Visualization of Coastal Carbonate Lithosomes: Color-Intensity Patterns and Georadar Imaging of a Semi-Lithified Strandplain, Eleuthera Island, The Bahamas

Abstract

Quaternary carbonate strandplains serve as archives of land–sea interaction, including the impacts of storms and tsunamis. Incipient lithification, especially of compound beach/dune ridges within the action zone of salt spray, presents challenges to geological research, which is often limited to exposures. This study combines aerial image analysis with geophysical datasets to assess the morphostratigraphy and internal structure of the Freedom Beach Strandplain along southern Eleuthera Island, The Bahamas. Color-intensity analysis of field photographs and satellite images revealed general patterns that can be used to distinguish between areas with different grayscale parameters (sand-covered surfaces, lithified ridges, vegetation, etc.). Cross-shore (dip-section) high-resolution (800 MHz) georadar images across ten ridges (A-J) documented the internal architecture of swash-aligned ridge–swale sets. Signatures attributed to storms include truncations in shore-normal radargrams, scour features in alongshore (strike-section) images, and an extensive accumulation of large mollusk shells along one of the oldest ridges (ridge J). Preliminary radiocarbon dating yielded ages of up to 600 years, suggesting intense storms with 50–60-year periodicity as a possible mechanism for ridge formation.

1. Introduction

Coastal strandplains (beach/dune ridge and swale systems) exist on all continents and are characterized by seaward accretion (progradation) during the Holocene. These systems serve as archives of morphodynamic processes punctuated by hydroclimatic events (e.g., intense storms), which are often responsible for the origin of the ridge morphology itself [1,2]. This aspect is particularly important for sea-level research, where care must be taken in using morphological elements as water-level indicators [3,4,5,6].

The Bahamian archipelago contains numerous strandplains with variable orientations, and there has been recent interest in understanding their origin and distribution [6,7]. However, to date, little research has focused on the morphostratigraphy (spatial ridge–swale relationships) and internal architecture of the Bahamian strandplains. The latter aspect is due largely to substantial challenges presented by semi-indurated limestone substrates and limited outcrops of low-lying ridge sets. Understanding key architectural elements of wave-built and aeolian facies not only has implications for locating similar structures in areas where morphology is lacking (e.g., graded surfaces) but also has important implications for recognizing analogous sequences in the rock record [8,9].

Using a small strandplain along the southern coast of Eleuthera Island as an example, the aims of this paper are as follows: (1) to characterize morphostratigraphy of vegetated ridges using color patterns on aerial images; (2) to assess the general erosional-accretionary patterns of ridge and swale systems using high-resolution geophysical images; and (3) to establish the preliminary chronology of the strandplain based on preliminary radiocarbon ages.

2. Study Site

The research area is located in Freetown, southern Eleuthera (“Freedom”) Island, the Bahamas (24°45′40″ N/76°15′45″ W; Figure 1). Here, the study area is referred to as the Freedom Beach Strandplain (FBS), a ridge-and-swale set bounded by a beachrock headland to the west and Plum Creek to the east (Figure 1C). This swash-aligned (concave-seaward or catenary) strandplain has a southerly exposure to Exuma Sound and spans >900 m in alongshore extent (modern beach), with a maximum width of ~300 m in its central section. This is a microtidal setting (mean tidal range: 0.6 m) with a ~2 h lag between the Exuma Sound and the east coast of Eleuthera exposed to the Atlantic Ocean (Figure 1B). The study site has low wave energy due to its leeward location and an embayed morphology bounded by headlands, with the greatest morphodynamic impact occurring during intense storms.

3. Materials and Methods

This study was part of a regional field research program conducted in June 2013 that included several sites on Eleuthera Island (Figure 1). The main 250-m-long topographic and geophysical profile is located in the eastern portion of the FBS (Figure 2A). Topographic leveling was accomplished using the Emery-style rod-and-level method, with a vertical accuracy of <2 cm. Key inflection points were geolocated with a hand-held Garmin GPS (accuracy: <1.5 m) and included seaward ridge base, ridge crest, and landward ridge base, with areas between adjacent ridge bases representing swales (e.g., swale D/E spans between the landward edge of ridge D and seaward edge of ridge E).

The light/dark patterns visible in satellite images (Google Earth; resolution ~0.3 m; Figure 1C and Figure 2A) and in field photographs were analyzed using ImageJ freeware (v. 1.54g). The image color-intensity (ICI) was expressed by variable-range spectral modes [10,11]. A 256-shade grayscale (GS) spectrum varied from lowest intensity black (GS = 0) to highest intensity white (GS = 255). A 16-color option rendered the light/dark regions as hotter (red)/cooler (blue) hues, respectively. The pattern of GS values was used to represent the relative ICI variations between light (bare carbonate sand) and dark (vegetation, beachrock) surfaces. At study sites where ridge tops have higher-intensity values (e.g., exposed light sand) or shallow/exposed surfaces are contrasted with dark underwater regions, a normal grayscale was used. This way, topographically low areas were represented by troughs or negative peaks. The scale bar was inverted or values were re-calculated as an inverse grayscale (IGS: 256-GS) to facilitate the visualization of patterns where darker values (e.g., vegetated ridges; Figure 2A) were expressed as positive peaks.

Geophysical surveys were collected using a digital MALÅ GPR system with a ProEx control unit, a Ramac XV11 monitor, and a shielded 800 MHz transceiver antenna with an odometer wheel for triggering and along-ground distance measurements (Figure 2B). The main shore-normal (dip-section) profile (combined lines 59 and 63) traversed all ridges from A to J, with profile 61 to the south spanning the youngest set (Figure 2A). Profiles 60, 62, and 64 were collected along the swales to offer strike-section visualization of the subsurface architecture. A signal velocity of 12 cm/ns was used for upper unsaturated parts of the carbonate lithosome (semi-consolidated and unconsolidated) based on empirical data from a number of Bahamian sites [8] and hyperbola fitting during surveys. Basic post-processing protocols were applied to representative radargrams (2D sections or B-scans) for technical aspects of data collection and processing [8,12,13,14,15]. The topographic profile collected along the main shore-normal transect served as the basis for the surface normalization of georadar images. Preliminary radiocarbon ages were obtained at the Beta Analytic Laboratory on whole-rock samples from ridges A, H, and J, as well as a mollusk shell from ridge J, following weathering rind removal and local reservoir correction’s application.

4. Results

Remotely sensed databases (satellite images, aerial photography; Figure 2) and field surveys (Figure 3) demonstrate a clear visual pattern of ridge–swale topography based on relative color-intensity characteristics. In the field, the image color-intensity varies from light (GS > 100) for exposed carbonate sand (modern beach and swales) to darker surfaces characteristic of lithified limestone and vegetation, both common for the ridges. It is important to note that this is a relative difference assessed for every field photograph (Figure 3); however, common ranges of grayscale values are sufficient to differentiate the key sub-environments.

Similarly, the spatial patterns of the gross ridge–swale geometry visible in satellite images (Figure 4A) were expressed in a 16-bit color rendition (Figure 4B), with cooler (blue) colors corresponding to vegetated ridges. The topographic profile along the geophysical survey line in the eastern part of the strandplain (Figure 4C) shows that most ridges range between 1.5 m (ridge I) and 2.7 m (ridge A) were above the higher high-water level. Grayscale profiles D (~topographic profile location) and E (central part of strandplain) show general similarity in ICI patterns that correlate with topography (Figure 4D,E).

This correlation is further demonstrated in Figure 5, with higher seaward (younger) ridges having darker colors and most of the older ridges yielding relatively light values (low inverse grayscale, IGS). Profiles D and E have similar trends, with a second-order (quadratic) polynomial producing a good fit.

The subsurface architecture of the ridges is expressed in shore-normal (dip-section; Figure 6 and Figure 7) and shore parallel (strike-section; Figure 8) GPR images. The second youngest ridge B reveals a complex pattern of seaward-dipping reflections (note: high-amplitude 2D scans are designed to show the fine resolution of the 800 MHz setup). In the landward portion (B/C swale), a sharp and deep seaward-dipping reflection is onlapped by landward-dipping bounding surfaces. The complex and chaotic bedding of the ridge B superstructure transitions into a series of seaward-dipping radar facies, with multiple hyperbolic diffractions (point-source targets) seen throughout the radargram (Figure 6).

Along the landward portion of the FBS, the oldest surveyed ridge J contains a similar pattern, with sharp truncations both landward and seaward of the ridge (Figure 7). Above the deeper zone of attenuated signal response, there are numerous hyperbolic diffractions underlying the offlapping sequence.

The shore-parallel survey collected along the seaward portion of ridge J reveals several concave-up structures, which extend to ~1 m below the ground surface (note: no topographic correction was necessary). Similar to shore-normal radargram segments, multiple point-source diffractions are visible at depth, with signal attenuation beginning at a ~2 m depth.

The first set of radiocarbon dates ranges between 570 ± 30 yBP (cal CE 1710 to post-1950) for ridge A, 370 ± 30 yBP (post-1950) for ridge H, and 970 ± 40 yBP (cal CE 1400–1600) for ridge J (2-sigma calibrated ages in parentheses). An extensive accumulation of mostly juvenile conch shells (Aliger sp.), exhibiting different degrees of breakage, abrasion, and weathering, occur along the seaward slope of ridge J. The 290 ± 30 yBP radiocarbon date for a large shell is calibrated to a “modern” age.

5. Discussion

This study demonstrates the applicability of combining a simple image analysis tool with geophysical imaging for investigating carbonate strandplain morphology and subsurface architecture, respectively (Figure 9). The Freedom Beach Strandplain is a classic swash-aligned beach/dune ridge-and-swale system developed through progradation within a low-energy embayment [2,6,7,16]. The ridges are commonly associated with erosional features, such as storm scarps, which likely nucleate their formation. This is supported by their catenary (concave-seaward) geometry mimicking the approach of refracted waves [17,18,19,20,21,22]. Whereas sea-level position is an important long-term constraint on beach/dune ridge elevation, most ridge sets are the result of sub-century storm-mediated periodicity rather than high-frequency sea-level fluctuations [3,4,5,23].

Image-color-intensity values offer a promising means of distinguishing between ridges and swales, in this case, darker colors resulting from vegetation and carbonate lithification (Figure 3, Figure 4 and Figure 5; Figure 9A–C). The latter is likely due to incipient diagenesis, bacterial staining, or vegetation debris (Figure 3), among other factors. To show that the higher topographic features have lower intensity (darker) ICI values, profiles can be plotted either in an inverted grayscale (GS = 0 [black] at top; Figure 4D,E) or as an inverse grayscale (IG = 256 [black] at top; Figure 5). At some sites, light-colored carbonate ridges and flooded swales with vegetation or algal mats would be represented by a normal grayscale (GS = 256 [white] at top).

Georadar images of carbonate facies, both semi-lithified and unconsolidated (oolitic and bioclastic sand), have generally similar pattern characteristics to their unconsolidated siliciclastic counterparts [8,24,25,26,27,28,29,30,31]. Smaller-scale complexity documented in FBS radargrams is likely due to a combination of diagenetic features, buried vegetation (or trunk molds), water/air escape features (up to large pipe structures), and a spectrum of biogenic structures, such as medium-to-large burrows [32,33,34,35,36,37,38,39,40]. Because larger biogenic structures, like sea turtle nests, likely exist along paleo-swales at many Caribbean sites, caution must be taken in differentiating them from small storm channels or anthropogenic excavations [41,42]. Large-scale depressions (Figure 8) are likely due to storm scour, but more research is needed. Rather than merely using them as “tie” lines between traditional shore-normal transects, we highlight the importance of conducting alongshore surveys even in areas where long-term paleo-channels are not likely (tidal prism access), but event-related scouring may be identified.

The anomalous accumulation of conch shells (Aliger sp.) along the seaward slope of ridge J is likely an event-generated accumulation [43,44] and was recently interpreted as a storm-generated thanatocoenosis of mostly local mollusks [45]. This was likely a moderate storm (seaward scarp in Figure 7) that did not affect the ridge J superstructure. Intense storms, especially following limited progradation, are likely to modify or erode low unlithified ridges. Based on preliminary radiocarbon dating results, the age of 970 ± 40 yBP (cal 1400–1600 CE) for ridge J suggests a 40–60-year/ridge time-averaged accumulation rate for the 10-ridge sequence. With a medium-term 50–60-year recurrence interval of tropical cyclones suggested as a hydroclimatic forcing at many sites, including the Caribbean basin [46,47,48,49,50,51,52], the preliminary chronology is consistent with storms as a mechanism for scarp-nucleated, catenary beach/dune ridge formation.

6. Conclusions

Our study demonstrates the viability of image-color-intensity analysis and georadar imaging for characterizing the surface morphology and subsurface structure of a carbonate strandplain, respectively. Color-intensity analysis provides a rapid and very effective means of performing both the qualitative and quantitative assessment of morphological and ground-cover elements. Lighter (higher GS) shades correspond to bare carbonate sand (beach and swales), whereas darker hues are more common for lithified limestone (e.g., ridges) and vegetated areas. Grayscale inversion, depicted as an inverted GS axis or IGS value, is used for visual comparison with topographic profiles. High-resolution GPR images, especially with high-amplitude rendition, reveal highly complex depositional patterns, which are further complicated by point-source reflections attributed to limestone rubble or biogenic structures. Shallow depressions in shore-parallel radargrams may be the result of storm scour. Clear truncations (paleo-scarps) immediately behind the ridges support the mechanism of storm erosion as a nucleus for subsequent swash-generated vertical accretion and aeolian aggradation. This easily explains the swash-aligned pattern of most strandplains, which resembles wave refraction patterns within embayments. The occurrence of exposed molluscan taphoceonoses is likely related to a moderate storm, as in the case of ridge J fronted by a younger scarp. Preliminary ages are generally consistent with the 50–60-year recurrence interval of tropical cyclones in this part of the Atlantic Ocean basin.