Citation
Constraining the Last Interglacial Sea Level Signal in the Bahamas

Material Information

Title:
Constraining the Last Interglacial Sea Level Signal in the Bahamas
Creator:
Li, Jin
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (111 p.)

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Geology
Geological Sciences
Committee Chair:
DUTTON,ANDREA LYNN
Committee Co-Chair:
ADAMS,PETER N
Committee Members:
JAEGER,JOHN M
Graduation Date:
5/3/2014

Subjects

Subjects / Keywords:
Coral reefs ( jstor )
Corals ( jstor )
Fossils ( jstor )
Ice ( jstor )
Ice sheets ( jstor )
Modeling ( jstor )
Reefs ( jstor )
Rubble ( jstor )
Sea level ( jstor )
Topographical elevation ( jstor )
Geological Sciences -- Dissertations, Academic -- UF
bahamas -- last-interglacial -- laurentide-ice-sheet -- sea-level
Virginia Key ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Geology thesis, M.S.

Notes

Abstract:
Given the climatic similarity between the last interglacial (LIG) period (ca. 125 ka ago) and climate model projections for the near future, the reconstruction of eustatic sea level (ESL) in the LIG period is relevant for understanding future sea level change in in the context of global warming. Age (using uranium and thorium isotopic analysis) and elevation data for fossil corals from this time period in the circumCaribbean region are not consistent with relative sea level predictions from glaciohydroisostatic modeling, indicating that modifications need to be made to the existing ice model, specifically for the last and penultimate glacial maxima. To determine which of these ice sheet models needs adjusting, elevations of two coeval stratigraphic horizons within the fossil reef were surveyed on San Salvador and Great Inagua. The result demonstrates a minimal gradient in elevation between the sites and a rate of sea level rise that is significantly lower than the model predictions. This indicates that the penultimate glacial Laurentide ice sheet needs to be reduced and some adjustment to the last glacial maximum Laurentide ice sheet may also be required to bring the model into agreement with the observational data. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2014.
Local:
Adviser: DUTTON,ANDREA LYNN.
Local:
Co-adviser: ADAMS,PETER N.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2016-05-31
Statement of Responsibility:
by Jin Li.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
5/31/2016
Resource Identifier:
908645652 ( OCLC )
Classification:
LD1780 2014 ( lcc )

Downloads

This item has the following downloads:


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E8P9C7SK5_UXH8AP INGEST_TIME 2014-10-03T22:24:11Z PACKAGE UFE0046777_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

CONSTRAINING THE LAST INTERGLACIAL SEA LEVEL SIGNAL IN THE BAHAMAS By JIN LI A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2014

PAGE 2

2014 Jin Li

PAGE 3

This document is dedicated to Dr. Andrea Dutton

PAGE 4

4 ACKNOWLEDGMENTS I would first like to thank my advisor, Dr. Andrea Dutton, for her patient and careful guidance and instructions in the whole project, as well as the kind encouragements and considerations particularly for an International transfer student. I will never forget her gentle comforts and encouragements that brought back me the hope, courage and the determination to go on firmly when I struggled with the graduate study and felt lost about the future at the beginning of my second semester at UF. I learned a lot from her, including not only the method to perform scientific research, but also a positive attitude towards research an interest in all kinds of related research area, not limited in one certain area, as well as an acute thinking at an y time. I am grateful for the opportunity she provided me to study as a graduate student here at UF. As for this precious opportunity, an appreciative thank you goes to Dr. John M. Jaeger, who is also one of my committee members, for his kind recommendatio n two years ago, so that I had access to know my advisor Dr. Andrea Dutton. And I would also like to thank my committee members Dr. John M. Jaeger and Dr. Peter N. Adams for their advice and review of the thesis. I also appreciate for the financial support that was provided by the NSF grant awarded to Dr. Andrea Dutton. I am also very grateful for Dr. Ellen E. Martin and Dr. Raymond Russo for their careful instructions and extra patience with an International transfer student like me both in and out of clas s. I have learned a lot in their classes, and felt well adjusted soon after arriving here in August, 2012. A special gratitude goes to my officemate at UF Kiernan Folz Donahue. He is such an excellent geology majoring student with richness geological knowledge and a

PAGE 5

5 good friend that is willing to help me with the study problems. Thanks to him, I could catch up soon with the class by making up for some back ground knowledge in geology. In particular, in the project, he helped me work out the problems together when I was stuck by certain points and look for certain background information by contacting the related institutions (e.g. NOAA and National Geodetic S urvey) I feel really lucky to have such a great friend like him. I also wa nt to thank Dr. Shu Gao, my advisor in undergraduate years as well as Dr. Yaping Wang from Nanjing University for all their untiring and sincere instructions, warmly encouragements and kindly help since my sophomore year (2009) Thanks to them, I began scientific research early in my undergraduate years and went on working firmly since then. Last but not least I would love to thank my parents and all my friends, in particular, the m ost significant old friend in my class at Nanjing University who is the first one to show me the way to work on scientific research, always ke eps my faith when I feel uncertain and cheers me up when I feel blue since we knew each other in 2009. I just want to show my dearest appreciation for all the things he has done for me. I am profoundly grateful to have such a soul mate friend by my side I also want to thank all my c lassmates in the Department of Geological S ciences at UF. They are so friendly and nice to me and we stayed together just like a big family. I will always keep in mind of what a wonderful time I have spent here in the geologic department.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF OBJECTS ................................ ................................ ................................ ....... 11 LIST OF ABBREVIATIONS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ..... 14 2 BACKGROUND ................................ ................................ ................................ ....... 24 Reconstructing sea level with U Th dated fossil corals ................................ ........... 24 Corals as sea level position indicators ................................ ............................. 24 U ................................ ......... 25 Glacio hydro isostatic effects ................................ ................................ ................ 30 Geologic Setting and LIG deposits in the two Bahamian Islands ............................ 34 Tectonic setting ................................ ................................ ................................ 36 LIG sedimentary record of the Bahamas ................................ .......................... 37 Sea level reconstruction from fossil corals ................................ ....................... 38 3 METHODS ................................ ................................ ................................ .............. 45 How to use a level instrument ................................ ................................ ................. 45 Level measurement in the field ................................ ................................ ............... 46 LIG deposition features ................................ ................................ .................... 46 Real time sea level ................................ ................................ ........................... 47 Correlation of elevation data to MLLW ................................ ................................ .... 48 Horizontal position calculation ................................ ................................ ................ 51 4 RESULTS ................................ ................................ ................................ ................ 59 ................................ .................... 59 Data points at Cockburn Town, San Salvador Island ................................ .............. 61 Age elevation data of LIG corals ................................ ................................ ............. 62 Error analysis ................................ ................................ ................................ .......... 63 5 DISCUSSION ................................ ................................ ................................ .......... 80

PAGE 7

7 Growth rates of corals ................................ ................................ ............................. 81 Peak sea level Position ................................ ................................ ........................... 82 Cockburn Town Reef, San Salvador Island ................................ ...................... 83 ................................ ............................ 85 Peak Sea Level ................................ ................................ ................................ 86 Unconformity surface (Reef I) ................................ ................................ ................. 87 Sea level during Reef I growth ................................ ................................ ......... 88 Rate of sea level change ................................ ................................ ........................ 89 Uncertainty in the estimation of sea level ................................ ................................ 89 Comparison of the observation with the model prediction ................................ ....... 90 6 CONCLUSIONS ................................ ................................ ................................ .... 100 APPENDIX ................................ ................................ ................................ .................. 102 LIST OF REFERENCES ................................ ................................ ............................. 103 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 111

PAGE 8

8 LIST OF TABLES Table page 2 1 Elevation of top in situ corals measured in the Bahamas ................................ ...... 43 3 4 1 Surveyed results: elevations of LIG sea level indicators in Great Inagua ................ 65 4 2 Surveyed results: elevations of LIG sea level indicators in San Salvador. .............. 66 4 3 References for horizontal position ................................ ................................ ........... 66 4 4 Elevation age data of fossil coral s ................................ ................................ ........... 67 5 1 Highest in situ corals in San Salvador ................................ ................................ ..... 92 5 2 Comparison of highest in situ corals of the same species in San Salvador and Great Inagua ................................ ................................ ................................ ....... 92 5 3 Estimations for LIG sea level in the two islands. ................................ ..................... 92

PAGE 9

9 LIST OF FIGURES Figure page 1 1 A cartoon representation of some of the glacio hydro isostatic effects. .................. 18 1 2 Predictions of global RSL change under the influence of glacial isost atic adjustment (GIA) in the present time. ................................ ................................ 19 1 3 LIG RSL predictions by a forward ice model. ................................ .......................... 20 1 4 Comparison between forward model prediction (Dutton & Lambeck, 2012) and observation (Che n et al., 1991) based on elevation age data of fossil corals in Great Inagua Island. ................................ ................................ ....................... 21 1 5 Comparison between forward model prediction (Dutton & Lambeck, 2012) and observation (Chen et al., 1991) based on elevation age data of fossil corals in San Salvador Island. ................................ ................................ ....................... 22 1 6 Map of the Bahaman archipelago ................................ ................................ ........... 23 2 1 Four stages of development of Bahamian is lands during a glacial/interglacial sea level cycle. ................................ ................................ ................................ ... 44 3 1 Standard level instrument ................................ ................................ ....................... 53 3 2 Cartoon of how level instrument works. ................................ ................................ .. 53 3 3 Cartoon of three readings from optical level. ................................ ........................... 54 3 ................................ .......... 55 3 5 Water depth recorder installation at Cockburn Town, San Salvador Island. ........... 56 3 6 Outline of the process of correlating surveyed LIG corals to MLLW. ....................... 57 3 7 The benchmark found at Cockburn Town, San Salvador Island. ............................ 58 4 1 Sea level relative to MLLW in Great Inagua (Mar. 27 and 28, 2013). ...................... 69 4 2 Measured sea levels in San Salvador (Apr. 2 and 3, 2013) with the water depth recorder every half a minute in the field. ................................ ............................. 69 4 3 Horizontal distribution of surveyed points in Great Inagua Island. .......................... 70 4 4 Horizontal distribution of surveyed points in San Salvador Island. .......................... 71 4 5 Surveyed elevations above MLLW for in situ coral heads in Great Inagua Island. .. 72

PAGE 10

10 4 6 Surveyed elevations for in situ coral heads in San Salvador Island. ....................... 73 4 7 Thicker rubble layer that lies above the platform in Great Inagua. .......................... 74 4 8 Overview of the extensive platform in the southeast of the surveyed reef in Great Inagua. ................................ ................................ ................................ ...... 74 4 9 Planed off Diploria (Reef I) on the platform in Great Inagua ................................ ... 75 4 10 Highest in situ Diploria head (Reef II) that lies above the rubble layer in Great Inagua. ................................ ................................ ................................ ............... 75 4 11 Highest in situ A. palmate (Reef II) (ID#83) in San Salvador Island. ..................... 76 4 12 Two highest corals at Cockburn Town, San Salvador Island ................................ 77 4 13 Highest in situ A. palmate below the unconformity (ID#96) that reaches a maximum height of 2.18 m. ................................ ................................ ................ 78 4 14 Age data of corals in Reef I and Reef II in San Salvador. ................................ ..... 79 4 15 Age data of corals in Reef I and Reef II in Great Inagua. ................................ ...... 79 5 1 Model predictions of LIG RSL in Great Inagua and San Salvador. ......................... 93 5 2 Generalized platform reef zonation cross section. ................................ .................. 94 5 3A View of a section of the fossil reef crest in San Salvador. ................................ .... 95 5 3B The same outcrop as in Figure 5 3A but looking west (towards the ocean). ........ 96 5 4 The approximate positions of the highest in situ corals (Reef II) in the paleogeographic map of San Salvador. ................................ .............................. 97 5 5 High Diploria (Reef II) covered by prograding beach sediment in Great Inagua. .... 98 5 6 Comparison between model prediction and field observation of LIG sea level in San Salvador and Great Inagua. ................................ ................................ ........ 99

PAGE 11

11 LIST OF OBJECTS Object page Supplementary data (.xlxs file 376 KB) ................................ ................................ ....... 102 Mat lab code for MLLW calculation ( .html file 24 KB) ................................ ................... 102

PAGE 12

12 LIST OF ABBREVIATIONS ESL Eustatic sea level, or global sea level, which is a function of continental ice volume and the shape of ocean basin. LIG Last interglacial period (~125 ka ago), also known as marine isotope stage (MIS) 5e, the most recent interglacial periods before the present one. MIS Marine isotope stages are alternating warm and cool periods in the Earth's paleoclimate, deduced from benthic oxygen isotopes in deep sea. Odd numbers represent warm period s and even numbers represent cold periods. MIS 2 and MIS 6 are the past two glacial maxima i.e. last glacial maximum and penultimate glacial maximum. MLLW Mean lower low water RSL Relative sea level, or local sea level, which is influenced not only by ice vol ume but also lithospheric response to ice/water loading changes that are caused by ice sheet decay/growth, i.e. glacio hydro effects.

PAGE 13

13 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science CONSTRAINING THE LAST INTERGLACIAL SEA LEVEL SIGNAL IN THE BAHAMAS By Jin Li May 2014 Chair: Andrea Dutton Major: Geological Sciences Given the climatic similarity between the last interglacial (LIG) period (~125 ka ago) and climate model projections for the near future, the reconstruction of eustatic sea level (ESL) in the LIG period is relevant for understanding future sea level change in in the context of global warming. U Th ages and elevation data for fossil corals from this time period in the circum Caribbean region are not consistent with relative sea level predictions from glacio hydro isostatic modeling, indicating that modificat ions need to be made to the existing ice model, specifically for the last and penultimate glacial maxima. To determine which of these ice sheet models needs adjusting, elevations of two coeval stratigraphic horizons within the fossil reef were surveyed on San Salvador and Great Inagua. The result demonstrates a minimal gradient in elevation between the sites and a rate of sea level rise that is significantly low er than the model predictions. This indicates that the penultimate glacial Laurentide ice sheet needs to be reduced and some adjustment to the last glacial maximum Laurentide ice sheet may also be required to bring the model into agreement with the observational data.

PAGE 14

14 CHAPTER 1 INTRODUCTION To project sea level change in the future requires a better understanding of ice sheet dynamics under sustained global warming conditions. One approach to acquire this knowledge is to study geologic records of sea level during a warm period in the past when the climatic condit ions were similar to today. The Last Interglacial (LIG) period, (~125 ka before present), also known as marine isotope stage (MIS) 5e, is the most recent period of global warmth that can be studied to understand the dynamics of polar ice sheets under warme r conditions. According to Mckay et al. (2011), the global mean sea surface temperature in the LIG period was only ~0.7 o C warmer than preindustrial conditions whereas the polar temperature was 3 5 o C warmer than preindustrial conditions. In addition, beca use the LIG period is the most recent interglacial before the present, there is more geologic evidence available compared to earlier interglacial periods. Although it is not a perfect analogue, the LIG period is still useful to understanding the behavior o f ice sheet and sea level in the future. One common method to reconstruct LIG sea level is to use elevations of fossil corals that can be dated accurately with the U Th geochemical method (Edwards et al., 2003; Stirling and Andersen, 2009). Despite numero us U Th dates of corals distributed globally during the LIG period, it remains challenging to reconstruct the true history of eustatic sea level (ESL) which varies in response to changes in the volume of land based ice and ocean basin shape. In particular, the typical approach of reconstructing relative sea level (RSL) at a single site as a direct proxy for ESL can be misleading because of geographic variability of the RSL signal that results from the superposition of glacio hydro isostatic effects ( Figure 1 1 and 1 2). These non eustatic effects include

PAGE 15

15 from the variable loading of ice (and water) on land ( and in the oceans) on glacial interglacial timescales. Quantitati ve consideration of these isostatic effects is therefore necessary in order to extract an ESL signal from RSL records, even at sites that are considered tectonically stable. An ice model spanning from MIS 6, the penultimate glacial maximum, to present has been previously employed to calculate isostatic effects at different sites during the LIG period (Dutton and Lambeck, 2012). At present, there is an inconsistency between the forward modeling results and field observations of RSL derived from fossil corals in the western North Atlantic and circum Caribbean region. A spatial gradient in the peak sea level across this region is predicted by the model ( Figure 1 3 ), yet, such a sea level change gradient has not been reported by previous researchers (Chen et al. 1991; Neumann and Hearty, 1996; White et al., 1998) ( Figure 1 4 and 1 5). These studies do disagree on the height of peak sea level across the region suggesting a possible geographic variability that has yet to be carefully documented. Also, the model p redictions for the rates of LIG sea level change at different sites are higher than the calculations from reconstructed sea level using geologic evidence such as fossil corals in the field. To resolve the discrepancy between the model predictions and field observations, the two most sensitive parameters to change in the model are the volume and geographic extent of ice sheet in the last and penultimate glacial maximum (MIS 2 and MIS 6, respectively) (Lambeck et al., 2012). Unfortunately, there is limited di rect observational data from the last glacial maximum (LGM) (Yokoyama et al., 2000;

PAGE 16

16 Lambeck et al., 2002; Peltier and Fairbanks, 2006), and even less for MIS 6. Most of the knowledge about the total volume of grounded ice at glacial maxima during the Quate rnary comes from the benthic 18 O signal, which is influenced by a combination of continental ice volume, seawater temperature and local hydrochemical conditions (Skinner and Shackleton, 2005; Sosdian and Rosenthal, 2012; Yu and Broecker, 2010). Although the majority of the benthic 18 O signal reflects changes in ice volume, it is difficult to separate these various influences apart to discern precise timing or extent of ice volume changes from this proxy. Alternatively, given that the LIG RSL in near to intermediate field sites with resp ect to the edge of ice sheet in the past glaciations (such as the Bahamas) is quite sensitive to the volume and geographic extent of grounded ice sheets in MIS 2 and MIS 6 as well as the viscosity of the upper mantle (further explained in Chapter 2 ), LIG R SL in these near to intermediate sites could be useful to constrain the ice sheet in the past two glacial maxima and the viscosity in the upper mantle. That is, the size and extent of the LGM ice sheet will affect the absolute elevation of the LIG shorelin e, whereas the rate of sea level rise due to forebulge relaxation is dependent on the size and extent of the ice sheet in MIS 6, and the lithospheric movement in response to the ice/water loading change due to the decay/ growth of continental ice sheet is determined by the viscosity of upper mantle. Thus, the goal of this research is (a) to determine whether or not the gradient in peak LIG sea level exists as predicted in the forward model, (b) if such a gradient exists, to quantify it, and (c) determine what the rate of sea level change is during LIG highstand across the Western North Atlantic/circum Caribbean region, in order to determine which parameter in the ice model needs to be revised and how to revise it.

PAGE 17

17 To reach these objectives, I surveyed the elevations of LIG corals with respect to a tid al datum on two Bahaman islands, San Salvador and Great Inagua islands ( Figure 1 6 ). Both islands are tectonically stable near to intermediate sites, and have well developed reefs in the LIG period. Numerous U Th age data exist for the se reefs, whereas few precise elevation data have been reported relative to known tidal datums (Freeman Lynde and Ryan, 1985; Chen et al., 1991; Thompson et al., 2011). Thus, a thorough survey of the reefs would be necessary to determine the peak sea level and rates of LIG sea level change in the two islands. These RSL data will provide invaluable information about the size and extent of the Laurentide ice sheet during MIS 6 and 2 as well as the viscosity in the upper mantle.

PAGE 18

18 Figure 1 1 A cartoon representation of some of the glacio hydro isostatic effects. The effects includ e Earth surface deformation and changes in the gravity field that result from the mass transfer of water between land based ice and the oceans on glacial interglacial timescales. As the ice sheet (pale yellow) melts and even after it is done melting, point A will rebound, Point C will subside, and the ocean basins will experience water loading. Additionally the sloped surface of sea level in the vicinity of the ice sheet margins that results from the g ravitational attraction between the water and the ice will relax due to the loss of ice mass.

PAGE 19

19 Figure 1 2 Predictions of global RSL change under the influence of glacial isostatic adjustment (GIA) in the present time. Warm colors represent RSL rise whe reas cool colors represent RSL fall. Note that near the former centers of ice sheets, RSL continues to fall due to crustal rebound and in the forebulge areas around the perimeter of the ice sheet, RSL is rising due to relaxation of Thi s f igure is reprint of Figure 3a in Page 30 from Tamisiea and Mitrovica, 2011 The moving boundaries of sea level change: Understanding the origins of geographic variability Oceanography 24 ( 2 ):24 39, doi:10.5670/oceanog.2011.25 Permission is granted for the reprint from t he Oceanography Society

PAGE 20

20 Figure 1 3 LIG RSL predictions by a forward ice model All the sites are located within Caribbean region and are considered near to intermediate sites with respect to the Laurentide ice sheet in the Last Glacial Maxima Data used in the figure are from Dutton, A., & Lambeck, K. (2012). Ice volume and sea level during the last interglaci al. Science 337(6091), 216 219 -10 -5 0 5 10 15 115 120 125 130 135 Sea level (m) Age (ka) San Salvador Great Inagua Yucatan Barbados ESL

PAGE 21

21 Figure 1 4 Comparison between forward model prediction (Dutton and Lambeck, 2012 same as Figure 1 3 ) and observation based on elevation age data of fossil corals in Great Inagua Island ( U Th age elevation data of fossil corals from supplementary data in Chen, J. H., Curran, H. A., White, B., & Wasserburg, G J. (1991). Precise chronology of the last interglacial period: 234U 230Th data from fossil coral reefs in the Bahamas. Geological Society of America Bulletin 103(1), 82 97 ) Either the gradient of sea level and/or the peak point change in forward model prediction needs to be changed to reach a consistency with the observation. -10 -5 0 5 10 15 115 120 125 130 135 Sea level (m) Age (ka) Model predicted RSL Elevations of corals ESL

PAGE 22

22 Figure 1 5 Comparison between forward model prediction (Dutton and Lambeck, 2012 same as Figure 1 3 ) and observation ( U Th age elevation data of fossil corals from supplemen tary data in Chen et al., 1991, same as Figure 1 4 ) based on elevation age data of fossil corals in San Salvador Island. -10 -5 0 5 10 15 115 120 125 130 135 Sea level (m) Age (ka) Model predicted RSL Elevation of corals ESL

PAGE 23

23 Figure 1 6 Map of the Bahaman archipelago ( A ) Map of the Bahama s Islands. Reprint of Figure 1 in Page 62 from in Mylroie, J.E. 2008 Late Quaternary sea level position: evidence from Bahamian carbonate deposition and dissolution cycles. Quaternary International 183(1), 61 75 Permission is granted by Elsevier A License Agreement between Jin Li and Elsevier has been signed (license n umber: 3347100562385 ) The locations of the Settlement Point (SP) and Virginia Key (VA) tide gauges are shown with red dots ( B ) M ap of San Salvador Island and the approximate location (near Cockburn Town) of reef s in the research. ( C ) M ap of Great Inagua I sland and the approximate location of reef units in the research. Figure 1 6 B and 1 6 C is the reprint of Figure 1 in Page 83 from Chen, J. H., Curran, H. A., White, B., and Wasserburg, G. J. (1991). Precise chronology of the last interglacial period: 234U 230Th data from fossil coral reefs in the Bahamas. Geological Society of America Bulletin 103(1), 82 97. P ermission s are granted from GSA to use the two single maps without fees or further requests to GSA

PAGE 24

24 CHAPTER 2 BACKGROUND Reconstructing sea level with U Th dated fossil corals Corals as sea level position indicators Fossil corals provide a useful sea level marker. Most corals mainly obtain energy from photosynthetic algae that requires sunlight in clear, shallow water. These corals can only survive within a limited water depth below sea surface where the depth range varies with species and can be a few meters to tens of meters. Thus, they can provide a minimum estimate of paleo sea level elevation. Hence, there is a range of uncertainty with respect to assigning an appropriate paleodepth to fossil corals. To narrow down this uncertainty and ensure the reliability of the elevatio n data from fossil corals, researchers usually choose species that live close to sea surface and have a limited range of depth. For example, Acropora palmate is a prominent shallow water reef builder in the Caribbean region that is often used in paleo sea level studies. A. palmate is typically considered to live at depth less than 5 m to water surface during low tide (Lighty et al., 1982), although this species has been found living in depths as great as 17 m (Goreau and Wells, 1967). It typically occupies the high energy reef crest zone. For instance, at Carrie Bow Cay, Belize Barrier Reef A. palmate lives at water depth of 0.5 to 2.0 m (Riitzler and Macintyre, 1982) and at Eleuthera Island, the Bahamas, reef lives at depth shallower than 3 m (Zankl and Sch roeder, 1972). Lighty et al. (1982) used in situ A. palmate samples from reefs within the tropical western Atlantic to construct a minimum sea level curve for the past 10,000 years with radiocarbon dating fossil corals and their results are consistent with sea level age plots by Curry et al. (1969), Moslow and Heron (1981), Scholl et al. (1969). This consistency was taken as proof of the

PAGE 25

25 reliability for A. palmate as an index for paleo sea level position. Additional information can also be used to narro w the possible water depth that the corals lived in and hence reduce the uncertainty in the sea level elevation reconstruction. This evidence may include the sedimentary facies in the sites where the coral was found or the assemblage of other organisms suc h as coralline algae, trace fossils (i.e. geological records of biologic activities such as burrows and borings). The other factor that makes fossil corals an ex cellent proxy to reconstruct pa l e o sea level position is that they can be dated precisely by ra diometric decay method. Coralline aragonite incorporates trace amounts of uranium from seawater during growth, making it possible to precisely date the timing of growth using U series isotopes (e.g. Chen et al., 1991; Stirling et al., 1995; Stirling et al. 1998; Stirling and Anderson, 2009) and thus to construct elevation age data of sea level in the past. U U Th dating, also called 230 Th dating, is a radiometric dating technique commonly used to determine the age of marine carbonate materials such as coralline aragonite (Barnes et al., 1956) and aragonitic carbonate bank sediments (Slowey et al., 1996). With this dating method, three pertinent nuclides in the 238 U decay chain are used: 238 U that has a half life of ~4.510 9 years (Jaffey, 1971), 234 U that has a half life of ~ 245,620260 years (Cheng et al. 2013) and 230 Th that has an even shorter half life of ~75,584110 years (Cheng et al. 2013). The 238 U initial parent nuclide decays via to short lived isotop es ( 234 Th with a half life of ~24.1 days and 234 Pa with a half life of ~6.7 hours) to the intermediate daughter 234 U product, which in time further decays to 230 Th. Fractionation occurs between uranium and thorium, since materials grow or precipitate from natural waters that contain only uranium in solution, and no thorium. This is

PAGE 26

26 because uranium, which is delivered to the ocean by weathering, is soluble in oxyge nated water whereas thorium is insoluble. Living corals incorporate uranium isotopes (mainly 238 U and 234 U) from seawater without fractionating between them and thus have an initial 234 U/ 238 U ratio equal to that of the ambient seawater (Cheng et al., 2000a ; initial condition of seawater and thus in corals lays the foundation of performing U Th dating on fossil corals. In addition to determining an age, the initial 234 U/ 238 U ratio (usually r eferred to 234 U i Th age and used to check the reliability of calculated age (see below). The incorporated uranium isotopes decay to 230 Th after the death of coral. Due to its shorter hal f life, 230 Th approaches a secular equilibrium with 234 U, i.e. the decay rate of 230 Th is equal to its production rate (decay rate of 234 U). The U Th dating technique calculates an age based on the degree to which secular equilibrium between 230 Th and 234 U has been restored after the initial fractionation. In theory, the ages dated by this method can range from as young as 3 years to more than 600 ka (thousands of years ago) (Edwards et al., 2003). The 230 Th age is derived from Equation ( 2 1) (Edwards et al ., 1987a): (2 1) (relationship between decay constant and half life ( ) is: subscript 238 refers to 238 U, 234 refers to 234 U and 230 refers to 230 Th), and activity ratio.

PAGE 27

27 As with other radiometric dating techniques, the calculated age will only be accurate if the sample has remained a closed system. The measurement precision has been dramatically improved with the use of TIMS (thermal ionization mass spectrometry) and MC ICPMS ( multiple collector inductively coupled plasma mass spectrometry) in U series isotopic analysis compared with the early alpha cou nting technology. So far a precision of 1 2 ppm (part per million) has been achieved in activity ratios (Stirling and Andersen, 2009). On the other hand, fossil corals with a perfect closed system of U Th isotopes are rare in nature (e.g. Edwards et al., 1 987b; Chen et al., 1991; Thompson et al., 2003; Thompson et al., 2011). Diagenetic effects after the death of corals can cause loss and/or gain of U series isotopes in non radioactive processes (Chen et al., 1991; Thompson et al., 2003). This is the so cal series dating in fossil corals To avoid samples that are altered by diagenetic effects, X ray diffraction (XRD) and visual examination with a microscope are performed prior to is otopic analysis and coral samples that have undergone serious diagenesis will be rejected. These examinations can identify the signs of recrystallization and secondary calcite in a preliminary level (e.g. Chen et al., 1991; Stirling et al., 1995; Israelson and Wohlfarth, 1999). The carbonate structures of corals include walls, septa and possible cements trapped in the septa. Wall structures are made up of dense aragonite that are more likely to be free of detrital and recrystallized material whereas septa o ften contain substantial amounts of secondary calcite and 232 Th concentration. Samples that are discovered to contain significant amount of calcite can then be excluded. A commonly

PAGE 28

28 used criterion is that samples with more than 99% aragonite, minimal calcit e infilling, ~3 ppm uranium in total and low 232 Th concentrations are likely to be a U Th closed system and will be used for U Th dating (e.g. Stirling et al. 1995; Israelson and Wohlfarth, 1999). After isotopic analysis and U Th age calculation, the initi al 234 U/ 238 234 U i ) of the sample is determined with the calculated U Th age using Equation ( 2 2). (2 2 ) w here 234 represents the decay constant for 234 234 U refers to the activity ratio of 234 U and 238 time ( when measurement is performed). The comparison of the calculated 234 U i value to that of modern seawate r has been commonly used to check the degree of open system remobilization and thus test the reliability of the U Th age (Edwards et al., 1987a, b) As mentioned above, the 234 U/ 238 U ratio at the time when corals were living should be close to that of ambi ent since living corals incorporate uranium isotopes from the surrounding seawater with no fractionation between 238 U and 234 U (Cheng et al., 2000a; Delanghe et al., 2002; Robinson et al., 2004a). Given that the uranium isotopic composition in the ocean ha changed significantly since at least 200 ka ago (Cheng et al., 2000a; Robinson et al., 2004a), if the back 234 U i 234 U value in 6), it should be reasonable to conclude that the coral sample has not behaved as a closed

PAGE 29

29 system of uranium isotopes and that the U Th age is not reliable (Gallup et al., 1994; Stirling et al., 1998; Esat et al., 1999; Yokoyama et al., 2001). A large numb er of age elevation data for paleo sea level in the last full glacial cycle (since ~130 ka ago to present, in particular, for the LIG period) have been obtained by performing U Th on fossil coral reefs around the globe (mainly in tropical and subtropical r egions) (e.g. Chen et al., 1991; Stirling et al., 1995; Stirling et al., 1998; Israelson and Wohlfarth, 1999; Thompson and Goldstein, 2005; Blanchon et al., 2009; Thompson et al., 2011; Muhs et al., 2011). Yet, because diagenetic effects become more signif icant as coral samples become older, thus, most of U Th ages from coral samples older than ~130 ka (i.e. time before the last glacial cycles) have been rejected based on the 234 U criteria, making few data available for that time (Gallup et al., 1994; Stirling et al., 1995; Esat et al., 1999; Stirling et al., 2001; Potter et al., 2004; Stirling and Anderson, 2009). The practical dating range is currently much shorter than theoretical one (as old as 600 k a Edwards et al., 2003). To avoid removing age da from altered corals into more reliable age estimates, many efforts have been made in recent years on an improved understanding of diagenetic mechanisms within fossil reef systems (e.g. Chen et al., 1991; F ruijtier et al., 2000; Henderson et al., 2001; Thompson et al., 2003; Villemant and Feuillet, 2003; Scholz et al., 2004; Scholz and Mangini, 2006; Thompson et al., 2011). Several open system models have been put forth, providing the possibility to convert inaccurate conventional U series age determinations into reliable open system U series modeled ages. Examples include the open system model built by Thompson et al. (2003) and the model built by Villemant and Feuillet (2003). One

PAGE 30

30 diagenetic mechanism has b een proposed by Chen et al. (1991), Henderson et al. (2001) and the recoil redistribution has been used as the primary diagenetic mechanism in both models. Another open system model is sys 2006). One specific application of open system model is: Thompson et al. (2011) correct the ages for diagenetic disturbed fossil corals with the open system model by Thompson et al. (2003) and the modeli ng results help resolve age difference between two LIG coral reef units in Great Inagua Island, the Bahamas. The accuracy of the open system model by Thompson et al (2003) depends on the choice of 234 U value in ocean, the variability of which is still bei ng debated. Overall, all these models sometimes work well for samples that are at LIG ages or younger than LIG ages: model corrected ages turn out to be consistent with conventionally derived ages for coral samples that are not significantly influenced by diagenetic effects from the same reef unit (Scholz and Mangini, 2006). But in other cases, the array of diagenetic compositions has been shown to be inconsistent with the open system models due to other modes of diagenesis or the superposition of multiple diagenetic pathways (Scholz and Mangini, 2006). Glacio hydro isostatic effects In contrast to ESL, which is mainly func tion of continental ice volume as well as the shape of ocean basin RSL can be influenced by vertical movements of continents, including tectonic activity and glac io hydro isostatic adjustment. In tectonically active regions, the tectonic component is subtracted from measured elevations of paleo sea level markers (e.g. fossil corals) in order to identify LIG sea level position ( Gallup et a l., 1994; Chappell et al., 1996) However, uplift (or

PAGE 31

31 subsidence) rate is calculated from the past time when sea level was assumed stable. Assuming that the uplift (or subsidence) rate does not change with time, the tectonic component is estimated by multi plying this uplift (or subside nce ) rate with the age. As a result, the problem becomes circular and the error in sea level reconstruction could be overamplified. This will impede an accurate reconstruction of sea level in tectonically active sites. In addi tion, considering that ESL in the LIG period is higher than that during other interglacial periods based on the deep ocean oxygen isotope record ( Imbrie et al. 1984; Chappell and Shackleton, 1986; Shackleton, 1987; Lisieki and Raymo, 2005), LIG sea level i ndicators such as fossil coral reefs are now exposed in the air and accessible for research even in tectonically stable sites. Consequently, it can be more advantageous to reconstruct LIG sea level in tectonically stable sites over tectonically active site s. Glacio hydro isostatic effects are caused by the deformation of the solid Earth surface and ocean surface as well as the gravity field in response to changes in ice and water loads during glacial interglacial cycles, and can influence RSL in time scale of hundreds of thousands of years (Lambeck and Chappell, 2001; Peltier and Fairbanks, 2006). As shown in Figure 1 1 during glacial periods there will be a depression beneath the ice sheets and a corresponding bulge in the front of ice sheet edge. Sites si tting on the bulge (Point C in Figure 1 1 ) will experience an obvious rising RSL during the interglacial period even when degla ciation ceases (no grounded ice volume change) and ESL stays stable. In other words, paleo sea level position of any previous interglacial is a function of not only ice volume during that interglacial period, but also ice volume of glacial periods preceding and following that interglacia l period (Potter and

PAGE 32

32 Lambeck, 2004). For instance, the position of paleo sea level in the LIG period should be dependent on ice volume of MIS 6 (penultimate glacial maxima), LIG and MIS 2 (last glacial maxima, LGM). The influence of ice volume before and a fter the interglacial, combining with the effect of water load change throughout the interglacial, comprises the glacio hydro isostatic effect. Glacio hydro isostatic effects vary with distance between main ice sheet center and the site of interest: the ef fect is remarkable in the near to intermediate field sites (e.g. San Salvador Island, Bahamas, is close to the edge of North American Ice Sheet) whereas becomes much weaker in far field sites (e.g. the Seychelles in the Indian Ocean) ( Figure 1 2 ). The LIG sea level reconstructed from different sites could be inconsistent with each other without correction of glacio hydro isostatic effects (Lambeck et al., 2012). In addition, a large number of age elevation data of LIG relative sea level (based on elevation and U Th dating of fossil corals) have been collected across the western North Atlantic and circum Caribbean region where the isostatic effects on RSL are significant and cannot be neglected (as shown in Figure 1 2 ). Thus, a quantitative understanding of g lacio hydro isostatic effect s is necessary to extract ESL signal from such RSL data. An ice model that spans MIS 6 to present and includes over 6000 data points has been utilized in Dutton and Lambeck (2012) to quantify the isostatic effects. This ice mod el combines ice models that are previously developed for the last two glacial cycles back to MIS 7 by Lambeck et al (2006) and Lambeck et al (2010). An iterative process will be performed to revise the model with observation data set (Lambeck et al., 2012) In the first iteration, the ice volume in the LIG period is assumed to be equal with that in

PAGE 33

33 modern time. This could be a reasonable assumption since the influence of ice volume during LIG period on the position of LIG sea level is much smaller than that during the glacial before and after the LIG, i.e. MIS 6 and MIS 2. The change gradient of LIG RSL is dependent upon the size and extent of the MIS 6 ice sheets and the viscosity in upper mantle, which lies just below the lithosphere and supports the litho spheric movement in response to the change of ice/water loading whereas the volume of ice sheets in MIS 2 dominate the absolute elevation of LIG RSL (Lambeck et al., 2012). To be noted, in the near to intermediate sites such as circum Caribbean region, th e absolute elevation of LIG RSL should be still declining, due to the residual relaxation of the periphery bulge formed in MIS 2 (LGM) that has yet to occur. Lambeck et al (2012) estimates the deviation from the condition in the first iteration (i.e. ESL o f LIG is equal to that of present time) to be at most 6 8%. Difference between the observations in the field (sea level reconstructed using corals and other sea level indicators) and the predictions from the model by correcting the glacio hydro isostatic e ffects should reflect the difference between ice volume (thus, ESL) of LIG and present time. In the second iteration, this difference would be distributed between individual ice sheets in the LIG period. The modeling process described above will be repeate d until the modeling result is consistent with the observation in the field. At this point, there is a discrepancy between the forward modeling results and the observational data from the Western North Atlantic and circum Caribbean region (Dutton and Lamb eck, 2012). As shown in Figure 1 3 the ice model predicts a spatial gradient in peak sea level across the circum Caribbean region, but no such gradient has been clearly identified (Fig ure 1 4 and 1 5). For example, peak sea level is estimated as

PAGE 34

34 ~+2 to 2. 5 m in the middle part of LIG in both Great Inagua and San Salvador Island, based on coral U Th age and elevation data and the assumption that corals grew just adjacent to sea surface (Neumann and Hearty, 1996; White et al., 1998). In contrast, Chen et al. (1991) propose that sea level stays at ~+5.5 m above present sea level in San Salvador Island and ~+5 m in Great Inagua Island for most time of LIG period, based on the assumption that corals grew at depth of 3 m below sea surface in the Bahamas. Thompson et al., (2011) propose that sea level peaked at +6 m in both San Salvador and Great Inagua, also invoking the 3 m paleodepth estimate. Flank margin caves positioned at +4 m have been used to argue for a +4 m peak sea level in the Bahamas during the LIG se a level highstand (Mylroie and Carew, 1990). However, only one of these studies reports elevations relative to a tidal datum so it is not clear whether the range of data is an artifact of poor elevation control or reflects some real variation. The rate of sea level change with time that is predicted by the model is also higher than that from observation ( Figure 1 3 ). To resolve the discrepancies between the modeling and observing results, two steps are needed. The first is to better constrain the elevations of the LIG reefs in the near to intermediate sites (sites within circum Caribbean region). The second is to revise the related parameters in the model the size and geographic extent of ice sheets in MIS 2 and/or MIS 6 and the value of viscosity in upper mantle based on the results of RSL reconstruction in this research to improve the fit between the model and the observations. Geologic Setting and LIG deposits in the two Bahamian Islands The Bahamian Archipelago ( Figure 1 6 ) began to form after the rifting of Pangea and the opening of Atlantic Ocean in the late Middle Jurassic, and the carbonate banks occupied foundered portions of rifted continental crust as well as extinct volcanic

PAGE 35

35 seamounts (Meyerhoff and Hatten, 1974). The surficial rocks across the Bahamas are entirely Quaternary carbonates. Rocks at lower elevations are mainly composed of carbonate eolianite and subtidal deposits (such as coral reefs), whereas rocks at higher elevation (above +6 m) are mostly eolian facies (Mylroie, 2008). It is the ESL change caused by glaciation decay/growth that dominates in the production of Bahamian carbonates throughout the Quaternary (since ca. 260 million years ago) (Mylroie and Carew, 1995; Kindler and Hearty, 1995; Mylroie, 2008). San Salvador Island ( 2 4 06 N 74 29 W ) ( Figure 1 6 B ) and Great Inagua Island ( 25 4 N 77 20 W ) ( Figure 1 6 C ) are located on discrete carbonate platforms of the Bahamian Archipelago. They are selected for the study because the platforms have been tectonically stable or slowly su bsiding at a rate of 14 mm/ka over the past 30 Ma (millions of years) (Freeman Lynde and Ryan, 1985). Both islands are located close to the edge of the North American ice sheet in the LGM (MIS 2) and are taken as near to intermediate sites, where RSL is qu ite sensitive to glacio hydro isostatic effects. Hence, LIG RSL observations for these two islands will be useful to constrain the ice sheet size and extent in MIS 2 and 6. In particular, since Great Inagua Island lies further from the former margins of N. American ice sheet than San Salvador does, accurate reconstruction of LIG RSL in the two islands can be used to check if the spatial gradient of peak sea level predicted in the ice model (Dutton and Lambeck, 2012) exists. There is already a great deal of U Th data from fossil corals in the two islands (Chen et al., 1991; Thompson et al. 2011). Yet, precise elevation measurements with identified tidal datum are scarce. Surficial depositional units of late Pleistocene (typically, LIG) in these islands are q uite well developed. In particular, large,

PAGE 36

36 well preserved fossil coral reefs with mid to late Pleistocene ages crop along the western coast of San Salvador, immediately northwest of the center of Cockburn Town, and along the west coast of Great Inagua at D related coral rubble facies, depositional units of LIG period also include large eolianite ridges, caps of well developed terra rossa paleosols and interbeds of loose, light protosols (Hearty and Kindler, 1997; Hearty 1998). Erosional features of LIG include sea cliffs, flank margin caves, notches and basal ramps (Carew and Mylroie, 1995b). Tectonic setting The tectonic regime of the Bahamas platform has been debated in the past. The southeastern portion of the Baham an Archipelago has been interpreted as tectonically active on the basis of seismic reflection data of the deep water carbonate bank that indicates its proximity to a tectonic plate boundary (Mullins and Hine, 1989). Carew and Mylroie (1995a) proposed tecto nic stability there since the late Quaternary (an interval including at least MIS 1, 3 and 5) based on the data from in situ fossil reefs and flank margin caves. Flank margin caves are developed rapidly in small lenses within a limited vertical range and t hus can be a good sea level indicator. Throughout the Bahamas, fossil coral reefs that have been found at 0 to ~+2.5 m above present sea level all turn out to be formed during the LIG highstand (e.g. Chen et al. 1991; Hearty et al., 2007). Similarly, the flank margin caves, which have been found at +1 to +7 m above present sea level, are supposed to form during the LIG period, based on the finding that the maximum U Th ages of stalagmites discovered in the caves are of less than 100,000 years (Carew et al. 1984; Carew and Mylroie, 1987). Assuming a sea level highstand of +6 m above present sea level, Carew and Mylroie (1995a) suggested that these data proves an overall tectonic stability with a possible isostatic subsidence at a rate of 1 2

PAGE 37

37 cm per thousand years at least since 300 ka ago. However, considering the geographic variability in the peak LIG sea level as mentioned before, it is not safe to make a pre a circular problem when one tries to use this subsidence rate to f ig ure out LIG sea level change in the Bahamas. Instead, subsidence rate estimated from well over the past 30 Ma (millions of years) are 14 mm/ka (Freeman Lynde and Ryan, 1985 ) LIG sedimentary record of the Bahamas As mentioned above, t he evolution of the Bahamian islands has been dependent largely on sea level changes since the beginning of Quaternary. Information of paleo sea level for most time of the Quaternary (from early until late Pleistocene, i .e. time prior to substage 5e, LIG) comes from deep ocean oxygen record (e.g. Imbrie et al. 1984; Chappell and Shackleton, 1986; Shackleton, 1987; Lisieki and Raymo, 2005) alone. The oxygen isotope record indicates that sea level elevation during LIG high stand was higher than any other time period throughout Quaternary. Most of U Th dating results of fossil corals on separate islands across the Bahamas also indicate a LIG age consistently (e.g. Great Inagua by Chen et al. (1991) and Thompson et al (2011); San Salvador Islands by Chen et al. (1991); Andros Island, Grand Bahama Island and New Providence Island by Neumann and Moore (1975)). Further, amino acid racemization (AAR) dating was used on whole rock samples to determine the relative ages so as to cor relate units within and/or between islands (Hearty and Kindler, 1995; Hearty and Kaufman, 2000; Hearty and Neumann, 2001). Hearty and Kaufman (2000) employ whole rock AAR method for regional correlations in the Bahamas and demonstrate consistency among mea n A/I ratios for dunes, chevron ridges, and sub boulder deposits on several islands. Based on absolute ages from U Th dating and relative ages from A/I

PAGE 38

38 ratios numerous surficial marine deposits across the Bahamas have been attributed to MIS 5e. Furthermor e, the depositional model for the LIG period on San Salvador Island appears to apply in a general sense to the whole Bahama Archipelago (Carew and Mylroie, 1995b). Sea level reconstruction from fossil corals Chen et al. (1991) provide 34 U Th ages for foss il LIG corals from Cockburn work was later supplemented by some 112 raw U Th analyses reported by Thompson et al. (2011) of corals from the same reefs studied by Chen et a l. (1991) after exclusion of samples with anomalous Th and U. The advantage of the more recent work is that the ages are much more precise and the dataset includes a more detailed insight into diagenesis through multiple subsampling of individual coral hea ds. Unfortunately no elevations were reported for the Thompson et al. (2011) dataset, so the only paired age elevation data available for this region is the Chen et al. (1991) dataset aside from a few analyses of corals from Abaco Island (Hearty et al., 20 07). Chen et al. (1991) interpret ed the beginning of reef construction at ~132 ka ago, initiated by the colonization of Acropora palmate on hard surface of previously formed carbonate rock and possibly other corals in the Cockburn Town reef, San Salvador Island. Patch reefs developed in large scale from 129 to 126 ka, reef crest was composed of reef building coral A. palmate and the associated patch reefs flourished along the flanks. The elevations of dated coral samples species including A. palmate, Mon tastrea annularis and Diploria strigosa range from 1.1 to 2.6 m above mean lower low water (MLLW) after correction or reported elevations to this tidal datum by Dutton and Lambeck (2012).

PAGE 39

39 int, Great Inagua, a regionally extensive erosional surface bracketed by fossil reefs can be traced for at least 4 km (Chen et al., 1991; White et al., 1998; Wilson et al., 1998). This erosional unconformity has been interpreted as a wave cut bench ( Figure 4 8 ). An erosional surface has also been observed o n San Salvador Island (Hearty and Kindler, 1993; White et al., 1998), Abaco Island (the Bahamas) (Hearty et al., 2007), Florida (Fruijtier et al., 2000) and the Yucatan (Blanchon et al., 2009). This erosional unconformity has been interpreted to represent an ephemeral ESL fall near the middle of the LIG period, separating two LIG sea level highstands as well as two generations of reef growth (White et al., 1998; Thompson et al., 2011). Thompson et al. (2011) resolve a 4 ka age difference between the two ree fs: the youngest fossil coral below the unconformity surface has an open system model age of 123 ka and the overlying reef has an interpreted age of 119.2 ka using the open system model of Thompson et al. (2003). In contrast, Chen et al. (1991) report clos ed system ages of the reef above unconformity at 122.0 to 123.8 ka and the reef below unconformity at 130.0 to 124.9 ka. Hence, the unconformity is inferred to be formed at ~125 ka with an uncertainty of ~2.1 ka using closed system ages or somewhere betwee n 119 and 123 ka using the open system model (more detailed discussion in Chapter 4 ). Neumann and Hearty (1996) and Hearty et al (2007) both suggest a rapid sea level rise to ~ +6 to 9 m at the end of the LIG sea level highstand based on notch formation, although the notches cannot be directly dated. Neumann and Hearty (1996) inferred that sea level for most of the LIG interval remained near +2 m, restraining reef

PAGE 40

40 growth, and that the notch found at +6 m represents a rapid and brief excursion just before t he close of the LIG period. They speculated that this peak in sea level did not last long since no fossil corals have been found higher than +3 m, i.e. no fossil coral has been found corresponding to a +6 m sea level. The subsequent sea level drop must hav e been rapid in order to prevent the notch profile from erosion (Neumann and Hearty, 1996). Chen et al. (1991) also inferred a rapid sea level decrease at the end of the highstand, shortly after 120 ka ago according to the closed system ages. They estimate the rate of sea level fall to be more than 3 m/ka, based on the lack of corals younger than 120 ka found in the fossil reef of San Salvador and Great Inagua Island. However, we note that the seaward portion of the reef at both locations has been eroded by Holocene sea level rise, such that the regressive phase of sea level at the end of the LIG period is most likely missing from the sedime ntary record that is preserved. In addition to the difference in the timing of LIG sea level highstand in the Bahamas ( Chen et al., 1991; Thompson et al., 2011), discrepancy exists in the elevations of fossil corals in the LIG and thus possible sea level elevations in the Bahamas among previous studies. Different results of highest in situ corals, which reflect the minimum elevation of peak RSL in the LIG period, have been reported ( Table 2 1 ). In addition to the geographic position, another probable reason for the discrepancy in the elevation of highest in situ fossil coral reported in dif ferent islands of the Bahamas is that except in a few cases (e.g. Chen et al., 1991), reference (i.e. datum, such as mean low water springs, MLWS, and mean water level, MWL) used in the measurement e Bahamas, this reference

PAGE 41

41 uncertainty could reasonably cause a discrepancy of at most ~0.7 m in the elevation result. In this project, surveys along the two reef horizons that are separated by the unconformity surface in Great Inagua and San Salvador Isla n d were performed with a standard level instrument with the focus on (1) the highest in situ corals at the top of the reef and (2) the elevation of corals just below the unconformity boundary of the two reef sequence. All the elevation data are linked to l ocal MLLW. As described above, this unconformity surface was inferred to form when sea level fell close to the middle of LIG period and reef formed prior to this sea l evel fall (referred to as Reef I ) was eroded (or planed off). Later sea level rose back t o and stayed at a high point until the end of LIG when sea level fell rapidly (Chen et al., 1991; White et al., 1998; Wilson et al., 1998; Neumann and Hearty, 1996; Hearty et al., 2007). The reef that developed on top of the uniformity surface during this period is referred to as Reef II The elevation of the highest in situ coral (from Reef II ) should represent the minimum position of peak sea level during the LIG period. The difference in peak sea level for San Salvador and Great Inagua (if there is any) would demonstrate the existence of spatial gradient of peak sea level that is predicted by the ice model (Dutton and Lambeck, 2012), and provide constraints to revise related parameters (ice volume in MIS 2 and MIS 6). With the elevations of corals just be low the unconformity (highest in situ coral from Reef I ), sea level position prior to the rapid sea level change event in the mid LIG can also be estimated. Combining the elevation of highest in situ coral from Reef II and Reef I, the rate of sea level cha nge during the LIG period can then be estimated, which is another significant parameter in the ice model.

PAGE 42

42 Carew and Mylroie (1995b) proposed a conceptual model of depositional process under the control of sea level changes in the LIG period by integrating the products and processes of subsidence, sea level change, subaerial diagenesis and carbonate sedimentation. Four phases of the development of the Bahama n Islands are delineated based on sea level change: a transgressive phase, a stillstand phase, a regre ssive phase and a lowstand phase. General processes in the four phases are illustrated in Figure 2 1 In the lowstand phase ( Figure 2 1 A) when sea level stays below the level of the platform (presently at 10 to 20m) pedogenesis and karst processes domina te and carbonate production on the bank stops. Terra rossa paleosols are produced when the bare c arbonate platform is exposed through very low sedimentation rates of aeolian dust derived from Africa (Carew and Mylroie, 1995b) The transgressive phase is th e early stage of sea level rise during deglaciation ( Figure 2 1 B). The carbonate platform begins to be flooded and carbonate sediment production initiates. The so called with their symbiotic algae, calcified green and red algae. Transgressive eolianites are formed and coral growth starts to build reefs. When sea level stays high for a period of time (still stand phase) eolianite, beach and subtidal deposits, including reef s and shoals, are deposited ( Figure 2 1 C). Reefs grow up to sea level and lagoons are filled with carbonate sediment. The transgressive eolianite undergoes attack by coastal processes. This is followed by the regressive phase, when ice sheets begin to form and expand, and sea level begins to fall ( Figure 2 1 D). Previously inundated carbonate banks become emergent may get partially removed to provide material for regressive eolianites.

PAGE 43

43 Table 2 1. Elevation of top in situ corals measured in the Bahama s Site Elevation of top coral Reference San Salvador up to 2.25 m relative MHT Chen et al., 1991 San Salvador San Salvador ~2 m (tidal datum not clear) xxxxx Carew and Mylroie, 1987; Carew and Mylroie, 1995a ; Hearty and Kindler, 1993 Thompson et al., 2011 Great Inagua Reef II: ~3 m; Reef I: ~ 1 m (assuming 1 m of erosion to form wave cut bench) Thompson et al., 2011 Great Inagua Reef II: 0.25 to 1.5 m; Reef I: 0 m Chen et al., 1991 Abaco Island below +3 m Hearty et al., 2007 Andros Island +1.5 m (lithified littoral rubble containing unaltered coral) Neumann and Moore, 1975 New Providence ~2 m Garrett and Gould, 1984 New Provid ence + 2 m, in situ coral head Neumann and Moore, 1978 Berry Islands +0.7 m, in situ Montastrea coral head Neumann and Moore, 1975 Moore's Island No coral reported, Only bioerosional notch, +5.6 to 5.9 m Neumann and Moore, 1977

PAGE 44

44 Figure 2 1 Four stages of development of Bahamian islands during a glacial/interglacial sea level cycle. (A) Low stand phase; (B) Transgressive phase (C) Still stand phase; (D) Regressive phase All the four figures are revise d from Figure 1 (Page 6 7) in Carew and Mylroie, 1995b Reprint of Figure 4 in Page 64 from in Mylroie, J.E. 2008, Late Quaternary sea level position: evidence from Bahamian carbonate deposition and dissolution cycles. Quaternary International 183(1), 61 75. Permission is granted by Elsevier. A License Agreement between Jin Li and Elsevier has been signed (license number: 3347100562385 )

PAGE 45

45 CHAPTER 3 M ETHODS Field work was conducted on Great Inagua and San Salvador Island. The main objective was to survey, i.e., using a level instrument to survey along the LIG reefs and tying the results to mean lower low water (MLLW) as a tidal datum. The following includes three p arts: (1) a brief introduction of how the surveying was carried out, (2) description of measurements, (3) relation of elevation data to MLLW, and (4) calculation of horizontal positions. How to use a level instrument A level instrument mainly comprises objectives), tripod (supporting the optical level) and a level staff (a leveling rod that stands on the objective for surve y) ( Figure 3 1 A and 3 1 B ). The instrument is used to measure the height of one objective relative to another by shooting them with the optical level separately. The survey results are elevation differences between different objectives. Figure 3 2 is a cart oon that briefly illustrates how a level instrument works. The accuracy of the level instrument is 1 centimeter. In addition to elevation measurement, the leveling instrument can be used to identify relative horizontal positions of objectives at the same t ime. From the screen of the optical level (the telescope), there are three reading lines: reading from the mid line represents the relative elevation, whereas the readings from the upper distance line and lower distance line are used to calculate the dista nce between the shooting objective and the survey station (where the optical level is installed). Based on some geometry principles ( Figure 3 3 ), the distance (in meters) is calculated as 100 times the difference between reading

PAGE 46

46 of upper distance line and lower distance line. The optical level used in this research (Leica NA724) is equipped with a horizontal circle, which is used to measure the angles between different objectives shot by the optical level. When shooting the first objective in every station, a bearing is taken with a compass as a reference. In this case, absolute bearings of all the objectives surveyed at the same station can be calculated from relative azimuths of each objective (readings of the horizontal circle on the optical level) and th e reference bearing. Generally, the operation of a level instrument includes 3 steps: installation, adjustment and surveying (reading). First, place the optical level (mounted on the tripod) in a proper position, which is not too far from the objective to ensure a clear reading. Second, leveling adjustment. The optical level is adjusted with the leveling screws and the scale is set exactly vertical with the bubble. Third, take a sight of the first objective, record all the data inside the eyepiece of the le vel instrument (data from the upper, middle and lower line), as well as the azimuths from the horizontal circle, and also record the bearing on the compass as mentioned above. Once this first sight is taken, the leveling screws should never be touched unti l the end of surveying. Subsequently, shoot and measure the elevations as well as horizontal positions of various objects by turning around the instrument. When the objective is too far from the station, the station (optical level) has to be moved. At this time, a reference is back shot in order to link all the data together. Level measurement in the field LIG deposition features As mentioned in Chapter 2 both Great Inagua and San Salvador Island preserve evidence of two coral reef growth stages, i.e., tw o fossil reef horizons that are separated

PAGE 47

47 by an unconformity. Surveys were performed to determine the elevation of specific stratigraphic features: (1) the highest in situ fossil coral as well as (2) the unconformity surface that is interpreted as evidence of an ephemeral drop in sea level during the LIG period. In the process of surveying as described above, a the conch shell that can be reoccupied was used repeatedly as a reference point in Great Inagua ( Figure 3 4 ) to tie sea level measurements to the su rficial features. Sea level measurements were made using the level, and returning to the same location at many times during the day to capture at least one complete tidal cycle each day. In San Salvador Island, no specific reference point was used so many times because we were able to install and use a water depth recorder to measure the position of sea level through the tidal cycles. In Great Inagua, we measured the elevation and horizontal position of 1 0 in situ corals that were identified as local high points, 7 measurements delineating the vertical extent of coral rubble layers, and 2 2 points on the unconformity. In San Salvador, we measured 19 in situ coral fossils that were local high points across the reef s urface and 8 additional points on the unconformity surface. We also found and measured the position of a permanent benchmark in both islands, which can be used to identify the absolute horizontal position and elevation of all the LIG deposition features fo r any future work at these sites. Unfortunately, we were unable to find any existing information about the elevation of these benchmarks, which have not been maintained. Real time sea level Real time sea level was measured on the two islands, so that the elevation data of the LIG deposition features can be referenced to local MLLW The purpose of the sea level measurements is to figure out the water level at high tide and low tide. On Great Inagua Island, water level was surveyed regularly using the level instrument with the

PAGE 48

48 method described above. Specifically, on March 27 and 28, 2013, near the time of the NOAA predicted high and low tides, tide level was surveyed every 15 minutes in order to accurately record the elevation of high and low tide. On San Sa lvador, a few data points of real time water level were measured in the same way as for Great Inagua, but real time sea level was mainly recorded at an interval of 30 seconds with an ONSET water depth recorder from March 31 (10:00 AM LST) to April 4 (2:43 PM LST) 2013 ( Figure 3 5 ). Correlation of elevation data to MLLW The goal is to relate the elevation data of the LIG deposition features to MLLW. This goal can be achieved by contrasting our measured water level with the observed sea level from the neares t tide gauge station. In other words, given the measured elevation difference between LIG deposits (that are on/close to sea level) and present real time water level using the level instrument, elevations of LIG sea level indicators can be linked to MLLW b y relating the real time water level to the observed sea level which is referenced to known tidal datum like MLLW ( Figure 3 6 A ). Because the predicted tides may differ from the actual observed tides due to weather and wave set up it is necessary to determi ne the offset between the predicted and observed tides in the region for the days when we measured sea level position s Ideally, I assumed that the influence of weather will be the same or at least similar among different sites nearby. I calculated the off set between the predicted and observed tides to assess the influence of weather in Great Inagua and San Salvador by using the observational data from the tide gauge at Settlement Point, the Bahamas (26 42.6' N, 78 59.8' W) (Figure 1 6A) This is the site that is used to generate regional tide predictions at the different islands in the Bahamas.

PAGE 49

49 Yet, another problem appears here: the observed water level data from the tide gauge at Settlement Point (provided by the University of Hawaii, Sea Level Center) in March and April, 2013 has never been tied to any tidal datums (such as MLLW). Hence, the observed tides (reported on a floating scale) cannot be compared to the prediction data from NOAA to calculate the weather difference. This complication arose beca use the NOAA tide predictions are given relative to the tidal datum established at a previous tide gauge installed at Settlement Point that was discontinued in 2001. After several years, a new tide gauge was installed at Settlement Point, but was not tied into the position of the previous tide gauge. Since the station has only been in operation since 2002, it is not possible to calculate local tidal datums on the new (floating) scale, which would require ~19 years worth of data. To circumvent this proble m, I employed a method called simultaneous comparisons as described by the Virginia Institution of Marine Sciences, the college of William and Mary ( http://web.vims.edu/physical/research/TCTutorial/ tidaldatum.htm ). This technique can be used to transfer tidal datums between stations without having to obtain measurements for 19 years. I selected the closest tide gauge at Virginia Key, Florida (25 43 .8' N, 80 9.7' W) as the primary station (Figure 1 6A) At the primary tide station NOAA tidal datum elevations for the current National Tide Datum Epoch (NTDE) are known, i.e. the 1983 2001 MLLW position is known. In the method, this MLLW at Virginia Key station is combined with tidal observations for a month (here I use March, 2013) at both Virginia Key station and Settlement Point to estimate MLLW at Settlement Point.

PAGE 50

50 T he whole process is outlined Figure 3 6B and t he steps in this calculation are shown as below: To calculate the Mean Range (MR) at Settlement Point, first, we need to calculate the Monthly Mean Range (MMR) at Virginia station and Settlement Point: MMR VA = MMHHW VA MMLLW VA (3 1 ) MMR SP = MMHHW SP MMLLW SP (3 2 ) where MMHHW is Monthly Mean Higher High Water whereas MMLLW is Monthly Mean Lower Low Water, and the subscript VA and SP refers to Virginia Key and Settlem ent Point station, separately. The range ratio (RR) between the stations is: RR = MMR SP /MMR VA (3 3 ) Thus, Mean Range at Settlement Point is: MR SP = RR*MR VA (3 4 ) Since the two stations are connected by a tidal waterway and it is acceptable to assume that both experience similar monthly deviations from mean tide level, then Mean Tide Level (MTL) at Settlement Point can be calculated as: MTL SP =MTL VA MMTL VA + MMTL SP (3 5 ) where MMTL refers to Monthly Mean Tide Level (MMTL) (here, I used the data from March, 2013, as mentioned above), subscript VA and SP are the same as above. Finally, MLLW at Settlement Point station is: MLLW SP = MTL SP 0.5* MR SP (3 6 ) Then as mentioned above, the influence of weather at Settlement Point station in March and April 2013 is calculated separately as the mean difference b etween tidal observation (by University of Hawaii Sea Level Center) and prediction (by NOAA). By

PAGE 51

51 subtracting the weather induced tide offset from the measured real time sea level in Great Inagua (March, 2013) and San Salvador (April, 2013) the measured se a level data can now be directly related to predicted tide (from NOAA), which is reported relative to known tidal datums (e.g., MLLW). Thus, all our measured elevation data of LIG deposits are connected to local MLLW. The calculations mentioned above were realized using MatLab program languages ( Appendix Object 2 ) Horizontal position calculation The raw data measured with the level instrument that are related to horizontal positions, i.e. d ata that can be used in the calculation of latitude and longitude for all the between each objective and the station, from which the elevation of the objective is surveyed ( equals to the difference of readings between upper an d lower line, shown in Figure 3 3). In local rectangular coordinate, x small and negligible compared with the radius of the Earth, the arc (or curve) between points within the research area could be assumed as straight line. In other words, the extension of the small piece of Earth (where the research area is located) can be considere d as a planar. Thus, it is acceptable to back step the latitude and longitude of each point with one to two points (whose latitude and longitude is known) as reference(s), assuming the expansion length per degree of latitude (or longitude) is known and con stant within a small area (calculat ion is provided in the website: http://www.fas.org/news /reference/calc/degree.html ). On San Salvador Island, the position where the water depth recorder (24.054833 o 74.5355 o ) is installed as well as the position of the

PAGE 52

52 benchmark (24.051303 o 74.533303 o ) can be approximately identified on Google Earth and thus they provide as the reference points (Figure 3 7) On Great Inagua Island, the positi on of th e conch shell (Figure 3 4) is estimated as (21.012286 o 73.699846 o ), and it is used to calculate the latitude and longitude for all the LIG survey points on the island. The purpose of computing the coordinates of the survey points was simply to underst and the relative position of the data points, and to provide an approximate position of the points that could be overlain on the ArcGIS world back map. There are more sophisticated means of determining precise coordinates that were not necessary for the pu rposes of this project.

PAGE 53

53 Figure 3 1 Standard level instrument ( A) Optical level and tripod ( B) Example of staff that stands on the objective The photos are taken Inagua by Jin Li and Dr. Andrea Dutton, respectively. Figure 3 2 Cartoon of how level instrument works. Elevation difference between point B and A is equal to the difference between readings of middle line on point B Fig ure 3 3). A B

PAGE 54

54 Figure 3 3 Cartoon of three readings from optical level That is, upper distance line, middle elevation line and lower distance line. Reading of middle line represents elevations, whereas readings from upper and lower line are related to horizontal distance from the survey station (where the optical level is installed) and the objective. upper distance line and lower distance upper visual line and middle visual line, which is set and known for different level instrument. Based on the geometry relationship in the cartoon, distance

PAGE 55

55 Figure 3 4 Survey r That is, t he conch shell that can be reoccupied. The photo was taken by Dr. Andrea Dutton at

PAGE 56

56 Figure 3 5 Water depth recorder installation at Cockburn Town, San Salvador Island. The photo was taken by Dr. Andrea Dutton at Cockburn Town, San Salvador.

PAGE 57

57 Figure 3 6 Outline of the process of correlating surveyed LIG cora ls to MLLW. (A ) A most intuitive idea to make the correlation: if the measured real time sea level could be correlated to tidal datum MLLW, the surveyed points could be correlated to MLLW, since the surveyed points are related to measured real time sea level in Great Inagu a (GI) and San S alvador (SS). (B ) The process to realize the idea explained in (A ). Boxes from the bottom to the top and from left to right display the correlation processes, with arrows pointing the directions. The converse directions (boxes from the top to the bottom an d from right to left) represent the train of thought in the correlation process as explained in the text. SP refers to Settlement Point, Grand Bahama Island.

PAGE 58

58 Figure 3 7 The benchmark found at Cockburn Town, San Salvador Island This benchmark, BM1, i by the US Department of commerce coast and geodetic survey. It was installed ~20 ft north of the nort h end of a row of steel piling. The elevation at installation was determined as 9.3 ft above m ean sea level Mean sea level at San Salvador was based on 12 months of records spanning July 1958 June 1959. Elevations of other tide planes referred to this datum include mean low water (MLW) at 1.15 m and mean high water at +1.15 m where 0 is MSL. The photo was taken by Dr. Andrea Dutton at Cockburn Town, San Salvador.

PAGE 59

59 CHAPTER 4 RESULTS The measured sea level relative to MLLW in Great Inagua and San Salvador are plotted in Figure 4 1 and Figure 4 2 The elevation Point Great Inagua ( the conch shell that can be reoccupied ) ( Figure 3 4 ) and Cockburn Town, San Salvador (local water surface at 10:04 AM on April 2, 2013) ( Figure 3 5 ) is 0. 80 m and 0. 33 m above local MLLW, respectively. They are added to the raw data f rom elevation survey in Great Inagua and San Salvador Island correspondingly. In Table 4 3 the estimated latitude and longitude for the water depth reco rder and benchmark (references i n San Salvador) and the conch shell that can be reoccupied (reference i n Great Inagua) using Google Earth as well as the length per degree of latitude and longitude (from website http://www.fas.org/news/reference /calc/degree.html ) are listed. The precise elev ation results and approximate horizontal positions are listed in Table 4 1 and 4 2. All the recorded raw data from fieldwork and related calculations a re listed in detail in Appendix. The horizontal distribution of LIG fossil corals and rubble layers in G reat Inagua and San Salvador Island are displayed in Figure 4 3 and 4 4 based on the latitude and longitude speculations. Figure 4 5 and 4 6 are the plots of elevation change with latitude for LIG corals and unconformity in Great Inagua and San Salvador I sland, respectively. Survey data from Great Inagua are reported in Table 4 1 All elevations of in the tables and throughout the text are reported relative to local MLLW. NO. 1 10 points are precise vertical positions and approximate horizontal positions of in situ corals at local high points in the sequence. NO. 11 17 points are elevations and approximate

PAGE 60

60 horizontal positions of the surfaces bounding coral rubble layers. The occurrence of in San Salvador Island. Two patch reefs (Patch Reef 01 an d 02) were surveyed in the field, both of which are mainly composed alternating layers of in situ corals and rubble Great Inagua is: very large well developed corals (mainly p laned off Diploria and Montastrea annularis ), rubble layer (~25 cm thick), massive coral heads (~40 cm thick), beach facies ( Figure 4 7 and 4 6). The top of the rubble layer for Patch Reef 01 and 02 are +2.38 m (NO. 13 in Table 4 1 ) and +2.11 m (NO. 17 in Table 4 1 ). The coral rubble is dominantly composed of branches of A palmate and A cervicornis that would have originated near the reef crest or reef slope and have been transported into the lagoon. These rubble deposits do not appear to be monospecific clusters of branches that have collapsed in situ but rather a mixture of taxa that have been transported and re deposited within the reef during storm events or associated with meter scale sea level changes during the LIG sea level highstand. It is not cl ear if the alternation between coral heads and rubble layer represents successive sea level oscillations (as interpreted by Thompson et al., 2011) because there is no independent evidence of subaerial exposure or sea level change. Because coral rubble can form in different ways, the presence of a coral rubble layer does not uniquely indicate a particular process (e.g., coral rubble layer that sits directly on top of t he wave cut platform (unconformity) ( Figure 4 7 and 4 6). NO. 18 39 are points surveyed on the surface of this platform, including top s of coral heads that have been planed off by waves during the rapid sea level

PAGE 61

61 oscillation in mid LIG period (mentioned in Chapter 2 ) and bottom of rubble layer that sits above the unconformity ( Figure 4 8 and 4 9 ). Above the rubble layer, the highest in situ corals in Reef II are dominantly Diploria in situ coral are two Diploria t hat sits close to each other at +2.43 m above MLLW ( ID#31 and ID# 33, Table 4 1 ) ( Figure 4 1 0 ). Data points at Cockburn Town, San Salvador Island Data from San Salvador Island are reported in Table 4 2 Measurements 1 19 are precise elevations and approxim ate horizontal positions of the 19 in situ fossil corals in San Salvador Island that we identified as potential candidates for the highest in situ coral in the sequence. This was conducted by surveying the surface of the highest coral in each section of th e reef (referred to as the local high point). Measurements 21 28 are precise elevations and approximate horizontal positions of the 8 points that lie on the unconformity surface separating the two reefs. These points include flattened coral heads of Reef I (reef below the unconformity) and bottoms of rubble layer or coral layer of Reef II (reef above the unconformity). The results indicate that the highest in situ coral is a Montastrea that reaches nearly 3 m ( ID#82, +2.98 m) above MLLW ( Figure 4 4 and Figure 4 1 2 ). The highest in situ A palmate is +2.91 m ( ID# 83), which lies not far from that ID#82 Montastrea ( Figure 4 4 and 4 1 1 ). Several in situ corals are found to be close to +3 m in San Salvador, including Diploria, Montastrea and A. palmate T he highest point on the unconformity surface (the in situ A. palmate just below the unconformity, ID#96, Table 4 2 ) is 2.18 m ( Figure 4 1 3 ). And the lowest point on the unconformity surface is 1.46 m (ID#77, in situ A. cervicornis ).

PAGE 62

62 Age elevation data of LIG corals Table 4 4 is the result of the age and elevation (if available) coupled data of fossil from Chen et al. (1991) and Thompson et al. (2011). The age data from Thompson et al. (2011) are the raw data directly from U Th analysis, without an open system model correction. Data used in the project ( Table 4 4 ) are data selected from these two references that meet three screening criteria, including: (1) 232 Th concentration sh ould be less than 0.4 ppb; (2) 238 U concentration should fall into a certain range for living corals. That is, 2.6 3.8 ppm (parts per million) for A palmate and A cervicornis and 1.9 3.6 ppm for other species (Thompson et al., 2011) (3) The calculated value of initial 234 U/ 238 U ratio ( 234 U i ) should fall within the range of 147 8 Andersen, 2006). Only the elevation data from Chen et al. (1991) is available, which has been correlated to MLLW by Dutton and Lambeck (2012). Thompson et al. (2011) did not report the elevation data for the dated samples thus, no elevation data could be directly associated with their U Th age data, except that they estimated the wave cut bench (platform) to be ~0 m above mean sea level ( relativ e to an unspecified tidal Figure 4 1 4 and 4 1 5 exhibits the results of age data of corals in Reef I and Reef II (no elevation implied except that Reef II is stratigraphically above Reef I ) in Great Inagua and San Salv ador, respectively. The timing of the unconformity formation is constrained between 124.8 0.3 and 124.2 1.5 ka in Great Inagua, and between 125.1 1.7 and 124.4 1.6 ka in San Salvador. Thus, the occurrence of the rapid ESL drop p ed in mid LIG is esti mated at ~125 ka.

PAGE 63

63 Error analysis The survey error using the standard level instrument is 1 mm, i.e. 10 3 m. For the measur ement of real time sea level the error could be 0.2 m based on the size of the waves during the sea level measurements. In contrast, measurements of real time sea level at Cockburn Town, San Salvador Island, were performed by the water depth recorder, whose error is 1 mm ( 10 3 m ) The water level at 1 0:04 am (April 2, 2013), which is used as the survey reference in San Salvador Island, was measured when the water was calm and tranquil. In the calculation of MLLW position at Settlement Point Station, no specific error has been assigned to account for u ncertainty associated with simultaneous comparisons technique employed to transfer the tidal datums ( http://web.vims.edu/ physical/research/TCTutorial/tidaldatum.htm ) It could be assumed that the error should be small (at most ~0.1 m) since the two stations (Settlement Point and Virginia Key) are connected by a tidal water way and experience similar monthly deviations from mean tide level. The standard deviation of monthl y weather influence at Settlement Point Station in March and April, 2013 are respectively: W M =0.191 m, W A =0.185 m, refers to April. Thus, the error of the height data for the LI G deposit features (using MLLW as a reference) mainly comes from the calculation of the weather influence in order to link the elevation of the LIG deposit features with MLLW tidal datum, which is less than 0.2

PAGE 64

64 m. Thus, the total surveying/measuring error of LIG corals relative to MLLW ( ) is calculated as the following equation: (4 1) fossil corals to MLLW as represented by different subscripts. the error when estimating the monthly average weather influence in Great Inagua and San Salvador. Thus, for Great Inagua, the surveying/measuring error ( h G ) equals to: For San Salvador, the surveying/measuring error ( h S ) equals to:

PAGE 65

65 Table 4 1 Surveyed results: elevations of LIG sea level indicators in Great Inagua NO. 1 ID. 2 Description 3 Elevation 4 Latitude Longitude 1 21 Top of in situ coral (Dp) Reef II 2.30 21.01287182 73.70080274 2 27 Top of in situ coral (Dp) Reef II 2.35 21.01322144 73.70099011 3 28 Top of in situ coral (Dp) Reef II 2.34 21.01325104 73.70106610 4 29 Top of in situ coral (Dp) Reef II 2.37 21.01320829 73.70118250 5 30 Top of in situ coral (Dp) Reef II 2.35 21.01316570 73.70115749 6 31 Top of in situ coral (Dp) Reef II 2.43 21.01314301 73.70114316 7 32 Top of in situ coral (Dp) Reef II 2.36 21.01310222 73.70098689 8 33 Top of in situ coral (Dp) Reef II 2.43 21.01309237 73.70098545 9 34 Top of in situ coral (Dp) Reef II 2.31 21.01315637 73.70101605 10 46 Top of in situ coral (Ma) Reef II 1.94 21.01199754 73.69987801 11 14 Top of rubble layer Reef II 1.75 21.01300445 73.70093600 12 24 Top of rubble layer Reef II 1.98 21.01301824 73.70103455 13 25 Top of rubble layer Reef II 2.38 21.01301824 73.70103455 14 54 Top of rubble layer Reef II 2.06 21.01314956 73.70110999 15 56 Top of rubble layer Reef II 1.99 21.01308995 73.70112746 16 57 Top of rubble layer Reef II 1.94 21.01312505 73.70102698 17 58 Top of rubble layer Reef II 2.11 21.01328853 73.70101037 18 1 Unconformity Reef I 0.71 21.01264528 73.70029948 19 2 Unconformity Reef I 0.68 21.01265173 73.70031602 20 3 Unconformity Reef I 0.75 21.01266564 73.70035147 21 5 Unconformity Reef I 0.89 21.01269350 73.70028363 22 6 Unconformity Reef I 1.05 21.01273224 73.70023011 23 7 Unconformity Reef I 1.15 21.01277253 73.70018593 24 12 Unconformity Reef I 1.43 21.01297661 73.70082117 25 13 Unconformity Reef I 1.33 21.01300658 73.70091579 26 17 Unconformity Reef I 0.73 21.01262809 73.70032796 27 18 Unconformity Reef I 0.69 21.01262280 73.70030411 28 19 Unconformity Reef I 1.17 21.01276596 73.70020369 29 48 Unconformity Reef I 1.07 21.01200054 73.69988731 30 49 Unconformity Reef I 1.74 21.01300291 73.70092645 31 50 Unconformity Reef I 1.38 21.01307274 73.70085391 32 59 Unconformity Reef I 1.13 21.01216450 73.69988135 33 62 Unconformity Reef I 1.46 21.01099782 73.69930629 34 63 Unconformity Reef I 1.57 21.01089507 73.69933456 35 64 Unconformity Reef I 1.55 21.01079422 73.69934557 36 65 Unconformity Reef I 1.50 21.01212762 73.69997658 37 15 Unconformity Reef I 1.63 21.01300409 73.70093253 38 10 Unconformity Reef I 1.22 21.01284766 73.70054382 39 22 Unconformity Reef I 1.25 21.01301192 73.70102517 Note (same as below): 1. NO. of data points in this table ; 2. ID. in Supplementary Data ( .xlsx file ) in Appendix ; 3. Coral species: (Ap) Acropora palmata (Ac) Acropora cervicornis (Ma) Montastraea annularis (Dp) Diploria (D. clivosa or D. strigosa) The abbreviations for coral species are the same throughout the whole document and will not be explained repeatedly below. 4. Elevation above MLLW. All the elevations are relative to MLLW

PAGE 66

66 Table 4 2 Surveyed results: elevations of LIG sea level indicators in San Salvador. NO. ID Description Elevation Latitude Longitude 1 72 Top of in situ coral (Ap) Reef II 2.76 24.05249150 74.53514853 2 83 Top of in situ coral (Ap) Reef II 2.91 24.05176645 74.53429410 3 98 Top of in situ coral (Ap) Reef II 2.80 24.05292772 74.53549060 4 69 Top of in situ coral (Dp) Reef II 2.53 24.05322721 74.53558815 5 70 Top of in situ coral (Dp) Reef II 2.74 24.05269397 74.53535266 6 71 Top of in situ coral (Dp) Reef II 2.81 24.05258127 74.53524358 7 73 Top of in situ coral (Dp) Reef II 2.50 24.05238714 74.53500284 8 74 Top of in situ coral (Dp) Reef II 2.53 24.05218102 74.53474405 9 78 Top of in situ coral (Dp) Reef II 2.93 24.05202482 74.53459143 10 81 Top of in situ coral (Dp) Reef II 2.82 24.05195952 74.53448721 11 87 Top of in situ coral (Dp) Reef II 2.69 24.05154005 74.53384997 12 91 Top of in situ coral (Dp) Reef II 2.90 24.05140464 74.53356332 13 82 Top of in situ coral (Ma) Reef II 2.98 24.05192421 74.53447739 14 85 Top of in situ coral (Ma) Reef II 2.09 24.05176220 74.53415760 15 86 Top of in situ coral (Ma) Reef II 2.76 24.05163352 74.53395743 16 88 Top of in situ coral (Ma) Reef II 2.64 24.05148396 74.53376700 17 89 Top of in situ coral (Ma) Reef II 2.90 24.05153634 74.53374539 18 90 Top of in situ coral (Ma) Reef II 2.80 24.05141012 74.53366112 19 92 Top of in situ coral (Ma) Reef II 2.97 24.05133372 74.53345214 20 97 Top of rubble layer Reef II 2.76 24.05209943 74.53462341 21 75 Unconformity Reef I 1.74 24.05212373 74.53469613 22 76 Unconformity Reef I 1.85 24.05208940 74.53464577 23 80 Unconformity Reef I 2.53 24.05196845 74.53449407 24 95 Unconformity Reef I 1.62 24.05122809 74.53293032 25 93 Unconformity Reef I 1.67 24.05143882 74.53375223 26 79 Unconformity Reef I 1.49 24.05196783 74.53446688 27 96 Unconformity Reef I 2.18 24.05209943 74.53462341 28 77 Unconformity Reef I 1.46 24.05209889 74.53464687 Table 4 3 References for horizontal position Reference Latitude ( o ) Longitude ( o ) Length of a Degree of Latitude (m) Length of a Degree of Longitude (m) Water depth recorder 24.054833 74.5355 110759 101709 Benchmark 24.051303 74.533303 110759 101711 The conch shell 21.012286 73.699846 110717 103962

PAGE 67

67 Table 4 4 Elevation age data of fossil corals NO 1 Location 2 Sample ID 3 Spe cies Age 4 Age error Elevation (m above MHT) 5 Elevation (m above MLLW) 6 Elevation error 7 1 GI (DP) 28 (87GI 4) Ap 130.7 1.4 0 0.7 1.5 2 GI (DP) 29 (86GI 4) Ma 125.3 2.1 0 0.7 1.5 3 GI (DP) 30 (87GI 1) Ma 128.8 1.2 0 0.7 1.5 4 GI (DP) 31 (87GI 2) Ds 125.8 1.7 0 0.7 1.5 5 GI (DP) 32 (87GI 3) Ma 124.2 1.5 0.5 1.2 1.5 6 GI (DP) 35 (86GI 7) Dc 123.7 1.5 1.5 2.2 1.5 7 GI (DP) 36 (87GI 6) Ma 123.2 1.6 1 1.7 1.5 8 GI (DP) 37 (87GI 7) Ds 124.1 1.1 0.5 1.2 1.5 9 GI (DP) GI 4 (GI A) Ma 119.5 0.5 10 GI (DP) GI 7 (GI A) Ma 126.4 0.5 11 GI (DP) GI 10 (GI A) Ma 122.8 0.4 12 GI (DP) GI 13 (GI A) Ma 125.2 0.3 13 GI (DP) GI 16 (GI B) Ma 125.7 0.8 14 GI (DP) GI 19 (GI D) Ma 123.9 0.7 15 GI (DP) GI 24 (GI E) Ma 127.5 0.3 16 GI (DP) GI 26 (GI E) Ds 126.6 0.3 17 GI (DP) GI 31 (GI E) Dc 128.2 0.3 18 GI (DP) GI 34 (GI E) Ma 124.8 0.3 19 GI (DP) GI 42 (GI F) Ap 125.7 0.6 20 GI (DP) GI 43 (GI F) Ap 125.5 0.8 21 GI (DP) GI 49 (GI H) Ap 125 0.8 22 GI (DP) GI 50 (GI H) Ap 127.3 0.6 23 GI (DP) GI 52 (GI H) Ma 126.2 0.6 24 GI (DP) GI 63 (GI J) Ma 126.1 0.4 25 SS (CT) 1 (6/86#5) Ma 120.2 1.4 2.25 2.6 1.5 26 SS (CT) 2 (87CT 1 S) Ma 121 1.9 2.25 2.6 1.5 27 SS (CT) 4 (87CT 3 S) Dc 123.5 2.1 1.5 1.9 1.5 28 SS (CT) 5 (87CT 4 S) Ma 125.1 1.7 2.25 2.6 1.5 29 SS (CT) 6 (88CT 1) Ma 122.7 1 2.5 2.9 1.5 31 SS (CT) 7 (87CT 5 S) Dc 130.3 2.3 1.5 1.9 1.5 33 SS (CT) 10 (87CT 7 S) Ds 121.4 1.6 2 2.4 1.5 34 SS (CT) 11 (87CT 8) Ap 123 1.4 1.5 1.9 1.5 35 SS (CT) 12 (87CT 9) Ap 124.4 1.6 1.75 2.1 1.5 36 SS (CT) 15 (88CT 3) Ap 124.2 1.7 2.25 2.6 1.5 37 SS (CT) 16 (88CT 12) Ap 125.6 1.7 1.5 1.9 1.5 38 SS (CT) 18 (88CT 14 S) Ap 127.9 2.1 1.5 1.9 1.5 39 SS (CT) 19 (87CT 15 L) Ds 126.9 1.4 1.5 1.9 1.5 40 SS (CT) 20 (87CT 16A S) Ma 126.3 2.9 1.25 1.6 1.5 41 SS (CT) 21 (88CT 8) Ma 128.9 1.1 0.75 1.1 1.5 42 SS (CT) SS 31 Ap 119.6 0.5 43 SS (CT) SS 32 Ds 119.9 0.7 44 SS (CT) SS 33 Ap 123.2 1.3 45 SS (CT) SS 34 Dc 121.9 0.2 46 SS (CT) SS 35 Dc 122.5 0.2 47 SS (CT) SS 38 Ap 122.3 0.9 48 SS (CT) SS 39 Ma 125.1 1.1

PAGE 68

68 Table 4 4 Continued NO 1 Location 2 Sample ID 3 Spe cies Age 4 Age error Elevation (m above MHT) 5 Elevation (m above MLLW) 6 Elevation error 7 49 SS (CT) SS 40 Ds 124.6 1 50 SS (CT) SS 42 Ap 123.4 0.3 51 SS (CT) SS 07 1 Ma 122.4 0.2 52 SS (CT) SS 07 3 Ma 121.1 0.3 53 SS (CT) SS 07 4 Ma 121.5 0.3 Note: 1. N O. compiled in this project, NO#1 8 and #25 41 data points are data from text ). NO#9 24 data points are data points (raw results as described in the text ) from Tho mpson et al., 2011, which meet the same three screening criteria. Cockburn Town, San Salvador Island. 3. Sample ID in the original references. 4. U Th ag es after sc reening (unit: ka). 5. Elevation (m) of corals reported in the original research by Chen et al. (1991). 6. Elevations after being corrected by Dutton and Lambeck (2012), which are correlated to local tidal datum (MLLW). 7 Errors of elevation data estimat ed by Dutton and Lambeck (2012), based on description, or lack thereof, of elevation measurement protocol; minimum value assigned defined as 1 meter.

PAGE 69

69 Figure 4 1 Sea level relative to MLLW in Great Inagua (M ar 27 and 28, 2013 ) Sea level was measured using the level instrument relative to the conch shell in the field, which was calculated to be 0.8 m above MLLW. This height offset was added to all the measured sea level as shown in the figure. Figure 4 2 Measured sea level s in San Salvador ( Ap r. 2 and 3, 2013 ) with the water depth recorder every half a minute in the field.

PAGE 70

70 Figure 4 3 Horizontal distribution of surveyed points in Great Inagua Island. Red star: the location of the conch shell that can be reoccupied ( Figure 3 4 ). The abbreviations of coral species are the same throughout the thesis as mentioned above : Dp, Diploria sp. (either D. strigosa or D. clivosa ); Ma, M. annularis ; Ac, A. cervicornis The ID# of the highest in situ corals (Reef II) that are used to deduce LIG pea k sea level (as well as sea level just before the rapid sea level regress in mid LIG period) in Chapter 5 are denoted in the figure.

PAGE 71

71 Figure 4 4 Horizontal distribution of surveyed points in San Salvador Island. Red asterisk represents the water depth re corder ( Figure 3 5 ) and the red Star is the location of the benchmark ( Figure 3 7 ). The ID# of the highest in situ corals (Reef II) that are used to deduce LIG peak sea level in Chapter 5 are denoted in the figure.

PAGE 72

72 Figure 4 5 Surveyed elevations above MLLW for in situ coral heads in Great Inagua Island. Following the terminology of Thompson et al. (2011) Reef I is early reef growth stage and Reef II is later reef growth stage. They are separated by the unconformity: the forme r sits below the unconformity whereas the latter and displayed as an extensive wave cut bench (or platform). This platform inclines a little towards the ocean, thus the elevati on of unconformity surface decreases seaward. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 21.0105 21.0110 21.0115 21.0120 21.0125 21.0130 21.0135 Elevation above MLLW (m) Latitude (degree) Dp (Reef II) Ma (Reef II) Top of rubble layer (Reef II) Unconformity (Reef I) S N

PAGE 73

73 Figure 4 6 Surveyed elevations for in situ coral heads in San Salvador Island. The most prominent unconformity has an undulating surface, and there appear to be multiple unconformities visible in sections of the reef, particularly towards the northern end of the preserved reef section. Note that the scale of y axis is 0 3.5 m in this figure, rather than 0 3.0 m in Figure 4 5. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 24.0510 24.0515 24.0520 24.0525 24.0530 24.0535 Elevation above MLLW (m) Latitude (degree) Dp (Reef II) Ma (Reef II) Ap (Reef II) Ap (Reef I) (Lower unconformity) Ac (Reef I) (Lower unconformity) Unconformity (lower) Unconformity (upper)

PAGE 74

74 Figure 4 7 Thicker rubble layer that lies above the platform in Great Inagua. Above the platform, sedimentary sequence from bottom to top is: rubble layer, coral heads and beach facies. The bottom of the rubble is surveyed at 1.63 m (ID#15, Table 4 1 ). It is the highest point on the unconformity. The figure is patch reef 02 ( PR02) Point, Great Inagua. Figure 4 8 Overview of the extensive plat form in the southeast of the surveyed reef in Great Inagua. Great Inagua.

PAGE 75

75 Fig ure 4 9 Planed off Diploria (Reef I) on the platform in Great Inagua The photo was Figure 4 1 0 Highest in situ Diploria head (Reef II) that lies above the rubble layer in Great Inagua. The Great Inagua.

PAGE 76

76 Figure 4 1 1 Highest in situ A. palmate (Reef II) (ID#83) in San Salvador Island. The photo was taken by Dr. Andrea Dutton at Cockburn Town, San Salvador.

PAGE 77

77 Figure 4 1 2 Two highest corals at Cockburn Town, San Salvador Island (A) M. annularis (ID# 82, 2.98 m, and also the highest of all in situ corals of any species that we surveyed at the Cockburn Town reef, San Salvador Island), (B) Diploria sp. (ID#79, 2.93 m), (C) A. palmate rubble. The photo was taken by Dr. Andrea Dutton at Cockburn Town, San Salvador.

PAGE 78

78 Figure 4 1 3 Highest in situ A. palmate below the unconformi ty ( ID#96) that reaches a maximum height of 2.18 m. The dashed line follows the unconformity surface approximat ely. The A. palmate mound is composed of in situ corals and rubble. The thickness of this A. palmate outcrop is ~1.5 m. The photo was taken by Dr. Andrea Dutton at Cockburn Town, San Salvador.

PAGE 79

79 Figure 4 1 4 Age data of corals in Reef I and Reef II in San Salvador No elevation scale is implied in the plot. The dashed line represents unconformity that separates the two coral growth stages. Data is from Chen et al., 1991, Curran et al., 1989 and Thompson et al ., 2011 (raw data after screening process). Figure 4 1 5 Age data of corals in Ree f I and Reef II in Great Inagua. No elevation scale is implied in the plot. The dashed line represents unconformity that separates the two coral growth stages. Data is from Chen et al. (1991), Curran et al. (1989) and Thompson et al., 2011 (raw data after screening process).

PAGE 80

80 CHAPTER 5 DISCUSSION At present, the forward glacio hydro isostatic model does not produce RSL predictions that agree with the existing data in the circum Caribbean region. Hence, the overarching goal of this project is to modify the ice model used in the glacio hydro isostat ic model so that the model predictions have better agreement with the observational data. To achieve this, the surveying undertaken of the fossil reefs in the Bahamas will help to determine: (1) the difference between peak sea levels in San Salvador and G reat Inagua Island during the LIG period, and (2) the rate of sea level change at each island (denoted as and in Figure 5 1 ). The combination of these results will help to identify the changes required to the Laurentide ice sheet during the last two gl acial maxima (MIS 6 and MIS 2) and revise the ice model in Dutton and Lambeck (2012). Based on personal observation and prior reports, three dominant coral genera are present in the LIG fossil reefs on Great Inagua and San Salvador Island : A palmate (reef builder in the crest zone ) with subordinate A cervicornis as well as Diploria sp ( D strigosa and D. clivosa ) and M annularis that are typically associated with patch reefs in the lagoon ( Figure 5 2 ). In order to translate surveyed elevations of the surfaces of these coral heads into past sea level position, it is first necessary to have a better understanding of the likely water depth during growth. Some generalizations can be made for each taxon, but are still associated with fairly large uncertainties given consideration of the full depth range of each taxon. Additional information regarding paleodepth can be inferred from other

PAGE 81

81 indicators in the reef facies, including associated fauna and sedimentar y features to improve the paleodepth estimate. Growth rates of corals T he dominant reef crest building coral in this region A. palmate can have a quite rapid growth rate of ~10 cm per year, i.e. ( ~100 m/ka ) which is more than 10 times that of other cora l species such as Montastrea sp. and Diploria sp (Lewis et al., 1968). Hudson (1981) found that the average growth rate of M. annularis in Floridian nearshore patch reef areas is ~8.2 mm per year, i.e. ( ~8.2 m/ka ) and Dodge and Vaisnys (1975) estimate d t he annual growth rate of D. strigosa in Bermuda water to be ~3.3 mm / a ( 3.3 m/ka ) T hese growth rates imply that they would be able to keep up with rapid sea level rise i.e. at least several meters per 1000 yrs. Yet, the observation that these taxa are ca pable of such high skeletal growth rates still implies that individual corals would be capable of keeping up with rapid sea level rise events in the mid LIG period ~2.6 m/ka as suggested by Thompson et al. (2011) After reaching maximum vertical extent co rals can develop laterally to form a reef flat under when sea level is stable and growth conditions are ideal Factors that promote optimal growth conditions include particular ranges of light intensity, temperature and salinity Observational data of the actual reef accretion rates of Holocene reefs in the Caribbean suggest that overall reef accretion is much slower than growth rates of individual corals such as A. palmate and M. annularis mentioned above ( Shinn et al., 1989 ). U pward development of th e reef can be impeded by adverse changes of the environment in lagoonal waters, such as fluctuating temperatures and salinities. Storms and hurricane can be another cause of slower growth rate of coral reef than individual corals. Yet, A. palmate has such a high growth rate that they are able to offset the

PAGE 82

82 breaks by storm in less than 5 years (Shinn et al., 1989). I t is not uncommon to find A. palmate dying due to lack of accommodation space due to their rapid growth rate (Shinn et al. 1989). Another facto r that may impair reef development is disease. In summary the observed fossil coral taxa in the Bahamas should be able to reach the possible peak point of their growth range in vertical direction during the LIG period (from ~1 31 to 119 ka, Chen et al., 19 91; Thompson et al., 2011; Dutton and Lambeck, 2012). Peak sea level Position The elevation of the highest in situ coral combined with its paleodepth under water surface is used to estimate the position of peak sea level in the LIG period in both islands. Lighty et al (1982) considered A. palmate to be a reliable sea level indicator in western Atlantic because (1) they have a limited growth depth that is restricted to within 5 meters (~1 5 m) below sea surface, (2) they are resistant to both post depositi onal transport and (3) compaction due to their massive size. Chen et al (1991) report ed that A. palmate as the reef crest building coral, can commonly grow to mean low tide level in the Bahamas with individual colonies reaching a 3 4 m in vertical extent. In contrast, M. annularis are often predominant between water depth of 7 and 25 m in lagoon region, although it also has been found to live close to water surface (Humann, 1993; Chen et al., 1991). Diploria usually live in shallow wat ers between 1 and 30 m, although it could be found in deep water up to 47 m. They could inhabit in many shallow water environment, including open reefs, lagoons in Turtle grass beds ( Thalassia testudinum ) and even on mangrove roots ( Humann, 1993; Fricke 1 985)

PAGE 83

83 Cockburn Town Reef, San Salvador Island At the Cockburn Town reef, San Salvador Island, the highest in situ A. palmate that was surveyed is 2.91 m (ID#83 A. palmate as referred in the following discussion) ( Figure 4 1 1 ). This individual ID#83 A. pal mate is found surrounded by corals of other species like Montastrea and Diploria which usually live in the back reef within the lagoon Only the top surface of the coral is visible which exposes the thin upper branches of the A. palmate colony. In contrast, the second highest in situ A. palmate surveyed (2.80 m, ID #98 A. palmate as referred in the following discussion ) has very thick and large branches and is surrounded by numerous large A. palmate corals and branches that appear to ha ve collapsed in situ ( Figure 5 3 A and B ). This outcrop of A. palmate cora ls and rubble is interpreted as representing part of the reef crest during the LIG period. These interpretations of the facies context for these two highest A. palmate in our survey i s consistent with those of Curran et al. (1988) according to their descriptions and paleogeographic map of the former reef ( Figure 5 4 ). In the context of their reconstruction, the A. palmate reef crest sits to the north, within the defined reef crest zone The A. palmate that has a slightly higher surveyed elevation sits in the well developed patch reefs of the lagoon, close to the southeastern end of the large bank/barrier reef. The location of these A. palmate corals in different reef zones, and the di ffering observations of the size and preservation of the colonies lead to slightly different interpretations of the paleodepth of these corals. The sizes of exposed branches of the ID#98 A. palmate and associated branches in the reef crest zone are substan tial. At the surface of this outcrop, the trunks of the A. palmate are clearly exposed indicating that

PAGE 84

84 the top probably fell off. The collapse of A. palmate corals is a typical observation for both fossil and modern corals in the Caribbean region. In the modern reef, some broken off branches of A. palmate that were still alive were observed on the lagoon floor beside a colony of A. palmate The height of the ID#98 A. palmate reef crest is ~1.5 m from the base, which sits on a sub horizontal surface, to th e top of the exposed, truncated trunk of the corals ( Figure 5 3 A and B ). Given the large size of the A. palmate in this outcrop and given that A. palmate can grow to 3 4 m high, we interpret that the actual height of this reef crest was probably some 1.5 t o 2.5 m higher than the surface we surveyed, which would bring the elevation of the reef crest up to 4.3 to 5.3 m. If we assume that these A. palmate specimens in the reef crest reached their maximum height (up to MLLW), this implies that mean sea level r eached at least 4.3 to 5.3 m above present mean sea level (the uncertainty in the estimation is discussed in the fifth part of Chapter 5 ) In contrast, we interpret that the ID#83 A. palmate in the patch reef had a slightly greater paleodepth or that it r ecords sea level position at a different time within the LIG sea level highstand, since the thin branches of the top of the coral are exposed at the surface, which sits at 2.91 m, a couple meters below the reef crest colonies described above. At the Cockbu rn Town reef, the highest in situ Montastrea (ID#82, 2.98 m) and the highest in situ Diploria (ID# 78, 2.93 m) are located very close to the ID #98 A. palmate (2.91 m) ( Figure 4 11 and 4 1 2). The dominant taxa in this portion of the fossil reef are very large colonies of Montastrea sp. and Diploria sp ., as well as some isolated occurrences of A. palmate (but not dominant). Given the various species, the outcrop that is composed of these corals is most likely in the back reef within the lagoon zone

PAGE 85

85 ( Figure 5 2 ). All the fossil coral heads in this area are very large (>1 m in diameter) and are closely packed so that there is very little accommodation space for coral growth. The surface elevation of the corals is very similar (ranging from about 2.65 to nearl y 3 m) as shown in Figure 4 1 2 and 4 6 Thus, these corals are interpreted as having effectively filled their accommodation space in a well developed back reef zone. If peak sea level were 4.3 to 5.3 m higher at this site in the LIG period based on the ele vation of highest A. palmate in reef crest, the highest in situ Diploria sp ( ID#78, 2.93 m) was ~1.5 to 2.5 m below MLLW. The coral taxa that are preserved in situ predomina ntly Diploria and Montastrea Corals that occur in the highest part are mainly Diploria and only one potentially highest in situ M. annularis was surveyed as shown in Table 5 2 The elevations of in situ Diploria sp mainly concentrate between 2.3 m and 2 .4 m, the highest two are both 2.43 m above MLLW ( Figure 4 1 0 and 4 5 ). As with the also closely packed, indicating that the corals filled the available accommodation sp ace ( Figure 4 7 ). In Reef II, however, the growth is sparser but still includes some large (~1 2 m diameter) coral heads. The nature of the unconformity on Great Inagua Island is also quite different. The contact between Reef I and Reef II is sharp and p lanar, and has previously been interpreted as a wave cut bench that formed during an ephemeral drop in sea level (Curran et al., 1989; Chen et al., 1991; Thompson et al., 2011). The stratigraphy of the LIG reef is easier to discern at Great Inagua, in par t because after the termination of growth in Reef I, there are isolated patch reefs that afford easy access to vertical

PAGE 86

86 sections of the outcrop. As shown in Figure 4 1 0, the generalized stratigraphy of the patch reefs sitting on top of the unconformity is : thick rubble layer ( ~25 cm ), massive coral heads (~4 0 cm thick ) and beach facies (from the bottom to the top). As mentioned before, the highest corals in the sequence are mostly Diploria On the landward section of the patch reef, these Diploria are overlain by prograding beach sands, similar to the final progradational sands observed at San Salvador ( Figure 5 5 ). Peak Sea Level As mentioned above, in contrary to Reef II in Cockburn Town, San Salvador where there are various coral species (e.g. A palmate Montastrea and Diploria ), highest corals are mainly Diploria compare the difference between peak levels of same species (here, only Diploria ) to estimate gradient in peak sea level, although A. palmate in reef crest could provide a better constraints on the absolute position of sea level in each site. The difference between elevation of the highest in situ Diploria on San Salvador (ID#78, 2.93 m) and Great Ina gua Island (ID# 6 and ID# 8, 2.43 m) is 0.5 m. This reflects a possible difference of 0.5 m in the LIG peak sea level between the two islands, assuming these corals grew at similar paleodepths. Since there are not so many M. annularis as Diploria sp. in highest part of Reef II in the patch r eefs that we surveyed in Great Inagua, it is not possible to provide a meaningful comparison of sea level between the two islands using this taxon. In summary, the spatial gradient of peak sea level in the LIG period between San Salv ador and Great Inagua I sland is 0.5 m as reflected by the elevation difference of the highest Diploria in Reef II in the two islands. And two presumptions are made here: (1) the two Diploria have equal paleodepth below sea surface and (2) both corals were living during the peak of the LIG sea level highstand. Based on the reef

PAGE 87

87 environment as described before, the two corals were both likely to have lived in shallow water in the back reef. Given the appearance of A. palmate coral heads near the ID#78 Diploria in San Salvador, the paleodepth of that Diploria should be within 5 m. With respect to the assumption (2), considering that ESL could have reached the peak point at the close of LIG period, corals should have grown to the highest points at that time (Chen et al., 1991; Stirlin g et al., 1995; Neumann and Hearty 1996; Thompson et al., 2011). In that case, peak sea level in Great Inagua should be ~0.5 m lo wer than that in San Salvador Since peak sea level in San Salvador is estimated as 4.3 5.3 m from ID#98 A. palmate in reef cr est, LIG peak sea level should be ~3.8 4.8 m in Great Inagua. The estimation s of LIG peak sea level in Great Inagua and San Salvador Island are displayed in Table 5 3 and the uncertainty involved in the estimation is discussed in the fifth part of Chapter 5 Unconformity surface (Reef I) As mentioned before, an unconformity has been observed in both Great Inagua and San Salvador Island, although it is displayed as an extensive wave cut platform as wide as hundreds of meters in Great Inagua Island, whereas i n San Salvador Island, the unconformity is more undulate and discontinuous. This unconformity is explained by some researchers as the evidence of a regressive and a transgressive sea level change that occurred rapidly in mid LIG period all over the globe ( Chen et al., 1991; White et al., 1998 ; Thompson et al., 2011). Petrologic evidence of grainstone from nearshore reefs provides independent confirmations for this scenario of regressive transgressive sea level change. White et al. (1997) found that grainsto commonly has a sequence of marine cements, followed by isopachous non marine cements, which is in some cases is superimposed by a final generation of marine high

PAGE 88

88 Mg calcite and aragonite This illustrates a change sequ ence starting from early marine environment (high sea level), changing to non marine environment (sea level drop) and finally returning to marine environment (sea level rise).The unconformity was formed during the sea regress and separated the two coral gr owth stages: R eef I (early) below the unconformity, and R eef II (late) above the unconformity following the nomenclature in Thompson et al., 2011 This regressive transgressive cycle was constrained to begin at ~125.1 ka and ended at ~124.2 ka, lasting for about several hundred years with an overall consideration of all the U Th ages of youngest coral above the unconformity and oldest coral below the unconformity ( Figure 4 1 4 and 4 1 5 ). Sea level during Reef I growth The wave is mainly composed of planed off coral heads (Reef I). The elevation of platform is surveyed to change gradually from ~1.1 to ~1.6 m over hundreds of meters, in both the northwest (the promontory where most of the potential corals of Reef II) and southeas t of the surveyed outcrop ( Figure 4 5 ). The highest point of Reef I is surveyed from the bottom of coral rubble that sits on the platform at 1.63 m (ID#15, Fig ure 4 7 ). During the rapid sea level fall and the subsequent lowstand of several hundred years, c orals from Reef I were eroded by wave action (and/or biological activities). It is not clear how high the reef reached before the ephemeral sea level fall, but based on the size of the planed off coral heads, we estimate that sea level was at least 0.5 m a bove the present unconformity. This puts the elevation of sea level at +2.1 3 m ( Table 5 3 ) At Cockburn Town, San Salvador Island, no extensive platform was observed. Rather, the unconformity surface is undulating. In the field, the highest in situ A. palmate

PAGE 89

89 on the unconformity (Reef I) was surveyed at 2.18 m ( Figure 4 1 3 ID#96). Based on the size and height of the A. palmate mound which has truncated branches at the top of the outcrop, an offset height of ~0.5 m is added to its elevation and thus, th e ID#96 A. palmate is 2.68 m above MLLW ( Table 5 3 ) Rate of sea level change Based on the estimations of sea level position just before the rapid oscillation and at the close of LIG period when sea level reached the peak point, the rate of sea level change ( v ) is estimated as: ( 5 1) peak sea level difference between peak sea level that rose back after the rapid drop in the mid LIG period. Thus, the sea level change rate ranges from 0. 29 to 0.4 8 m/ka to in San Salvador, and 0.30 0.48 m/ka in Great Inagua. Uncertainty in the estimation of sea level In addition to the surveying error of fossil corals ( 0.29 m for Great Inagua and 0.21 m for San Salvador as calculated in the fourth part of Chapter 4 ), the most significant uncertainty in the estimation of LIG peak sea level as well as sea level just before the rapid sea level drop in mid LIG period comes from the uncertainty of paleodepth when coral was alive below water surface as well as the heights of corals that have been truncated off. These uncertainties dominate the total error when translating the elevation of fossil corals into the elevation of LIG sea level and are

PAGE 90

90 Table 5 3 The paleodepth uncertainty is likely to be 1 m in terms of pinpointing th e absolute position of peak sea level in San Salvador (4.3 5.3 m) due to the uncertain height offset of ID#98 A. palmate (1.5 2.5 m). When using the height difference of the highest in situ Diploria in the two islands (0.5 m) to estimate LIG peak sea leve l in Great Inagua (3.8 4.8 m) it is assumed that the paleodepth uncertainty should be the same for the same species. The uncertainty here could be totally 2 m by adding a 1 m paleodepth uncertainty of Diploria to the uncertainty of 1 m height offset of ID#98 A. palmate thus, the range of LIG peak sea level in Great Inagua is expanded to 3.3 5.3 m. When estimating sea level just before the rapi d sea level drop ~125.1 ka ago a minimum offset height of 0.5 m was added to the elevation of the ID#96 A. palmate (highest in situ coral below the unconformity) to estimate the lower limit of paleo sea level as 2.68 m in San Salvador And the paleodepth uncertainty is at most 1.5 m so that the maximum estimation of sea level 125.1 ka ago was 3.68 m in San Salvador. In a similar way, in Great Inagua, a minimum offset height of 0.5 m was added to the highest coral just below the unconformity ( ID#15 Diploria ) to estimate the lower limit of sea level position as 2.13 m, and the paleodepth uncertainty i s at most 2 m, making the upper limit of sea level 125.1 ka ago to be ~3.63 m. All the estimations of sea level and related uncertainties are displayed in Table 5 3 as mentioned before. Comparison of the observation with the model prediction The comparison between the model prediction and field observation for LIG sea level at the end and middle of LIG period is shown in Figure 5 6 The rate of sea level change seems much slo wer than predicted by the forward model in both islands: 0.29 0.4 8 m/ka versus 1 m/ ka in San Salvador, and 0.3 0.48 versus 0.73 m/ka. S o this would

PAGE 91

91 require a change in the ice sheet in MIS 6, that is, Laurentide ice sheet in MIS 6 should be smaller. This change will also affect the final position of the peak sea level, which is also lowe r than the model predict ions: 4.3 5.3 m versus 11.26 m in San Salvador, and 3.3 5.3 m versus 8.33 m in Great Inagua If this change is not enough then the size of the Laurentide ice sheet in MIS 2 would have to be changed, that is, the size needs to be ma de smaller in order to lower the position of the peak sea level model prediction.

PAGE 92

92 Table 5 1 Highest in situ corals in San Salvador ID# Species Elevation (m) 83 A. palmate 2.91 98 A. palmate 2.80 78 Diploria sp. 2.93 82 M. annularis 2.98 Table 5 2 Comparison of highest in situ corals of the same species in San Salvador and Great Inagua Species ID# Site Elevation (m) Diploria sp 78 San Salvador 2.93 31, 33 Great Inagua 2.43 M. annularis 82 San Salvador 2.98 46 Great Inagua 1.94 Table 5 3 Estimations for LIG sea level in the two islands San Salvador Great Inagua Age (ka) Elevation (m) Error (m) Elevation (m) Error (m) Survey ( ) Paleodepth Survey ( ) Paleodepth 119.5 4.3 5.3 0.21 1 3. 3 5.3 m 0.29 2 125.1 2.68 3.68 0.21 1 .5 2.13 3.63 0.29 2 Note: The age data are from Chen et al., 1991 and raw data from Thompson et al., 2011 after screening process as described above. Errors include two parts: survey/measuring error and paleodepth error, as discussed in the text

PAGE 93

93 Figure 5 1 Model predictions of LIG RSL in Great Inagua and San Salvador In the model ESL is set equal to zero (present sea level) for the duration of the LIG period in the forward model ( data of predicted LIG sea level is from Dutton and Lambeck, 2012 same as in Figure 1 3 ). A significant difference has been displayed in the magnitude of RSL between the two islands as well as ESL. As marked in the figure, the goal in the project is to revise the model by calculating the gradient between peak sea levels in the two islands and the rates of sea level change in the two islands.

PAGE 94

94 Figure 5 2 Generalized platform reef zonation cross section. Reef crest zone is dominated by A. palmate whereas the back reef and lagoon z one are usually occupied by more various coral species such as Diploria sp. Montastrea sp. and Acropora sp. like A. cervicornis and A. palmate.

PAGE 95

95 Figure 5 3 A View of a section of the fossil reef crest in San Salvador The section is composed of 100% A. palmate at this location. Large, thick trunks of A. palmate are truncated and surrounded by large branches of the corals that appear to have collapsed in situ Given that A. palmate dominates the high energy reef crest zone, this region is interpreted as the reef crest in LIG period. The photo was taken by Dr. Andrea Dutton at Cockburn Town, San Salvador.

PAGE 96

96 Figure 5 3 B The s ame outcrop as in Figure 5 3 A but looking west (towards the ocean). Large A. palmate corals that have collapsed in situ (ID#98, 2.80 m) The thickness of this A. palmate outcrop is ~1.5 m. The photo was taken by Dr. Andrea Dutton at Cockburn Town, San Salvador.

PAGE 97

97 Figu re 5 4 The approximate positions of the highest in situ corals (Reef II) in the paleogeographic map of San Salvador Map on the left is the reprint of Figure 3 in Page 110 of Curran, H. A., White, B., Chen, J. H., & Wasserburg, G. J. (1989). Comparative morphologic analysis and geochronology for the developmen t and decline of two Pleistocene coral reefs, San Salvador and Great Inagua Islands. In Proc 4th Symp Geol Bahamas. Bahamian Field Station, San Sal vador (pp. 107 117) Permission is granted for the reprint by the Bahamian Field Station (a.k.a. Gerace Resea rch Centre ). In the snip of Google Earth map on the right side, the distribution of the h ighest corals of different species in Reef II include ID#83 A. palmate in back reef (2.91 m), ID#78 Diploria (2.93 m) and ID#82 Montastrea (2.98 m) are displayed And ID#98 A. palmate is the highest in the reef crest zone (2.80 m).

PAGE 98

98 Figure 5 5 High Diploria (Reef II) covered by prograding beach sediment in Great I n a gua. Inagua.

PAGE 99

99 Figure 5 6 Comparison between model prediction and fie ld observation of LIG sea level in San Salvador and Great Inagua. The discrete points are the elevations of corals that are used to interpret the positions of LIG peak sea level and mid LIG sea level just before the rapid sea level drop ~125.1 ka ago (orange square for San Salvador and purple diamond for Great Inagua). The times of these two sea level events are similar in both islands, and the separation of time in the figure is just for visual convenience. The s hort vertical dashed lines above the points are the whole ranges of LIG sea level in San Salvador (orange, 4.3 5 .3 m at 119.5 ka, and 2.68 3. 6 8 m at 125.1 ka) and Great Inagua (purple, 3.3 5.3 m at 119.5 ka, and 2.13 3. 6 3 m at 125.1 ka) (due to the paleode pth uncertainty as discussed in Section 5.5). T he error bars on the points represent the survey error as discussed in the fourth part of Chapter 4 which are too small to discern in the figure compared with the symbol as well as the paleodepth uncertainty. The model prediction are from data in Dutton and Lambeck, 2012 (same as Figure 1 3).

PAGE 100

100 CHAPTER 6 CONCLUSIONS By surveying the potentially highest in situ fossil corals in the two reefs that are separated by the unconformity in Great Inagua and San Salvado r and connecting the elevation data with the tidal datum MLLW, we provide a relatively accurate estimation for (a) LIG peak sea level and (b) the rate of sea level change in the two islands. These data can provide valuable constraints on volume of Laurenti de ice sheet in MIS 6 and MIS 2 in order to adjust the ice model and resolve the inconsistence between the predictions in the model and observations in the field. (1) In San Salvador, based on (i) the elevation of the A. palmate (ID#98, 2.80 m) that is on the top of highly packed A. palmate mound in reef crest, (ii) the common height of 3 4 m of A. palmate in the Bahamas, and (iii) the large size of this coral, which is ~1.5 m tall after the collapse, as well as large A. palmate and branches that appeared t o collapse in situ in the surrounding, it is suggested that this A. palmate should have reached the limitation of its growing height, i.e., water surface at low tide. Thus, LIG peak sea level is estimated at 4.3 5.3 m by adding 1.5 2.5 m of height offset o f the coral. (2) The elevation difference between the highest in situ corals (Reef II) of the same coral species ( Diploria ) is 0.5 m (2.43 m in Great Inagua and 2.93 m in San Salvador). Considering that both corals should have reached the peak point of thei r growing limitation in height and that the paleodepth of the two highest Diploria could probably the same, the gradient of peak sea level is 0.5 m in the two islands.

PAGE 101

101 (3) LIG peak sea level in Great Inagua is est imated as 3. 3 5.3 m by combining the highest Diploria height difference (0.5 m) between the two islands and the 1 m paleodepth uncertainty of the fossil Diploria in Great Inagua. (4) There is an unconformity separated the two reefs (Reef I and R eef II) in both the two islands, which is interpreted to form during the short lowstand in mid LIG period. Sea level dropped rapidly at ~125.1 ka and returned to high level rapidly after several hundred years at 124.2 ka, based on U Th age of corals from the two reef s in the two islands (Chen et al., 1991; Thompson et al., 2011) Minimum s ea lev el is estimated as ~ 2. 6 m just before the rapid regress ion at ~125.1 ka in Great Inagua, based on the elevation of highest point in the platform (2.13 m) with an additional 0.5 m estimated that was eroded away during the sea level lowstand (5) Sea level is estimated to be at least 2.68 m before the rapid sea level fall in San Salvador, given the elevation of highest in situ A. palmate (ID#96, 2.18 m) and estimations of 0.5 m he ight offset. (6) The rate of sea level rise is estimated as 0. 29 to 0. 48 m/ka in San Salvador and 0. 30 to 0. 4 8 m/ka in Great Inagua. (7) The estimated rate of sea level change is much lower than the model prediction in both islands. The volume of Laurentid e ice sheet in MIS 6 needs to be made smaller in the model to resolve this inconsistence. This change would also cause a decrease in the prediction of peak sea level in both islands. If the resultant decrease is not large enough to eliminate the gap betwee n the model prediction and observation, then the volume of Laurentide ice sheet in MIS 2 needs to be made smaller as well.

PAGE 102

102 APPENDIX Object A 1. Supplementary data (.xlsx file 376 KB) Object A 2 Matlab code for MLLW calculation (.html file 24 KB)

PAGE 103

103 LIST OF REFERENCE S Andersen, M.B. (2006). The precise measurement of uranium series isotopes in the marine environment. ETH Ph.D. Thesis Zrich. And ersen, M. B., Stirling, C. H., Potter, E. K., Halliday, A. N., Blake, S. G., McCulloch, M. T., Ayling, B.F. and O'Leary, M. (2008). High precision U series measurements of more than 500,000 year old fossil corals. Earth and Planetary Science Letters 265(1 ), 229 245. Barnes, J.W., Lang, E.J., and Potratz, H.A. (1956). Ratio of ionium to uranium in coral limestone. Science 124, 175 176. Carew, J. L. and Mylroie, J. E. (1987). A refined geochronology for San Salvador Island, Bahamas. In: Curran, H. A. (ed.), Proceedings of the 3rd Symposium on the Geology of the Bahamas, pp. 35 44. CCFL Bahamian Field Station. Fort Lauderdale, Florida. Carew, J.L. and Mylroie, J.E. (1995a). A stratigraphic and depositional model for the Bahama Islands. In: Curran, H.A. and W hite, B. (Eds.), Geological Society of America Special Paper 300, Terrestrial and Shallow Marine Geology of the Bahamas and Bermuda, pp. 5 31. Carew, J. L., and Mylroie, J. E. (1995b). Depositional model and stratigraphy for the Quaternary geology of the B ahama Islands (pp. 5 32). Geological Society of America Carew, J. L., Mylroie, J. E., Wehmiller, J. F. and Lively, R. A. (1984). Estimates of Late Pleistocene sea level highstands from San Salvador, Bahamas. ln: Teeter, J. W. (ed.) Proceedings of the 2nd Symposium on the Geology of the Bahamas, pp. 153 175. CCFL Bahamian Field Station, Fort Lauderdale, Florida. Chappell, J. and Shackleton, N.J., 1986.Oxygen isotopes and sea level. Nature 324: 137 140 Chen, J.H., Curran, H.A., White, B., Wasserburg, G.J. ( 1991). Precise chronology of the last interglacial period: 234U 230Th data from fossil coral reefs in the Bahamas. Geological Society of America Bulletin 103, 82 97. Cheng, H., Adkins, J., Edwards, R.L. and Boyle, E.A. (2000a). U Th dating of deep sea cor als. Geochimica et Cosmochimica Acta 64, 2401 2416. Cheng, H., Edwards, R.L., Hoff, J., Gallup, C.D., Richards, D.A., Asmerom, Y. (2000b). The half lives of uranium 234 and thorium 230. Chemical Geology 169, 17 33.

PAGE 104

104 Cheng, H., Edwards, R. L., Wang, X. F. Woodhead, J., Hellstrom, J., Wang, Y. J., & Kong, X. G. (2008). A new generation of 230Th dating techniques: Tests of precision and accuracy. Geochimica et Cosmochimica Acta Supplement 72, 157. Cheng, H., Lawrence Edwards, R., Shen, C. C., Polyak, V. J., Asmerom, Y., Woodhead, J., ... & Wang, X. (2013). Improvements in 230 Th dating, 230 Th and 234U half life values, and U Th isotopic measurements by multi collector inductively coupled plasma mass spectrometry. Earth and Planetary Science Letters 371, 82 91. Chen, J. H., Curran, H. A., White, B., & Wasserburg, G. J. (1991). Precise chronology of the last interglacial period: 234U 230Th data from fossil coral reefs in the Bahamas. Geological Society of America Bulletin 103(1), 82 97. Curran, H. A., Whi te, B., Chen, J. H., & Wasserburg, G. J. (1989). Comparative morphologic analysis and geochronology for the development and decline of two Pleistocene coral reefs, San Salvador and Great Inagua Islands. In Proc 4th Symp Geol Bahamas. Bahamian Field Station San Salvador (pp. 107 117). Curray, J.R., Emmel, F.J. and Crampton, P.J.S. (1969) Holocene history of a strand plain, lagoonal coast, Nayarit, Mexico. In: Coastal lagoons, a symposium UNAM UNESCO, Mexico, DF, 1967. Mexico Univ Nac Autonoma, pp 63 100. Delanghe, D., Bard, E., Hamelin, B., 2002. New TIMS constraints on the uranium 238 and uranium 234 in seawaters from the main ocean basins and the Mediterranean Sea. Marine Chemistry 80, 79 93. h banding as environmental recorder. Nature, 258: 706 708. Dutton, A., and Lambeck, K. (2012). Ice volume and sea level during the last interglacial. Science 337(6091), 216 219. Edwards, L.R., Chen, J.H., Wasserburg, G.J. (1987a). 238U 234U 230Th 232Th sy stematics and the precise measurement of time over the past 500 000 years. Earth Planet Science Letter 81, 175 192. Edwards, R.L., Chen, J.H., Ku, T.L., Wasserburg, G.J. (1987b). Precise timing of the last interglacial period from mass spectrometric deter mination of 230Th in corals. Science 236, 1547 1553. Edwards, R. L., Gallup, C. D., & Cheng, H. (2003). Uranium series dating of marine and lacustrine carbonates. Reviews in Mineralogy and Geochemistry 52(1), 363 405.

PAGE 105

105 Esat, T.M., McCulloch, M.T., Chappel l, J., Pillans, B. and Omura, A. (1999). Rapid fluctuations in sea level recorded at Huon Peninsula during the penultimate deglaciation. Science 283, 197 201. Esat, T.M. and Yokoyama,Y. (2006). Variability inthe uraniumisotope composition of the oceans ov er glacial interglacial timescales. Geochimica et Cosmochimica Acta 70, 4140 4150. Fricke, H. and Meischner, D. (1985) Depth limits of Bermudian scleractinian corals: a submersible survey. Marine Biology (Historical Archive) 88, 175 187. Fruijtier, C., E lliott, T., Schlager,W. (2000). Mass spectrometric 234U 230Th ages from the Key Largo Formation, Florida Keys, United States: constraints on diagenetic age disturbance. Geological Society of America Bulletin 112, 267 277. Gallup, C.D., Edwards, R.L., John son, R.G. (1994). The timing of high sea levels over the past 200,000 years. Science 263, 796 800. Garrett, P., & DUCKLOW, H. (1975). Coral diseases in Bermuda. Nature, 235: 349 350 Goreau, T. F., & Wells, J. W. (1967). The shallow water Scleractinia of J amaica: revised list of species and their vertical distribution range. Bulletin of Marine Science 17(2), 442 453. Hearty, P.J., 1998. The geology of Eleuthera Island, Bahamas: a Rosetta stone of Quaternary stratigraphy and sea level history. Quaternary S cience Reviews 17, 333 355. Hearty, P.J., Kaufman, D.S., 2000. Whole rock aminostratigraphy and Quaternary sea level history of the Bahamas. Quaternary Research 54 (4), 163 173. Hearty, P. J., & Kindler, P. (1993). New perspectives on Bahamian geology: S an Salvador Island, Bahamas. Journal of Coastal Research, 577 594.Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L. and Shackleton, N.J., 1984. The orbital theory of Pleistocene climate: support from a revised chronology of the marine180 record. In: A.I. Berger et al. (Editors), Milankovitch and Climate, Part I. Reidel, Dordrecht, pp. 269 305. Henderson, G.M., Slowey, N.C., Fleisher, M.Q., 2001. U Th dating of carbonate platform and slope sediments. Ge ochimica et Cosmochimica Acta 65, 2757 2770. Hallock, P., & Schlager, W. (1986). Nutrient excess and the demise of coral reefs and carbonate platforms. Palaios 389 398.

PAGE 106

106 Hudson, J. H. (198 1). Growth rates in Montastraea annularis : a record of environmenta l change in Key Largo Coral Reef Marine Sanctuary, Florida. Bulletin of Marine Science 31 (2), 444 459. Humann, P., 1993. Reef Coral Identification Florida Caribbean Bahamas, (ed. N. Deloach). New World Publications, Inc., Paramount Miller Graphics, Inc. Jacksonville, Florida. Israelson, C., and Wohlfarth, B. (1999). Timing of the last interglacial high sea level on the Seychelles Islands, Indian Ocean. Quaternary Research 51(3), 306 316. Jaffey, A. H., Flynn, K. F., Glendenin, L. E., Bentley, W. T., an d Essling, A. M. (1971). Precision Measurement of Half Lives and Specific Activities of^{235} U and^{238} U. Physical Review C 4(5), 1889 1906. Kindler and Hearty (1996). Carbonate petrography as an indicator of climate and sea level changes: new data fro m Bahamian Quaternary units. Sedimentology 43, 381 399 Kindler, P., Reyss, J. L., Cazala, C., & Plagnes, V. (2007). Discovery of a composite reefal terrace of middle and late Pleistocene age in Great Inagua Island, Bahamas. Implications for regional tecto nics and sea level history. Sedimentary Geology 194(1), 141 147. Lambeck, K. and Chappell, J. (2001). Sea level change through the last glacial cycle. Science 292, 679 686. Lambeck, K., Purcell, A., & Dutton, A. (2012). The anatomy of interglacial sea levels: the relationship between sea levels and ice volumes during the Last Interglacial. Earth and Planetary Science Letters 315, 4 11. Lambeck, K., Purcell, A., Funder, S., KJR, K. U. R. T., Larsen, E ., & Moller, P. E. R. (2006). Constraints on the Late Saalian to early Middle Weichselian ice sheet of Eurasia from field data and rebound modelling. Boreas 35(3), 539 575. Lambeck, K., Purcell, A., Zhao, J., & SVENSSON, N. O. (2010). The Scandinavian ice sheet: from MIS 4 to the end of the Last Glacial Maximum. Boreas 39(2), 410 435. Lambeck, K., Yokayama, Y. and Purcell, T. (2002) Into and out of the Last Glacial Maximum: sea level change during Oxygen Isotope Stage 3 and 2. Quaternary Science Reviews 21: 346 360. Lewis, J. B., Axelsen, F., Goodbody, I., Page, C., & Chislett, G. (1968). Comparative growth rates of some reef corals in the Caribbean (No. MS 10). McGill University, MONTREAL (QUEBEC) MARINE SCIENCES CENTRE, pp. 1 10.

PAGE 107

107 Lighty, R. G., Macintyre, I. G., & Stuckenrath, R. (1982). Acropora palmate reef framework: a reliable indicator of sea level in the western Atlantic for the past 10,000 years. Coral reefs 1(2), 125 130. Lisiecki, L. E., and Raymo, M. E. (2005). A Pliocen e Pleistocene stack of 57 globally distributed benthic 18O records. Paleoceanography 20 (1). Mckay, N.P., Overpeck, J.T. and Otto Bliesner, B.L. (2011). The role of ocean thermal expansion in Last Interglacial sea level rise. Geophysical Research Lette r 38, L14605. Meyerhoff, A. A., & Hatten, C. W. (1974). Bahamas salient of North America: Tectonic framework, stratigraphy, and petroleum potential. AAPG Bulletin 58(6), 1201 1239. Moslow, T.F. and Heron, S.D. Jr (1981) Holocene depositional history of a microtidal cuspate foreland cape: Cape Lookout, North Caroline. Marine Geology 41:251 270 Muhs, D. R., Simmons, K. R., Schumann, R. R., & Halley, R. B. (2011). Sea level history of the past two interglacial periods: new evidence from U series dating of r eef corals from south Florida. Quaternary Science Reviews 30(5), 570 590. Mullins, H. T., Dolan, J., Breen, N., Andersen, B., Gaylord, M., Petruccione, J. L., Wellner, R.W., Melillo, A.J. and Jurgens, A. D. (1991). Retreat of carbonate platforms: response to tectonic processes. Geology 19(11), 1089 1092. Mullins, H. T., & Hine, A. C. (1989). Scalloped bank margins: Beginning of t he end for carbonate platforms? Geology 17(1), 30 33. Mylroie, J. E. (2008). Late Quaternary sea level position: evidence from Bahamian carbonate deposition and dissolution cycles. Quaternary International 183(1), 61 75. Mylroie, J. E., & Carew, J. L. (1990). The flank margin model for dissolution cave development in carbonate platforms. Earth Surface Processes and Landforms 15( 5), 413 424. Neumann, A. C., & Hearty, P. J. (1996). Rapid sea level changes at the close of the last interglacial (substage 5e) recorded in Bahamian island geology. Geology 24(9), 775 778. Neumann, A. C. and Moore, W. S. (1975). Sea level events and Plei stocene coral ages in the northern Bahamas. Quaternary Research 5, 215 224. Newell, N. D., & Rigby, J. K. (1957). Geological studies on the Great Bahama Bank. Regional aspects of carbonate deposition. Special Publication Society of Economic Paleontologist s and Mineralogists 5, 15 72.

PAGE 108

108 Peltier, W.R. and Fairbanks, R.G. (2006). Global glacial ice volume and Last Glacial Maximum duration from an extended Barbados sea level record. Quaternary Science Review 25, 3322 3337. Potter, E.K., Esat, T.A., Schellmann G., Radtke, U., Lambeck, K. and McCulloch, M.T. (2004). Suborbital period sea level oscillations during marine isotope substages 5a and 5c. Earth and Planetary Science Letters 225, 191 204. Riitzler, K. and Macintyre, I.G. (1982). The habitat distribut ion and community structure of the barrier reef complex at Carrie Bow Cay, Belize. In: Riitzler K, Macintyre IG (eds) The Atlantic barrier reef eco system at Carrie Bow Cay, Belize. Scientific Reports 1, Smithsonian Contributions to the Marine Sciences no 12, Smithsonian Institution Press, Washington DC, pp 9 46 Robinson, L.F., Belshaw, N.S., Henderson, G.M. (2004a). U and Th concentrations and isotope ratios in modern carbonates and waters from the Bahamas. Geochimica et Cosmochimica Acta 68, 1777 1789. R obinson, L.F., Henderson, G.M., Hall, L., Matthews, I. (2004b). Climatic control of riverine and seawater uranium isotope ratios. Science 305, 851 854. Scholl, D.W., Craighead, F.C. and Stulver, M. (1969) Florida submergence curve revised: its relation to coastal sedimentation rates. Science 163, 562 564 Scholz, D. and Mangini, A. (2006). Estimating the uncertainty of coral isochron U Th ages. Quaternary Geochronology 1, 279 288. Scholz, D., Mangini, A. and Felis, T. (2004). U series dating of diagenetic ally altered fossil reef corals. Earth and Planetary Science Letters 218, 163 178. Shackleton, N.J., 1987. Oxygen isotopes, ice volume, and sea level. Quaternary Science Review 6, 183 190. Shinn, E. A., Lidz, B. H., Halley, R. B., Hudson, J. H., & Kindinger, J. L. (1989). Reefs of Florida and the Dry Tortugas: Miami to Key West, Florida American Geophysical Union 176, 1 53 Slowey, N. C., Henderson, G. M., and Curry, W. B. (1996). Direct U Th dating of marine sediments from the two most recent interglacial periods. Nature 383(6597), 242 244. Skinner, L. C., & Shackleton, N. J. (2005). An Atlantic lead over Pacific deep water change across Termination I: Implications for the application of the marine isotope stage stratigraphy. Quaternary Science Reviews 24(5), 571 580. Sea Temperature and Ice Volume Changes Across the Pliocene Pleistocene Science 328, 1480.

PAGE 109

109 Stirling, C. H., and Andersen, M. B. (2009). Uranium series dating of fossil coral reefs: Extending the sea level record beyond the last glacial cycle. Earth and Planetary Science Letters 284(3), 269 283. Stirling C. H., Esat T. M., Lambeck K., McCulloch M. T., Blake S. G., Lee D. C., and Halliday A. N. (2001). Orbital forcing of the marine isotope stage 9 interglacial. Science 291(5502), 290 293. Stirling, C. H., Esat, T. M., Lam beck, K., and McCulloch, M. T. (1998). Timing and duration of the Last Interglacial: evidence for a restricted interval of widespread coral reef growth. Earth and Planetary Science Letters 160(3), 745 762. Stirling, C. H., Esat, T. M., McCulloch, M. T., a nd Lambeck, K. (1995). High precision U series dating of corals from Western Australia and implications for the timing and duration of the Last Interglacial. Earth and Planetary Science Letters 135(1), 115 130. Storr, J. F. (1964). Ecology and oceanograph y of the coral reef tract, Abaco Island, Bahamas. Geological Society of America Special Papers 79, 1 94. Tamisiea, M. E., & Mitrovica, J. X. (2011). The moving boundaries of sea level change: Understanding the origins of geographic variability. Oceanograp hy 24(2):24 39, http://dx.doi.org/10.5670/oceanog.2011.25. Thompson, W. G., Curran, H. A., Wilson, M. A., & White, B. (2011). Sea level oscillations during the last interglacial highstand recorded by Bahamas corals. Nature Geoscience 4 (10), 684 687. Thom pson, W.G., Spiegelman, M.W., Goldstein, S.L. and Speed, R.C. (2003). An open system model for U series age determinations of fossil corals. Earth and Planetary Science Letters 210, 365 381. Villemant, B. and Feuillet, N. (2003). Dating open systems by th e 238U 234U 230Th method: application to Quaternary reef terraces. Earth and Planetary Science Letter 210, 105 118. White, B., Curran, H. A., & Wilson, M. A. (1998). Bahamian coral reefs yield evidence of a brief sea level lowstan d during the last intergl acial. Carbonates and Evaporites 13(1), 10 22. Yokoyama, Y., Lambeck, K., De Deckker, P., Johnston, P. and Fifiled, L.K. (2000) Timing of the Last Glacial Maximum from observed sea level minima. Nature 406, 713 716. Yu, J. and Broecker, W.S. (2010) Com Sea Temperature and Ice Volume Changes Across the Pliocene Science 328, 1480.

PAGE 110

110 Zankl, H., & Schroeder, J. H. (1972). Interaction of genetic processes in Holocene reefs off North Eleuthera Island, Bahamas. Geologische Rundschau 61(2), 520 541.

PAGE 111

111 BIOGRAPHICAL SKETCH Jin Li was a n International student from northeast China. She finished elementary through high school in her hometown Benxi, Liaoning province. She was admitted by Nanjing University in September 2008 after graduating from high school. S he was selected as an honors level Geography Major sponsored by the Natu ral Science Foundation of China. At NJU she beg an to learn about earth sciences which lead to an interest in coastal research. I n her sophomore year, she led a National Undergraduate Innovation Program sponsored by the National Ministry of Education of China, which aims to help undergraduates who have innovative ideas to conduct scientific research. Her project aimed to quantify interactions between the morphology of the marsh grass Spartina Alterniflora and sediment deposition between the coasts of Jiangsu and Fujian provinces, as caused by changes in environmental factors She developed a method for estimating deposition rates on tidal flats using the morphological data of Spartina alterniflora plants, which she published in 2013 in the Journal of Coastal Research ( Volume 29, Issue 6: pp. 1452 1463) After obtaining a B achelor of Science at NJU in 2012, Jin Li began her graduate study at the University of Florida in the D e partment of Geological S ciences, where her research is focused on the reconstruction of paleo sea level during the last interglacial (LIG) period in the Bahamas using fossil corals. She intends to graduate in the spring of 2014 with a Master of Science