Depositional Analysis of a Holocene Carbonate Strand Plain Using High Resolution GPR, Sediment Analysis, and C-14 Dating

Brigham Young University BYU ScholarsArchive

Theses and Dissertations

2019-12-01

Kaleb Robert Markert Brigham Young University

Follow this and additional works at: https://scholarsarchive.byu.edu/etd

Part of the Physical Sciences and Mathematics Commons

BYU ScholarsArchive Citation Markert, Kaleb Robert, "Depositional Analysis of a Holocene Carbonate Strand Plain Using High Resolution GPR, Sediment Analysis, and C-14 Dating" (2019). Theses and Dissertations. 7751. https://scholarsarchive.byu.edu/etd/7751

This Thesis is brought to you for free and open access by BYU ScholarsArchive. It has been accepted for inclusion in Theses and Dissertations by an authorized administrator of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. Depositional Analysis of a Holocene Carbonate Strand Plain Using High Resolution GPR, Sediment Analysis, and C-14 Dating

Kaleb Robert Markert

A thesis submitted to the faculty of Brigham Young University in partial fulfillment of the requirements for the degree of

Master of Science

John H. McBride, Chair Scott Ritter Sam Hudson

Department of Geological Sciences

Brigham Young University

Depositional Analysis of a Holocene Carbonate Strand Plain Using High Resolution GPR, Sediment Analysis, and C-14 Dating

Kaleb Robert Markert Department of Geology, BYU Master of Science

Understanding modern carbonate depositional systems gives valuable insights into the interpretation of ancient carbonate systems. Ancient carbonate strand plains have the potential to act as productive hydrocarbon reservoirs because of their relatively high porosity. Unfortunately, they are difficult to identify in the rock record because of the lack of work done on modern analogues. San Salvador Island in the Bahamas hosts a well-exposed and easily accessible Holocene strand plain, ideal as a modern analogue. Sandy Hook, located on the southeast part of the island, consists of approximately 35 ridge-swale features that show signs of four distinct zones of deposition.

In this study, 20 sediment samples were collected from one meter in depth, and throughout each zone. The sediment analysis reveals remarkably consistent grain-size distributions across the four zones. Carbon-14 ages were taken from the sediment samples. They revealed ages decreasing from 2617 ± 188 YBP to modern moving seaward through zone 4. The ages reveal depositional rates of 0.08 to 0.29 meters of accretion per year. Four pseudo-3D GPR surveys were acquired in Zones B, C, and D, and a 2D GPR survey was acquired that spanned the three zones. The pseudo-3D surveys revealed consistent reflectors through the width of the survey indicating that the 2D surveys represent more than the single line. The 2D survey reveals semi-parallel seaward- dipping reflectors (representing fair-weather deposits) that are truncated by sigmoidal seaward- dipping reflectors (high-energy storm deposits). Indicating that Sandy Hook was built through both fair-weather deposits and high-energy storm events.

Keywords: ground penetrating radar, sediment analysis, carbon-14, San Salvador Island, Sandy Hook, strand plain deposition ACKNOWLEDGEMENTS

This project would not have been possible without the help of my graduate committee and other faculty of the BYU Geology department. I would like to thank my committee chair,

Dr. John H. McBride, for all the tutelage and support that he provided. He allowed me opportunities to work through the process and grow as a scientist. I would like to thank Scott

Ritter for helping me navigate the world of carbonates and for his support in collecting the data. I would like to thank Sam Hudson for his help revising my thesis. Thank you to David Tingey for the time that he spent collecting and preparing the samples on San Salvador and for his help exploring the meaning of the data. I am thankful for Kevin Rey’s guidance as I prepared and processed the carbon-14 samples, and for Bill Keach’s help with importing my GPR data into

Geoprobe. The data processing and visualization were possible thanks to software grants from

Landmark (Halliburton) University Grant Program and from the HIS Kingdom Educational

Grant Program.

Lastly, I would like to thank my wonderful wife, Isabella, for her love and support through the process. She encouraged me through hard times and gave me the final push needed to complete this project. TABLE OF CONTENTS TITLE PAGE...... i ABSTRACT ...... ii ACKNOWLEDGEMENTS ...... iii TABLE OF CONTENTS ...... iv LIST OF FIGURES ...... v INTRODUCTION ...... 1 BACKGROUND ...... 2 Previous Research ...... 4 Beach-Ridge Systems...... 4 Sandy Hook ...... 5 METHODS ...... 6 Sediment Analysis ...... 8 RESULTS ...... 10 Sediment Analysis Results ...... 10 C-14 Results...... 11 GPR Results...... 11 DISCUSSION...... 13 Radar Facies ...... 13 Radar Facies Distribution ...... 15 Sea-Level Fluctuations ...... 17 Storm vs. Fair-Weather ...... 18 CONCLUSION ...... 19 REFERENCES ...... 20

iv LIST OF FIGURES Figure 1...... 23 Figure 2...... 24 Figure 3...... 25 Figure 4...... 26 Figure 5...... 27 Figure 6...... 28 Figure 7...... 29 Figure 8...... 30 Figure 9...... 32 Figure 10...... 34 Figure 11...... 35 Figure 12...... 36

v INTRODUCTION Carbonate strand plains are important to understand because they have the potential to be

excellent oil and gas reservoirs (Ward and Brady, 1979), they are the dominant geologic feature

in the Bahamas (Carew and Mylroie 1997), and they account for the majority of the Pleistocene-

Holocene interglacial island development (Aurell et al. 1995; Garrett and Gould, 1984; Hearty

and Kindler, 1993; Hearty and Kindler, 1997; Mylroie and Carew, 2008). Despite their

prominence in the Pleistocene and Holocene epochs, strand plains are rare in the ancient record

(Scholle et al., 1983). Because they are rare and difficult to identify in the rock record, placing

constraints on deposition rates and extent is difficult and requires additional study. Furthermore,

a more in-depth study of carbonate strand plains will increase our knowledge of Bahamian Island

development in general and our ability to locate strand plains in the rock record.

Strand plains are ridge-swale features that form parallel or semi-parallel to the coastline.

Prior studies of carbonate strand plains (Ward and Brady, 1979; Harris, 1979; Carney et al.,

1993; Kim, 2001) is not a prolific as studies of clastic strand plains (Tanner, 1995; Taylor and

Stone, 1996; Van Heteran et al., 1996; Otvos, 2000; Tamura, 2012; Gontz et al., 2014). Although connections between clastic and carbonate strand plains exist, the main driving mechanism behind the development of the two strand plains, the sediment source, is different. Carbonate strand plains form from sediment created offshore that is accreted onto the beach (Evans, 1970;

Rivers et al., 2019), where clastic strand plains are formed from allochthonous sediment that is transported and stored via marine-driven mechanisms such as longshore drift (Boyd et al., 1992).

In spite of the differences of the sediment source, the depositional patterns of carbonates and clastics are remarkably similar (Evans, 1970); therefore, models of clastic deposition are applicable to this study.

1 This study chose to focus on carbonate strand plains because carbonates are more susceptible to GPR. Since carbonates are monomineralic (calcium carbonate), the reflections are caused by changes in porosity, which make it very easy to interpret. This makes processing the data simple because the dielectric constant remains constant. Finally, carbonates are ideal for

GPR surveys because GPR signals penetrate well through carbonates. This is because there are no clays that cause attenuation like in their clastic counterparts.

San Salvador is an ideal location for studying carbonate strand plains. It harbors a

Holocene carbonate strand plain in the southeastern corner that is easily accessible. San Salvador is also an excellent location because it is used as a model of development for all of the Bahamian island (Carew and Mylroie, 1997); therefore, the results easily obtained from San Salvador about island development can be applied to the other Bahamian, and even Caribbean Islands such as the Turks and Caicos Islands.

The purpose of this study is to determine if there is a diagnostic GPR signature that can be found within carbonate strand plains. To accomplish this, we integrated carbon-14 dating, sediment analysis, and ground penetrating RADAR (GPR).

BACKGROUND The Bahama Islands is the 1400 km northern portion of a NW-SE trending archipelago located in the western North Atlantic Ocean (21o to 27o30’N and 69o to 80o30’W). The northwestern Islands protrude out of the ocean from two large carbonate platforms, Little

Bahama Bank and Great Bahama Bank (Figure 1). The southeastern Bahamas has a series of small isolated carbonate platforms. Many of these smaller carbonate platforms are topped by islands that cover the majority of the area available. The water depth on both the larger and smaller platforms is generally less than 10 meters (Mylroie and Carew, 1995). The individual

2 platforms are separated by troughs and channels that reach depths of 2000 meters. San Salvador

Island is located on one of these smaller isolated carbonate platforms and is surrounded by

deeper water (Figure 1). Therefore, is an ideal location to study carbonate island development as

it can act as a model for the entirety of the Bahamian Islands (Carew and Mylroie 1997).

The depositional history of San Salvador occurred in four major depositional phases

(figure 2; Hearty and Kindler, 1993). Phase I occurred 380,000 to 180,000 years ago with the

development of small ridges. The ridges acted as anchors that influenced the remainder of the

island deposition. Phase II: large ooid dunes began to develop as the sea level rose. These ooid

dunes connected the previous anchors forming large catenary features (ancient strand plains) that

dominate the landscape of San Salvador. Phase III: the eastern shoreline of San Salvador formed

during the Late Sangamonian as the sea retreated. Evidence suggests that the Phase III deposits

were emplaced during a period of time when the sea level was lower than present. Phase IV: catenary Holocene sand deposits resulting from a change in the rate of sea-level rise 3,500 years ago. Several strand-plain systems formed during this phase of deposition, the most significant of them being Sandy Hook. The landforms created during Phase IV are generally lower in elevation than the landforms created during Phase II (Hearty and Kindler, 1993).

Sandy Hook, located in the southeastern corner of San Salvador Island in the Bahamas, is a modern strand plain. It is comprised of 35 ridge-swale features (Figure 3). It was deposited in

Phase IV of San Salvador’s development and is anchored to Phase III deposits (Figure 2). It is the best location for studying strand plains in San Salvador because it is flat and easily accessible. Sandy Hook has a network of roads and cul-de-sacs running both parallel and perpendicular to the strand plain’s ridges (Figure 3). This road system creates an ideal environment to conduct GPR surveys because the pavement is only few millimeters thick and is

3 flat and debris free. The cul-de-sacs offer similarly ideal circumstances to conduct pseudo-3D surveys. The road systems are flat because the swales were filled in with sediment. This is important because it preserved the original topography of the strand plain under a thin (~0.5 m) layer of sediment. They also offer easy access to dune ridges that were sampled for both carbon-

14 dating and grain size analysis. Without the road systems Sandy Hook could not have been used for two major reasons: 1) the island is covered in a dense foliage. Without the roads we would not have been able to conduct the surveys, and 2) the surveys would need extensive topographic corrections. Because of these conditions Sandy Hook is the best location to study modern carbonate strand plain development.

Previous Research Beach-Ridge Systems Beach-ridge systems were first studied in 1852 by Redman (Redman, 1852; 1864) who concluded that beach ridge systems were formed by storm events. This was the accepted depositional model for all beach ridges until Johnson (1919) showed that beach ridges could be deposited in normal low energy. The question then became how to build relief without storm waves. Davies (1957) determined that the berm is the only ridge formed by waves, and that relief ultimately came from vegetation trapping eolian sediments. The idea of storm-driven beach-ridge systems was not abandoned altogether. Taylor and Stone (1996) determined that storms did form beach-ridge systems, but the beach-ridge systems formed by storms were gravel beach ridges.

They also concluded that it was unlikely that storms would build sandy beach-ridge systems.

The most commonly accepted sandy beach-ridge depositional model is “beachface progradation

under fair-weather waves without the necessity of sea-level oscillations” (Tamura, 2012 and

Tanner, 1995).

4 Carter (1986) studied the beach-ridge formation of Magilligan, Northern Ireland. He found two distinct modes of deposition in the beach-ridge system. Mode 1 was a fair-weather model of deposition similar to the model proposed by Davies (1957). Mode 2 was a storm-driven sandy beach-ridge model. “Sediment is initially stored in nearshore bars, then is moved along shore causing bar elongation. The bar moves into the intermediate and even reflective wave domains. Eventually narrowing the bar leads to severance of a down-drift portion, usually at or near the major drift pulse to the northwest of the foreland. Normally this sediment-starved remnant welds progressively onto the beach face in the direction of the dominant process gradient, but occasionally locally generated counteractive seas move sediment onshore employing the bar as a traverse conduit.”

Harris (1979) completed an in-depth study of Joulters Cay’s ooid shoals. He touched briefly on how the shoals developed into the strand plain observable of Joulters Cay today. He found that strand plains “formed as sediments were redistributed by tidal and storm currents.” He discussed the formation of South Joulters Cay through the following process. 1) Ridges formed parallel to the shore initially. 2) A tidal channel cut through, truncating previous ridges and caused ridges forming while the channel was active to curve bankward. He stated that the tidal channel which formed was likely formed by a storm event. 3) Ridge growth continued eventually blocking the tidal channel, and 4) Spit growth continued, forming what is observed today.

For this study we will focus on the two major competing models: strand plains created due to sea-level fluctuations, and strand plains created by storm events.

Sandy Hook The most recent research conducted on Sandy Hook was by Mattheus et al. (2018). They completed an integrated geomorphological study coupled with GPR surveys. They found that

5 ridge separation decreases from 50 meters to 30 meters moving seaward. Their GPR surveys

revealed that accretion was interrupted by ravinement surfaces. Their study concluded that the

decrease in ridge spacing resulted from an increase in accommodation as accretion moved

seaward.

Kim (2001) completed a study of the different depositional environments surrounding

Sandy Hook. Kim (2001) studied a strand plain (Sandy Hook), Holocene and Pleistocene dunes,

a high energy lagoon (Snow Bay), and a tidally influenced lagoon (Pigeon Creek). Locations of

the studied areas can be found in figure 3 and 4. Kim (2001) concluded that because of how the

environments interact, it would be difficult to distinguish between the different environments in

the rock record.

Some early GPR work was completed by Carney et al. (1993) and Dominic et al. (1995).

Carney et al. (1993) completed a more in-depth survey where they reconstructed Sandy Hook using GPR and sediment analysis. Their paper focused primarily on the sediment analysis and used GPR to show that Sandy Hook contained dipping reflectors that dipped at similar angles to the modern beach. Dominic et al. (1995)’s work was limited to testing the capabilities of GPR.

They included a 2D GPR survey taken from Sandy Hook in their explanation of some of the capabilities of GPR.

Hearty and Kindler (1993) conducted an in-depth sedimentologic study of San Salvador.

Their study included a brief discussion about the Rice Bay Formation, the stratigraphic unit that composes Sandy Hook. However, they did not study or attempt to understand the deposition of

Sandy Hook.

METHODS

6 The study of Sandy Hook, San Salvador was accomplished in three integrated ways: 1) sampling from natural exposure of granular carbonate sediments for carbon-14 dating; carbon-14 age determinations were used to constrain the timing and rate of accretion of the sediments. 2)

Holocene stratigraphic profiles two meters in depth were collected and 3) use of 2D and 3D GPR acquisition. In order to obtain a more complete picture of the development of the Sandy Hook strand plain, Sandy Hook was divided into four zones (Figure 4), based on the vegetation patterns observed in aerial photos. It was observed while collecting the data that the vegetation grew primarily in the swales, and the ridges were exposed. So, by marking the clearings we were able to map the ridges. The four zones were created based on the map of the ridges. Psuedo-3D and 2D GPR surveys were acquired in three of the zones (Zones B, C, and D), and sediment samples were taken from each of the four zones, as well as several locations within each zone.

Carbon-14 Dating

Sediment samples were collected from dune ridges, using a four-inch diameter hand auger. Five complete stratigraphic profiles (~2 meters in length) were sampled approximately every 10 centimeters. In addition, 14 other smaller sediment samples were collected sampling approximately every 10 centimeters from the depths of 80 centimeters to 110 centimeters. Larger shell fragments and foraminifera were taken from ~100 centimeters from six sample locations.

An additional sample of shells was collected from the modern beach front to be used as a control.

It was found that a mild prewash completely dissolved the foraminifera. Therefore, clean shells and foraminifera were picked to avoid needing to prewash them to remove excess meteoric calcite. The shells and foraminifera were dissolved in an 85 percent phosphoric acid solution, releasing the CO2 contained within the calcium carbonate. The CO2 was then sealed in glass tubes and sent to the University of Georgia for Carbon-14 age dating.

7 Sediment Analysis A sediment analysis on the three complete two-meter stratigraphic profiles (SH-2 from

Zone B, SH-3 from Zone A, and SH-5 from Zone D; Figure 4) and the remaining 16 sediment

samples taken from approximately 100 cm was conducted by using a sieve and point-count

analysis. The samples were sieved and the weight percentage of sediment in each sieve size were

recorded. It was found that very few individual grains were larger than 25 mesh, and that 25

mesh sieves would remove the larger clumps of sediment. Since the analysis is to determine the

energy of the depositional environment based on individual grains, a 25-mesh sieve was used as the first sieve size, guaranteeing that individual grains of the given size were weighed. Sieves of mesh size 30, 35, 40, 45, 50, 60, 70, 80, 100, and 120 were chosen as the other dividers. The weight percentages were plotted in histograms and compared to each other to see variations throughout the island.

Point counts were conducted to determine the concentrations of grain types. Photos of thin sections made from loose and cemented sediment taken from the auger locations were used for the point counts. Five photos of each thin section were taken in plain light conditions and using a 10x lenses. Three-hundred points from each photo (1,500 points from each thin section) were picked using a grid style counting method, and the grains were identified. This method was used to conduct point counts on samples SH-2, SH-4, SH-8, SH-11, SH-14, SH-16, and SH-20.

Samples were strategically chosen from each of the 4 zones of Sandy Hook deposition.

GPR

Four pseudo-3D GPR surveys were conducted using GSSI (Geophysical Survey System,

Inc.) antenna with a center frequency of 400 MHz. Obtaining full-resolution 3D surveys requires a line spacing of approximately a quarter of the wavelength (Grasmueck et al., 2005). Using the

GSSI 400 MHz the line spacing would have to be 0.03 to 0.06 m (Hazard et al., 2017). Because

8 of time constraints, and lack of precision positioning equipment a line spacing of 0.3048 m was selected. Tests were conducted using different parameters until optimal parameters were found.

Data were collected in continuous mode with a rate of 2048 samples/trace over 100 nanoseconds

(20.5 samples/ns), resulting in approximately 79 scans per meter. The acquisition frequency filter was set at 100 to 800 MHz. The survey areas were of various sizes: Survey 1 7.0 m x 16.5 m,

Survey 2 6.1 m x 29.9 m, Survey 3 6.1 m x 10.4 m, and Survey 4 11.0 m x 6.1 m. An exploratory survey was conducted each 3D survey location to determine where the best reflectivity was. The dimensions of the surveys were limited by the available space of the cul-de-sac or road section, and the location of the reflectivity. The trace direction was approximately west to east, moving north through the survey area. There was no need for an elevation correction because the surveys were conducted in flat cul-de-sacs and on wide flat portions of the road. Various 2D GPR surveys were conducted using GSSI antennas with center frequencies of 400 MHz and 200 MHz.

The 2D surveys using the 400 MHz antenna were acquired in continuous mode using a rate of

1024 samples/trace over 100 nanoseconds (10.3 samples/ns), resulting in approximately 39 traces per meter. The other parameters were the same as the pseudo-3D surveys. The 2D surveys using the 200 MHz were acquired again in continuous listening mode with a rate of 1024 samples/trace over 300 nanoseconds (3.4 samples/ns), resulting in approximately 39 traces per meter. The acquisition frequency filter for the 200 MHz antenna was set at 50 to 600 MHz. The

2D surveys were conducted both perpendicular and along strike to the clinoforms.

The data were processed in GSSI’s RADAN 6.6 software, utilizing background removal, gain restoration, 2D Kirchhoff migration, customized exponential gain, automatic gain control, and time-to-depth conversion. Kirchhoff 2D migration was completed using information gathered from diffraction analysis and the findings of Hazard et al. (2017). A velocity of 0.07

9 m/ns (a dielectric constant of 18.4) was used for the migration. The processed files were

exported using REFLEX2DQuick software and imported into Halliburton’s Geoprobe seismic

interpretation software.

RESULTS Sediment Analysis Results From the three complete, two meter deep, auger locations we observed the color of the sediment changed from gray to a yellowish color to white as the auger descended. We also found that a cemented layer that lies approximately one meter below the surface on Sandy Hook.

Samples of the cemented section were sent to Wagner Petrographics and were made into thin sections. The thin sections reveal meniscus cementation around some of the grains (Figure 5).

There appears to be a lack of pendant structures present. It was found that there was only one

distinct type of foraminifera in the sediment samples: Arachaias angulatus. Some bivalves and

gastropods shells were present.

The sediment analysis shows a general trend that is consistent with the findings of

Mattheus et al. (2018). They found that there was a decreasing ooid concentration and an

increasing skeletal grain concentration in more seaward samples (Figure 5).

The granulometric analysis of the three complete two meter stratigraphic profiles, and the

16 samples collected at one meter depth reveal similar patterns. Histograms based on weight

percentages of the complete profiles are left skewed with maxima in the 50-70 mesh size (300 to

210 microns) range. Histograms based on weight percentages of the 19 sediment samples

collected from one meter in depth also reveal all of the samples but 3 (SH-4, SH-11, and SH-16)

have a left skewed profile and maxima between 50-70 mesh. SH-4, SH-11, and SH-16 are

relatively symmetric and have a maxima at 45-50 mesh. In spite of these little variation, the

overall pattern of the histograms is remarkably similar (Figure 6).

10 C-14 Results There was some ambiguity in the results of the carbon-14 dating results for the first three,

the eastern and older, zones of deposition (Figure 4). The C-14 dating results for Zone D, the youngest zone, were straight forward decreasing from 2617 ± 188 ybp at the landward edge of

Zone D to 1139 ± 146 ybp for sediment taken from a swale adjacent to the beach (Figure 7).

Modern shell fragments were also dated and were found to be modern in age. With these ages

Zone D shows depositional rates ranging from 0.29 to 0.08 meters per year. The C-14 ages

calculated in the Zone D sediments correspond to a period of storm “hyperactivity” in the Gulf of

Mexico that lasted from 1000 to 3400 YBP identified by Liu Fearn (2000).

GPR Results Four pseudo-3D surveys were examined, and one long 2D line that ran from Zone D

through Zone B. They all uniformly show seaward-dipping clinoforms (Figures 9-11). There are

also two bright distinct reflectors present throughout the entirety of the collected data (2D and

pseudo-3D). The upper most reflector (S1) is located between 0.2 and 0.5 meters below the road

surface. S1 undulates as the surveys progress. The S1 reflection is long and continuous. It

appears to have no significant reflections that cut through it, and there appears to be no

significant reflections in sediment lying immediately above S1 (identified in Figure 10).

The second bright reflector (S2) found throughout all of the acquired GPR lines is located

between 1- and 1.25-meters’ depth. It varies slightly in depth, but the variance does not appear to

be connected to the changes in S1’s elevation. In contrast to S1, S2 appears to be an area of

enhanced reflectivity. It is composed almost entirely dipping reflectors that continue to extend

below and above the increased contrast area. The strength of the reflections often dims as they

move off of S2 area (identified in Figure10).

11 The long 2D transects show both clinoform patterns clearly (Figure 9). The upper inflection of the clinoform pattern begins at S2. The inflections vary in depth as the surveys progress. They do not appear to be related to S1. In areas of good reflectivity, it is possible to trace the clinoforms long distances (up to 10 meters in some areas). They frequently cross S2 sometimes extending to S1 and terminate at it. Some of the reflectors that extend towards S1 appear to be truncated and capped with a small package of steeper dipping reflectors that extend from S1 and downlap onto the extending reflectors. This small package of dipping reflectors dips landward rather than seaward as the other reflectors dip.

Analysis of the dipping reflectors contained within the pseudo-3D surveys reveal the reflectors dip between 3 to 15 degrees, with the majority of the reflectors dipping between 6 and

10 degrees (Figure11). The longest continuous reflectors extend 6 meters with the majority of the reflectors being discontinuous and extending only a few meters. Survey 4 located in Zone B had an average reflection dip of 6.8 degrees with the average reflector extending 3.0 meters (pseudo-

3D volume displayed in Figure 10b). Survey 3 located in Zone C had an average reflection dip of

9.4 degrees with an average reflector extending 1.3 meters. Survey 2 located in the northern portion of Zone D had an average dip of 10.2 degrees and an average reflector extending 2.5 meters. Survey 1 located in the southern portion of Zone D had an average reflection dip of 8.8 degrees and an average reflector extending 2.3 meters (pseudo-3D volume displayed in Figure

10a). The 2D survey reveals that the reflectors occur in packages that are frequently separated by areas of low reflectivity.

The horizontal time slices of the pseude-3D surveys show that the orientation of the reflectors changes from being oblique to north-south between 1 and 2 meters in depth (Figure

10). This change in orientation is consistent throughout the survey. The time slices also display

12 continuous lines that indicate that the dipping reflector is consistent through the width of the

survey.

A few hyperbolic diffractions are present in the data after the migration was completed.

These diffractions lie along S1. They are most likely out of the plane diffractions and will be

ignored.

DISCUSSION Based on the patterns of the GPR reflectors the data can be divided into two radar

surfaces (S1 and S2), and six radar packages (F1-F5). Because of the monomineralic nature of the study area, the reflections are caused by differences in porosity and cementation that occur along bedding planes and bounding surfaces. (Hazard, 2018; Rust and Russell, 2001;

Cunningham 2004; Franseen et al., 2007).

Radar Facies Radar surface S1 is the brightest and most continuous reflector found in all of the surveys. S1 has more variability undulating as it moves across the survey. It varies between 0.2 and 0.5 meters. S1 corresponds to the original topography of the study area, with the zone above

S1 being fill that was used to create the level surface of the road. This interpretation was confirmed by conducting a simple off-road to on-road survey. The reflection was not present in the off-road portion of the survey, but gradually appeared and increased in depth as the GPR unit increased in elevation moving onto the road. The brightness and continuous nature of the reflection is the result of allochthonous carbonate sediment being placed above the original topography to create the level road.

Radar surface S2 is a bright reflective zone found in all of the radar surveys. The zone is primarily comprised of multiple reflections that are bright and then fade as they move away from

13 the zone. S2 has a very low-angle seaward dip (~0.3o) and varies in depth between 1 and 1.25

meters. This zone corresponds to a hard-cemented layer hit by the auger in all of the sample locations. The thin section analysis of the cemented layer revealed meniscus cementation that is associated with water table diagenesis. Indicating that S2 is the water table. S2 is not as continuous as S1. It is not present in the area of low reflectivity.

Radar facies F1 is characterized by long, sigmoidally dipping clinoforms (A visualization of the radar facies is found in Figure 8). The clinoforms pass through S2. They can be frequently traced from S1 to the reflection free zone towards the bottom of the survey. They relatively

parallel to each other but do experience some onlaping. They have steep dip angles ranging from

3 to 15 degrees, with the majority of the reflectors dipping between 6 to 8 degrees. The traceable

clinoforms extend between 2 and 10 meters with the majority of them being extending 4 to 8

meters. The reflectors in a particular F1 grouping stack with the point of inflection increasing

seaward (Figure 8). This pattern is commonly identified as a forced regression pattern in the

falling stage systems tract. This pattern is an indication of a drop in relative sea-level, moving the

accommodation seaward (Coe, 2003). F1 deposits appear to be have no pattern in their

deposition (Figure 9).

F2 deposits are similar to F1 deposits. They have steep dipping reflectors ranging from 3

to 15 degrees with the majority of them dipping between 6 to 8 degrees. The traceable reflectors

extend between 2 and 20 meters with the majority of them extending 4 to 10 meters. F2 differs

from F1 in that the inflection points contained within a given F2 grouping do not stack (Figure

8). F2 reflectors are relatively parallel to one another and do not appear to truncate the other

packages. They are, however, often truncated by and lap onto F1 deposits. The pattern resembles

14 progradation in which there is no relative change in sea-level, but there is a steady source of

sediment.

F3 is comprised of landward dipping reflectors. F3 is usually a smaller package that is

located below S1 and above S2 (Figure 9). It varies between 0.5 and 0.8 meters in depth. The

reflectors vary in length from 0.5 to 10 meters in length. The shorter reflectors are steeper and

offlap from S1 and onto F2 deposits. The longer reflectors connect into F2 deposits and show the

turnover of the sediment landward. They usually correspond to areas where S1 is shallower,

indicating a ridge is present. F3 is interpreted to be eolian dune deposits (Gomez-Ortiz et al.2009).

F4 is a high attenuation zone. F4 is primarily located at the bottom of the surveys where

the GPR signal attenuates. F4 zones are also found on the leeward side of some F3 deposits. This

correlation could be coincidental, but these areas were predictably found in the swales as the data

were being acquired. This could indicate that that the swales were harder for the signal to

penetrate than the ridges. This could be the result of a number of things such as an increase in

salt concentration, or an increase in fine particles. In order to conclusively state whether these

zones were the result of the F3 deposits sediment sampling would need to be taken along the

ridge into the swale and show an increase in such particles.

F5 is a chaotic zone with no significant reflections. It lies above S1 and is allochthonous and anthropogenic in nature.

Radar Facies Distribution The dominant radar facies contained in all of the zones is F2. It appears from the data that

F2 is a fair-weather deposition pattern. The reflectors travel from deep in the survey up through the water table and roll over landward. This is interpreted as being a shoreface that transitions

15 into a back-dune environment, F3 facies. This pattern is punctuated at times with F1 deposits.

These F1 deposits appear to cut into and truncate the F2 deposits (Figure 9). F1 deposits are

interpreted as storm deposits. These facies likely form in very short time spans in compared to their F2 counterparts. They appear to be constructed of sediment that is eroded from the F2.

Interestingly, the F3 deposits are most prominent above the F1 deposits. This could be because the rapid deposition of F1 creates an environment directly above sea-level that allows for dunes to form more easily. One way this could occur is when the storm surges abate, sediment is left above normal sea-level. The sediment then dries and becomes mobile forming dunes.

There appears to be no zone-specific radar facies, F1-5 appear to be present in every zone. This is easily visualized in the pseudo-3D surveys. The pseudo-3D surveys show consistency between the 3 zones sampled, and laterally within Zone D. Figure 10 shows time slices of the survey 1. It shows the consistency of the reflectors within the survey. It clearly shows continuous reflectors that extend the width of the survey and move seaward deeper into the survey. Figure 11 displays one line from each of the pseudo-3D surveys and displays it in its location on Sandy Hook. It shows the similarities in the reflector dip and direction, and the thickness of the reflective packages. There are slight variations in the dip of the reflectors between zones ~1 degree. These subtle variations in the dip and length of reflectors are likely the result of changes in the amount of accommodation available as deposition occurred (Mattheus et al., 2018).

The lack of variance between the zones does indicate that strand plains form in a repeatable pattern. This pattern appears to be able to be interrupted by erosional events separating the four zones (observe the overall orientation of the interpreted ridges of Figure 4) that can change the orientation of deposition, without having a large impact on the process.

16 Sea-Level Fluctuations Ground-penetrating radar surveys across beach ridge deposits have the potential to contain evidence of past relative sea-level (Tamura et al. 2008; Nielson and Clemmensen, 2009).

Tamura et al. (2008) identified downlapping reflectors as markers indicating sea-level ~1 m below the mean sea-level. This method has been established as one method of identifying relative sea-level fluctuations from ground-penetrating radar surveys (Nielson and Clemmensen,

2009; Clemmensen and Nielson, 2010; Hede et al., 2013; Hede et al., 2015). An analysis of the

2D GPR line transecting zones B-D shows that sea-level fluctuated minimally through the development of Sandy Hook. The downlapping reflector’s points of inflection varied from 0.78 to 1.1 meters below the surface in Zone B, from 0.9 to 1.38 meters below the surface in Zone C, and 1.1 to 1.38 meters below the surface in Zone D. The average depth of the inflection points in

Zone B was 0.94 m, in Zone C was 1.13 m, and in Zone D was 1.22 m. The depth of inflections decreases as the island moves from west to east. This indicates that there has been a decrease in overall sea level through the Holocene.

The inflection points also appeared to be correlated with S2 the bright reflection identified earlier as the top of the water table. It is possible that the correlation between the inflection point and the water table are due to the difference in the dielectric constant of water and air (the interaction would result in steeper dipping reflections in the water as opposed to the air). We do not feel this is the case here as there does not appear to be an overall steepening resulting in change, but it appears to occur on distinct reflectors. Thus, the inflections would be evidence of slight variations in relative sea-level.

The F1 deposits would be interpreted as FSST deposits if looked at through the lens of sequence stratigraphy. This would imply that large, long-term, relative sea-level falls were present. In this case, the frequency of occurrence is an order of magnitude smaller than the

17 flooding surfaces of Van Wagoner et al. (1990). The FSST deposits identified here were termed storm surfaces by Bristow et al. (2000) and are likely the result of deposition after storm erosion

(Bristow et al. 2000).

Storm vs. Fair-Weather The difficulty in definitively stating whether or not Sandy Hook was formed through a series of storm events, or gradually over time in fair-weather conditions is that Sandy Hook is entirely calcium carbonate. This is problematic because the primary method of identifying storm deposits is by identifying heavy mineral lags (Dougherty et al., 2004), which are not present in carbonate systems. Therefore, a new method of identifying storm deposits is put forth. We believe that sigmoidal FSST signatures that have higher frequency than the flooding surfaces identified by Van Wagoner et al. (1990) and that truncate the normal pattern can be used as evidence of storm deposits.

We believe that Sandy Hook formed through a series of fair-weather/high-energy storm cycles (Figure 12). Under fair weather conditions, deposition occurs as sediment is stored on the beachface. Small ridges have the potential to form by fair-weather waves building a berm

(Davies, 1957; Tamura, 2012; Figure 12a). Deposition continues in this manner until times of high energy storm periods. During stormy periods, the storm surge raise the normal wave level.

This causes a reworking of sediment deposited during fair-weather conditions and carrying the sediment landward (Figure 12b). When the storms abate, the wave level falls back to fair- weather conditions. Because the sediment is primarily medium to fine in grain size, the grains are carried back seaward and deposited quickly in the accommodation created by the change in wave level (Figure 12c). Fair-weather condition deposition begins again. Eolian deposits

18 commence as the storm deposits and recently deposited fair-weather dry and become mobile

(Figure 12d).

CONCLUSION We believe that Sandy Hook was developed partially through periods of fair-weather accumulation, and partially through abrupt periods of drastic deposition (Figure 12). Evidence of fair-weather accumulation and storm deposits is found in the GPR data. The GPR signature of

Sandy Hook is remarkably similar moving from zone to zone across erosional boundaries. This implies that even though deposition was interrupted, it continued in the same way as before.

Additional work can be done to confirm whether or not this depositional pattern is present in other carbonate strand plains, or whether it is Sandy Hook specific.

The C-14 results indicate that Sandy Hook is Holocene in age. In the last 2600 years,

Sandy Hook accreted at rates of 0.29 to 0.08 meters per year or 212 - 891 m3 per year of sediment accumulation. Combining C-14 and the 2D GPR line we were able to conclude that the sea level fluctuated by 10 centimeters in the last 2600 years, and that the sea-level fluctuation was not related to the ridge height. Rather it appears that ridge height is correlated to the beach deposits directly below the ridge. Smaller, more shallow ridges form on fair weather beach deposits; whereas, larger ridges form on top of the storm deposits. The reasoning for this correlation is unknown, but it is theorized that this occurs because the erosion associated with the storm deposit creates conditions that enhance the development of ridges.

19 REFERENCES

Aurell, M., McNeill, D. F., Guyomard, T., and Kindler, P., 1995, Pleistocene shallowing-upward sequences in New Providence, Bahamas: Signature of high-frequency sea-level fluctuations in shallow carbonate platforms: Journal of Sedimentary Research, v. 65, p. 170-182. Boyd, R., Dalrymple, R., and Zaitlin, B.A., 1992, Classification of clastic coastal depositional environments: Sedimentary Geology, v. 80, p. 139-150. Bristow, C.S., Chronston, P.N., and Bailey, S.D., 2000, The structure and development of foredunes on a locally prograding coast: insights from ground-penetrating radar surveys, Norfolk, UK: Sedimentology, vol. 47, p. 923-944. Bristow, C. S. and Pucillo K., 2006, Quantifying rates of coastal progradation from sediment volume using GPR and OSL: the Holocene fill of Guichen Bay, south-east South Australia: Sedimentology, vol. 53, p. 769-788 Carew, J.L., and Mylroie, J.E., 1997, Geology of the Bahamas. P. 91-139. In: Vacher, H.L., and Quinn, T.M., (eds) Geology and Hydrogeology of Carbonate Islands: Developments in Sedimentology 54. Elsevier, Amsterdam Carney, C., Stoyka, G.S., Boardman, M.R., and Kim, N., 1993, Depositional history and diagenesis of a Holocene strand plain, Sandy Hook, San Salvador, Bahamas, in White, B., ed., The Proceedings of the Sixth Symposium on the Geology of the Bahamas, p. 35-45 Carter, R.W.G., 1986, The morphodynamics of beach-ridge formation: Magilligan, Northern Ireland: Marine Geology v. 73, p. 191-214. Clemmensen, L.B., and Nielson, L., 2010, Internal architecture on a raised beach ridges system (Anholt, Denmark) resolved by ground-penetrating radar investigation: Sedimentary Geology, vol. 223, p. 281-290. Coe, A.L. ed., 2003, The sedimentary record of sea-level change: Cambridge, Cambridge University Press Cunningham, K.J., 2004, Application of ground-penetrating radar, digital optical borehole images, and cores for characterization of porosity hydraulic conductivity and paleokarst in the Biscayne aquifer, southeastern Florida, USA: Journal of Applied Geophysics, v. 55, p. 61–76. Davies, J.L., 1957, The importance of cut and fill in the development of sand beach ridges: Australian Journal of Science, vol. 20, p. 105-111. Dominic, D.F., Egan, K., Carney, C., Wolfe, P.J., and Boardman, M.R., 1995, Delineation of shallow stratigraphy using ground penetrating radar: Journal of Applied Geophysics, v. 33, p. 167-175. Dougherty, A.J., FitzGerald, D.M., and Buynevich, I.V., 2004, Evidence for storm-dominated early progradation of Castle Neck barrier, Massachusetts, USA: Marine Geology, v. 210, p. 123-134.

20 Evans, G., 1970, Coastal and Nearshore Sedimentation: a Comparison of Clastic and Carbonate Deposition: Proceedings of the Geologist’ Association, v. 87, p. 493-508 Franseen, E.K., Byrnes, A.P., Xia, J., and Miller, R.D., 2007, Improving resolution and understanding controls on GPR response in carbonate strata: implications for attribute analysis: The Leading Edge, v. 26, p. 984–993. Garrett, P., and Gould, S. J., 1984, Geology of New Providence Island, Bahamas: Geological Society of America Bulletin, v. 95, p. 209-220. Gomez-Ortiz, D., Martin-Crespo, T., Rodriguez, I., Sanchez, M.J., and Montoya, I., 2009, The internal structure of modern barchan dunes of the Ebro River Delta (Spain) from ground penetrating radar: Journal of Applied Geophysics, vol. 68, p. 159-170. Gontz, A.M., Moss, P.T., and Wagenknecht, E.K., 2014, Stratigraphic Architecture of Regressive Strand plain, Flinders Beach, North Stradbroke Island, Queensland, Australia: Journal of Coastal Research, v.30, Iss 3, p. 575-585. Grasmueck, M., Weger, R., and Horstmeyer, H., 2005, Full-resolution 3D GPR imaging: Geophysics, v. 70, p. K12-K19. Harris, P.M., 1979, Facies anatomy and Diagenesis of a Bahamian Ooid Shoal; Sedimenta VII: Miami, Florida, Comparative Sedimentology Laboratory p. 163 Hazard, C.S., Ritter, S.M., McBride, J.H., Tingey, D.G., Keach, R.W., 2017, Ground- penetrating-radar characterization and porosity evolution of an upper Pleistocene oolite- capped depositional cycle, Red Bays, Northwest Andros Island, Great Bahama Bank, Journal of Sedimentary Research, v. 87, p. 523-545 Hede, M.U., Bendixen, M., Clemmensen, L.B., Kroon, A., Nielson, L., 2013, Joint interpretation of beach-ridge architecture and coastal topography show the validity of sea-level markers observed in ground-penetrating radar data: The Holocene, v.23 p. 1238-1246 Hede, M.U., Sander, L., Clemmensen, L.B., Kroon, A., Pejrup, M., and Neilson, L., 2015, Changes in Holocene relative sea-level and coastal morphology: A study of a raised beach ridge systems on Samso, southwest Scandinavia: The Holocene, v. 25, p. 1402-1414. Hearty, P. J., and Kindler, P., 1993, New perspectives on Bahamian geology: San Salvador Island, Bahamas: Journal of Coastal Research, p. 577-594. Hearty, P. J., and Kindler, P., 1997, The stratigraphy and surficial geology of New Providence and surrounding islands, Bahamas: Journal of Coastal Research, p.798 Johnson, D.W., 1919, Shore Processes and Sedimentation, Prentice-Hall, Englewood Cliffs, NJ. p. 429 Kim, N., 2001, Petrographic characteristics of a Holocene carbonate grainstones in southeast San Salvador, Bahamas: Implications for the sediment reworking in a high energy lagoon: Geosciences Journal, v. 5, p. 349-359. Liu, K.B and Fearn, M.L., 2000, Reconstruction of prehistoric landfall frequencies of catastrophic hurricanes in northwestern Florida from lake sediment records: Quaternary Research, vol. 54, p. 238-245

21 Mattheus, C.R., Farhan, S. A., and Fowler, J. K., 2018, Geomorphology of the Late Holocene Sandy Hook Strandplain, San Salvador Island, Bahamas, Gerace Research Centre, p.15 Mylroie, J.E. and Carew, J.L., 1995, Geology and karst geomorphology of San Salvador Island, Bahamas: Carbonates and Evaporites, v. 10, iss. 2, p. 193-206. Mylroie, J. E., and Carew, J. L., 2008, Field guide to the geology and karst geomorphology of San Salvador Island: San Salvador, Bahamas, Gerace Research Centre, p. 89. Neilson, L., and Clemmensen, L.B., 2009, Sea-level markers identified in ground-penetrating radar data collected across a modern beach ridges system in a microtidal regime: Terra Nova v. 21, p. 474-479. Otvos, E.G., 2000, Beach ridges--definitions and significance: Geomorphology, v. 32, p. 83-108 Redman, J.B., 1852, on the alluvial formations, and the local changes of the south coast of England: Minutes of Proceedings Institution of Civil Engineers, v. 11, p. 162-223 Redman, J.B., 1864, the east coast between Thames and the Wash estuaries: Minutes of Proceedings Institution of Civil Engineers, v. 23, p. 186-257. Rivers, J., Engel, M., Dalrymple, R., Yousif, R., Strohmenger, C. J. and Al-Ahaikh, I., 2019, Are carbonate barrier islands mobile? Insights from a mid to late-Holocene system, AL Ruwais, northern Qatar: Sedimentology. doi: 10.1111/sed.12653 Rust, A.C., AND Russell, J.K., 2001, Mapping porosity variation in a welded pyroclastic deposit with signal and velocity patterns from ground-penetrating radar surveys: Bulletin of Volcanology, v. 62, p. 457–463. Scholle, P.A., Bebout, D.G. and Moore, C.H. eds., 1983, Carbonate Depositional Environments: AAPG Memoir 33, No. 33, AAPG. Tamura, T., Murakami, F., Nanayama, F., Watanabe, K., Saito, Y., 2008. Ground-penetrating radar profiles of Holocene raised-beach deposits in the Kujukuri strand plain, Pacific coast of eastern Japan: Marine Geology, v. 248, p. 11-27 Tamura, T., 2012, Beach ridges and prograded beach deposits as palaeoenvironment records: Earth-Science Reviews, v. 114, p. 279-297 Tanner, W.F., 1995. Origin of beach ridges and swales: Marine Geology v. 129, p. 149-161. Taylor, M.J., and Stone, G., W., 1996, Beach-Ridges: a review: Journal of Coastal Research v. 12, p. 612-621. Van Heteran, S., FitzGerald, D.M., McKinlay, P.A., and Buynevich, I.V., 1998. Radar facies of paraglacial barrier systems: coastal New England, USA: Sedimentology, v. 45, p. 181-200. Van Wagoner, J.C., Mitchum, R.M., Campion, K.M. and Rahmaniam, V.D., 1990, Siliciclastic sequence stratigraphy in well logs, cores, and outcrops: AAPG Methods in Exploration Series, no. 7, Tulsa, OK, 55 pp. Ward, W.C., Brady, M. J., 1979, Strandline Sedimentation of Carbonate Grainstones, Upper Pleistocene, Yucatan Peninsula, Mexico: AAPG Bulletin vol. 63, No. 3, p.362-369

22 nd. nd. A) Location of the Bahama Islands. San Salvador, and isolated carbonate platform, marked by a white box. B) San Salvador Isla San Salvador a white box. B) by marked platform, carbonate isolated and San Salvador, Islands. Bahama the of Location A) — Figure 1 Earth. Google of copyright Image an arrow. by identified corner, the in southeastern located Hook Sandy The

23 A B

Figure 2—Geologic Map of San Salvador Island, modified from Hearty and Kindler, (1993).

24 Figure 3—Sandy Hook. The locations of the pseudo-3D GPR surveys are marked by orange rectangles. The 2D GPR survey is shown as a red line. The auger locations are marked by yellow dots. The C-14 locations are marked by a black dot contained within the yellow auger dots. Image copyright of Google Earth.

25 Figure 4—Line drawn image of Sandy Hook. The lines were drawn based on vegetation patterns. The orientation of the lines was used to separate Sandy hook into four zones, identified by different color lines. The data collection locations are indicated as in figure 3.

26 Figure 5—Thin section photos of the cemented layer interpreted as the top of the water table. Meniscus cementation is identified by white arrows. Background image copyright of Google Earth.

27 Figure 6—Histograms with the auger locations from which they were acquired. The histograms show the similarities in the sediment from the different zones. Background image copyright of Google Earth.

28 Figure 7—Isochron map displaying the C-14 ages. The ages decrease from 2617±188 YBP to -13±1YBP (modern) in 460-375 meters. Background image copyright of Google Earth.

29 Figure 8—Radar facies diagram.

30 A

31 C

Figure 9—Four uninterpreted (upper) and interpreted (lower) 400 MHZ profiles (A-D). The profiles depict the interaction of the radar facies, and surfaces. The interpretations were made based on the Radar facies diagram 32 (Figure 8). The vertical exaggeration of the profiles is the same as the vertical exaggeration of figure 8’s figures. Background image copyright of Google Earth. A

33 B

Figure 10—A) Horizontal time slices of Survey 1. B) Horizontal time slices of Survey 4. The time slices decrease in depth moving from left to right down the figure. The top left image is the shallowest time slice, and the bottom right is the deepest. Similarities between the changes in dip direction and lateral continuity of the reflectors is identified.

34 Figure 11—2D lines taken from the pseudo-3D surveys in their locations. The 2D lines were selected as the best representation of the reflectivity found within the surveys. The 2D lines are displayed with zero vertical exaggeration, depicting the similarities in dip between the zones.

35 ediments ediments quickly in the quickly ing weather deposits from. weatherEolian deposits from. deposits - weather conditions deposition occurs as sediment is stored on the on the stored is as sediment occurs deposition conditions weather - ther return and conditions fair sediments are brought back seaward, deposit seaward, back brought are sediments wea - -weather deposits dry and become mobile. mobile. become and dry deposits -weather air . D) Fair D). A) fair D). Under - , and the levels weather - to fair lain deposition (Alain deposition its and recently deposited f deposited recently its and falls level wave he Model of strand p — Figure 12 Figure of s carrying and of reworking causes This level. wave the normal raise surges storm periods storm large During B) beachface. level in wave the change by created accommodation depos as storm commence landward. C)landward. T