 cm-thick lense of flat rip-up clasts, interpreted as a transgressive lag, The flat rip-up clasts were formed from erosion and reworking of the partially cemented, fine-grained beds, immediately underlying the ‘Switch”. A core drilled through the very bottom of the bioherm confirms this interpretation

(Droxler et al., 2016); at its very bottom a few small rip-up clasts mixed with large trilobite fragments are mixed with the initial growth of the stromatolitic columns. In summary, microbial growth was initiated on low-relief lenses of flat rip-up clasts interpreted as a transgressive lag and occurred contemporaneously to the deposition of an inter-reef bioclastic grainstone bed. Similar substrates for growth have been observed in other Upper Cambrian microbial reefs (Chen, 2014a; Cowan and James, 1993; Lee et al., 2015). In the Bahamas, analogous modern subtidal microbialites of the Exuma Cays have been documented growing on flat pebble deposits or on large shells (Dill et a., 1986; Feldman et al., 1998; Reid et al., 1999;

Ginsburg and Planavsky, 2008).

Figure 3.11 Fallen Block – Mill Creek displaying the base of the microbial buildups ;

(A) Fallen microbial block lying on its side and displaying the flat base of the buildup; B) View from below the base of the buildup. A flat lense of rip-up clasts is outlined, visible in detail in C). D) Cliff from which the fallen block detached shows the depression where the buildup was initially attached. Growth Phases 1, 2, and 3 are identified with blue, red, and black transparancies on A and are explaied in details on Fig. 3.12 (B). 100

Figure 3.12 The three-phase growth model displayed by the microbial buildups; A)

Oblique drone image of “Mitch Herm” within the Zesch Cliff above the Llano River;

B) Growth phase interpretation of Mitch Herm and coeval sediments showing strata 101 below ‘the switch,” the main microbial interval with its distinct phases, and the overlying strata. The blue line represents the flat substrate over which microbial buildups nucleate. Phase 1- ‘colonizing Phase’ is the first microbial growth represented by a rind covering the complete phase 1 growth. It is marked with sky blue color. Phase 1 grows coeval to the inter reef grainstones (colored in orange) but do not interact with the well developed thrombolitic external rind, only with the stromatolitic columns in the interior of Phase 1 reefs. Phase 1 is onlapped by mixed carbonate siliciclastic silty layer. The Phase 2, colored in red, nucleates on top of phase 1 and grows upwards and outwards and intimately interacts with the adjacent coeval grainstones. Phase 3 represents the last phase and is similar to the

Phase 1 thrombolitic rinds and covers the complete microbial buildups, in addition to grow also as individual mounds on top of the inter reef grainstone (colored in black).

102

3.4.2.3. Phase 1: “Colonizing Phase”

In the initial colonizing growth phase, the microbial buildups are usually clearly defined. They formed 3-5 m high, oblong in shape reliefs covering areas ranging from 10 to 300 square meters (Fig. 3.12). Sufficient accommodation allowed the microbial buildups to develop those appreciable synoptic reliefs. Phase 1 buildup margins are well defined, consisting of purely microbial thrombolytic rinds, contemporaneously cemented and lithified while they were growing. The rind literally encloses the Phase 1 buildups. The Phase 1 microbial buildup interiors are composed of multiple orange-colored stromatolite columns, 10’s of cm in width and height. Spaces between the columns are filled with dominantly fine-grained carbonate sediments. The margins of individual columns consist of thin gray well- cemented thrombolytic rinds.

The James River pavement and adjacent cliff presents a spectacular locality where microbial outcrops cover a portion of the river floor, 600 m long and 200 m wide, and along one side of the river an adjacent cliff spans 600 m length and ranges in thickness up to 50 meters. The James River floor is a cut through the first phase of microbial growth and hence provides arguably the best location to investigate the buildups and inter-buildup sediments of the first phase in great details (Fig. 3.5).

Four different scales of clustering are identified within the first phase: A) The smallest scale of a few decimeters is represented by circular to oblong features having a micritic rind on the outside. B) The smallest scale clusters into the medium scale, which is a few meters in length and has its own rind. C) The medium scale 103 features cluster into the large scale that is a few tens of meters in longest axis and has its own rind. D) The large scale is observed to cluster into the largest scale called as a ‘complex’, which is few hundreds of meters long and has its own rind. The orientation of the complex, large, and medium is generally towards NE-SW and the smallest scale orientation axis varies.

Inter-reef sediments contemporaneously accumulating with the first growth phase consist of highly resistant, 1 to 2 m-thick beds dipping towards buildup rinds.

The beds thin immediately around the bioherm and completely pinch out beneath it

(Fig. 3.13). The dip angle of the inter reef beds are enhanced by the differential compaction of the sediments underlying the microbial buildup. Once the inter reef bed dip is corrected for the effect of differential compaction, its true depositional geometry becomes sub horizontal and its base lines up with the flat bottom of the substratum lenses made of rip up clasts immediately beneath the buildups. A silty bed accumulated over the first grainstone bed and clearly onlaps the phase 1 buildup rind, and therefore marks the end of phase one growth. 104

Figure 3.13 First phase morphology and contemporaneous inter reef sediments; A)

Unnamed and ;B) Porcino Buildup, respectively, along James River;C) Wayne buildup - Shepard Cliff, displaying the first phase morphology and dipping beds. ;D)

Three scales of the first phase of growth: large is few tens of meter in diameter, medium is few meters in diameter and smaller is a few decimeter in diameter; E)

Largest scale, called as ‘complex’, is a few hundred of meters across. D and E are extracted from the digital outcrop model 105

Figure 3.14 Phase 2 and 3 growth of the microbial buildups; A) Part of Shepard Cliff displaying Mark buildup and the Granddaughters buildups. The phase 1 of microbial growth is onlapped by a silty unit, and over the same silty unit, microbial growth in granddaughters is observed; B) Growth of Phase 2 building upwards and outwards of Phase 1; C) Granddaughter buildups displaying backstepping morphology indicating sub-Phases 2A and 2B in Phase; D) Western edge of Wayne buildup displaying sub-Phase 2A and 2B interacting with adjacent grainstone in very high frequency; E) Andrea buildups displaying flank sediments in Phase 2 inter-fingering with inter reef sediments. Phase 3 is observed capping Phase 2; F) Drox Rock at Mill 106 creek displaying the rind and interior of Phase 1, followed by Phase 2 growth over it.

‘ Honey comb’ structures are observed in phase 2, capped by Phase 3.

107

3.4.2.4. Phase 2: “Aggrading and Laterally Expanding Phase”

The microbial bioherms continued to grow in Phase 2 contemporaneously with deposition of inter-reef bioclastic grainstones. Growth was characterized by juxtaposed columnar stromatolites. Incremental pauses in sea level rise periodically constrained vertical growth and induced lateral expansion beyond the margins of the underlying Phase 1 growth. These expansions, coupled with low synoptic relief of the bioherm, allowed intimate interaction of the reef margin with inter-reef sediment, creating an interfingering nature between boundstone and surrounding bioclastic grainstones and inhibiting the formation of a well-defined bioherm rind.

Termination of phase 2 microbial growth resulted from a deepening event and is marked by the end of deposition of bioclastic grainstone.

Phase 2 microbial buildup growth preferentially nucleates over stable and hard substratum produced by the first phase (Fig. 3.14). Phase 2 growth is defined by stromatolitic columns that grow upwards and outwards, often to extend the diameter of the buildup beyond that of the underlying Phase 1, leading to an appearance which shows Phase 2 growth directly over first silty beds which onlap

Phase 1 growth (Fig. 3.14). Phase 2 growth can be further subdivided into growth

Sub-Phase 2A and 2B on the basis of the morphology observed on Shepard Cliff. Fig.

14 clearly shows the backstepping of growth Sub-Phase 2A onto 2B. A special geomorphic feature observed in several examples is where the columns are characterized by a series of rectangular box structures, referred to as ‘a honey- comb’ structure, defined by radially oriented, stacked microbial columns each 108 separated by their own thin homogeneous micrite rind. This geomorphic feature is peculiar for parts of Phase 2, which interacts with the adjacent grainstones being deposited contemporaneously, and due to differential weathering of the flank facies, this feature displays a honey-comb morphology.

Inter-reef sediments consist of three to four, 1 to 2 m-thick grainstones beds

(Fig. 3.7-3.10). The inter reef bed thickness varies from Shepard Cliff (thinner grainstone beds) to Zesch cliff (thicker grainstone beds) as evidenced from outcrop observations (Fig. 3.7-3.10). Five sections (2-5 m thick) were measured along the

James River to interrogate the lithology and continuity of inter-reef beds (Fig. 3.15).

The grainstone beds contain skeletal grains (trilobites and brachiopods), peloids, and ooids and vary in thickness within these five sections. The grainstone beds are separated by semi-continuous to continuous, thin carbonate-siliciclastic beds

(heterolithic unit). Gamma ray logs collected at these measured sections indicate the highest values within the first heterolithic unit above the first grainstone bed, as well as the last heterolithic unit above the uppermost oolitic/peloidal grainstone which can be identified in most of these sections.

109

Figure 3.15 – The image is an oblique view of the James River looking south taken from a drone. The microbial buildups observed on the James River Cliff are colored in blue. White dotted lines represent the sections measured for the Inter reef strata.

The cliffs are not completely vertical but jump gradually up section as one move laterally. Measured sections and their Gamma ray data is correlated across to observe the continuity of the inter reef beds. Around the base of most of the sections a high gamma ray value is observed which could be correlated across. This bed represents the mixed carbonate siliciclastic unit onlapping Phase 1 of the microbial buildups on the Cliff along Llano River as observed in Fig. 3.12. A high gamma value is also observed for a mixed carbonate clastic bed which is equivalent to the bed onlapping Phase 3 on the cliffs along James and Llano River, Fig. 3.12. An ooid 110 grainstone whose thickness varies from few centimeters upto few meters is observed below the high gamma bed.

111

3.4.2.5. Phase 3: “Capping Phase”

Phase 3 of microbial buildup growth is characterized by a series of juxtaposed, well-defined bioherms, and acts as a “capping phase” for the full bioherm unit. Increased accommodation after phase two, allowed the buildups to develop significant synoptic relief, once again permitting the formation of a well- defined microbial rind. This phase is similar to the Phase 1 rind in terms of overall morphology as well as the composition (thrombolitic fabric). In some cases, Phase 3 is observed to develop better on west side of the buildups. Inter-reef sedimentation apparently ceased during Phase 3 and the demise of the microbial reef complex is marked by siltstones onlapping the buildups.

3.4.3. Upper Point Peak Microbial Buildups, Macro, Meso, and Micro Scale

Microbial buildup morphology has been investigated at different scales, referred as macro, meso, and micro (Lee et al., 2015 and references therein). Macro- scale defines the overall morphology of the buildup/buildup, meso-scale defines the internal architecture of the buildup or buildup, and micro-scale defines the micro elements which contribute to the internal architecture of the buildup/buildup.

The microbial buildups of this study reveal mounded macro-scale morphology, and stromatolitic and thrombolitic meso-morphology. Micro-scale morphology is not the focus of this paper, although previous studies indicate the presence of calci-microbes such as Girvanella, Tarthinia, Tarthinia, Epiphyton, and

Renalcis (Spincer, 1997). 112

113

Figure 3.16 Palaogeographic controls on depositional environments; A)

Paleogeographic map of Upper Cambrian time period (Furongian – 497-485.4 ma) modified from Blakey (2013), displaying the direction of Trade Winds over

Laurentia and the location of field area in central Texas; B) James River Pavement

(Orthophotograph – 600x200 m), displaying Phase 1 microbial buildups mapped on the orthophotograph. The microbial buildups were mapped manually and then color coded based on their major direction of orientation in Arc. Map 10.1. A rose diagram is plotted to demonstrate the major direction of orientation for all the buildups; C)

A zoom-in of the James River Pavement orthophotograph looking at Pi Buildup, colored in blue, in a horizontal cross section. Microbial debris (green) is observed on northwest and south east part of Pi buildup. Additionally microbial debris is also observed in northwest and southeast side of other microbial buildups in the James

River as well as shown in C; D) A zoom-in of the orthophotograph of the James River looking at Twin herms. Inter reef grainstones are exposed in detail at Twin herms and different grainstone beds are colored with different colors, the older is orange and the younger is red. At the edge of the microbial buildup a moat is observed colored in yellow. The grainstone units have preserved asymmetric arcuate ripples

(dotted black lines), and larger features display dune geometry; E) This represents

James River orthophotograph as in B, but the focus here is the inter-buildup channel pattern. The microbial buildups are colored with blue diamond pattern and the channel is colored with orange.

114

3.5. Discussion

Previous regional studies show the study area to be a part of a low-latitude mixed shelfal carbonate-siliciclastic system, characterized by the lateral juxtaposition and vertical stacking of neritic carbonates and siliciclastics (Lochman-

Balk, 1970; Golonka, 2007, Morgan, 2012). The depositional setting has been previously interpreted as a rimmed margin platform (Barnes and Bell, 1977), although in absence of metazoans it would be difficult for the microbialites to build barrier reefs. Therefore, to reconstruct the depositional model of the Upper

Cambrian Microbial buildups, an understanding of the paleogeography, stratigraphy, modern mixed systems, and modern microbial systems is required

(Lochman-Balk, 1970; Purdy and Gischler, 2003; Reid et al, 2003; Andres and Reid,

2006; Tcherepanov et al., 2008; Droxler and Jorry, 2013; Lee et al., 2015).

The unique finding of this study is recognition and documentation of the three-phase evolution of the microbial unit. Two major factors significantly influencing growth of the buildups are sea level fluctuations and siliclastic input. In mixed carbonate-siliciclastic settings, siliciclastics commonly move seawards on the shelf due to sea –level falls and thereby shut down the carbonate factory. In contrast, during sea –level rises siliciclastics are pushed landward up the shelf, allowing the carbonate factory to reestablish on top of the previous lowstand siliclastic deposits (Droxler and Jorry, 2013). The influx of siliciclastics could also be increased due to change in climate (monsoon). Water depth variation can be reasonably inferred by morphological differences between the growth phases of the 115 buildups. Phase 1 reefs (oblong, 3-5 m high, rind on outer surface), indicating at least 3-5 m water depth (relief of the first phase), were immediately succeeded by the arrival of siliclastics into the system which onlap the buildups and mark the end of phase 1. During the growth Phase 2, when stromatolitic columns grow upwards and outwards and interacting with grainstones, the siliciclastics in the inter-reef sediments are interbedded with grainstones indicating a very shallow environment, not too far from the coast. During Phase 3 buildup growth, the absence of siliciclastics and grainstones in the inter-reef area indicate a relative deepening.

3.5.1. Paleogeographic Setting

3.5.1.1. Trade Winds

Winds and wind-driven currents most likely played a role in defining the dominant NE-SW orientation of the microbial buildups along the James River outcrops. During the Upper Cambrian, the direction of the orientation would be SW-

NE (since the North American continent was tilted 90 degree to the east). Fig. 3.16 A shows that the major axis of the buildups is parallel to the direction of trade winds

(SW-NE). Additionally, asymmetric ripples have been observed in inter-buildup skeletal (trilobite and brachiopod), peloid, and ooid grainstones perpendicular to the major axis of the microbial buildups. Winds could have also played an important role in bringing fine grained terrigenous sediments to the shallow seas of the Great

American Carbonate Bank. 116

3.5.1.2. Tides

Chafetz (1973) and Cornish (1975) have suggested the presence of tidal influence in this area, whereas Spincer (1999) suggested none. The presence of microbial hash (defined as fined grained microbial debris) on the northeast and southwest edges of the buildups indicates a bi-directional tidal flow, suggesting that the former hypothesis is correct (Fig. 3.16 C). Additionally, fine-debris on the NE edge is smaller than that of the SW indicating that the major flow direction was from

NE to SW (James River) or from SE towards NW in the Upper Cambrian Time. Tides play an important role in burying and exposing the microbialites with ooid grainstones in a cyclic manner at Highborne Caye (Andres, and Reid, 2006).

Mapping the microbial buildups in the James River, a channel like feature is observed between the buildups (Fig. 3.16 E). This channel is interpreted on the basis of presence and absence of the microbial buildups in the James River that was influenced by tidal currents in transporting and redistributing the grainstones as well as the microbial hash.

117

118

Figure 3.17 Composite Depositional model for Upper Cambrian microbial buildups with two separate carbonate factories on a mixed system ramp. There are two separate units, the Lower and the Upper Point Peak units. The Lower Point Peak in the study area represents inboard platform facies and the Upper Point Peak represents outboard platform facies. Therefore a composite depositional model is developed where the upper Point Peak is used as analog to the downdip facies of the lower Point peak inboard platform facies. The inboard strata mostly consist of smaller microbial buildups (three different types – less influence by siliciclastics downdip on the ramp) and mixed carbonate clastic sediments with some sands. The outboard platform facies consists of the larger microbial unit which gets as thick as

14 m.

119

3.5.1.3. Storms and Waves

The presence of flat pebble conglomerates just below the microbial unit indicates the strength of currents during a transgression, prior to nucleation of the microbial buildups.

3.5.2. Depositional Model

The outcrops examined in this study expose the complete thickness of the

Point Peak Member of the Wilberns Formation, which includes the lower and upper microbial units. These units together represent a composite section of inboard and outboard microbial facies, stacked on top of each other, and are used to develop a depositional model of the area. Lower Point Peak outcrops in this locality consists of smaller microbial biostromes and bioherms (<1 m in height), associated with heterolithic facies, glauconitic siltstones, detrital sands indicating very shallow subtidal to intertidal environments. The Shark Bay domal stromatolites are similar in their morphology and size to the Lower Point Peak microbial buildups and therefore could be used as analogs (Reid et al., 2003). The shark bay domal stromatolites grow in supratidal-shallow subtidal environments which could be used as environment of deposition for the different types of the Lower Point Peak microbial buildups. The Upper Microbial unit, up to 14 m thick, is associated with thick skeletal and oolitic grainstones indicating subtidal environments (with little fluctuations in base level bringing siliciclastics within the system).

Based on the facies associations of smaller microbial buildups from the literature and field studies, heterolithic facies are interpreted to represent the most 120 inboard facies in the depositional model of Fig. 3.17. These facies change downdip to skeletal (trilobite and brachiopod), peloid, and ooid grainstone shoals associated with tidal channels. Small-scale bioherms and biostromes develop both in front of and behind the grainstone shoals. Downdip from these facies exists a relatively open shelf where glauconitic silty beds and heterolithic units are deposited. Eocrinoids are also observed indicating more open marine conditions. These facies grades downdip into skeletal (trilobite and brachiopod) grainstone shoals and ooid shoals, associated with larger microbial buildups. The most downdip facies is the heterolithic facies associated with the interclastic conglomerate.

The composite depositional model (Fig. 3.17) displays an extensive ramp-like shelf greater than 100 km, consistent with the regional setting documented for the

Upper Cambrian in this area (Lochman-Balk 1970a; Golonka, 2007; Morgan, 2012).

The study area examined does not extend updip to the paleo shoreline, although, regional studies indicate the shoreline to be towards West/ North-West (Lochman-

Balk 1970a; Morgan, 2012). Inner shelf facies are siliciclastic influenced as evidenced in regional depositional maps (Lochman-Balk 1970a; Golonka, 2007).

Progressing downdip, the first carbonate environment is recognized, potentially developing shallow shoals formed by ooid and skeletal grainstones. Three types of smaller microbial buildups are found associated with the ooid/skeletal grainstone shoal facies belt (more updip, around the middle, and more downdip) and form the

Lower Microbial Unit within the Point Peak. In general, the updip smaller microbial heads incorporate more siliclastic sediments the siliciclastic influence gets less in the second type, and no siliciclastics are present in the downdip examples. Downdip 121 of the shoals and small buildups is an extensive shelf where low sedimentation rates are causing the development of glauconites (Potential casues for low sedimentation rate might be a flexure in the broad shelf and/or the shelf is more current swept downdip). Additionally, the presence of eocrinoid mounds is observed downdip of the shoal-buildup belt indicates a more open ocean conditions and suggests the absence of any barrier downdip. It is difficult to fully discern the downdip facies of the lower microbial unit, although, the observations through regional sections indicate the microbial buildups get thick as much as 30 m.

Regional transgressive-regressive cycle maps (Lochman-Balk, 1970a) indicate a major transgression during the time period that the Pointpeak Member was deposited, supporting that the downdip facies should have transgressed and stepped towards the shoreline, placing it above the inboard shelf strata in the study area examined. Therefore in the “composite” generalized depositional model (Fig.

3.17), the upper microbial unit is used to approximate the downdip facies of lower microbial unit to complete the depositional . Potentially, this downdip microbial belt could be anywhere from 40-60 kms from the updip one. The downdip microbial belt

(Upper Micfrobial Unit) is represented by microbial buildups displaying a major axis of orientation NE-SW, are as much as 150 m in length, and have microbial debris around the both ends of their major axes. Ooid grainstones and skeletal grainstones are also a major inter-reef sediment. These microbial buildups are up to 14 m thick and detailed observations indicate three different phases phases that were previously described but are not depicted in the generalized depositional model. 122

123

Chapter 4

Spatial Statistical Analysis Utilizing High-Resolution Drone Imagery: Upper Cambrian Microbial Buildups, Central Texas

The following chapter is a reproduction of Khanna, P., Pyrcz, M., Droxler,

A.W., Lehrmann, D., Harris, P.M., (in prep.).

Abstract

Quantification statistical methods help understanding biogenic reef morphologies, their growth patterns, trends, and scaling relationships. A new Arc

G.I.S tool - Spatial statistical analysis, with general statistical techniques, is used to provide meaningful geological information on fossil microbial reef growth. Upper

Cambrian microbial buildups, in the Point Peak Member of the Wilberns Formation

(South Mason County, Central Texas) display a distinct 3 phase morphological 124 evolution. A world class pavement outcrop, 600 x 200 m in surface area along the banks of the James River, displays a horizontal section through Phase 1 of the upper

Point Peak Microbial buildup unit. High resolution drone imagery is used to develop a digital outcrop model of the James River Pavement. A total of four scales, small

(few dm), medium (few m), large (few tens of m’s), and complex (few hundreds of m) microbial buildups have been identified on the James River Pavement. A total of nine attributes – length, width, area, perimeter, anisotropy, circularity, eccentricity, and compactness are studied for small, medium and large microbial buildup scales.

The two microbial buildups at the complex scale on the James River Pavement cannot be quantified in detail due the limited number of their occurrences.

Univariate, multivariate, multi-dimensional scaling, grouping analysis, Ripley’s K analysis, and Getis Ord GI statistics are used to identify meaningful attributes, trends, and clustering relationships. The major results of the statistical analyses point out that, as the size of the mapped polygon feature increases, the feature shape becomes more complex, less circular, and orients towards NE-SW. On the other hand, fractal or multi fractal relationships do not apply to the overall growth of Phase 1 buildup growth.

4.1. Introduction

Microbialites, a term coined by Burne and Moore (1987), is used to refer to stromatolies, thrombolites, and similar biogenic sedimentary accumulation created by the interaction of microbes/bacteria and sediment particles through trapping 125 and biding, in addition to direct precipitation (Riding, 2011). They are benthic microbial communities which represent the signatures of early life on earth captured in the rock record as well as the oldest form of life to evolve and survive until present (Allwood et al., 2006; Riding, 2011). Localities with fossil and modern microbial buildups/reefs are the Nature’s best repositories to probe their overal morphologies at different scales to understand their microbial growth patterns and their spatial architecture (Jeffery et al., 1996; Reid et al., 2003; Burns et al., 2004;

Allwood et al., 2006; Andres and Reid, 2006; Chagas et al., 2016; Chirayath and

Earle, 2016; Lee et al., 2016). Although modern microbial buildups provide excellent

3D exposures of microbial growth morphologies (Chagas et al., 2016; Chirayath and

Earle, 2016), the ancient microbial growth morphologies and patterns would be the best analogs in learning about early life evolution on Earth and even how initial extra-terrestrial life could be recorded (McKay and Stoker, 1989; Allwood et al.,

2006).

Spatial statistical analysis of microbialite accumulation is required to understand their 2D/3D distributions, patterns, and relationships, and ultimately to define the different processes involved in their growth (Scott and Janikas, 2010).

The analysis identifies statistically significant spatial clusters or spatial outliers, patterns of clustering or dispersion, group features, group scale, and spatial relationships. Latest technology (Drones) and access to world-class outcrops are the two main requirements to conduct detailed spatial statistical analysis of any microbial buildup setting. High-resolution drone imagery collected in 2014 over newly accessible world-class outcrops of Upper Cambrian microbial reefs from 126

Figure 4.1 Stratigraphic column and measured sections of Point Peak Member,

Wilberns Formation across Llano uplift.

A) Stratigraphic column for Cambrian strata Llano Uplift (modified from Barnes and

Bell,1977 and Kyle and Mcbride, 2014). The microbial buildups within Point Peak

Member of Wilberns Formation are the focus of this study. Legend: Blue color –

Limestone, Purple- Dolomite, Yellow- Sandstones, Green – Shales, Granites – Red; B)

Location of the filed area, Mason County, Central Texas. James and Llano Rivers, and

Mill Creek expose Upper Cambrian Microbial buildups. The red rectangle on James

River is the focus of the study exposing microbial buildups in a pavement view.

127

Central Texas provides a unique opportunity to interrogate these microbial buildups with spatial statistical analysis tools (Fig. 4.1-4.2). Along the banks of the James and

Llano Rivers and Mill Creek in southern Mason County, these reefs are spectacularly exposed in 3D on pavements as large 600 m x 200 m in surface areas, and also in 2D along several 400-500 m wide and 40-50 m high cliffs (Fig. 4.3-4.4).

The main objectives of this study are to utilize drone imagery to map the

Upper Cambrian microbial reefs, interrogate in depth their spatial relationships/statistics, and to define patterns and trends in their growth based on spatial statistics. Different types of spatial statistical analyses such as Ripley’s K analysis, Getis Ord Gi Statistics, Incremental Spatial Autocorrelation, Multi- dimensional scaling, and Grouping Analysis (Scott and Janikas, 2010) are used to identify their growth patterns, trends, relationships within and in between different scales, and fractal relationships.

4.2. Geological Background and Stratigraphic Context

4.2.1. Upper Cambrian Central Texas

The Llano Uplift in Central Texas relates to a structural dome of Precambrian igneous and metamorphic rocks exposed as a structural window, about 100 km east/west and 70 km north/south (Garrison et al., 1979). Middle Cambrian to early

Ordovician sedimentary strata, referred to as the Moore Hollow Group, were unconformably deposited on the Llano Precambrian basement, (Bridge et al., 1947;

Ahr 1967, 1971; Barnes and Bell, 1977; Morgan 2012; Miller et al., 2012). The 128

Moore Hollow Group consists of two Formations – the Riley and Wilberns

Formations. Riley Formation is made up of three members – the Hickory Sandstone,

Cap Mountain Limestone, and Lion Mountain sandstone Members. The Wilberns

Formation overlies the Riley Formation and includes four members – the Welge sandstone, Morgan creek limestone, Point Peak, and San Saba Members (Fig. 4.1A;

Paigne, 1911; Bridge et al., 1947). The Moore Hollow Group is overlain by the

Ordovician Ellenburger Formation (Cloud and Barnes, 1948).

4.2.2. Upper Point Peak Microbial Unit – Wilberns Formation

The microbial reefs examined in this study lie within the Point Peak Member of the Wilberns Formation. The Point Peak Member has been mapped within the

Llano uplift area by Bridge et al., (1947) and Ahr (1967). Two distinct, lower and upper, microbial reef units are present within the Pointpeak Member (Chapter 3;

Khanna et al., in prep.). The upper microbial reef unit, spectacularly exposed along the James and Llano Rivers, and Mill Creek as pavement and cliff outcrops, is the focus of this study. The microbial buildups,, up to 14 m in height, are surrounded with thick skeletal and oolitic grainstone beds, intercalated with thin mixed carbonate siliciclastic silty beds. These microbial buildups are interpreted to have grown during a sea level transgression in subtidal environments; small fluctuations in base level brought, at different times siliciclastics into the microbial reef system.

Systematic and comparative analyses of the multiple microbial buildups and their inter reef sediments, exposed along Shepard and Zesch Cliffs, allow identifying their morphology and vertical and lateral facies, in addition to similarities and 129 heterogeneities among them within the microbial unit. Moreover pavement outcrops provide unique opportunities to observe the morphology and facies lateral variability among, within, and between several buildups. A systematic three-phase growth evolution of the microbial buildup became obvious first in direct field observations, and then on drone imagery, combined with the qualitative and quantitative analysis of the buildup morphology using photogrammetry data (Fig.

4.2, 4.3).

4.2.3. Three phase evolution of Upper Point Peak microbial unit

The Shepard and Zesch Cliffs along the Llano River expose the strata underlying and overlying the microbial buildup interval and, therefore, provide a full record of the depositional environments that existed prior to the initial establishment, growth, and final demise of the microbial buildups. The sedimentary strata underlying the upper Point Peak microbial reef unit consist of a succession of siltstone, sandstone, grainstone and packstone beds of mixed carbonate/siliciclastic composition (Sliger, 1957). A marked change from mixed carbonate-clastic to pure carbonate systems , referred to as the ‘switch’, is defined by a discontinuous transgressive lag of rip up clasts from the lower strata (Droxler et al., 2016). When cemented, lenses of rip up clasts became the nucleation ground or substratum for the microbes and bacteria to initiate the microbial buildup growth. ‘Mitch Herm’

(Fig. 4.3), exposed along the eastern side of Zesch Cliff on the Llano River, clearly illustrates the overall and general evolution of the buildups, from their establishment, clear three-phase growth, and ultimate demise. 130

The colonizing growth phase 1 was initiated on low-relief lenses of flat rip-up clasts, interpreted as a transgressive lag and occurred contemporaneously to the deposition of an inter-reef bioclastic grainstone bed (Droxler et al., 2016). Phase 1 buildup reefs are typically 3-5 m high, oblong in shape, and display well defined margins referred to as thrombolytic rind, enclosing numerous stromatolite columns in the buildup interior (Chapter 3; Khanna et al., in prep.). The James River outcrop represents a spectacular locality where the Upper Point Peak microbial buildups cover a portion of the river banks as pavement and cliff about 600 x 200 m wide.

The James River pavement is horizontally cut through the first phase of microbial buildup growth and hence provides arguably the best location to investigate the buildups and inter-buildup sediments of the first phase in great detail (Fig. 4.3).

Four different scales of clustering within the first phase have been identified with the James River digital outcrop (Fig. 4.4). The smallest scale (few decimeters in dimensions), represented by numerous circular to oblong microbial columns, cluster to develop medium scale grouped columns (few meters in dimensions).

Medium scale cluster to form large scale (few tens of meters in dimensions) groupings, and associated large scale clusters form the largest scale referred to as a

‘complex’ (few hundreds of m in dimensions). All the scales have their own rind separating each grouping from the other. These scaling relationships within the first microbial buildup growth phase are the focus of this paper.

131

Figure 4.2 Microbial Buildup three phase evolution model – Mitch Herm- Zesch Cliff is used to demonstrate the three different phases based on morphology.

3 growth phase interpretation of Mitch Herm and coeval sediments showing lower

Point Peak below ‘the switch,” (dark orange line) the upper Point Peak microbial interval, and the overlying San Saba Limestone.

132

The inter reef sediments, contemporaneous to phase 1, consist of a highly resistant, 1 to 2 m-thick bed dipping towards the buildup rinds, forming a discontinuous moat around the bioherm, and pinching out underneath the buildup

(Fig. 4.3). A silty bed accumulated over the first grainstone bed and clearly onlaps the Phase 1 buildup rind and, therefore, marks the end of Phase 1 growth

Phase 2 buildup growth, initiated on top of the Phase 1 reefs, is characterized by the vertically aggrading and laterally expanding growth of the reefs in direct interaction with the inter reef ooid and bioclastic grainstone and packstone sediment (Fig. 4.3). Phase 3 buildups grew as well defined individual reefs with well-developed rinds, capping the complete microbial reef complexes and even expanding further on top of Phase 2 inter reef sediment. Phase 3 buildups lack of inter reef sediment, contrasting with the first two growth phases during which inter reef sediment are always contemporaneously accumulating. A second major silty unit is observed onlapping the third phase, marking the ultimate demise of the microbial reef growth (Fig. 4.3).

4.3. Data Collection, Biases, and Errors

An aerial survey was conducted in February 2014 to collect digital photographs over eleven outcrops, three pavements and seven cliffs including the

Morgan Creek Limestone, Point Peak and San Saba Members. 133

4.3.1. Data Acquisition

Before the drone survey, for the pavement outcrops, markers were placed in the field and the precise locations for several markers were collected using DGPS

(Differential GPS). A Trimble Total Station unit was used to determine the location coordinates of all the markers in a local coordinate system with an accuracy of 10 cm. DGPS accuracy of marker location, with respect to each other, is less than what could be achieved from Trimble Total Station, therefore both DGPS and Total Station were used.

For James River outcrop, the drone was assigned to fly 40 m above the ground, and the flight path was automated. During the data collection 90% overlap was kept and a total of 540 photographs were collected in three flights.

4.3.2. Data Processing

The collected digital photographs were imported to Agisoft 1.0 to build virtual outcrop models. After adding the photographs, camera positions, saved in

134

Figure 4.3 – Orthophotograph of James River Pavement displaying displaying horizontal cross section (pavement) through Phase 1 growth of Upper Point Peak microbial buildups, Wilberns Formation.

135

Figure 4.4 – Scaling relationships of the Phase 1 microbial buildups.

Three scales of Phase 1 growth (upper three photographs); Smaller scale, few decimeter in diameter, medium scale, few meters in diameter, and Large scale, few tens of meter in diameter, Largest scale (lower photograph), refered to as ‘complex’ is represented by Don buildup on James River pavement (few hundred of meters in dimensions). 136

EXIF file format (Exchangeable Image File Format), were also loaded in Agisoft.

Further, the Align Photograph tool was used to refine the camera position for each photo and to build an initial 3D point cloud model. To improve the accuracy of the location of the 3D point cloud model, ground control points (GCP’s) were integrated with the dataset. Since the original point cloud is v sparse, a Build Mesh tool is selected to create a polygonal mesh which uses the 3D point cloud to make edges, vertices, and faces. After building the mesh, the location data for the markers is used as ground control points and is imported in Agisoft. These ground control points help in accurately locating the initial 3D point cloud. Further, a dense point cloud is created using a Build Dense Cloud tool and, in the final step, another mesh is created with the dense 3D point cloud. The virtual outcrop model is exported in 3 different file formats, TIFF file format (exports an Orthophotograph of the outcrop), DEM

(Digital Elevation Model in xyz format), and KMZ (google earth readable file format).

4.3.3. Mapping 1st Phase

First phase microbial buildups are exposed on the James River Pavement

(Fig. 4.1B). Due to the similarity in color as well as reflectivity of the different facies, it is difficult to conduct an unsupervised classification to map the outlines of the first phase buildups. Therefore, the buildups have been mapped manually (Fig. 4.5) in

Arc G.I.S. 10.1.

137

Figure 4.5 – Small , Medium, and Large scale polygons mapped on James River

Pavement.

138

4.3.4. Sources of error and Biases

There are two main sources of error: First – distortion in the orthophotograph of the James River which is developed based on Agisoft. Second –

Manual interpretation of the first phase buildups exposed in James River.

The errors and uncertainties are calculated in the Agisoft after the successful processing of each dataset. The error estimates for James River are on the order of less than 1%. Therefore, these uncertainties are not included in the statistical analysis.

The estimated error based on manual interpretation is projected to be less than 5%.

4.4. Spatial Statistical Analysis

The James River outcrop is a 600 m long and 200 m wide pavement which spectacularly exposes the Upper Cambrian microbial buildup unit in the Upper Point

Peak member and the basis for this study (Fig. 4.3). This pavement represents a horizontal slice through growth Phase 1 of the microbial buildups. This study deals with the spatial statistics of the Phase 1 only. The major observations based on the

James River Pavement indicate that there are at least 4 different scales that can be mapped within the microbial unit (Fig. 4.4). James River pavement displays circular to oblong shaped microbial buildups. To identify these different scales, maximum length/width is used; four groups are established: the smallest scale is few dm, 139 medium scale is few m, large scale is few 10’s of m, and the largest complex scale is few 100’s of m (Fig. 4.4). The presence of large scale can be confirmed by visiting

Zesch and Shepard Cliffs where the maximum exposed width/length of the oblique cut through these microbial buildups represents 65-90 m. Since only two examples of largest scale are observed in James River, the detailed spatial statistics of the largest scale have not been calculated.

4.4.1. Univariate Analysis

Univariate analysis is the simplest and straightforward way of describing the data.

A total of nine attributes- width, length, area, perimeter, orientation, anisotropy, circularity, eccentricity, and compactness - have been analyzed for the three different scales of the Phase 1 microbial buildups. P10, P50, and P90 are displayed in Table 1 for the nine attributes for small, medium and large scales. P10, P50, and P90 are statistical confidence level of an estimate, where, P90 represents that 90 percent of values exceed this number, P50 represents the middle estimate, and P10 represents that only 10 percent values exceed this number. P50 for the small, medium, and large scales indicates that the attributes for small scale (length or width) are a few decimeters, medium scale (length or width) are a few meters, and large scale (length or width) are a few tens of meters in dimensions.

140

Small

Width Length Area Perimeter Orientation Anisotropy Circularity Eccentricity Compactness (m) (m) (m2) (m) (degree)

P90 0.12 0.19 0.52 0.02 23.59 1.22 0.76 0.57 13.59

P50 0.19 0.30 0.81 0.04 91.20 1.51 0.86 0.75 14.55

P10 0.30 0.46 1.20 0.10 154.73 2.09 0.93 0.88 16.54

Medium

Width Length Area Perimeter Orientation Anisotropy Circularity Eccentricity Compactness (m) (m) (m2) (m) (degree)

P90 2.48 3.93 7.98 10.76 11.35 1.22 0.78 0.58 13.41

P50 4.23 6.58 23.88 18.81 50.28 1.56 0.87 0.77 14.46

P10 7.05 11.84 62.92 30.21 153.98 2.05 0.94 0.87 16.17

Large

Width Length Area Perimeter Orientation Anisotropy Circularity Eccentricity Compactness (m) (m) (m2) (m) (degree)

P90 10.68 17.63 142.74 45.32 21.54 1.46 0.74 0.73 14.22

P50 15.39 29.14 318.06 69.92 34.50 1.73 0.83 0.82 15.21

P10 22.94 38.97 579.78 95.96 78.08 2.36 0.89 0.91 17.02

Table 4.1 –The P10, P50, and P90 values of nine attributes (width, length, area, perimeter, orientation, anisotropy, circularity, eccentricity, and compactness) are listed.

141

Figure 4.6 – Small , Medium, and Large scale polygons mapped on James River

Pavement color coded on the basis of anisotropy a), and orientation, b). The frequency plots are also plotted for anisotropy and rose diagrams for orientation.

142

Mapping the small, medium, and large scales, and interrogating each attribute individually provides first order information about the growth patterns and trends of the microbial first phase growth. Fig. 4.6 shows color coded anisotropy distribution for small, medium, and large scales, and the frequency distribution of the anisotropy values. We define anisotropy as length divided by width. The first observations indicate random distribution of anisotropy values. The maps for orientation display strong NE-SW trends of orientation for large and medium scales, although for small scale a very light signal towards NW-SE is observed. Additional analysis of small scale features (Fig. 4.7), which form clusters and make up the interior of the medium scale, indicates that each cluster of small scale has a trend. The trends for these small scale clusters (representing interior of different medium scale) vary in direction from N-S, NE-SE, NW-SW, and E-W.

4.4.2. Multivariate Analysis

Multivariate analysis is conducted to identify linear dependencies of multiple attributes on each other at the same time (Fig. 4.8). By plotting correlation matrixes, the multivariate statistical analysis is conducted. Correlation matrixes have all the nine attributes (width, length, area, perimeter, orientation, anisotropy, circularity, eccentricity, and compactness) on the x axis and y axis each. In the correlation matrix there are a total of 81 boxes. Each box represents comparison of two attributes. Nine boxes by default 143

Figure 4.7 Small scale polygons mapped on James River Pavement form clusters and make up the interior of the medium scale. For each cluster of small scale, a rose diagram is plotted related to their orientation.

144

Multivariate correlation - Small Scale Multivariate correlation - Medium Scale

Multivariate correlation - Large Scale

Figure 4.8 Correlation matrix for Small, Medium, and Large scale polygons.

Correlation matrixes display scatter plots and correlation coefficients.

145 represent comparison of an attribute with itself, represented by dark green color, which should be discarded. The remaining 36 boxes on either side of the dark green band represent identical comparisons of two attributes. Therefore, one set of 36 boxes is displayed with the correlation coefficients and the other set of 36 boxes is represented by scatter plots (Fig. 4.8). The correlation coefficients have a value from

-1 to 1. If the value is 1, -1 or near to them, then the relationship between the two attributes is linear. If the value is 0 or close to zero then there is no direct relationship between the attributes.

Fig. 4.8 displays the correlation matrix for small, medium and large scales. The multivariate correlation analysis demonstrates that across scales (small, medium, and large), there are three major multivariate relationships. First, there is a direct relationship between width, length, area, and perimeter. Second, there is a direct relationship between anisotropy, eccentricity, and compactness. Third, circularity has an inverse relationship with anisotropy, eccentricity, and compactness. With this analysis no direct relationships of orientation were found dependent on any other attribute.

4.4.3. Multi-dimensional Scaling (MDS)

Multidimensional Scaling (MDS) statistical analysis is conducted to identify whether the attributes are redundant or if they provide new information (Fig. 4.9).

On the basis of similarities and dissimilarities between the attributes, the MDS arranges the data in space, plotted on two or three dimensions. In this study the

MDS results are plotted on 2D. 146

Small Scale (MDS) Medium Scale

Large Scale (MDS)

Figure 4.9 Multi- Dimensional Scaling analysis for Small, Medium, and Large scale polygons (the 3 axis are plotted according to all the attributes and are dimensionless). 147

The MDS analysis indicates that there are four groups that can be identified, each in small, medium, and large scales. The attributes width, length, area, and perimeter in first group; orientation in second group; anisotropy, eccentricity, and compactness in third group; and circularity in fourth group. The groups have been divided based on similarities and dissimilarities, therefore, only one attribute from each group provides meaningful information, and attributes within the same group are redundant. Therefore, the attributes - length, orientation, circularity, and compactness could be used to represent all the meaningful information. The grouping analysis will be conducted for these variables.

4.4.4. Grouping Analysis

MDS analysis demonstrated that only four (length, orientation, circularity, and compactness) out of the nine attributes would provide meaningful information.

The grouping analysis is conducted to identify groups on the basis of these four attributes for the small, medium, and large scale (Scott and Janikas, 2010; Fig. 4.10).

The grouping analysis uses similarities and dissimilarities to identify groups on the basis of these four attributes. The optimum numbers of groups identified are two for all the three scales. Box plots are used to display the grouping analysis results (the axis on box plots show standardized values). Compactness or shape complexity increases as its values increases. Circularity is represented by values ranging from 0 to 1 and as the value approaches 0, the object represents more circular shape.

148

Figure 4.10 – Grouping analysis for Small, Medium, and Large scale polygons based on length, orientation, circularity, and compactness.

149

For all three scales (small, medium, and large) in general the shape complexity increases as shape length is increases, and circularity decreases. The large scale and medium scale features display NE-SW orientation as the length and complexity are increasing, whereas circularity decreases. For small scale no orientation trends are demonstrated by grouping analysis.

4.4.5. Ripley’s K Analysis

Small, medium, and large scale features mapped on the James River pavement are tested for clustering or dispersion with Ripley’s K analysis (Scott and

Janikas, 2010; Fig. 4.11). Ripley’s K analysis is a method to identify how features/points within a gives are occur (clustered/dispersed).

Ripley’s K analysis generates K function plots. Fig. 4.11 is used to explain what

K function plot represents. For small and medium scale statistically significant clustering is observed at smaller distances with statistically significant dispersion at larger distances.

For the large scale clustering is observed at smaller distances but statistically significant dispersion is observed at large distances.

4.4.6. Getis Ord Gi Statistics

Getis Ord Gi Statistics is a hot spot analysis tool (Scott and Janikas, 2010; Fig.

4.12). This tool identifies clustering of similar values and locates the hot spots. A feature with high value if found surrounded by similar values would become a hotspot in comparison to a high value feature surrounded by smaller values. This analysis is conducted for all the scales and interesting results are observed. 150

Figure 4.11 – Ripley’s K analysis for Small , Medium, and Large scale polygons.

151

Figure 4.12 – Getis Ord Gi Statistics- hot spot analysis for Small , Medium, and Large scale polygons. The color scale is dependedent on the degree of clustering. It does not take into account the size/shape of the attribute.

152

For small, medium and large several hot spots are observed indicating clustering of similar scale features.

4.5. Discussion

Spatial statistical analysis conducted on drone imagery of Upper Cambrian microbial buildups has characterized their depositional patterns in a unique manner and provided added insight into their development. Meaningful information is extracted from the statistical analysis to characterize growth patterns, trends, size, and shape relationships of the buildups. Univariate analysis provides information about three different scales of buildup clustering that are mapped over the first phase buildups in the

James River pavement. The three scales vary in shapes and in sizes from decimeter, to meters, to tens of meter scale. The largest scale mapped on the James River pavement has dimensions of a few hundreds of meters (the two complex scales from James River show major axis length to be 110 and 170 m). The largest scale of clustering is confirmed by visiting a cliff outcrop on Zesch cliff, where the Gene buildup has an oblique exposure of about 85-90 meters comparable to the largest scale of clustering in the James River.

These results confirm three levels of scaling and provide direct evidence for the fourth scale.

To understand the relationships between the different attributes, the multivariate analysis is conducted. This analysis has been used to demonstrate linear relationships through correlation coefficients and scatter plots. For all the scales (small, medium, and large), microbial growth patterns indicate linear relationships between these two sets of attributes; first set - length, width, area, perimeter, second set - anisotropy, circularity, 153 eccentricity, and compactness. Random relationship between orientation and the two sets of attributes is observed. These results indicate that out of the nine attributes several of them provide similar information and, therefore, are redundant. MDS analysis (multidimensional analysis) confirms and point out that length, orientation, circularity, and compactness provide meaningful information and the other attributes (width, area, perimeter, anisotropy, and eccentricity) are redundant.

Utilizing the similarities and dissimilarities of length, orientation, circularity, and compactness, the grouping analysis is conducted for all the scale. This analysis demonstrated that as length is increasing, the shape complexity is increasing and circularity is decreasing. This demonstrates that as the length within a scale is increasing, the shape is becoming more complex and less circular. The smallest features within each scale are therefore more circular and they develop into more complex form as the size of the features increases due to clustering. The direction of orientation is observed to be oriented more towards NE-SW as the length and complexity are increasing.

To identify and demonstrate clustering of different scale features, Ripley’s K analysis was utilized. It clearly demonstrated the clustering within the three different scales mapped on James River. Getis Ord GI analysis was used to identify clustering of similar features within a scale. Within small scale features, medium scale features, and large scale features, clustering of similar size features was observed. Either because of the biological controls or due to any of the environmental controls such as tides, wind, waves, and storms, the clustering of similar features occurred. 154

References

Abbey, E.A., Webster, J.M., Beaman, R.J., 2011. Geomorphology of submerged reefs on the shelf edge of the Great Barrier Reef: The influence of oscillating Pleistocene sea-levels. Mar. Geol. 288, 61–78.

Acker, K.L., 1987. The Carbonate-siliciclastic Facies Transition in the Modern Sediments Off the Northeast Coast of Barbados, [M.S. thesis]: Montreal, Quebec, McGill University, 230 p.

Agassiz, A., 1903. The coral reefs of the Maldives. Mem. Museum Comp. Zool. Harvard Coll. Vol XXIX 29, 1–168.

Ahr, W.M., 1967, Origin and palaeoenvironment of some Upper Cambrian algal reefs, Mason County area, Texas: Rice University, Houston, [PhD Theis]: Houston, Texas, Rice University, 130 p.

Ahr, W.M, 1971. Paleoenvironment, Algal structures, and fossil aglae in the Upper Cambrian of Central Texas, Journal of Sedimentary Petrology, Vol. 41, No,1, p.205- 216.

Alley, R., B., Clark, P., U., Huybrechts, P., Joughin, 2005. Ice-sheet and sea-level changes. Science 310:456–60.

Alley, R. B., Mayewski, P. A., Sowers, T., Stuiver, M., Taylor, K. C., Clark P. U, 1997. Holocene climatic instability: A prominent, widespread event 8200 yr ago. Geology, 25 (6): 483–486.

Allwood, A.C., Walter M.R., Kamber, B.S., Marshall, C.P., and Burch I.W., 2006. Stromatolite reef from the early Archaean era of Australia. Nature, Bol. 441, 8.

Anderson, R.C., 1998. Submarine topography of Maldivian atolls suggests a sea level of 130 metres below present at the last glacial maximum. Coral Reefs 17, 339–341.

Andréfouët, S., 2012. Biodiversity, ressources, and conservation of Baa Atoll (Repubic of Maldives) a UNESCO Man and Biosphere Reserve, Atoll Research Bulletin, 590 : 1-247. 155

Andres, M.S. and Reid, R.,P., 2006. Growth morphologies of modern marine stromatolites: A case study from Highborne Cay, Bahamas. Sedimentry Geology 185, pp. 319-328.

Aubert, O., Droxler, A., 1996. Seismic stratigraphy and depositional signatures of the Maldive carbonate system (Indian Ocean). Mar. Pet. Geol. 13, 503–536.

Aubert, O., Droxler, A.W., 1992. General Cenozoic evolution of the Maldives carbonate system (equatorial Indian Ocean). Bull. Des Centres Rech. Explor. Elf Aquitaine 16, 113–136.

Azzopardi, E., Sayer, M., 2012. Estimation of depth and temperature in 47 models of diving decompression computer. Underwater Technology, 31, 3–12.

Bamber, J.L and Aspinall, W.P, 2013. An expert judgement assessment of future sea- level rise from the ice sheets. Nature Climate Change 3, 424–427.

Bard, E, et al., 1996. Deglacial sea-level record from Tahiti corals and the timing of global meltwater discharge. Nature 382:241–244.

Bard, E, Hamelin, B, Delanghe-Sabatier, D. 2010. Deglacial meltwater pulse 1B and Younger Dryas sea-levels revisited with boreholes at Tahiti. Science 327(5970): 1235–1237.

Bard, E., Hamelin, B., Fairbanks, R.G., Zindler, A., 1990a. Calibration of the 14C timescale over the past 30,000 years using mass spectrometric U-Th ages from Barbados corals. Nature 345, 405– 410.

Barnes V.E. & Bell W.C., 1977. The Moore Hollow Group of central Texas, Univ. Texas, Austin, Bur. Econ. Geol., Report of Investigations 88; Parts 1-3.

Barrett, W., 1962. Emerged and submerged shorelines of the Dominican Republic. Rev. Geogr. 51–77.

Belopolsky, A. V, Droxler, A.W., 2004b. Seismic expressions of prograding carbonate bank margins: Middle Miocene, Maldives, Indian Ocean. American Association Petroleum Geologists Memoir, 81 (2004), pp. 267-290.

Belopolsky, A.V., Droxler, A.W., 1999. Uppermost Pleistocene transgressive coralgal reefs on the edge of the South Texas shelf: analogs for reefal reservoirs buried in siliciclastic shelf. In Advanced Reservoir Characterization for the 21st Century, ed. TF Hentz, pp. 41–50. Houston: Gulf Coast Sect. SEPM. 156

Betzler, C., Fürstenau, J., Lüdmann, T., Hübscher, C., Lindhorst, S., Paul, A., Reijmer, J.J.G., Droxler, A.W., 2013. Sea-level and ocean-current control on carbonate- platform growth, Maldives, Indian Ocean. Basin Res. 25, 172–196.

Bianchi, C.N., Ardizzone, G.D., Belluscio, A., Colantoni, P., Diviacco, G., Morri, C., Tunesi, L., 2004. Benthic cartography. Biol. Mar. Mediterr. 11, 347–370.

Bianchi, C.N., Colantoni, P., Geister, J., Morri, C., 1997. Reef geomorphology, sediments and ecological zonation at Felidu Atoll, Maldive islands (Indian Ocean), in: Proceedings of the 8th International Coral Reef Symposium. pp. 431–436.

Bianchi, C.N., Morri, C., Lasagna, R., Montefalcone, M., Gatti, G., Parravicini, V., Rovere, A., 2016. Resilience of the Marine Animal Forest, in: Rossi, S., Bramanti, L., Gori, A., del Valle, C. (Eds.), Marine Animal Forests. Springer International Publishing, Cham, pp. 1–30.

Bianchi, C.N., Morri, C., Pichon, M., Benzoni, F., Colantoni, P., Baldelli, G., Sandrini, M., 2006. Dynamics and pattern of coral recolonization following the 1998 bleaching event in the reefs of the Maldives, in: Proc 10th Int Coral Reef Symp. pp. 30–37.

Bird, M. I. et al., 2007. An inflection in the rate of early mid-Holocene eustatic sea- level rise: A new sea-level curve from Singapore. Estuarine Coastal Shelf Sci 71:523– 536.

Bird, M.I., et al., 2010. Punctuated eustatic sea-level rise in the early mid-Holocene. Geology 38:803–806.

Blanchon P. 2011. Geomorphic zonation. In: Hopley D, ed. Encyclopedia of modern coral reefs: structure, form and process. Dordrecht: Springer, 469–486.

Blanchon, P., and Shaw, J., 1995. Reef drowning during last deglaciation: Evidence for catastrophic sea-level rise and ice-sheet collapse. Geology, Vol23; no.1; p. 4-8.

Blanchon, P., Jones, B., 1995. Marine-planation terraces on the shelf around Grand Cayman: a result of stepped Holocene sea-level rise. J. Coast. Res. 1–33.

Blanchon, P., Jones, B., Ford, D.C., 2002. Discovery of a submerged relic reef and shoreline off Grand Cayman: Further support for an early Holocene jump in sea level. Sediment. Geol. 147, 253–270. 157

Boucot, A., 1990. Phanerozoic extinctions: how similar are they to each other? In: Kauffman, E.G., Walliser, O.H. (Eds.), Extinction Events in Earth History. Springer, Berlin, pp. 5–30.

Braithwaite, C.J.R., Taylor, J.D., Kennedy, W.J., 1973. The evolution of an atoll: the depositional and erosional history of Aldabra. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 307–340.

Bridge J., Barnes V.E. & Cloud P.E., 1947. Stratigraphy of the Upper Cambrian, llano uplift, Texas, Geol. Soc. Am. Bull. 58; 109-124.

Bright, T. J. and Rezak, R., 1976. A biological and geological reconnaissance of selected topographical features on the Texas continental shelf. A report to Bureau of Land Management.

Brown, B.E., Dunne, R.P., 1988. The environmental impact of coral mining on coral reefs in the Maldives. Environ. Conserv. 15, 159–165.

Burne, R.V., Moore, L.S., 1987. Microbialites; organosedimentary deposits of benthic microbial communities. PALAIOS ; 2 (3): 241–254.

Burns, B.P., Goh, F., Allen, M., and Neilan, 2004. Microbial diversity of extant stromatolites in the hypersaline marine environment of Shark Bay, Australia. Environmental Microbiology, 6 (10), 1096-1101.

Cabioch, G., Montaggioni, L., Frank, N., SEAWAY, Sallé, E., Payri, C., Pelletier, B., Paterne, M., 2008. Successive reef depositional events along the Marquesas foreslopes (French Polynesia) since 26 ka. Mar. Geol. 254, 18–34.

Camoin, G.F., Seard C., Deschampes, P., Webster, J.M., Abbey, E., Braga, J.C., Iryu, Y., Durand, N., Bard, E., Hamelian, B., Yokohama, Y., Thomas, L., Henderson, G.M., Dussouillez, P., 2012. Reef response to sea-level and environmental changes during the last deglaciation: Integrated Ocean Drilling Program Expedition 310, Tahiti Sea- level: Geology 40, no. 7, p. 643–646.

Carter, R. M., Carter, L., Johnson, D. P., 1986. Submergent shorelines in the SW pacific: evidence for an episodic post-glacial transgression: Sedimentology 33, p. 629-649.

Casalbore, D., Falese, F., Martorelli, E., Romagnoli, C., Chiocci, F.L., 2017. Submarine depositional terraces in the Tyrrhenian Sea as a proxy for paleo-sea level 158

reconstruction: Problems and perspective. Quat. Int. Volume 439, Part A, Pages 169- 180.

Chafetz, H.S. and Reid, A., 2000. Syndepositional shallow water precipitation of glauconite minerals. Sedimentary Geology, 136: 29–42.

Chagas, A. A. P., Webb, G.E., Burne, R. V., and Southam, G., 2016. Modern lacustrine microbialites: Towards a synthesis of aqueous and carbonate geochemistry and mineralogy. Earth-Science Reviews, Vol 162, 338-363.

Chappell, J, Polach, H., 1991. Post-glacial sea-level rise from a coral record at Huon Peninsula, Papua New Guinea. Nature 349:147–149.

Chen, J., 2014a. Surface and subsurface reworking by storms on a Cambrian carbonate platform: Evidence from limestone breccias and conglomerates: Geologos, v. 20, p. 13–23.

Chen, X., Zhang X., Church J. A., Watson, C.S., King M. A., Monselesan D., Legresy, B., Harig, C., 2017. The increasing rate of global mean sea-level rise during 1993-2014. Nature climate change, 7, 492-495.

Chevalier, J.P., Denizot, M., Mougin, J.-L., Plessis, Y., SALVAT, B., 1968. Etude géomorphologique et bionomique de l’atoll de Mururoa (Tuamotu). Cah. du pacifique 1–144.

Chirayath, V. and Earle, S., A., 2016. Drones that see through waves - preliminary results from airborne fluid lensing for centimetre-scale acquatic conservation. Aquatic Conservation: Marine and Freshwater Ecosystems, 26, (Suppl. 2): 237-250.

Ciarapica, G., Passeri, L., 1993. An overview of the Maldivian coral reefs in Felidu and North Malé Atoll (Indian Ocean): Platform drowning by ecological crises. Facies 28, 33–65.

Clark P.U., et al., 2001. Freshwater forcing of abrupt climate change during the last glaciation. Science 293(5528):283–287.

Cloud P.E. & Barnes V.E., 1948. The Ellenburger Group of Central Texas, Univ Texas, Austin Publ 4621, 473pp.

Colantoni, P., Baldelli, G., Bianchi, C.N., Capaccioni, B., Morri, C., Sandrini, M., Tassi, F., 2003. A cave flooded by marine water with hydrogen sulphide highlights the recent 159

evolution of the Maldives (Indian Ocean): preliminary notes. Le grotte d’Italia 4, 29– 37.

Collins, L.B., 1988. Sediments and history of the Rottnest Shelf, southwest Australia: a swell-dominated, non-tropical carbonate margin. Sediment. Geol. 60, 15–49.

Correa, T.B.S., Eberli, G.P., Grasmueck, M., Reed, J.K., Verwer, K., Purkis, S.J., 2012. Variability of cold-water coralmounds in a high sediment input and tidal current regime, Straits of Florida. Sedimentology 59, 1278–1304.

Cowan, C.A., and James, N.P., 1993, The interactions of sea-level change, terrigenous- sediment influx, and carbonate productivity as controls on Upper Cambrian Grand Cycles of western Newfoundland, Canada, Geological Society of America Bulletin, v., 109, p., 16-42.

Cullen, D.J., 1967. Submarine evidence from New Zealand of a rapid rise in sea level about 11,000 years BP. Palaeogeogr. Palaeoclimatol. Palaeoecol. 3, 289–298.

Darwin, C., 1842. The Structure and Distribution of Coral Reefs. Being the first part of the geology of the voyage of the Beagle, under the command of Capt. Fitzroy, R.N. during the years 1832 to 1836, London: Smith Elder and Co.

Davies, P.S., Stoddart, D.R., Sigee, D.C., 1971. Reef forms of Addu Atoll, Maldives Islands. Symp zol Soc L. 28, 217–259.

De Deckker, P, Yokoyama, Y., 2009. Micropalaeontological evidence for Late Quaternary sea-level changes in Bonaparte Gulf, Australia. Global Planet Change 66:85–92.

DeConto, R.M.m Pollard, D., 2016. Contribution of Antarctica to past and future sea- level rise Nature, 531, 591- 597.

Deschamps, P., Durand, N., Bard, E., Hamelin, B., Camoin, G., Thomas, A.L, Henderson, G.M., Okuno, J., Yokohama Y., 2012. Ice-sheet collapse and sea-level rise at the Bølling warming 14,600 years ago. Nature, 483, 559–564.

Digerfeldt, G., Hendry, M.D., 1987. An 8000 year Holocene sea-level record from Jamaica: implications for interpretation of Caribbean reef and coastal history. Coral Reefs 5, 165–169.

Dill, R.F., Shinn, E.A., Jones, A.T., Kelly, K. and Steinen, R.P., 1986. Giant subtidal stromatolites forming in normal salinity waters, Nature, no. 324, p. 55–58. 160

Dravis, J.J., 1983. Hardened Subtidal Stromatolites, Bahamas. Science, Vol 219, Issue 4583, pp. 385-386.

Droxler, A. W., Hopson, H. H., Khanna, P., Proctor, J. M., Lehrmann, D. J., Harris, P.M., 2016.Transgressive Lag of Flat Rip-up Clasts – Substratum for Initial Growth of Upper Cambrian Large Microbial Bioherms. Session – Geobiology of Carbonate Systems AAPG – ACE meeting, Calgary, Canada.

Droxler, A.W, Jorry S.J., 2013. Deglacial Origin of Barrier reefs Along Low-Latitude Mixed Siliclastic and Carbonate Continental Shelf Edges, Annu. Rev. Mar. Sci. 5:23.1- 23.36.

Duce, S., Vila-Concejo, A., Hamylton, S.M., Webster, J.M., Bruce, E., Beaman, R.J., 2016. A morphometric assessment and classification of coral reef spur and groove morphology. Geomorphology 265, 68-83.

Dullo, W.C., Camoin, G.F., Blomeier, D., Colonna, M., Eisenhauer, A., Faure, G., Casanova, J., Thomassin, B.A., 1998. Morphology and Sediments of the Fore‐Slopes of Mayotte, Comoro Islands: Direct Observations from a Submersible. Reefs carbonate platforms Pacific Indian Ocean. 217–236.

Düsterhus, A., Rovere, A., Carlson, A.E., Horton, B.P., Klemann, V., Tarasov, L., Barlow, N.L.M., Bradwell, T., Clark, J., Dutton, A., Gehrels, W.R., Hibbert, F.D., Hijma, M.P., Khan, N., Kopp, R.E., Sivan, D., Törnqvist, T.E., 2016. Palaeo-sea-level and palaeo-ice- sheet databases: problems, strategies, and perspectives. Clim. Past 12, 911–921.

Emery, K.O., 1958. Shallow submerged terraces off southern California. Geol. Soc. Amer. Bull. 69, 39–60.

Engelhart, S.E., Horton, B.P., 2012. Holocene sea level database for the Atlantic coast of the United States. Quat. Sci. Rev. 54, 12–25.

Fairbanks, R.G., 1989. A 17,000-year glacioeustatic sea-level record: Influence of glacial melting rates on Younger Dryas event and deep-ocean circulation. Nature 342, 637-642.

Fairbridge, R.W., Stewart Harris B, J., 1960. Alexa Bank, a drowned atoll on the Melanesian Border Plateau. Deep Sea Res. 7, 100–116.

Feldmann, M., and McKenzie, J.A., 1998. Stromatolite–thrombolite associations in a modern environment, Lee Stocking Island, Bahamas. Palaios, no. 13, p. 201–212. 161

Fletcher, C.H., Sherman, C.E., 1995. Submerged Shorelines on O’ahu, Hawai’i: Archive of Episodic Transgression During theDeglaciation? J. Coast. Res. 141–152.

Flower, B.P., Hastings, D.W., Hill, W.H., Quinn, T.M., 2004. Phasing of deglacial warming and Laurentide ice sheet meltwater in the Gulf of Mexico. Geology, v. 32; no. 7; p. 597–600.

Focke, J.W., 1978. Holocene development of coral fringing reefs, leeward off Curacao and Bonaire (Netherlands Antilles). Mar. Geol. 28, M31–M41.

Ford D. & Golonka J., 2003. Phanerozoic paleogeography, paleoenvironment and lithofacies maps of the circum-Atlantic margins. In: Golonka J. (ed.), Thematic set on paleogeographic reconstruction and hydrocarbon basins: Atlantic, Caribbean, South America, Middle East, Russian Far East, Arctic, Marine and Petroleum Geology, 20, 249–285.

Fricke, H.W., Landmann, G., 1983. On The Origin Of Red-Sea Submarine Canyons - Observations By Submersible. Naturwissenschaften 70, 195–197.

Fürstenau, J., Lindhorst, S., Betzler, C., Hübscher, C., 2009. Submerged reef terraces of the Maldives (Indian Ocean). Geo-Marine Lett. 30, 511–515.

Gardiner John Stanley, 1902. The fauna and geography of the Maldive and Laccadive archipelagoes : being the account of the work carried on and of the collections made by an expedition during the years 1899 and 1900. Cambridge University Press. Vol II. Part I.

Garrison, J. R., Jr., Long, L. E., and Richmann, D. L., 1979. Rb-Sr and K-Ar geochronologic and isotopic studies, Llano uplift, central Texas: Contributions to Mineralogy and Petrology, v. 69, p. 361-374.

Ginsburg, R.N., 1982. Actualistic depositional models for the great American Bank (Cambro-Ordivician). International Association o Sedimentologists Bulletin, 55, 340.

Ginsburg, RN, Planavsky, NJ, 2008. Diversity of Bahamian stromatolite substrates. in, Links Between Geological Processes, Microbial Activities & Evolution of Life. pg. 177-195. eds., Dilek Y, Furnes H, Muehlenbachs K. Springer academic press. Amsterdam. 162

Gischler, E., 2010. Indo-Pacific and Atlantic spurs and grooves revisited: the possible effects of different Holocene sea-level history, exposure, and reef accretion rate in the shallow fore reef. Facies 56: 173.

Gischler, E., Hudson, J.H., Pisera, A., 2008. Late Quaternary reef growth and sea level in the Maldives (Indian Ocean). Mar. Geol. 250, 104–113.

Gischler, E., Storz, D., Schmitt, D., 2014. Sizes, shapes, and patterns of coral reefs in the Maldives, Indian Ocean: The influence of wind, storms, and precipitation on a major tropical carbonate platform. Carbonates and Evaporites 29, 73–87.

Gladfelter, E. H., Monahan, R.K., Gladfelter, W.B., 1978. Growth Rates of Five Reef- Building Corals in the Northeastern Caribbean. Bulletin of Marine Science, Volume 28, number 4.

Glynn, P.W., Veron, J.E.N., Wellington, G.M., 1996. Clipperton Atoll (eastern Pacific): oceanography, geomorphology, reef-building coral ecology and biogeography. Coral Reefs 15, 71–99.

Golonka J., 2000. Cambrian-Neogene Plate Tectonic Maps. Wydawnictwa Uniwersytetu Jagielloñskiego, Kraków.

Golonka J., 2002. Plate-tectonic maps of the Phanerozoic. In: Kiessling W., Flügel E. & Golonka J. (eds), Phanerozoic reef patterns, SEPM (Society for Sedimentary Geology) Special Publication, 72, 21–75.

Golonka J., 2007.Phanerozoic Paleoenvironment and Paleolithofacies maps. Early Paleozoic. Geologia, Tom 35, Zeszyt 4, 589-654.

Golonka J., Bocharova N.Y., Ford D., Edrich M.E., Bednarczyk J. & Wildharber J., 2003. Paleogeographic reconstructions and basins development of the Arctic. In: Golonka J. (ed.), Thematic set on paleogeographic reconstruction and hydrocarbon basins: Atlantic, Caribbean, South America, Middle East, Russian Far East, Arctic, Marine and Petroleum Geology, 20, 211–248.

Golonka J., Gahagan L., Krobicki M., Marko F., Oszczypko N. & Œl1czka A., 2006a. Plate Tectonic Evolution and Paleogeography of the Circum-Carpathian Region. In: Golon- ka J. & Picha F. (eds), The Carpathians and their foreland: Geology and hydrocarbon resources, American Association of Petroleum Geologists Memoir, 84, 11–46. 163

Golonka J., Krobicki M., Paj1k J., Nguyen Van Giang & Zuchiewicz W., 2006b. Global plate tectonics and paleogeography of Southeast Asia. Faculty of Geology, Geophysics and Environmental Protection, AGH University of Science and Technology, Arkadia, Krakow.

Golonka J., Ross M.I and Scotese C.R., 1995. Phanerozoic paleographic and paleoclimate modelling maps, in Embry A.F., Beauchamp B. & Glass D.J. (eds.), Pangea: Global Environment and Resources, CSPG Memoir 17; 1-48.

Goreau, T.F., 1974. Fore Reef Morphology and Depositional Processes North Jamaica. Soc. Econ. Paleontol. Mineral. 77–89.

Grant, K.M., Rohling, E.J., Ramsey, C.B., Cheng, H., Edwards, R.L., Florindo, F., Heslop, D., Marra, F., Roberts, A.P., Tamisiea, M.E., others, 2014. Sea-level variability over five glacial cycles. Nat. Commun. 5.

Green, A.N., Cooper, J.A.G., Salzmann, L., 2014. Geomorphic and stratigraphic signals of postglacial meltwater pulses on continental shelves. Geology 42, 151–154.

Gvirtzman, G., 1994. Fluctuations of sea level during the past 400 000 years: the record of Sinai, Egypt (northern Red Sea). Coral Reefs 13, 203–214.

Halley, J.M., Hartley, S., Kallimanis, A.S., Kunin,W.E., Lennon, J.J., Sgardelis, S.P., 2004. Uses and abuses of fractal methodology in ecology. Ecology Letters 7, 254–271.

Hanebuth, T., Stattegger, K., Grootes, P., M., 2000. Rapid flooding of the Sunda Shelf: A late-glacial sea-level record. Science 288(5468):1033–1035.

Harper, R.A., 2011. Geology of the Mcmillan Ranch in Mason, Texas; An assessment of the nature of normal faults in the Mason area. M.S. Thesis, Texas A. & M. University.

Hay, C., Mitrovica, J.X., Gomez, N., Creveling, J.R., Austermann, J., Kopp, R.R., 2014. The sea-level fingerprints of ice-sheet collapse during interglacial periods. Quat. Sci. Rev. 87, 60–69.

Hengesh, J. V, Whitney, B.B., Rovere, A., 2011. A Tectonic Influence on Seafloor Stability along Australia’s North West Shelf, in: 21st International Offshore and Polar Engineering Conference. Maui, HI, pp. 1–9. 164

Hijma, M.P., Engelhart, S.E., Törnqvist, T.E., Horton, B.P., Hu, P., Hill, D.F., 2015. A protocol for a geological sea-level database, in: Shennan, I., Long, A.J., Horton, B.P. (Eds.), Handbook of Sea-Level Research. Wiley Online Library, pp. 536–553.

Hine, A.C., Neumann, A.C., 1977. Shallow carbonate-bank-margin growth and structure, Little Bahama Bank, Bahamas. Am. Assoc. Pet. Geol. Bull. 61, 376–406.

Hoffman P.E., 1991. Did the breakout of Laurentia turned Gondwanaland inside-out? Science, 252, 1409–1412.

Houbolt, J.J., 1957. Surface sediments of the Persian Gulf near the Qatar Peninsula, Dissertation, Univ. Utrecht, 113 pp. The Hague: Mouton and Co.

Hubbard, D.K., Dullo, W., 2016. Coral Reefs at the Crossroads – An Introduction. Coral Reefs at Crossroads, Coral Reefs of the World, 6.

James, N.P., Ginsburg, R.N., 1979. The geological setting of Belize reefs. Seaward Margin Belize Barrier Atoll Reefs Morphol. Sedimentol. Org. Distrib. Late Quat. Hist. 1–14.

James, N.P.,Wood, R., 2010. Reefs. In: James, N.P., Dalrymple, R.W. (Eds.), FaciesModels 4. Geological Association of Canada, St. John's, pp. 421–447.

Jeffery, D.L. and Stanton, R.J., 1996. Growth history of lower missipian waulsortian mounds: Distribution, stratal patters, and geometries, New Mexico. Facies, 35, 29- 58, PI. 6-19.

Johns, R.A., Dattilo, B.F., Spincer, B., 2007. Neotype and redescription of the Upper Cambrian anthaspidellid sponge, Wilbernicyathus donegani Wilson, 1950. J. Paleontol. 81, 435–444.

Johnson, D.P., Searle, D.E., 1984. Post‐glacial seismic stratigraphy, central Great Barrier Reef, Australia. Sedimentology 31, 335–352.

Kaye, C.A., 1959. Coastal Geology of Puerto Rico. US Government Printing Office.

Kaye, C.A., 1959. Geology of Isla Mona and notes on the age of Mona Passage., U. S. Geological Survey Professional Paper 317-CSP - 141.

Keating, B., 1985. Submersible observations on the flanks of Johnston island (Central Pacific Ocean), in: Proceedings of the Fifth International Coral Reef Symposium. pp. 413–418. 165

Kelleher, C., Lhemann, A., Swartz, H., Yazbek, L. Dl., Mercado, L., Lehrmann, D., Hopson, H., Khanna, P., Droxler, A., 2016. Sedimentology and depositional environments of cyclic microbial bearing strata of the Cambrian Point Peak Member of the Wilberns Fm., Exposed along Mill Creek the Llano River, and James River, Mason County, TX: South Central Section of the Geological Society of America Meeting, Baton Rougue, LA, Geological Society of America Abstracts with Programs. Vol. 48, No. 1.

Kench, P.S., Brander, R.W., Parnell, K.E., McLean, R.F., 2006. Wave energy gradients across a Maldivian atoll: Implications for island geomorphology. Geomorphology 81, 1–17.

Kench, P.S., Smithers, S.G., McLean, R.F., Nichol, S.L., 2009. Holocene reef growth in the Maldives: evidence of a mid-Holocene sea-level highstand in the central Indian Ocean. Geology 37, 455–458.

Khan, N.S., Ashe, E., Shaw, T.A., Vacchi, M., Walker, J., Peltier, W.R., Kopp, R.E., Horton, B.P., 2015. Holocene Relative Sea-Level Changes from Near-, Intermediate-, and Far- Field Locations. Curr. Clim. Chang. Reports 1, 247–262.

Khanna, P, Droxler, A.W., Nittrouer, J.A., Tunnell, J.W., Shirley, T.C., 2017. Coralgal Reefs morphology records punctuated sea-level rise during last deglaciation. Nature Communications, 8: 1047.

Kiessling, W., 2009. Geologic and biologic controls on the evolution of reefs. Annu. Rev. Ecol. Evol. Syst. 40, 173–192.

Klostermann, L., Gischler, E., Storz, D., Hudson, J.H., 2014. Sedimentary record of late Holocene event beds in a mid-ocean atoll lagoon, Maldives, Indian Ocean: Potential for deposition by tsunamis. Mar. Geol. 348, 37–43.

Komukai, R., 1956. On the bottom topography and sediments in the western passage of Tugaru Strait. Hydrographic office.

Kühlmann, D.H.H., 1970. Die Korallenriffe Kubas. I. Genese und Evolution. Int. Rev. der gesamten Hydrobiol. und Hydrogr. 55, 729–756.

Lambeck, K., Anzidei, M., Antonioli, F., Benini, A., Esposito, 2004. A. Sea-level in Roman time in the Central Mediterranean and implications for recent change. Earth Planet Sci Lett 224:563–575. 166

Lambeck, K., Rouby, H., Purcell, A., Sun, Y. Sambridge, M., 2014. Sea-level and global ice volumes from the Last Glacial Maximum to the Holocene. Proc. Natl. Acad. Sci. U.S.A. 111, 15296–15303.

Lasagna, R., Albertelli, G., Colantoni, P., Morri, C., Bianchi, C.N., 2010a. Ecological stages of Maldivian reefs after the coral mass mortality of 1998. Facies 56, 1–11.

Lasagna, R., Albertelli, G., Giovannetti, E., Grondona, M., Milani, A., Morri, C., Bianchi, C.N., 2008. Status of Maldivian reefs eight years after the 1998 coral mass mortality. Chem. Ecol. 24, 67–72.

Lasagna, R., Albertelli, G., Morri, C., Bianchi, C.N., 2010b. Acropora abundance and size in the Maldives six years after the 1998 mass mortality: patterns across reef typologies and depths. J. Mar. Biol. Assoc. United Kingdom 90, 919–922.

Lasagna, R., Gnone, G., Taruffi, M., Morri, C., Bianchi, C.N., Parravicini, V., Lavorano, S., 2014. A new synthetic index to evaluate reef coral condition. Ecol. Indic. 40, 1–9.

Lee, J., Chen, J., Chough, S. K., 2015. The middle-late Cambrian reef transition and related geological events: A review and new view. Earth Science Reviews 145, pp. 66-84.

Liddell, W.D., Ohlhorst, S.L., Boss, S.K., 1984. Community patterns on the Jamaican fore reef (15-56m). Paleont Am. 54, 385.

Lidz, B.H. and Shinn, E.A., 1991. Paleoshorelines, Reefs, and a Rising Sea: South Florida, U.S.A. Journal of Coastal Research Vol. 7, No.1,pp.203-229.

Lindquist, P., 1978. Geology of the South Texas Shelf Banks. M.S. thesis, Texas A&M University, College Station, Texas.

Liu, J., Milne, G. a., Kopp, R.E., Clark, P.U., Shennan, I., 2015. Sea-level constraints on the amplitude and source distribution of Meltwater Pulse 1A. Nat. Geosci. 9, 6–12.

Lochman-Balk C., 1970a. Upper Cambrian faunal patterns on the craton, Geol. Soc. Am. Bull. 81;3197-3224

Locker, S.D., Hine, A.C., Tedesco, L.P., and Shinn, E.A., 1996. Magnitude and timing of episodic sea-level rise during the last deglaciation. Geology, v.24; no. 9; p. 827-830.

Logan, B., 1962. Submarine topography of the Yucatan Platform. New Orleans Geological Society. 167

Macintyre, I.G., 1967. Submerged coral reefs, west coast of Barbados, West Indies. Can. J. Earth Sci. 4, 461–474.

Mallinson, D., Hine, A., Naar, D., Locker, S., Donahue, B., 2014. New perspectives on the geology and origin of the Florida Middle Ground carbonate banks, West Florida Shelf, USA. Mar. Geol. 355, 54–70.

McCausland P.J.A., Voo R. van der & Hall C.M., 2007. Circum-Iapetus paleogeography of the Precambrian-Cambrian transition with a new paleomagnetic constraint from Lau- rentia. Precambrian Research, 156, 125–152.

Miller, J.F., Loch, J.D., Taylor, J.F., 2012. Biostratigraphy of Cambrian and Lower Ordovician Strata in the Llano Uplift, Central Texas in J. R. Derby, R. D. Fritz, S. A. Longacre, W. A. Morgan, and C. A. Sternbach, eds., The great American carbonate bank: The geology and economic resources of the Cambrian – Ordovician Sauk megasequence of Laurentia: AAPG Memoir 98, p. 187 – 202.

Milne, G. A., Mitrovica, J., 2008. Searching for eustasy in deglacial sea-level histories. Quat Sci Rev 27:2292–2302.

Moawad, M.B., 2013. Detection of the submerged topography along the Egyptian Red Sea Coast using bathymetry and GIS-based analysis. Egypt. J. Remote Sens. Sp. Sci. 16, 35–52.

Montaggioni, L.F., 2005. History of Indo-Pacific coral reef systems since the last glaciation: Development patterns and controlling factors. Earth Sci. Rev. 71, 1–75.

Montaggioni, L.F., Braithwaite, C.J.R., 2009. Quaternary coral reef systems: history, development processes and controlling factors. Elsevier Science.

Morelock, J., Grove, K., Hernandez, M.L., 1983. Oceanography and patterns of shelf sediments, Mayaguez, Puerto Rico. J. Sediment. Res. 53, 371–381.

Morgan, William A., 2012. Sequence stratigraphy of the great American carbonate bank, in J. R. Derby, R. D. Fritz, S. A. Longacre, W. A. Morgan, and C. A. Sternbach, eds., The great American carbonate bank: The geology and economic resources of the Cambrian—Ordovician Sauk megasequence of Laurentia: AAPG Memoir 98, p. 37– 79. 168

Morri, C., Aliani, S., Bianchi, C.N., 2010. Reef status in the Rasfari region (North Malé Atoll, Maldives) five years before the mass mortality event of 1998. Estuar. Coast. Shelf Sci. 86, 258–264.

Morri, C., Bianchi, C.N., Aliani, S., 1995. Coral reefs at gangehi (north ari atoll, maldive islands). Publ. du Serv. Géologique du Luxemb. 29, 3–12.

Morri, C., Montefalcone, M., Lasagna, R., Gatti, G., Rovere, A., Parravicini, V., Baldelli, G., Colantoni, P., Bianchi, C.N., 2015. Through bleaching and tsunami: Coral reef recovery in the Maldives. Mar. Pollut. Bull. 98, 188–200.

Nair, R.R., 1974. Holocene sea-levels on the western continental shelf of India. Proc. Indian Acad. Sci. B 79, 197–203.

Naseer, A., Hatcher, B.G., 2004. Inventory of the Maldives coral reefs using morphometrics generated from Landsat ETM+ imagery. Coral Reefs 23, 161–168.

Nash, H, Furiness, S.J., Tunnell, Jr. W.T., 2013. What is known about species richness and distribution on the outer-shelf south Texas Banks? Gulf and Caribbean Research Vol 25, 9.18.

Neumann, A. C., Macintyre, I., 1985. Reef response of sea-level rise: keep-up, catch- up or give-up. In Proceedings of the fifth international coral reef congress 3, pp. 105–110.

Newell, N.D., 1956. Geological Reconnaissance of Raroia (Kon Tiki) Atoll, Tuamotu Archipelago: A Revision of Bulletin 31 and Bulletin 36 of the Atoll Research Bulletin, Pacific Science Board, Washington, DC, November 30, 1954. Am. Museum of Natural History.

North Greenland Ice Core Project members, North Greenland Ice Core Project Oxygen Isotope Data: IGBP PAGES/World Data Center for Paleoclimatology Data Contribution Series # 2004-059. NOAA/NGDC Paleoclimatology Program, Boulder CO, USA (2004).

Paigne S., 1911. Mineral resources of the Llano-Burnet region, Texas with an account of the Pre-Cambrian Geology, Bull. USGS 450; 103pp.

Palmer, A. R., 1981b. Subdivision of the Sauk sequence, in M. E. Taylor, ed., Short papers for the 2nd International Symposium on the Cambrian System: U.S. Geological Survey Open File Report 81-743, p. 160–162. 169

Parke, R.P., 1953. Geology of the Southwest Mason-Llano River Area, Texas, M.S. Thesis, Texas Tech University.

Peltier W. R., Fairbanks R. G., 2006. Global glacial ice volume and Last Glacial Maximum duration from an extended Barbados sea-level record. Quat Sci Rev 25:3322–3337.

Pisarevsky S.A., Murphy J.B., Cawood P.A. & Collins A.S., 2008. Late Neoproterozoic and early Cambrian palaeogeography: Models and problems. Geological Society Spe- cial Publication, 294, 9–31.

Poag, C.W., 1973. Late Quaternary sea levels in the Gulf of Mexico. Thesis, Texas A & M University.

Polk, T.P., 1952. Geology of the West Mason Area. Texas, M.S. Thesis, Texas A. & M. University.

Portnoy, B.S., 1987. A study of Two isolated Cambrian stromatolitic outcrops, Mason and Mcculloch Counties, Texas. Texas, M.S. Thesis, Texas A. & M. University.

Precht, W., Deslarzes, K.J.P., Hickerson, E.L., Schmahl, G.P., Nuttall, M.F., Aronson, R.B., 2014. Back to the future: The history of acroporid corals at the Flower Garden Banks, Gulf of Mexico, USA. Marine Geology 349, 152-161.

Pretorius, L., Green, A., Cooper, A., 2016. Submerged shoreline preservation and ravinement during rapid postglacial sea-level rise and subsequent “slowstand.” Geol. Soc. Am. Bull. B31381.1.

Purdy, E.G., Bertram, G.T., 1993. Carbonate concepts from the Maldives, Indian Ocean. AAPG Special Volumes. American Association of Petroleum Geologists, Volume 34.

Purdy, E.G., Gischler, E., 2003. The Belize margin revisited: 1. Holocene marine facies: International Journal of Earth Sciences, v. 92, p. 532–551.

Purkis, S., Casini, G., Hunt, D., Colpaert, A., 2015. Morphometric patterns in modern carbonate platforms can be applied to the ancient rock record: similarities between modern Alacranes Reef and Upper Palaeozoic platforms of the Barents Sea. Sediment. Geol. 321, 49e69.

Quartau, R., Hipólito, A., Romagnoli, C., Casalbore, D., Madeira, J., Tempera, F., Roque, C., Chiocci, F.L., 2014. The morphology of insular shelves as a key for understanding 170

the geological evolution of volcanic islands: Insights from Terceira Island (Azores). Geochemistry, Geophys. Geosystems 15, 1801–1826.

Read, J. F., and J. E. Repetski, 2012. Cambrian – Lower Middle Ordovician passive carbonate margin, southern Appala- chians, in J. R. Derby, R. D. Fritz, S. A. Longacre, W. A. Morgan, and C. A. Sternbach, eds., The great American carbonate bank: The geology and economic resources of the Cambrian – Ordovician Sauk megasequence of Laur- entia: AAPG Memoir 98, p. 357–382.

Reches, Z., Erez, J., Garfunkel, Z., 1987. Sedimentary and tectonic features in the northwestern Gulf of Elat, Israel. Tectonophysics 141, 169–180.

Reid, R.P., Macintyre, I.G., Steneck, R.S., 1999. A microbialite/algal ridge fringing reef complex, High- borne Cay, Bahamas. Atoll Res Bulletin, no. 466, p. 1–18.

Reid, R P; Visscher, P T; Decho, A W; Stolz, J F., Bebout, B.M., Dupraz, C., Macintyre, I.G., Paerl, H.W., Pinckney, J.L., Prufert-Bebouti, L., Steppe, T.F., and DesMarals, D.J.,2000. The role of microbes in accretion, lamination and early lithification of modern marine stromatolites. Nature; London, 406.6799, 989-92.

Reid, R.P., James, N.P., Macintyre, I.G., Dupraz, C.P., and Burne, R.V., 2003. Shark Bay stromatolites: Microfabrics and reinterpretation of origins. Facies, 49:299.

Reiss, Z., Hottinger, L., 1984. The Sea Bottom—a Mosaic of Substrates. Gulf Aqaba 139–202.

Rezak R, Bright T. J., 1976. Coral reefs and banks of Texas outer continental shelf. AAPG Bull. 60:713–14.

Rezak, R., Bright, T. J., and McGrail, D. W., 1983. Reefs and Banks of the Northwestern Gulf of Mexico: Their geological, biological, and physical dynamics. A report submitted to U.S Department of Interior, Minerals Management Service, Outer Continental Shelf office, New Orleans.

Riding, R., 2011. Microbialites, stromatolites, and thrombolites. In J. Reitner and V. Thiel (eds), Encyclopedia of Geobiology. Encyclopedia of Earth Science Series, Springer, Heidelberg, pp. 635-654.

Riding, R., 2006b. Microbial carbonate abundance compared with fluctuations in metazoan diversity over geological time. Sediment. Geol. 185, 229–238. 171

Rigby, J.K., Roberts, H.H., 1976. Grand Cayman Island: Geology, Sediments and Marine Communities. Brigham Young University Geological Studies, Special Publication, 4 (1976), pp. 1-122

Risk, M.J., Dunn, J.J., Allison, W.R., Horrill, C., 1994. Reef monitoring in Maldives and Zanzibar: low-tech and high-tech science. Proc. Colloq. Glob. Asp. coral reefs, Miami, 1993 9, 66–72.

Risk, M.J., Sluka, R., 2000. The Maldives: A nation of atolls, in: Coral Reefs of the Indian Ocean: Their Ecology and Conservation. pp. 325–351.

Romagnoli, C., Casalbore, D., Bortoluzzi, G., Bosman, A., Chiocci, F.L., D’Oriano, F., Gamberi, F., Ligi, M., Marani, M., 2013. Bathy-morphological setting of the Aeolian Islands. Geol. Soc. London, Mem. 37, 27–36.

Rovere, A., Parravicini, V., Vacchi, M., Montefalcone, M., Morri, C., Bianchi, C.N., Firpo, M., 2010. Geo-environmental cartography of the Marine Protected Area “Isola di Bergeggi” (Liguria, NW Mediterranean Sea). J. Maps 6, 505–519.

Rovere, A., Raymo, M.E., Vacchi, M., Lorscheid, T., Stocchi, P., Gómez-Pujol, L., Harris, D.L., Casella, E., O’Leary, M.J., Hearty, P.J., 2016a. The analysis of Last Interglacial (MIS 5e) relative sea-level indicators: Reconstructing sea-level in a warmer world. Earth-Science Rev. 159, 404–427.

Rovere, A., Stocchi, P., Vacchi, M., 2016b. Eustatic and Relative Sea Level Changes. Curr. Clim. Chang. Reports 2, 221–231.

Rovere, A., Vacchi, M., Firpo, M., Carobene, L., 2011. Underwater geomorphology of the rocky coastal tracts between Finale Ligure and Vado Ligure (western Liguria, NW Mediterranean Sea). Quat. Int. 232, 187–200.

Rufin-Soler, C., Mörner, N.-A., Laborel, J., Collina-Girard, J., 2013. Submarine Morphology in the Maldives and Holocene Sea-Level Rise. J. Coast. Res. 30, 30–40.

Saponari, L., Montano, S., Seveso, D., Galli, P., 2015. The occurrence of an Acanthaster planci outbreak in Ari Atoll, Maldives. Mar. Biodivers. 45, 599–600.

Sarano, F., Pichon, M., 1988. Morphology and ecology of the deep fore reef slope at Osprey Reef, (Coral Sea). Proc. 6th Int. Coral Reef Symp. 2, 607–611. 172

Scheer, G., 1972. Investigation of coral reefs in the Maldive Islands with notes on lagoon patch reefs and the method of coral sociology, in: Proceedings of the First International Coral Reef Symposium. pp. 87–120.

Scheer, G., 1974. Investigation of coral reefs at Rasdu Atoll in the Maldives with the quadrat method according to phytosociology, in: Proceedings of the 2nd International Coral Reef Symposium, Brisbane. pp. 655–670.

Schlager, W., 2005. Carbonate sedimentology and sequence stratigraphy. SEPM concepts in sedimentology and paleontology. SEPM Society for Sedimentary Geology Volume 8.

Schubert, J.K., Bottjer, D.J., 1992. Early Triassic stromatolites as post-mass extinction disaster forms. Geology 20, 883.

Schulz K.J., Stewart D.B., Tucker R.D., Pollock J.C. & Ayuso R.A., 2008. The Ellsworth terrane, coastal Maine: Geochronology, geochemistry, and Nd-Pb isotopic composition – Implications for the rifting of Ganderia. Geological Society of America Bulletin, 120, 1134–1158.

Shennan, I., Long, A.J., Horton, B.P., 2015. Handbook of sea-level research, in Handbook of Sea-Level Research (eds I. Shennan, A. J. Long and B. P. Horton), John Wiley & Sons, Ltd, Chichester, UK.

Siddall, M., Chappell, J., Potter, E.K., 2007. Eustatic sea level during past interglacials. Dev. Quat. Sci. 7, 75–92.

Siddiquie, H.N., 1975. Submerged terraces in the Laccadive Islands, India. Mar. Geol. 18, M95–M101.

Sliger, K.L., 1955. Geology of the Lower James River Area, M.S. Thesis, Texas A. & M. College

Sloss,L.L., Krumbein, W.C., and Dapples, E.C., 1949. Integrated facies analysis. In: C. R. Longwell, Editor, Sedimentary Facies in Geologic His- tory, Memoir vol. 39, Geological Society of America, pp. 91–124.

Spincer, B.R., 1997. The palaeoecology of some Upper Cambrian reefs from central Texas, The great basin and Colorado, USA, Unpublished Thesis-University of Cambridge. 173

Stanley, D.J., Swift, D.J.P., 1968. Bermuda’s reef-front platform: bathymetry and significance. Mar. Geol. 6, 479–500.

Steffensen, J. P., Andersen, K. K., Bigler, M., Clausen, H. B., Dahl-Jensen, D., Fischer, H., Goto-Azuma, K., Hansson, M., Johnsen, S. J., Jouzel, J., Masson-Delmotte, V., Popp, T., Rasmussen, S. O., Rothlisberger, R., Ruth, U., Stauffer, B., Siggaard-Andersen, M.-L., Sveinbjörnsdóttir, A. E., Svensson, A., White, J. W. C., 2008. High-Resolution Greenland Ice Core Data Show Abrupt Climate Change Happens in Few Years. Science 321, pp 681-684.

Stoddart, D.R., 1969a. Geomorphology of the Marovo elevated barrier reef, New Georgia. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 255, 383–402.

Stoddart, D.R., 1969b. Ecology and morphology of recent coral reefs. Biol. Rev. 44, 433–498.

Tayama, R., 1950. Submarine topography and distribution of bottom sediment on eastern part of Sagami Bay: Bull, Hydrogr. Office, Japan.

Tcherepanov, E.N., Droxler, A.W., Lapointe, P., Dickens, G.R., Bentlesy, S.J., Beaufort, L., Peterson, L.C., Daniell, J., and Opdyke, B.N., 2008. Neogene evolution of the mixed carbonate-siliciclastic system in the Gulf of Papua, Papua New Guinea: Journal of Geophysical Research, v. 113, F01S21.

Törnqvist, T. E. and Hijma, M.P., 2012. Links between early Holocene ice-sheet decay, sea-level rise and abrupt climate change. Nature Geoscience Vol 5, 601-606.

Tracey Jr, J.I., 1968. Reef features of the Carolline and Marshall Islands. Prof. Pap. US Geol. Surv. 600, 1–80.

Tracey, J.I., Ladd, H.S., Hoffmeister, J.E., 1948. Reefs of Bikini, Marshall Islands. Bull. Geol. Soc. Am. 59, 861–878.

Trettin H.P & Balkwill H.R., 1979. Contributions to the tectonic history of the Innuitian Province, Arctic Canada. Canadian Journal of Earth Sciences, 16, 748–769.

Trowbridge, A.C., 1930. Building of the Mississippi delta: Am. Assoc. Petroleum Geologists, V. 14, 867-901 p.

Vacchi, M., Rovere, A., Marriner, N., Morhange, C., Spada, G., Fontana, A., 2016. Multiproxy assessment of Holocene relative sea-level changes in the western 174

Mediterranean: variability in the sea-level histories and redefinition of the isostatic signal., in: EGU General Assembly Conference Abstracts. Vienna, AT.

Van Der Voo R., 1988. Paleozoic paleography of North America, Gondwana, and intervening displaced terranes: comparisons of paleomagnetism with paleoclimatology and biogeographic patterns, Geol. Soc. Am. Bull. 100;311-324.

Van Der Voo R., French R.B., Williams D.W., 1976. Paleomagnetism of the Wilberns Formation (Texas) and the Late Cambrian Paleomagnetic field for North America, J. Geophys. Res. 81;5633-5638.

Vora, K.H., Almeida, F., 1990. Submerged Reef Systems on the Central Western Continental-Shelf of India. Mar. Geol. 91, 255–262.

Wagle, B.G., Vora, K.H., Karisiddaiah, S.M., Veerayya, M., Almeida, F., 1994. Holocene submarine terraces on the western continental shelf of India; Implications for sea- level changes. Mar. Geol. 117, 207–225.

Wallace, C.C., Zahir, H., 2007. The “Xarifa” expedition and the atolls of the Maldives, 50 years on coral reefs. Coral Reefs, 26: 3.

Weight, R. W. R. Anderson, J. B., Fernandez, R., 2011. A. Rapid mud accumulation on the central Texas shelf linked to climate change and sea-level rise. J. Sediment. Res. 81:743–61.

Wenke, F., Caiwang, B., Junren, C., Xitao, Z., 1982. Preliminary study on submarine relief of the northern South China Sea. Acta Oceanol. Sin. 4, 462–471.

Wilber, R.J., Milliman, J.D., Halley, R.B., 1990. Accumulation of bank-top sediment on the western slope of Great Bahama Bank: rapid progradation of a carbonate megabank. Geology 18, 970–974.

Williams, D.G., 1994. Marine habitats of the Cocos (Keeling) Islands. Atoll Res. Bull. 406, 1–11.

Winker, C.D., 1979. Late Pleistocene fluvial-deltaic deposition: Texas coastal plain and shelf: M.A. thesis, The University of Texas at Austin, Austin, Texas, 187 p.

Wolfson, M.L., Naar, D.F., Howd, P.A., Locker, S.D., Donahue, B.T., Friedrichs, C.T., Trembanis, A.C., Richardson, M.D., Wever, T.F., 2007. Multibeam observations of mine burial near Clearwater, FL, including comparisons to predictions of wave- induced burial. Ocean. Eng. IEEE J. 32, 103–118. 175

Woo, J., Chough, S.K., 2010. Growth patterns of the Cambrian microbialite: phototropism and speciation of Epiphyton. Sediment. Geol. 229, 1–8.

Woodroffe, C.D., 1992. Morphology and evolution of reef islands in the Maldives, in: Proceedings of the 7th International Coral Reef Symposium. pp. 1217–1226.

Woodroffe, C.D., 2005. Late Quaternary sea-level highstands in the central and eastern Indian Ocean: A review. Glob. Planet. Change 49, 121–138.

Woodroffe, C.D., Stoddart, D.R., Harmon, R.S., Spencer, T., 1983. Coastal morphology and late quaternary history, Cayman Islands, West Indies. Quat. Res. 19, 64–84.

Woodroffe, C.D., Webster, J.M., 2014. Coral reefs and sea-level change. Mar. Geol. 352, 248–267.

Wright, D.J., Donahue, B.T., Naar, D.F., 2002. Seafloor mapping and GIS coordination at America’s remotest national marine sanctuary (American Samoa). Undersea with GIS 33–63.

Yokohama, Y., Lambeck, K., Johnston, P., Fifield, L.K., Fifield, L., De Deckker, P., 2000. Timing of the Last Glacial Maximum from observed sea-level minima. Nature 406(6797): 713–716.

Zecchin, M., Ceramicola, S., Gordini, E., Deponte, M., Critelli, S., 2011. Cliff overstep model and variability in the geometry of transgressive erosional surfaces in high- gradient shelves: The case of the Ionian Calabrian margin (southern Italy). Mar. Geol. 281, 43–58.

Zecchin, M., Ceramicola, S., Lodolo, E., Casalbore, D., Chiocci, F.L., 2015. Episodic, rapid sea-level rises on the central Mediterranean shelves after the Last Glacial Maximum: A review. Mar. Geol. 369, 212–223.

Zimmer, B., Precht, W., Hickerson, E., Sinclair, J., 2006. Discovery of Acropora palmata at the Flower Garden Banks National Marine Sanctuary, northwestern Gulf of Mexico. Coral Reefs 25, Issue 2, p 192. 176

Appendix A - South Texas banks

Figure A.1. - Baker Bank- A) Bathymetric Map: About 19 m of relief of the bank is exposed. Seismic lines A-B, C-D, E-F, G-H I-J and K-L are located. B) 3.5 kHz profiles: uninterpreted and interpreted seismic lines displaying locations of exposed and buried terraces. The bank is buried by TMB which is clearly observed through the seismic lines. 177

Figure A.2. South Baker Bank - Bathymetric Map: About 16 m of bank relief is exposed. Seismic line A-B is located. B) 3.5 kHz profiles: uninterpreted and interpreted seismic lines displaying locations of exposed terrace. 178

Figure A.3. Aransas Bank - A) Bathymetric Map: About 16 m of the bank relief is exposed. Seismic line A-B, C-D, and E-F are located. B) 3.5 kHz profiles: uninterpreted and interpreted seismic lines displaying locations of exposed and buried terraces. 179

Figure A.4. North Hospital Bank- A) Bathymetric Map: About 17 m of bank relief is exposed. Seismic lines A-B, C-D, E-F, G-H, I-J, and K-L are located. B) 3.5 kHz profiles: uninterpreted and interpreted seismic lines displaying locations of exposed and buried terraces.

180

Figure A.5. Hospital Bank -A) Bathymetric Map: About 27 m of bank relief is exposed. Seismic lines A-B, C-D, E-F, G-H, I-J, and K-L are located. B) 3.5 kHz profiles: uninterpreted and interpreted seismic lines displaying locations of exposed and buried terraces. 181

Figure A.6. Dream Bank - A) Bathymetric Map: About 22 m of bank relief is exposed. Seismic lines A-B, C-D, E-F, G-H, I-J, and K-L are located. B) 3.5 kHz profiles: uninterpreted and interpreted seismic lines displaying locations of exposed and buried terraces.

182

Figure A.7. Harte Bank Bathymetric Map: About 21 m of bank relief is exposed. 183

Figure A.8. Big and Small Adam Bank Bathymetric Map: About 10 m of bank relief is exposed. 184

Appendix B – 3D Point Clouds Upper Cambrian Microbial Buildups, Mason, Central Texas

Figure B.1. – 3D Point Cloud – Shepard Cliff – Jack Buildup 185

Figure B.2. – 3D Point Cloud – Shepard Cliff – Granddaughters 186

Figure B.3. – 3D Point Cloud – Shepard Cliff – Mark Buildup 187

Figure B.4. – 3D Point Cloud – Shepard Cliff – PK Buildup 188

Figure B.5. – 3D Point Cloud – Shepard Cliff – Wayne Buildup 189

Figure B.6. –3D Point Cloud – Shepard Cliff – James Lee Buildup 190

Figure B.7. – 3D Point Cloud – Shepard Cliff – Tony Buildup 191

Figure B.8. – 3D Point Cloud – Shepard Cliff – Tonya Buildup 192

Figure B.9. – 3D Point Cloud – Shepard Pavement – 408 photographs used to build the 3D model using Agisoft 1.0.

193

Figure B.10. –3D Point Cloud – Zesch Cliff – GeneBuildup 194

Figure B.11. – 3D Point Cloud – Zesch Cliff – Patsy Buildup 195

Figure B.12. – 3D Point Cloud – Zesch Cliff – Mitch Buildup

196

Figure B.13. – 3D Point Cloud – Zesch Pavement – 333 photographs used to build the 3D model using Agisoft 1.0. 197

Figure B.14. Goat’s Graveyard, Llano River, 3D Point Cloud – western section. 198

Figure B.15. Goat’s Graveyard, Llano River, 3D Point Cloud – eastern section.

199

Figure B.16. 3D Point Cloud – Upper Mill Creek– 35 photographs used to build the 3D model using Agisoft 1.0.

200

Figure B.17. 3D Point Cloud – Fallen Block Mill Creek – 18 photographs used to build the 3D model using Agisoft 1.0.

201

Figure B.18. 3D Point Cloud – Drox Rock Mill Creek – 40 photographs used to build the 3D model using Agisoft 1.0. 202

Figure B.19. 3D Point Cloud – Lauren Section Mill Creek (Upper Part only) – 33 photographs used to build the 3D model using Agisoft 1.0. 203

Figure B.20. 3D Point Cloud – James River Pavement – 560 photographs used to build the 3D model using Agisoft 1.0.

204

Figure B.21. – 3D Point Cloud of Faulted Buildup, James River built using 62 photographs in Agisoft 1.0.